pubmed_id
stringlengths 41
43
| abstract
stringlengths 3
18.8k
|
|---|---|
http://www.ncbi.nlm.nih.gov/pubmed/33364802 | 1. Diabetes Metab Syndr Obes. 2020 Dec 15;13:4981-4992. doi:
10.2147/DMSO.S283949. eCollection 2020.
Reporting and Methods in Developing Prognostic Prediction Models for Metabolic
Syndrome: A Systematic Review and Critical Appraisal.
Zhang H(1), Shao J(1), Chen D(1), Zou P(2), Cui N(3), Tang L(1), Wang D(1), Ye
Z(1).
Author information:
(1)Department of Nursing, Zhejiang University School of Medicine Sir Run Run
Shaw Hospital, Hangzhou, Zhejiang, People's Republic of China.
(2)Department of Scholar Practitioner Program, School of Nursing, Nipissing
University, Toronto, Ontario, Canada.
(3)Department of Nursing, The Second Affiliated Hospital Zhejiang University
School of Medicine, Hangzhou, Zhejiang, People's Republic of China.
PURPOSE: A prognostic prediction model for metabolic syndrome can calculate the
probability of risk of experiencing metabolic syndrome within a specific period
for individualized treatment decisions. We aimed to provide a systematic review
and critical appraisal on prognostic models for metabolic syndrome.
MATERIALS AND METHODS: Studies were identified through searching in English
databases (PubMed, EMBASE, CINAHL, and Web of Science) and Chinese databases
(Sinomed, WANFANG, CNKI, and CQVIP). A checklist for critical appraisal and data
extraction for systematic reviews of prediction modeling studies (CHARMS) and
the prediction model risk of bias assessment tool (PROBAST) were used for the
data extraction process and critical appraisal.
RESULTS: From the 29,668 retrieved articles, eleven studies meeting the
selection criteria were included in this review. Forty-eight predictors were
identified from prognostic prediction models. The c-statistic ranged from 0.67
to 0.95. Critical appraisal has shown that all modeling studies were subject to
a high risk of bias in methodological quality mainly driven by outcome and
statistical analysis, and six modeling studies were subject to a high risk of
bias in applicability.
CONCLUSION: Future model development and validation studies should adhere to the
transparent reporting of a multivariable prediction model for individual
prognosis or diagnosis (TRIPOD) statement to improve methodological quality and
applicability, thus increasing the transparency of the reporting of a prediction
model study. It is not appropriate to adopt any of the identified models in this
study for clinical practice since all models are prone to optimism and
overfitting.
© 2020 Zhang et al.
DOI: 10.2147/DMSO.S283949
PMCID: PMC7751606
PMID: 33364802
Conflict of interest statement: The authors have no conflicts of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/25314315 | 1. PLoS Med. 2014 Oct 14;11(10):e1001744. doi: 10.1371/journal.pmed.1001744.
eCollection 2014 Oct.
Critical appraisal and data extraction for systematic reviews of prediction
modelling studies: the CHARMS checklist.
Moons KG(1), de Groot JA(1), Bouwmeester W(1), Vergouwe Y(1), Mallett S(2),
Altman DG(3), Reitsma JB(1), Collins GS(3).
Author information:
(1)Julius Center for Health Sciences and Primary Care, UMC Utrecht, Utrecht, The
Netherlands.
(2)Department of Primary Care Health Sciences, New Radcliffe House, University
of Oxford, Oxford, United Kingdom.
(3)Centre for Statistics in Medicine, University of Oxford, Botnar Research
Centre, Windmill Road, Oxford, United Kingdom.
Carl Moons and colleagues provide a checklist and background explanation for
critically appraising and extracting data from systematic reviews of prognostic
and diagnostic prediction modelling studies. Please see later in the article for
the Editors' Summary.
DOI: 10.1371/journal.pmed.1001744
PMCID: PMC4196729
PMID: 25314315 [Indexed for MEDLINE]
Conflict of interest statement: The authors have declared that no competing
interests exist. |
http://www.ncbi.nlm.nih.gov/pubmed/35493941 | 1. Front Aging Neurosci. 2022 Apr 7;14:840386. doi: 10.3389/fnagi.2022.840386.
eCollection 2022.
Prediction Models for Conversion From Mild Cognitive Impairment to Alzheimer's
Disease: A Systematic Review and Meta-Analysis.
Chen Y(1), Qian X(2), Zhang Y(1), Su W(1), Huang Y(1), Wang X(1), Chen X(1),
Zhao E(1), Han L(1)(3), Ma Y(1)(4).
Author information:
(1)Evidence-Based Nursing, School of Nursing, Lanzhou University, Lanzhou,
China.
(2)Department of Neurology, Second Hospital of Lanzhou University, Lanzhou,
China.
(3)Department of Nursing, Gansu Provincial Hospital, Lanzhou, China.
(4)First School of Clinical Medicine, Lanzhou University, Lanzhou, China.
BACKGROUND AND PURPOSE: Alzheimer's disease (AD) is a devastating
neurodegenerative disorder with no cure, and available treatments are only able
to postpone the progression of the disease. Mild cognitive impairment (MCI) is
considered to be a transitional stage preceding AD. Therefore, prediction models
for conversion from MCI to AD are desperately required. These will allow early
treatment of patients with MCI before they develop AD. This study performed a
systematic review and meta-analysis to summarize the reported risk prediction
models and identify the most prevalent factors for conversion from MCI to AD.
METHODS: We systematically reviewed the studies from the databases of PubMed,
CINAHL Plus, Web of Science, Embase, and Cochrane Library, which were searched
through September 2021. Two reviewers independently identified eligible articles
and extracted the data. We used the Critical Appraisal and Data Extraction for
Systematic Reviews of Prediction Modeling Studies (CHARMS) checklist for the
risk of bias assessment.
RESULTS: In total, 18 articles describing the prediction models for conversion
from MCI to AD were identified. The dementia conversion rate of elderly patients
with MCI ranged from 14.49 to 87%. Models in 12 studies were developed using the
data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). C-index/area
under the receiver operating characteristic curve (AUC) of development models
were 0.67-0.98, and the validation models were 0.62-0.96. MRI, apolipoprotein E
genotype 4 (APOE4), older age, Mini-Mental State Examination (MMSE) score, and
Alzheimer's Disease Assessment Scale cognitive (ADAS-cog) score were the most
common and strongest predictors included in the models.
CONCLUSION: In this systematic review, many prediction models have been
developed and have good predictive performance, but the lack of external
validation of models limited the extensive application in the general
population. In clinical practice, it is recommended that medical professionals
adopt a comprehensive forecasting method rather than a single predictive factor
to screen patients with a high risk of MCI. Future research should pay attention
to the improvement, calibration, and validation of existing models while
considering new variables, new methods, and differences in risk profiles across
populations.
Copyright © 2022 Chen, Qian, Zhang, Su, Huang, Wang, Chen, Zhao, Han and Ma.
DOI: 10.3389/fnagi.2022.840386
PMCID: PMC9049273
PMID: 35493941
Conflict of interest statement: The authors declare that the research was
conducted in the absence of any commercial or financial relationships that could
be construed as a potential conflict of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/35633976 | 1. Front Pediatr. 2022 May 12;10:856159. doi: 10.3389/fped.2022.856159.
eCollection 2022.
Prediction Models for Bronchopulmonary Dysplasia in Preterm Infants: A
Systematic Review.
Peng HB(1), Zhan YL(1), Chen Y(1), Jin ZC(1), Liu F(1), Wang B(2), Yu ZB(3)(4).
Author information:
(1)Department of Neonatology, Affiliated Shenzhen Baoan Women's and Children's
Hospital, Jinan University, Shenzhen, China.
(2)Department of Pediatrics, The Affiliated Suqian First People's Hospital of
Nanjing Medical University, Suqian, China.
(3)Department of Neonatology, Shenzhen People's Hospital, The Second Clinical
Medical College, Jinan University, Shenzhen, China.
(4)The First Affiliated Hospital, Southern University of Science and Technology,
Shenzhen, China.
OBJECTIVE: To provide an overview and critical appraisal of prediction models
for bronchopulmonary dysplasia (BPD) in preterm infants.
METHODS: We searched PubMed, Embase, and the Cochrane Library to identify
relevant studies (up to November 2021). We included studies that reported
prediction model development and/or validation of BPD in preterm infants born at
≤32 weeks and/or ≤1,500 g birth weight. We extracted the data independently
based on the CHecklist for critical Appraisal and data extraction for systematic
Reviews of prediction Modelling Studies (CHARMS). We assessed risk of bias and
applicability independently using the Prediction model Risk Of Bias ASsessment
Tool (PROBAST).
RESULTS: Twenty-one prediction models from 13 studies reporting on model
development and 21 models from 10 studies reporting on external validation were
included. Oxygen dependency at 36 weeks' postmenstrual age was the most
frequently reported outcome in both development studies (71%) and validation
studies (81%). The most frequently used predictors in the models were birth
weight (67%), gestational age (62%), and sex (52%). Nearly all included studies
had high risk of bias, most often due to inadequate analysis. Small sample sizes
and insufficient event patients were common in both study types. Missing data
were often not reported or were discarded. Most studies reported on the models'
discrimination, while calibration was seldom assessed (development, 19%;
validation, 10%). Internal validation was lacking in 69% of development studies.
CONCLUSION: The included studies had many methodological shortcomings. Future
work should focus on following the recommended approaches for developing and
validating BPD prediction models.
Copyright © 2022 Peng, Zhan, Chen, Jin, Liu, Wang and Yu.
DOI: 10.3389/fped.2022.856159
PMCID: PMC9133667
PMID: 35633976
Conflict of interest statement: The authors declare that the research was
conducted in the absence of any commercial or financial relationships that could
be construed as a potential conflict of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/31762119 | 1. Head Neck. 2020 Apr;42(4):763-773. doi: 10.1002/hed.26025. Epub 2019 Nov 24.
A critical appraisal of the clinical applicability and risk of bias of the
predictive models for mortality and recurrence in patients with oropharyngeal
cancer: Systematic review.
Palazón-Bru A(1), Mares-García E(1), López-Bru D(2), Mares-Arambul E(3),
Folgado-de la Rosa DM(1), Carbonell-Torregrosa MLÁ(1)(4), Gil-Guillén VF(1).
Author information:
(1)Department of Clinical Medicine, Miguel Hernández University, San Juan de
Alicante, Alicante, Spain.
(2)Department of Otolaryngology, General University Hospital of Elche, Elche,
Alicante, Spain.
(3)Department of Otolaryngology, General University Hospital of Elda, Elda,
Alicante, Spain.
(4)Emergency Service, General University Hospital of Elda, Elda, Alicante,
Spain.
The use of predictive models is becoming widespread. However, these models
should be developed appropriately (CHecklist for critical Appraisal and data
extraction for systematic Reviews of prediction Modeling Studies [CHARMS] and
Prediction model Risk Of Bias ASsessment Tool [PROBAST] statements). Concerning
mortality/recurrence in oropharyngeal cancer, we are not aware of any systematic
reviews of the predictive models. We carried out a systematic review of the
MEDLINE/EMBASE databases of those predictive models. In these models, we
analyzed the 11 domains of the CHARMS statement and the risk of bias and
applicability, using the PROBAST tool. Six papers were finally included in the
systematic review and all of them presented high risk of bias and several
limitations in the statistical analysis. The applicability was satisfactory in
five out of six studies. None of the models could be considered ready for use in
clinical practice.
© 2019 Wiley Periodicals, Inc.
DOI: 10.1002/hed.26025
PMID: 31762119 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/35321760 | 1. Diagn Progn Res. 2022 Mar 24;6(1):4. doi: 10.1186/s41512-022-00119-9.
Multivariable prediction models for health care spending using machine learning:
a protocol of a systematic review.
Huang AW(1), Haslberger M(2), Coulibaly N(3), Galárraga O(3), Oganisian A(4),
Belbasis L(5), Panagiotou OA(3)(6).
Author information:
(1)Department of Health Services, Policy and Practice, Brown University School
of Public Health, Rhode Island, Providence, USA. [email protected].
(2)QUEST Center, Berlin Institute of Health, Charité-Universitätsmedizin Berlin,
Berlin, Germany.
(3)Department of Health Services, Policy and Practice, Brown University School
of Public Health, Rhode Island, Providence, USA.
(4)Department of Biostatistics, Brown University School of Public Health,
Providence, Rhode Island, USA.
(5)Meta-Research Innovation Center Berlin, QUEST Center, Berlin Institute of
Health, Charité-Universitätsmedizin Berlin, Berlin, Germany.
(6)Center for Evidence Synthesis in Health, Brown University School of Public
Health, Providence, Rhode Island, USA.
BACKGROUND: With rising cost pressures on health care systems, machine-learning
(ML)-based algorithms are increasingly used to predict health care costs.
Despite their potential advantages, the successful implementation of these
methods could be undermined by biases introduced in the design, conduct, or
analysis of studies seeking to develop and/or validate ML models. The utility of
such models may also be negatively affected by poor reporting of these studies.
In this systematic review, we aim to evaluate the reporting quality,
methodological characteristics, and risk of bias of ML-based prediction models
for individual-level health care spending.
METHODS: We will systematically search PubMed and Embase to identify studies
developing, updating, or validating ML-based models to predict an individual's
health care spending for any medical condition, over any time period, and in any
setting. We will exclude prediction models of aggregate-level health care
spending, models used to infer causality, models using radiomics or speech
parameters, models of non-clinically validated predictors (e.g., genomics), and
cost-effectiveness analyses without predicting individual-level health care
spending. We will extract data based on the Checklist for Critical Appraisal and
Data Extraction for Systematic Reviews of Prediction Modeling Studies (CHARMS),
previously published research, and relevant recommendations. We will assess the
adherence of ML-based studies to the Transparent Reporting of a multivariable
prediction model for Individual Prognosis Or Diagnosis (TRIPOD) statement and
examine the inclusion of transparency and reproducibility indicators (e.g.
statements on data sharing). To assess the risk of bias, we will apply the
Prediction model Risk Of Bias Assessment Tool (PROBAST). Findings will be
stratified by study design, ML methods used, population characteristics, and
medical field.
DISCUSSION: Our systematic review will appraise the quality, reporting, and risk
of bias of ML-based models for individualized health care cost prediction. This
review will provide an overview of the available models and give insights into
the strengths and limitations of using ML methods for the prediction of health
spending.
© 2022. The Author(s).
DOI: 10.1186/s41512-022-00119-9
PMCID: PMC8943988
PMID: 35321760
Conflict of interest statement: OAP has received personal fees from
International Consulting Associates Inc. unrelated to the submitted work. All
other authors declare that they have no competing interests. |
http://www.ncbi.nlm.nih.gov/pubmed/32453803 | 1. PLoS One. 2020 May 26;15(5):e0233575. doi: 10.1371/journal.pone.0233575.
eCollection 2020.
Systematic review of prediction models in relapsing remitting multiple
sclerosis.
Brown FS(1), Glasmacher SA(1), Kearns PKA(1), MacDougall N(2), Hunt D(1)(3)(4),
Connick P(1)(4), Chandran S(1)(4)(5).
Author information:
(1)Anne Rowling Regenerative Neurology Clinic, University of Edinburgh,
Edinburgh, United Kingdom.
(2)Institute of Neurological Sciences, Glasgow, United Kingdom.
(3)MRC Institute of Genetics and Molecular Medicine, University of Edinburgh,
Edinburgh, United Kingdom.
(4)Centre for Clinical Brain Sciences, University of Edinburgh, Edinburgh,
United Kingdom.
(5)UK Dementia Research Institute, University of Edinburgh, Edinburgh, United
Kingdom.
The natural history of relapsing remitting multiple sclerosis (RRMS) is variable
and prediction of individual prognosis challenging. The inability to reliably
predict prognosis at diagnosis has important implications for informed decision
making especially in relation to disease modifying therapies. We conducted a
systematic review in order to collate, describe and assess the methodological
quality of published prediction models in RRMS. We searched Medline, Embase and
Web of Science. Two reviewers independently screened abstracts and full text for
eligibility and assessed risk of bias. Studies reporting development or
validation of prediction models for RRMS in adults were included. Data
collection was guided by the checklist for critical appraisal and data
extraction for systematic reviews (CHARMS) and applicability and methodological
quality assessment by the prediction model risk of bias assessment tool
(PROBAST). 30 studies were included in the review. Applicability was assessed as
high risk of concern in 27 studies. Risk of bias was assessed as high for all
studies. The single most frequently included predictor was baseline EDSS (n =
11). T2 Lesion volume or number and brain atrophy were each retained in seven
studies. Five studies included external validation and none included impact
analysis. Although a number of prediction models for RRMS have been reported,
most are at high risk of bias and lack external validation and impact analysis,
restricting their application to routine clinical practice.
DOI: 10.1371/journal.pone.0233575
PMCID: PMC7250448
PMID: 32453803 [Indexed for MEDLINE]
Conflict of interest statement: I have read the journal's policy and the authors
of this manuscript have the following competing interests: NM has received
hospitality for educational events from Biogen, Novartis, Genzyme, Merck and
Roche over the past 5 years. NM has received honoraria for talks or advisory
boards from Biogen, Novartis, Genzyme, Merck and Roche over the same time
period. This does not alter our adherence to PLOS ONE policies on sharing data
and materials. |
http://www.ncbi.nlm.nih.gov/pubmed/34687634 | 1. Lancet Neurol. 2021 Nov;20(11):895-906. doi: 10.1016/S1474-4422(21)00263-5.
Safety and efficacy of losartan for the reduction of brain atrophy in clinically
diagnosed Alzheimer's disease (the RADAR trial): a double-blind, randomised,
placebo-controlled, phase 2 trial.
Kehoe PG(1), Turner N(2), Howden B(2), Jarutyte L(3), Clegg SL(4), Malone IB(4),
Barnes J(4), Nielsen C(4), Sudre CH(5), Wilson A(6), Thai NJ(6), Blair PS(2),
Coulthard E(3), Lane JA(2), Passmore P(7), Taylor J(2), Mutsaerts HJ(8), Thomas
DL(4), Fox NC(9), Wilkinson I(10), Ben-Shlomo Y(2); RADAR investigators.
Collaborators: Harkness K, Kuruvilla T, McShane R, Connelly P, Duncan G, Calvert
L, Lawrie A, Sheridan M, Jackson E, Udeze B, Pearson S, Langheinrich T, Wagle S,
Butchart J, Macharouthu A, Donaldson A, Neil W, Pattan V, Findlay D, Thomas A,
Barber R, Byrne A, Dalvi M, Negi R, McGuinness B.
Author information:
(1)Dementia Research Group, University of Bristol, Bristol, UK. Electronic
address: [email protected].
(2)Translational Health Sciences, Population Health Sciences, University of
Bristol, Bristol, UK; Bristol Trials Centre, University of Bristol, Bristol, UK.
(3)Dementia Neurology Research Group, University of Bristol, Bristol, UK.
(4)Dementia Research Centre, University College London, London, UK; UCL Queen
Square Institute of Neurology, University College London, London, UK.
(5)MRC Unit for Lifelong Health and Ageing at UCL, and Centre for Medical Image
Computing, University College London, London, UK; School of Biomedical
Engineering and Imaging Sciences, Kings College London, UK.
(6)Faculty of Health Sciences, Bristol Medical School, Clinical Research Imaging
Centre, University of Bristol, Bristol, UK.
(7)Institute of Clinical Sciences, Queens University Belfast, Royal Victoria
Hospital, Belfast, UK.
(8)Amsterdam University Medical Centers, Amsterdam Neuroscience, Amsterdam,
Netherlands.
(9)Dementia Research Centre, University College London, London, UK; UK Dementia
Research Institute, University College London, London, UK; UCL Queen Square
Institute of Neurology, University College London, London, UK.
(10)Clinical Pharmacology Unit, School of Clinical Medicine, University of
Cambridge, Addenbrookes Hospital, Cambridge, UK.
Comment in
Lancet Neurol. 2021 Nov;20(11):878-879. doi: 10.1016/S1474-4422(21)00340-9.
BACKGROUND: Drugs modifying angiotensin II signalling could reduce Alzheimer's
disease pathology, thus decreasing the rate of disease progression. We
investigated whether the angiotensin II receptor antagonist losartan, compared
with placebo, could reduce brain volume loss, as a measure of disease
progression, in clinically diagnosed mild-to-moderate Alzheimer's disease.
METHODS: In this double-blind, multicentre, randomised controlled trial,
eligible patients aged 55 years or older, previously untreated with angiotensin
II drugs and diagnosed (National Institute of Neurological and Communicative
Disorders and Stroke and the Alzheimer's Disease and Related Disorders
Association criteria) with mild-to-moderate Alzheimer's disease, and who had
capacity to consent, were recruited from 23 UK National Health Service hospital
trusts. After undergoing a 4-week, open-label phase of active treatment then
washout, participants were randomly assigned (1:1) oral over-encapsulated
preparations of either 100 mg losartan (after an initial two-dose titration
stage) or matched placebo daily for 12 months. Randomisation, minimised by age
and baseline medial temporal lobe atrophy score, was undertaken online or via
pin-access service by telephone. Participants, their study companions, and study
personnel were masked to group assignment. The primary outcome, analysed by the
intention-to-treat principle (ie, participants analysed in the group to which
they were randomised, without imputation for missing data), was change in whole
brain volume between baseline and 12 months, measured using volumetric MRI and
determined by boundary shift interval (BSI) analysis. The trial is registered
with the International Standard Randomised Controlled Trial Register
(ISRCTN93682878) and the European Union Drug Regulating Authorities Clinical
Trials Database (EudraCT 2012-003641-15), and is completed.
FINDINGS: Between July 22, 2014, and May 17, 2018, 261 participants entered the
open-label phase. 211 were randomly assigned losartan (n=105) or placebo
(n=106). Of 197 (93%) participants who completed the study, 171 (81%) had
complete primary outcome data. The mean brain volume (BSI) reduction was 19·1 mL
(SD 10·3) in the losartan group and 20·0 mL (10·8) in the placebo group. The
difference in total volume reduction between groups was -2·29 mL (95% CI -6·46
to 0·89; p=0·14). The number of adverse events was low (22 in the losartan group
and 20 in the placebo group) with no differences between treatment groups. There
was one treatment-related death per treatment group.
INTERPRETATION: 12 months of treatment with losartan was well tolerated but was
not effective in reducing the rate of brain atrophy in individuals with
clinically diagnosed mild-to-moderate Alzheimer's disease. Further research is
needed to assess the potential therapeutic benefit from earlier treatment in
patients with milder cognitive impairment or from longer treatment periods.
FUNDING: Efficacy and Mechanism Evaluation Programme (UK Medical Research
Council and National Institute for Health Research).
Copyright © 2021 The Author(s). Published by Elsevier Ltd. This is an Open
Access article under the CC BY 4.0 license. Published by Elsevier Ltd.. All
rights reserved.
DOI: 10.1016/S1474-4422(21)00263-5
PMCID: PMC8528717
PMID: 34687634 [Indexed for MEDLINE]
Conflict of interest statement: Declaration of interests PGK reports grants from
the National Institute of Health Research (NIHR), during the conduct of the
study; being a non-funded co-investigator of the ongoing Alzheimer's
Association-funded HEART phase 1b study of telmisartan and its use as an
intervention against the renin–angiotensin system in African American people at
risk of developing dementia by parental history; and having previously
undertaken advisory work for Novartis in their development and intended
trialling of dual acting inhibitors of Angiotensin Receptor Blockers and
neprilysin (LCZ696) for the treatment of heart failure. NT, BH, LJ, SLB, IBM,
CN, NJT, PSB, JAL, PP, JT, DLT, IW, and YBS report grants from the NIHR, during
the conduct of the study. JB reports grants from Alzheimer's Research UK,
outside the submitted work. CHS reports grants for the Alzheimer's Society,
during the conduct of the study. EC reports grants from the NIHR, during the
conduct of the study; and received payment from Biogen, Novartis, and Union
Chimique Belge for providing educational resources or consultancy around
Alzheimer's Disease trials. NCF reports other funding from Roche, personal fees
from Biogen, and non-financial support from Lilly and Ionis, outside the
submitted work. AW and H-JM declare no competing interests. |
http://www.ncbi.nlm.nih.gov/pubmed/35196337 | 1. PLoS One. 2022 Feb 23;17(2):e0263215. doi: 10.1371/journal.pone.0263215.
eCollection 2022.
Value of D-dimer in predicting various clinical outcomes following
community-acquired pneumonia: A network meta-analysis.
Li J(1), Zhou K(1), Duan H(1), Yue P(1), Zheng X(1), Liu L(1), Liao H(1), Wu
J(1), Li J(1), Hua Y(1), Li Y(1).
Author information:
(1)Department of Pediatrics, Key Laboratory of Birth Defects and Related
Diseases of Women and Children of MOE, West China Second University Hospital,
Sichuan University, Chengdu, Sichuan, China.
BACKGROUND: Whether high D-dimer level before treatment has any impact on poor
outcomes in patients with community-associated pneumonia (CAP) remains unclear.
Therefore, we conducted the first meta-analysis focusing specifically on
prognostic value of high D-dimer level before treatment in CAP patients.
METHODS: Pubmed, Embase, the Cochrane Central Register of Controlled Trials and
World Health Organization clinical trials registry center were searched up to
the end of March 2021. Randomized clinical trials (RCT) and observational
studies were included to demonstrate the association between the level of
D-dimer and clinical outcomes. Data were extracted using an adaptation of the
Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of
Prediction Modeling Studies (CHARMS-PF). When feasible, meta-analysis using
random-effects models was performed. Risk of bias and level of evidence were
assessed with the Quality in Prognosis Studies tool and an adaptation of Grading
of Recommendations Assessment, Development, and Evaluation. Data were analyzed
using STATA 14.0 to complete meta and network analysis.
MAIN OUTCOMES AND MEASURES: Besides d-dimer levels in CAP patients with poor
outcomes, we also analyzed proportion of patients with or without poor outcomes
correctly classified by the d-dimer levels as being at high or low risk. The
poor outcome includes severe CAP, death, pulmonary embolism (PE) and invasive
mechanical ventilators.
RESULTS: 32 studies with a total of 9,593 patients were eventually included.
Pooled effect size (ES) suggested that d-dimer level was significantly higher in
severe CAP patients than non-severe CAP patients with great heterogeneity (SMD =
1.21 95%CI 0.87-1.56, I2 = 86.8% p = 0.000). D-dimer level was significantly
elevated in non-survivors compared to survivors with CAP (SMD = 1.22 95%CI
0.67-1.77, I2 = 85.1% p = 0.000). Prognostic value of d-dimer for pulmonary
embolism (PE) was proved by hierarchical summary receiver operating
characteristic curve (HSROC) with good summary sensitivity (0.74, 95%CI,
0.50-0.89) and summary specificity (0.82, 95%CI, 0.41-0.97). Network
meta-analysis suggested that there was a significant elevation of d-dimer levels
in CAP patients with poor outcome than general CAP patients but d-dimer levels
weren't significantly different among poor outcomes.
CONCLUSION: The prognostic ability of d-dimer among patients with CAP appeared
to be good at correctly identifying high-risk populations of poor outcomes,
suggesting potential for clinical utility in patients with CAP.
DOI: 10.1371/journal.pone.0263215
PMCID: PMC8865637
PMID: 35196337 [Indexed for MEDLINE]
Conflict of interest statement: The authors have declared that no competing
interests exist. |
http://www.ncbi.nlm.nih.gov/pubmed/32583899 | 1. Stat Med. 2020 Oct 15;39(23):3207-3225. doi: 10.1002/sim.8660. Epub 2020 Jun
25.
A general presentation on how to carry out a CHARMS analysis for prognostic
multivariate models.
Palazón-Bru A(1), Martín-Pérez F(1), Mares-García E(1), Beneyto-Ripoll C(2),
Gil-Guillén VF(1), Pérez-Sempere Á(1), Carbonell-Torregrosa MÁ(1)(2).
Author information:
(1)Department of Clinical Medicine, Miguel Hernández University, Alicante,
Spain.
(2)Emergency Service, General University Hospital of Elda, Alicante, Spain.
The CHARMS (critical appraisal and data extraction for systematic reviews of
prediction modelling studies) checklist was created to provide methodological
appraisals of predictive models, based on the best available scientific evidence
and through systematic reviews. Our purpose is to give a general presentation on
how to carry out a CHARMS analysis for prognostic multivariate models, making
clear what the steps are and how they are applied individually to the studies
included in the systematic review. This tutorial is aimed at providing such a
resource. In addition to this explanation, we will apply the method to a real
case: predictive models of atrial fibrillation in the community. This
methodology could be applied to other predictive models using the steps provided
in our review so as to have complete information for each included model and
determine whether it can be implemented in daily clinical practice.
© 2020 John Wiley & Sons, Ltd.
DOI: 10.1002/sim.8660
PMID: 32583899 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/31441567 | 1. Eur J Cancer Care (Engl). 2019 Nov;28(6):e13157. doi: 10.1111/ecc.13157. Epub
2019 Aug 23.
A systematic review of predictive models for recurrence and mortality in
patients with tongue cancer.
Palazón-Bru A(1), Mares-García E(1), López-Bru D(2), Mares-Arambul E(3),
Gil-Guillén VF(1), Carbonell-Torregrosa MÁ(1)(4).
Author information:
(1)Department of Clinical Medicine, Miguel Hernández University, San Juan de
Alicante, Spain.
(2)Department of Otolaryngology, General University Hospital of Elche, Elche,
Spain.
(3)Department of Otolaryngology, General University Hospital of Elda, Elda,
Spain.
(4)Emergency Service, General University Hospital of Elda, Elda, Spain.
Comment in
Eur J Cancer Care (Engl). 2020 Mar;29(2):e13211. doi: 10.1111/ecc.13211.
INTRODUCTION: Predictive models must meet clinical/methodological standards to
be used in clinical practice. However, no critique of those models relating to
mortality/recurrence in tongue cancer has been done bearing in mind the accepted
standards.
METHODS: We conducted a systematic review evaluating the methodology and
clinical applicability of predictive models for mortality/recurrence in tongue
cancer published in MEDLINE and Scopus. For each model, we analysed (domains of
CHARMS, Critical Appraisal and Data Extraction for Systematic Reviews of
Prediction Modelling Studies) the following: source of data, participants,
outcome to be predicted, candidate predictors, sample size, missing data, model
development, model performance, model evaluation, results and interpretation and
discussion.
RESULTS: We found two papers that included eight prediction models, neither of
which adhered to the CHARMS recommendations.
CONCLUSION: Given the quality of tongue cancer models, new studies following
current consensus are needed to develop predictive tools applicable in clinical
practice.
© 2019 John Wiley & Sons Ltd.
DOI: 10.1111/ecc.13157
PMID: 31441567 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/33407048 | 1. Lupus. 2021 Mar;30(3):421-430. doi: 10.1177/0961203320983462. Epub 2021 Jan 6.
Predictive models of infection in patients with systemic lupus erythematosus: A
systematic literature review.
Restrepo-Escobar M(1), Granda-Carvajal PA(2), Aguirre DC(3), Hernández-Zapata
J(4), Vásquez GM(1), Jaimes F(1).
Author information:
(1)Department of Internal Medicine, Universidad de Antioquia, Medellín,
Colombia.
(2)Department of Internal Medicine and Subspecialties, Hospital Pablo Tobón
Uribe, Medellín, Colombia.
(3)Medical Research Institute, Universidad de Antioquia, Medellín, Colombia.
(4)Department of Pediatrics, Universidad de Antioquia, Medellín, Colombia.
INTRODUCTION: Having reliable predictive models of prognosis/the risk of
infection in systemic lupus erythematosus (SLE) patients would allow this
problem to be addressed on an individual basis to study and implement possible
preventive or therapeutic interventions.
OBJECTIVE: To identify and analyze all predictive models of prognosis/the risk
of infection in patients with SLE that exist in medical literature.
METHODS: A structured search in PubMed, Embase, and LILACS databases was carried
out until May 9, 2020. In addition, a search for abstracts in the American
Congress of Rheumatology (ACR) and European League Against Rheumatism (EULAR)
annual meetings' archives published over the past eight years was also
conducted. Studies on developing, validating or updating predictive prognostic
models carried out in patients with SLE, in which the outcome to be predicted is
some type of infection, that were generated in any clinical context and with any
time horizon were included. There were no restrictions on language, date, or
status of the publication. To carry out the systematic review, the CHARMS
(Critical Appraisal and Data Extraction for Systematic Reviews of Prediction
Modelling Studies) guideline recommendations were followed. The PROBAST tool (A
Tool to Assess the Risk of Bias and Applicability of Prediction Model Studies)
was used to assess the risk of bias and the applicability of each model.
RESULTS: We identified four models of infection prognosis in patients with SLE.
Mostly, there were very few events per candidate predictor. In addition, to
construct the models, an initial selection was made based on univariate analyses
with no contraction of the estimated coefficients being carried out. This
suggests that the proposed models have a high probability of overfitting and
being optimistic.
CONCLUSIONS: To date, very few prognostic models have been published on the
infection of SLE patients. These models are very heterogeneous and are rated as
having a high risk of bias and methodological weaknesses. Despite the widespread
recognition of the frequency and severity of infections in SLE patients, there
is no reliable predictive prognostic model that facilitates the study and
implementation of personalized preventive or therapeutic measures.Protocol
registration number: PROSPERO CRD42020171638.
DOI: 10.1177/0961203320983462
PMID: 33407048 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/31439598 | 1. BMJ Open. 2019 Aug 21;9(8):e027192. doi: 10.1136/bmjopen-2018-027192.
A systematic review of methodological quality of model development studies
predicting prognostic outcome for resectable pancreatic cancer.
Bradley A(1)(2), Van Der Meer R(3), McKay CJ(2).
Author information:
(1)Management Science, University of Strathclyde Business School, Glasgow, UK
[email protected].
(2)West of Scotland Pancreatic Unit, Glasgow Royal Infirmary, Glasgow, UK.
(3)Management Science, University of Strathclyde Business School, Glasgow, UK.
OBJECTIVES: To assess the methodological quality of prognostic model development
studies pertaining to post resection prognosis of pancreatic ductal
adenocarcinoma (PDAC).
DESIGN/SETTING: A narrative systematic review of international peer reviewed
journals DATA SOURCE: Searches were conducted of: MEDLINE, Embase, PubMed,
Cochrane database and Google Scholar for predictive modelling studies applied to
the outcome of prognosis for patients with PDAC post resection. Predictive
modelling studies in this context included prediction model development studies
with and without external validation and external validation studies with model
updating. Data was extracted following the Checklist for critical Appraisal and
data extraction for systematic Reviews of prediction Modelling Studies (CHARMS)
checklist.
PRIMARY AND SECONDARY OUTCOME MEASURES: Primary outcomes were all components of
the CHARMS checklist. Secondary outcomes included frequency of variables
included across predictive models.
RESULTS: 263 studies underwent full text review. 15 studies met the inclusion
criteria. 3 studies underwent external validation. Multivariable Cox
proportional hazard regression was the most commonly employed modelling method
(n=13). 10 studies were based on single centre databases. Five used prospective
databases, seven used retrospective databases and three used cancer data
registry. The mean number of candidate predictors was 19.47 (range 7 to 50). The
most commonly included variables were tumour grade (n=9), age (n=8), tumour
stage (n=7) and tumour size (n=5). Mean sample size was 1367 (range 50 to 6400).
5 studies reached statistical power. None of the studies reported blinding of
outcome measurement for predictor values. The most common form of presentation
was nomograms (n=5) and prognostic scores (n=5) followed by prognostic
calculators (n=3) and prognostic index (n=2).
CONCLUSIONS: Areas for improvement in future predictive model development have
been highlighted relating to: general aspects of model development and
reporting, applicability of models and sources of bias.
TRIAL REGISTRATION NUMBER: CRD42018105942.
© Author(s) (or their employer(s)) 2019. Re-use permitted under CC BY-NC. No
commercial re-use. See rights and permissions. Published by BMJ.
DOI: 10.1136/bmjopen-2018-027192
PMCID: PMC6707674
PMID: 31439598 [Indexed for MEDLINE]
Conflict of interest statement: Competing interests: None declared. |
http://www.ncbi.nlm.nih.gov/pubmed/31090660 | 1. Pancreas. 2019 May/Jun;48(5):598-604. doi: 10.1097/MPA.0000000000001312.
Personalized Pancreatic Cancer Management: A Systematic Review of How Machine
Learning Is Supporting Decision-making.
Bradley A, van der Meer R, McKay C(1).
Author information:
(1)West of Scotland Pancreatic Cancer Unit, Glasgow Royal Infirmary, Glasgow,
Scotland.
This review critically analyzes how machine learning is being used to support
clinical decision-making in the management of potentially resectable pancreatic
cancer. Following PRISMA (Preferred Reporting Items for Systematic Reviews and
Meta-analyses) guidelines, electronic searches of MEDLINE, Embase, PubMed, and
Cochrane Database were undertaken. Studies were assessed using the checklist for
critical appraisal and data extraction for systematic reviews of prediction
modeling studies (CHARMS) checklist. In total 89,959 citations were retrieved.
Six studies met the inclusion criteria. Three studies were Markov
decision-analysis models comparing neoadjuvant therapy versus upfront surgery.
Three studies predicted survival time using Bayesian modeling (n = 1) and
artificial neural network (n = 1), and one study explored machine learning
algorithms including Bayesian network, decision trees, k-nearest neighbor, and
artificial neural networks. The main methodological issues identified were
limited data sources, which limits generalizability and potentiates bias; lack
of external validation; and the need for transparency in methods of internal
validation, consecutive sampling, and selection of candidate predictors. The
future direction of research relies on expanding our view of the
multidisciplinary team to include professionals from computing and data science
with algorithms developed in conjunction with clinicians and viewed as aids, not
replacement, to traditional clinical decision-making.
DOI: 10.1097/MPA.0000000000001312
PMID: 31090660 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/35934199 | 1. Clin Microbiol Infect. 2023 Apr;29(4):434-440. doi: 10.1016/j.cmi.2022.07.019.
Epub 2022 Aug 4.
How to conduct a systematic review and meta-analysis of prognostic model
studies.
Damen JAA(1), Moons KGM(2), van Smeden M(3), Hooft L(2).
Author information:
(1)Cochrane Netherlands, University Medical Center Utrecht, Utrecht University,
the Netherlands; Julius Center for Health Sciences and Primary Care, University
Medical Center Utrecht, Utrecht University, the Netherlands. Electronic address:
[email protected].
(2)Cochrane Netherlands, University Medical Center Utrecht, Utrecht University,
the Netherlands; Julius Center for Health Sciences and Primary Care, University
Medical Center Utrecht, Utrecht University, the Netherlands.
(3)Julius Center for Health Sciences and Primary Care, University Medical Center
Utrecht, Utrecht University, the Netherlands.
BACKGROUND: Prognostic models are typically developed to estimate the risk that
an individual in a particular health state will develop a particular health
outcome, to support (shared) decision making. Systematic reviews of prognostic
model studies can help identify prognostic models that need to further be
validated or are ready to be implemented in healthcare.
OBJECTIVES: To provide a step-by-step guidance on how to conduct and read a
systematic review of prognostic model studies and to provide an overview of
methodology and guidance available for every step of the review progress.
SOURCES: Published, peer-reviewed guidance articles.
CONTENT: We describe the following steps for conducting a systematic review of
prognosis studies: 1) Developing the review question using the Population, Index
model, Comparator model, Outcome(s), Timing, Setting format, 2) Searching and
selection of articles, 3) Data extraction using the Critical Appraisal and Data
Extraction for Systematic Reviews of Prediction Modelling Studies (CHARMS)
checklist, 4) Quality and risk of bias assessment using the Prediction model
Risk Of Bias ASsessment (PROBAST) tool, 5) Analysing data and undertaking
quantitative meta-analysis, and 6) Presenting summary of findings, interpreting
results, and drawing conclusions. Guidance for each step is described and
illustrated using a case study on prognostic models for patients with COVID-19.
IMPLICATIONS: Guidance for conducting a systematic review of prognosis studies
is available, but the implications of these reviews for clinical practice and
further research highly depend on complete reporting of primary studies.
Copyright © 2022 The Author(s). Published by Elsevier Ltd.. All rights reserved.
DOI: 10.1016/j.cmi.2022.07.019
PMCID: PMC9351211
PMID: 35934199 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/32978532 | 1. Evid Based Dent. 2020 Sep;21(3):84-86. doi: 10.1038/s41432-020-0115-5.
Can medical practitioners rely on prediction models for COVID-19? A systematic
review.
Shamsoddin E(1).
Author information:
(1)National Institute for Medical Research Development, Tehran, Iran.
Aim This systematic review sought to assess and scrutinise the validity and
practicality of published and preprint reports of prediction models for the
diagnosis of coronavirus disease 2019 (COVID-19) in patients with suspected
infection, for prognosis of patients with COVID-19, and for identifying
individuals in the general population at increased risk of infection with
COVID-19 or being hospitalised with the illness.Data sources A systematic,
online search was conducted in PubMed and Embase. In order to do so, the authors
used Ovid as the host platform for these two databases and also investigated
bioRxiv, medRxiv and arXiv as repositories for the preprints of studies. A
public living systematic review list of COVID-19-related studies was used as the
baseline searching platform (Institute of Social and Preventive Medicine's
repository for living evidence on COVID-19).Study selection Studies which
developed or validated a multivariable prediction model related to COVID-19
patients' data (individual level data) were included. The authors did not put
any restrictions on the models included in their study regarding the model
setting, prediction horizon or outcomes.Data extraction and synthesis Checklists
of critical appraisal and data extraction for systematic reviews of prediction
modelling studies (CHARMS) and prediction model risk of bias assessment tool
(PROBAST) were used to guide developing of a standardised data extraction form.
Each model's predictive performance was extracted by using any summaries of
discrimination and calibration. All these steps were done according to the
aspects of the transparent reporting of a multivariable prediction model for
individual prognosis or diagnosis (TRIPOD) and preferred reporting items for
systematic reviews and meta-analyses (PRISMA).Results One hundred and forty-five
prediction models (107 studies) were selected for data extraction and critical
appraisal. The most common predictors of diagnosis and prognosis of COVID-19
were age, body temperature, lymphocyte count and lung imaging characteristics.
Influenza-like symptoms and neutrophil count were regularly predictive in
diagnostic models, while comorbidities, sex, C-reactive protein and creatinine
were common prognostic items. C-indices (a measure of discrimination for models)
ranged from 0.73 to 0.81 in prediction models for the general population, from
0.65 to more than 0.99 in diagnostic models, and from 0.68 to 0.99 in the
prognostic models. All the included studies were reported to have high risks of
bias.Conclusions Overall, this study did not recommend applying any of the
predictive models in clinical practice yet. High risk of bias, reporting
problems and (probably) optimistic reported performances are all among the
reasons for the previous conclusion. Prompt actions regarding accurate data
sharing and international collaborations are required to achieve more rigorous
prediction models for COVID-19.
DOI: 10.1038/s41432-020-0115-5
PMCID: PMC7517064
PMID: 32978532 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/33492724 | 1. Int J Clin Pract. 2021 Aug;75(8):e14044. doi: 10.1111/ijcp.14044. Epub 2021
Feb 4.
A critical appraisal of the prognostic predictive models for patients with
sepsis: Which model can be applied in clinical practice?
Beneyto-Ripoll C(1), Palazón-Bru A(2), Llópez-Espinós P(3), Martínez-Díaz AM(4),
Gil-Guillén VF(2), de Los Ángeles Carbonell-Torregrosa M(2)(3).
Author information:
(1)Emergency Services, General Hospital of Almansa, Almansa, Albacete, Spain.
(2)Department of Clinical Medicine, Miguel Hernández University, San Juan de
Alicante, Alicante, Spain.
(3)Emergency Services, General University Hospital of Elda, Elda, Alicante,
Spain.
(4)Emergency Services, University Hospital of Puerta del Mar, Cádiz, Cádiz,
Spain.
BACKGROUND: Sepsis is associated with high mortality and predictive models can
help in clinical decision-making. The objective of this study was to carry out a
systematic review of these models.
METHODS: In 2019, we conducted a systematic review in MEDLINE and EMBASE
(CDR42018111121:PROSPERO) of articles that developed predictive models for
mortality in septic patients (inclusion criteria). We followed the CHARMS
recommendations (Critical Appraisal and Data Extraction for Systematic Reviews
of Prediction Modelling Studies), extracting the information from its 11 domains
(Source of data, Participants, etc). We determined the risk of bias and
applicability (participants, outcome, predictors and analysis) through PROBAST
(Prediction model Risk Of Bias ASsessment Tool).
RESULTS: A total of 14 studies were included. In the CHARMS extraction, the
models found showed great variability in its 11 domains. Regarding the PROBAST
checklist, only one article had an unclear risk of bias as it did not indicate
how missing data were handled while the others all had a high risk of bias. This
was mainly due to the statistical analysis (inadequate sample size, handling of
continuous predictors, missing data and selection of predictors), since 13
studies had a high risk of bias. Applicability was satisfactory in six articles.
Most of the models integrate predictors from routine clinical practice.
Discrimination and calibration were assessed for almost all the models, with the
area under the ROC curve ranging from 0.59 to 0.955 and no lack of calibration.
Only three models were externally validated and their maximum discrimination
values in the derivation were from 0.712 and 0.84. One of them (Osborn) had
undergone multiple validation studies.
DISCUSSION: Despite most of the studies showing a high risk of bias, we very
cautiously recommend applying the Osborn model, as this has been externally
validated various times.
© 2021 John Wiley & Sons Ltd.
DOI: 10.1111/ijcp.14044
PMID: 33492724 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/35395724 | 1. BMC Med Res Methodol. 2022 Apr 8;22(1):101. doi: 10.1186/s12874-022-01577-x.
Methodological conduct of prognostic prediction models developed using machine
learning in oncology: a systematic review.
Dhiman P(1)(2), Ma J(3), Andaur Navarro CL(4)(5), Speich B(3)(6), Bullock G(7),
Damen JAA(4)(5), Hooft L(4)(5), Kirtley S(3), Riley RD(8), Van Calster
B(9)(10)(11), Moons KGM(4)(5), Collins GS(3)(12).
Author information:
(1)Centre for Statistics in Medicine, Nuffield Department of Orthopaedics,
Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, OX3
7LD, UK. [email protected].
(2)NIHR Oxford Biomedical Research Centre, Oxford University Hospitals NHS
Foundation Trust, Oxford, UK. [email protected].
(3)Centre for Statistics in Medicine, Nuffield Department of Orthopaedics,
Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford, OX3
7LD, UK.
(4)Julius Center for Health Sciences and Primary Care, University Medical Center
Utrecht, Utrecht University, Utrecht, The Netherlands.
(5)Cochrane Netherlands, University Medical Center Utrecht, Utrecht University,
Utrecht, The Netherlands.
(6)Basel Institute for Clinical Epidemiology and Biostatistics, Department of
Clinical Research, University Hospital Basel, University of Basel, Basel,
Switzerland.
(7)Nuffield Department of Orthopaedics, Rheumatology, and Musculoskeletal
Sciences, University of Oxford, Oxford, UK.
(8)Centre for Prognosis Research, School of Medicine, Keele University,
Staffordshire, ST5 5BG, UK.
(9)Department of Development and Regeneration, KU Leuven, Leuven, Belgium.
(10)Department of Biomedical Data Sciences, Leiden University Medical Center,
Leiden, the Netherlands.
(11)EPI-centre, KU Leuven, Leuven, Belgium.
(12)NIHR Oxford Biomedical Research Centre, Oxford University Hospitals NHS
Foundation Trust, Oxford, UK.
BACKGROUND: Describe and evaluate the methodological conduct of prognostic
prediction models developed using machine learning methods in oncology.
METHODS: We conducted a systematic review in MEDLINE and Embase between
01/01/2019 and 05/09/2019, for studies developing a prognostic prediction model
using machine learning methods in oncology. We used the Transparent Reporting of
a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD)
statement, Prediction model Risk Of Bias ASsessment Tool (PROBAST) and CHecklist
for critical Appraisal and data extraction for systematic Reviews of prediction
Modelling Studies (CHARMS) to assess the methodological conduct of included
publications. Results were summarised by modelling type: regression-,
non-regression-based and ensemble machine learning models.
RESULTS: Sixty-two publications met inclusion criteria developing 152 models
across all publications. Forty-two models were regression-based, 71 were
non-regression-based and 39 were ensemble models. A median of 647 individuals
(IQR: 203 to 4059) and 195 events (IQR: 38 to 1269) were used for model
development, and 553 individuals (IQR: 69 to 3069) and 50 events (IQR: 17.5 to
326.5) for model validation. A higher number of events per predictor was used
for developing regression-based models (median: 8, IQR: 7.1 to 23.5), compared
to alternative machine learning (median: 3.4, IQR: 1.1 to 19.1) and ensemble
models (median: 1.7, IQR: 1.1 to 6). Sample size was rarely justified (n = 5/62;
8%). Some or all continuous predictors were categorised before modelling in 24
studies (39%). 46% (n = 24/62) of models reporting predictor selection before
modelling used univariable analyses, and common method across all modelling
types. Ten out of 24 models for time-to-event outcomes accounted for censoring
(42%). A split sample approach was the most popular method for internal
validation (n = 25/62, 40%). Calibration was reported in 11 studies. Less than
half of models were reported or made available.
CONCLUSIONS: The methodological conduct of machine learning based clinical
prediction models is poor. Guidance is urgently needed, with increased awareness
and education of minimum prediction modelling standards. Particular focus is
needed on sample size estimation, development and validation analysis methods,
and ensuring the model is available for independent validation, to improve
quality of machine learning based clinical prediction models.
© 2022. The Author(s).
DOI: 10.1186/s12874-022-01577-x
PMCID: PMC8991704
PMID: 35395724 [Indexed for MEDLINE]
Conflict of interest statement: The authors declare no conflict of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/34413828 | 1. Front Neurol. 2021 Aug 3;12:718438. doi: 10.3389/fneur.2021.718438.
eCollection 2021.
Emerging Role of Carotid MRI for Personalized Ischemic Stroke Risk Prediction in
Patients With Carotid Artery Stenosis.
Nies KPH(1)(2), Smits LJM(3), Kassem M(1)(2), Nederkoorn PJ(4), van Oostenbrugge
RJ(1)(5), Kooi ME(1)(2).
Author information:
(1)Department of Radiology, CARIM School for Cardiovascular Diseases, Maastricht
University, Maastricht, Netherlands.
(2)Department of Radiology and Nuclear Medicine, Maastricht University Medical
Center, Maastricht, Netherlands.
(3)Department of Epidemiology, Maastricht University, Maastricht, Netherlands.
(4)Department of Neurology, Amsterdam University Medical Center, Amsterdam,
Netherlands.
(5)Department of Neurology, Maastricht University Medical Center, Maastricht,
Netherlands.
Rupture of a vulnerable carotid plaque is an important cause of ischemic stroke.
Prediction models can support medical decision-making by estimating individual
probabilities of future events, while magnetic resonance imaging (MRI) can
provide detailed information on plaque vulnerability. In this review, prediction
models for medium to long-term (>90 days) prediction of recurrent ischemic
stroke among patients on best medical treatment for carotid stenosis are
evaluated, and the emerging role of MRI of the carotid plaque for personalized
ischemic stroke prediction is discussed. A systematic search identified two
models; the European Carotid Surgery Trial (ECST) medical model, and the
Symptomatic Carotid Atheroma Inflammation Lumen stenosis (SCAIL) score. We
critically appraised these models by means of criteria derived from the CHARMS
(CHecklist for critical Appraisal and data extraction for systematic Reviews of
prediction Modeling Studies) and PROBAST (Prediction model Risk Of Bias
ASsessment Tool). We found both models to be at high risk of bias. The ECST
model, the most widely used model, was derived from data of large but relatively
old trials (1980s and 1990s), not reflecting lower risks of ischemic stroke
resulting from improvements in drug treatment (e.g., statins and anti-platelet
therapy). The SCAIL model, based on the degree of stenosis and positron emission
tomography/computed tomography (PET/CT)-based plaque inflammation, was derived
and externally validated in limited samples. Clinical implementation of the
SCAIL model can be challenging due to high costs and low accessibility of
PET/CT. MRI is a more readily available, lower-cost modality that has been
extensively validated to visualize all the hallmarks of plaque vulnerability.
The MRI methods to identify the different plaque features are described.
Intraplaque hemorrhage (IPH), a lipid-rich necrotic core (LRNC), and a thin or
ruptured fibrous cap (TRFC) on MRI have shown to strongly predict stroke in
meta-analyses. To improve personalized risk prediction, carotid plaque features
should be included in prediction models. Prediction of stroke in patients with
carotid stenosis needs modernization, and carotid MRI has potential in providing
strong predictors for that goal.
Copyright © 2021 Nies, Smits, Kassem, Nederkoorn, Oostenbrugge and Kooi.
DOI: 10.3389/fneur.2021.718438
PMCID: PMC8370465
PMID: 34413828
Conflict of interest statement: The authors declare that the research was
conducted in the absence of any commercial or financial relationships that could
be construed as a potential conflict of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/28284255 | 1. Int J Tuberc Lung Dis. 2017 Apr 1;21(4):405-411. doi: 10.5588/ijtld.16.0059.
A systematic review of prediction models for prevalent pulmonary tuberculosis in
adults.
Van Wyk SS(1), Lin HH(2), Claassens MM(1).
Author information:
(1)Department of Paediatrics and Child Health, Desmond Tutu Tuberculosis Centre,
Stellenbosch University, Cape Town, South Africa.
(2)Institute of Epidemiology and Preventive Medicine, National Taiwan
University, Taiwan.
A systematic review was conducted to describe the quality and characteristics of
prediction models for prevalent pulmonary tuberculosis (PTB) in adults at
routine TB care settings. A prediction model was defined as the combination of
two or more clinical predictors designed to estimate the probability of having
TB. Studies using culture-confirmed PTB as reference standard were included.
Models for in-patients, children or specific patient populations were excluded.
PubMed, Scopus and the Cochrane Library and abstracts from the International
Union Against Tuberculosis and Lung Disease, American Thoracic Society and
European Respiratory Society conferences were searched. The CHARMS (Critical
Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling
Studies) checklist was used for data extraction and quality assessment. From
13 671 identified records, six were included for data extraction; three assessed
smear-negative, culture-positive PTB as outcome and three focused on human
immunodeficiency virus infected individuals only. Reporting of model
development, performance and evaluation was poor. In four studies, predictive
performance was evaluated using the development data set (apparent performance),
one study did an internal validation and one study did an external validation.
Results were not pooled due to heterogeneity. Existing prediction models for
estimating prevalent PTB in adults at primary care level are poorly reported and
validated and are not useful for TB screening. The World Health Organization
symptom screen is recommended.
DOI: 10.5588/ijtld.16.0059
PMID: 28284255 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/36375644 | 1. Metabolism. 2023 Jan;138:155344. doi: 10.1016/j.metabol.2022.155344. Epub 2022
Nov 12.
Isoform changes of action potential regulators in the ventricles of
arrhythmogenic phospholamban-R14del humanized mouse hearts.
Rogalska ME(1), Vafiadaki E(2), Erpapazoglou Z(3), Haghighi K(4), Green L(5),
Mantzoros CS(6), Hajjar RJ(7), Tranter M(5), Karakikes I(8), Kranias EG(4),
Stillitano F(9), Kafasla P(3), Sanoudou D(10).
Author information:
(1)Centre for Genomic Regulation (CRG), The Barcelona Institute of Science and
Technology, Barcelona 08003, Spain.
(2)Molecular Biology Division, Biomedical Research Foundation of the Academy of
Athens, 11527 Athens, Greece.
(3)Institute for Fundamental Biomedical Research, B.S.R.C. "Alexander Fleming",
16672 Athens, Greece.
(4)Department of Pharmacology and Systems Physiology, University of Cincinnati
College of Medicine, Cincinnati, OH 45267, USA.
(5)Department of Medicine, University of Cincinnati College of Medicine,
Cincinnati, OH 45267, USA.
(6)Division of Endocrinology, Diabetes and Metabolism, Beth Israel Deaconess
Medical Center, Harvard Medical School, Boston, MA 02215, USA; Section of
Endocrinology, Boston VA Healthcare System, Harvard Medical School, Boston, MA
02215, USA.
(7)Flagship Pioneering, Cambridge, MA 02142, USA.
(8)Department of Cardiothoracic Surgery and Cardiovascular Institute, Stanford
University School of Medicine, 240 Pasteur Dr, Stanford, CA 94304, USA.
(9)Division Heart and Lung, Department of Cardiology, University Medical Center
Utrecht, 3584, CX, Utrecht, the Netherlands.
(10)Molecular Biology Division, Biomedical Research Foundation of the Academy of
Athens, 11527 Athens, Greece; Clinical Genomics and Pharmacogenomics Unit, 4(th)
Department of Internal Medicine, Attikon Hospital, Medical School, National and
Kapodistrian University of Athens, 11527 Athens, Greece; Center for New
Biotechnologies and Precision Medicine, Medical School, National and
Kapodistrian University of Athens, 11527 Athens, Greece. Electronic address:
[email protected].
Arrhythmogenic cardiomyopathy (ACM) is characterized by life-threatening
ventricular arrhythmias and sudden cardiac death and affects hundreds of
thousands of patients worldwide. The deletion of Arginine 14 (p.R14del) in the
phospholamban (PLN) gene has been implicated in the pathogenesis of ACM. PLN is
a key regulator of sarcoplasmic reticulum (SR) Ca2+ cycling and cardiac
contractility. Despite global gene and protein expression studies, the molecular
mechanisms of PLN-R14del ACM pathogenesis remain unclear. Using a humanized
PLN-R14del mouse model and human induced pluripotent stem cell derived
cardiomyocytes (iPSC-CMs), we investigated the transcriptome-wide mRNA splicing
changes associated with the R14del mutation. We identified >200 significant
alternative splicing (AS) events and distinct AS profiles were observed in the
right (RV) and left (LV) ventricles in PLN-R14del compared to WT mouse hearts.
Enrichment analysis of the AS events showed that the most affected biological
process was associated with "cardiac cell action potential", specifically in the
RV. We found that splicing of 2 key genes, Trpm4 and Camk2d, which encode
proteins regulating calcium homeostasis in the heart, were altered in PLN-R14del
mouse hearts and human iPSC-CMs. Bioinformatical analysis pointed to the
tissue-specific splicing factors Srrm4 and Nova1 as likely upstream regulators
of the observed splicing changes in the PLN-R14del cardiomyocytes. Our findings
suggest that aberrant splicing may affect Ca2+-homeostasis in the heart,
contributing to the increased risk of arrythmogenesis in PLN-R14del ACM.
Copyright © 2022 Elsevier Inc. All rights reserved.
DOI: 10.1016/j.metabol.2022.155344
PMID: 36375644 [Indexed for MEDLINE]
Conflict of interest statement: Declaration of competing interest The authors
have no conflicts of interest to declare. |
http://www.ncbi.nlm.nih.gov/pubmed/31381525 | 1. J Neuromuscul Dis. 2019;6(3):349-359. doi: 10.3233/JND-190402.
Molecular and Clinical Characteristics of a National Cohort of Paediatric
Duchenne Muscular Dystrophy Patients in Norway.
Annexstad EJ(1)(2), Fagerheim T(3), Holm I(2)(4), Rasmussen M(1)(5).
Author information:
(1)Oslo University Hospital, Unit for Congenital and Inherited Neuromuscular
Disorders, Oslo, Norway.
(2)University of Oslo, Faculty of Medicine, Oslo, Norway.
(3)University Hospital of North Norway, Department of Medical Genetics, Tromso,
Norway.
(4)Oslo University Hospital, Division of Orthopaedic Surgery, Section of
Research, Oslo, Norway.
(5)Oslo University Hospital, Department of Clinical Neurosciences for Children,
Oslo, Norway.
BACKGROUND: As new gene-related treatment options for Duchenne muscular
dystrophy (DMD) are being developed, precise information about the patients'
genetic diagnosis and knowledge about the diversities of natural history in DMD
is vital.
OBJECTIVE: To obtain detailed insight into the genetic and clinical
characteristics of paediatric DMD in Norway.
METHODS: 94 boys with DMD, aged 0-18 years, were identified over a period of 3.5
years, yielding a national prevalence of 13.5×10-5 boys. 73 boys (78%) were
recruited to full genetic and clinical or limited (genetic only) evaluation.
RESULTS: Molecular analysis disclosed 64% deletions, 18% duplications and 18%
point mutations. The mean age of diagnosis was 3.9±2.0 years. 78% were treated
with glucocorticoids from age 5.8±1.5 years. 23 boys (35%) had lost ambulation
at an age of 10.7±2.0 years. 17% were treated for left ventricular dysfunction
from age 12.1±3.0 years and 12% had received night-time non-invasive positive
pressure ventilation from age 13.0±2.5 years.
CONCLUSIONS: The distribution of mutation types and sites was similar to
previous studies but with more duplications and fewer point mutations. Any
genotype-phenotype correlations were not uncovered. The boys were diagnosed
early but there is still diagnostic delay among boys presenting with late motor
development. Glucocorticoid treatment was widespread, especially among the
younger boys. The clinical results of this comprehensive nationwide study
highlight the large variability of disease progression in DMD.
DOI: 10.3233/JND-190402
PMID: 31381525 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/21402533 | 1. Yi Chuan. 2011 Mar;33(3):251-4. doi: 10.3724/sp.j.1005.2011.00251.
[Association of mutation types and distribution characteristics of dystrophin
gene with clinical symptoms in Chinese population].
[Article in Chinese]
Li SY(1), Sun XF, Li Q, Zhang HM, Wang XM.
Author information:
(1)Guangzhou Key Laboratory of Reproductive and Genetics, the Third Affiliated
Hospital of Guangzhou Medical University, Guangzhou 510000, China.
[email protected]
Duchenne muscular dystrophy (DMD) is X-linked disorder caused by mutations in
the dystrophin gene. To investigate mutation types and distribution
characteristics of dystrophin gene in Chinese DMD patients, we used Multiplex
Ligation-Dependent Probe Amplification (MLPA) to analyze the dystrophin gene in
720 DMD patients, their mothers, and 20 normal adult males. Results showed that
detection rate was 64.9% (467/720) in all the patients, gene deletion rate was
54.3% (391/720), and gene duplication rate was 10.6% (76/720). The rate of
deletion mutant occurred in Exon 45-54 was 71.9% (281/391) in all gene deletion
patients; meanwhile, the rate of gene duplication occurred in Exon 1-40 was
82.9% (63/76) in all gene duplication ones. In all the patients with gene
deletion and duplication, the rate of DMD and IMD was 90.6% (423/467), and BMD,
9.4% (44/467). This indicates that the main reason of duchenne muscular
dystrophy is dystrophin gene deletion mutation, which would occur in any gene
unevenly with hot spots of mutation. The location and fragment length of gene
deletion and duplication cannot decide the severity of clinical symptoms
directly.
DOI: 10.3724/sp.j.1005.2011.00251
PMID: 21402533 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/26284620 | 1. PLoS One. 2015 Aug 18;10(8):e0135189. doi: 10.1371/journal.pone.0135189.
eCollection 2015.
DMD Mutations in 576 Dystrophinopathy Families: A Step Forward in
Genotype-Phenotype Correlations.
Juan-Mateu J(1), Gonzalez-Quereda L(1), Rodriguez MJ(2), Baena M(2), Verdura
E(2), Nascimento A(3), Ortez C(3), Baiget M(1), Gallano P(1).
Author information:
(1)Genetics Department, Hospital de la Santa Creu i Sant Pau, U705 CIBERER,
Barcelona, Spain.
(2)Genetics Department, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain.
(3)Neuromuscular Unit, Hospital Sant Joan de Deu, Esplugues de Llobregat, Spain.
Recent advances in molecular therapies for Duchenne muscular dystrophy (DMD)
require precise genetic diagnosis because most therapeutic strategies are
mutation-specific. To understand more about the genotype-phenotype correlations
of the DMD gene we performed a comprehensive analysis of the DMD mutational
spectrum in a large series of families. Here we provide the clinical,
pathological and genetic features of 576 dystrophinopathy patients. DMD gene
analysis was performed using the MLPA technique and whole gene sequencing in
blood DNA and muscle cDNA. The impact of the DNA variants on mRNA splicing and
protein functionality was evaluated by in silico analysis using computational
algorithms. DMD mutations were detected in 576 unrelated dystrophinopathy
families by combining the analysis of exonic copies and the analysis of small
mutations. We found that 471 of these mutations were large intragenic
rearrangements. Of these, 406 (70.5%) were exonic deletions, 64 (11.1%) were
exonic duplications, and one was a deletion/duplication complex rearrangement
(0.2%). Small mutations were identified in 105 cases (18.2%), most being
nonsense/frameshift types (75.2%). Mutations in splice sites, however, were
relatively frequent (20%). In total, 276 mutations were identified, 85 of which
have not been previously described. The diagnostic algorithm used proved to be
accurate for the molecular diagnosis of dystrophinopathies. The reading frame
rule was fulfilled in 90.4% of DMD patients and in 82.4% of Becker muscular
dystrophy patients (BMD), with significant differences between the mutation
types. We found that 58% of DMD patients would be included in single exon-exon
skipping trials, 63% from strategies directed against multiexon-skipping exons
45 to 55, and 14% from PTC therapy. A detailed analysis of missense mutations
provided valuable information about their impact on the protein structure.
DOI: 10.1371/journal.pone.0135189
PMCID: PMC4540588
PMID: 26284620 [Indexed for MEDLINE]
Conflict of interest statement: Competing Interests: The authors have declared
that no competing interests exist. |
http://www.ncbi.nlm.nih.gov/pubmed/35064276 | 1. Pediatr Cardiol. 2022 Apr;43(4):855-867. doi: 10.1007/s00246-021-02797-6. Epub
2022 Jan 22.
Diversity of Dystrophin Gene Mutations and Disease Progression in a Contemporary
Cohort of Duchenne Muscular Dystrophy.
Gambetta KE(1), McCulloch MA(2), Lal AK(3), Knecht K(4), Butts RJ(5), Villa
CR(6), Johnson JN(7), Conway J(8), Bock MJ(9), Schumacher KR(10), Law SP(11),
Friedland-Little JM(12), Deshpande SR(13), West SC(14), Lytrivi ID(15),
Wittlieb-Weber CA(16).
Author information:
(1)Division of Pediatric Cardiology, Ann & Robert H. Lurie Children's Hospital
of Chicago, Northwestern University Feinberg School of Medicine, 225 E Chicago
Avenue, Chicago, IL, 60605, USA. [email protected].
(2)University of Virginia Children's Hospital, Charlottesville, VA, USA.
(3)Primary Children's Hospital, University of Utah, Salt Lake City, UT, USA.
(4)Arkansas Children's Hospital, University of Arkansas for Medical Sciences,
Little Rock, AR, USA.
(5)Children's Medical Center of Dallas, UT Southwestern Medical Center, Dallas,
TX, USA.
(6)The Heart Institute, Cincinnati Children's Hospital Medical Center,
Cincinnati, OH, USA.
(7)Mayo Clinic Children's Center, Rochester, MN, USA.
(8)Stollery Children's Hospital, University of Alberta, Edmonton, AB, Canada.
(9)Loma Linda University Children's Hospital, Loma Linda, CA, USA.
(10)C.S. Mott Children's Hospital, University of Michigan, Ann Arbor, MI, USA.
(11)Morgan Stanley Children's Hospital of New York Presbyterian, New York, NY,
USA.
(12)Seattle Children's Hospital, Seattle, WA, USA.
(13)Children's National, Washington, DC, USA.
(14)Children's Hospital of Pittsburgh of UPMC, Pittsburgh, PA, USA.
(15)Mount Sinai Medical Center, New York, NY, USA.
(16)The Children's Hospital of Philadelphia, Philadelphia, PA, USA.
Abnormal dystrophin production due to mutations in the dystrophin gene causes
Duchenne Muscular Dystrophy (DMD). Cases demonstrate considerable genetic and
disease progression variability. It is unclear if specific gene mutations are
prognostic of outcomes in this population. We conducted a retrospective cohort
study of DMD patients followed at 17 centers across the USA and Canada from 2005
to 2015 with goal of understanding the genetic variability of DMD and its impact
on clinical outcomes. Cumulative incidence of clinically relevant outcomes was
stratified by genetic mutation type, exon mutation location, and extent of exon
deletion. Of 436 males with DMD, 324 (74.3%) underwent genetic testing.
Deletions were the most common mutation type (256, 79%), followed by point
mutations (45, 13.9%) and duplications (23, 7.1%). There were 131 combinations
of mutations with most mutations located along exons 45 to 52. The number of
exons deleted varied between 1 and 52 with a median of 3 exons deleted (IQR
1-6). Subjects with mutations starting at exon positions 40-54 had a later onset
of arrhythmias occurring at median age 25 years (95% CI 18-∞), p = 0.01. Loss of
ambulation occurred later at median age of 13 years (95% CI 12-15) in subjects
with mutations that started between exons 55-79, p = 0.01. There was no
association between mutation type or location and onset of cardiac dysfunction.
We report the genetic variability in DMD and its association with timing of
clinical outcomes. Genetic modifiers may explain some phenotypic variability.
© 2021. The Author(s), under exclusive licence to Springer Science+Business
Media, LLC, part of Springer Nature.
DOI: 10.1007/s00246-021-02797-6
PMID: 35064276 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/29973226 | 1. Orphanet J Rare Dis. 2018 Jul 4;13(1):109. doi: 10.1186/s13023-018-0853-z.
Comprehensive genetic characteristics of dystrophinopathies in China.
Ma P(1), Zhang S(1), Zhang H(1), Fang S(1), Dong Y(2), Zhang Y(3), Hao W(3), Wu
S(4), Zhao Y(5).
Author information:
(1)Department of Neurology, the General Hospital of Chinese People's Armed
Police Force, Beijing, China.
(2)Department of Magnetic Resonance, the General Hospital of Chinese People's
Armed Police Force, Beijing, China.
(3)Department of Precision Medicine Laboratory, the General Hospital of Chinese
People's Armed Police Force, Beijing, China.
(4)Department of Neurology, the General Hospital of Chinese People's Armed
Police Force, Beijing, China. [email protected].
(5)Research Institute of Neuromuscular and Neurodegenerative Diseases and
Department of Neurology, Qilu Hospital, Shandong University, Jinan, Shandong,
China. [email protected].
Erratum in
Orphanet J Rare Dis. 2021 Jun 2;16(1):252. doi: 10.1186/s13023-021-01853-x.
BACKGROUND: Dystrophinopathies are a set of severe and incurable X-linked
neuromuscular disorders caused by mutations in the dystrophin gene (DMD). These
mutations form a complex spectrum. A national registration network is essential
not only to provide more information about the prevalence and natural history of
the disease, but also to collect genetic data for analyzing the mutational
spectrum. This information is extremely beneficial for basic scientific
research, genetic diagnosis, trial planning, clinical care, and gene therapy.
METHODS: We collected data from 1400 patients (1042 patients with confirmed
unrelated Duchenne muscular dystrophy [DMD] or Becker muscular dystrophy [BMD])
registered in the Chinese Genetic Disease Registry from March 2012 to August
2017 and analyzed the genetic mutational characteristics of these patients.
RESULTS: Large deletions were the most frequent type of mutation (72.2%),
followed by nonsense mutations (11.9%), exon duplications (8.8%), small
deletions (3.0%), splice-site mutations (2.1%), small insertions (1.3%),
missense mutations (0.6%), and a combination mutation of a deletion and a
duplication (0.1%). Exon 45-50 deletion was the most frequent deletion type,
while exon 2 duplication was the most common duplication type. Two deletion
hotspots were calculated-one located toward the central part (exon 45-52) of the
gene and the other toward the 5'end (exon 8-26). We found no significant
difference between hereditary and de novo mutations on deletion hotspots.
Nonsense mutations accounted for 62.9% of all small mutations (197 patients).
CONCLUSION: We built a comprehensive national dystrophinopathy mutation database
in China, which is essential for basic and clinical research in this field. The
mutational spectrum and characteristics of this DMD/BMD group were largely
consistent with those in previous international DMD/BMD studies, with some
differences. Based on our results, about 12% of DMD/BMD patients with nonsense
mutations may benefit from stop codon read-through therapy. Additionally, the
top three targets for exon-skipping therapy are exon 51 (141, 13.5%), exon 53
(115, 11.0%), and exon 45 (84, 8.0%).
DOI: 10.1186/s13023-018-0853-z
PMCID: PMC6032532
PMID: 29973226 [Indexed for MEDLINE]
Conflict of interest statement: ETHICS APPROVAL AND CONSENT TO PARTICIPATE: The
study was approved by research ethics committee and medical ethics committee,
General Hospital of Chinese People’s Armed Police Forces. All information used
in the study were obtained from the General Hospital of Chinese People’s Armed
Police Forces, and were used for diagnostic purposes with written informed
consent. CONSENT FOR PUBLICATION: Not applicable COMPETING INTERESTS: The
authors declare that they have no competing interests. PUBLISHER’S NOTE:
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations. |
http://www.ncbi.nlm.nih.gov/pubmed/35939578 | 1. N Engl J Med. 2022 Jul 28;387(4):321-331. doi: 10.1056/NEJMoa2118024.
Trial of Anti-BDCA2 Antibody Litifilimab for Cutaneous Lupus Erythematosus.
Werth VP(1), Furie RA(1), Romero-Diaz J(1), Navarra S(1), Kalunian K(1), van
Vollenhoven RF(1), Nyberg F(1), Kaffenberger BH(1), Sheikh SZ(1), Radunovic
G(1), Huang X(1), Clark G(1), Carroll H(1), Naik H(1), Gaudreault F(1), Meyers
A(1), Barbey C(1), Musselli C(1), Franchimont N(1); LILAC Trial Investigators.
Collaborators: Alvarez A, Borgia A, Crespi G, García Carrasco M, Kerzberg E, Del
Valle Lucero E, Magariños G, Mannucci Walter PA, Moreno JLC, Savio V, Spindler
AJ, Tate P, Velasco Zamora JL, Demerdjieva Z, Goranov I, Gospodinov D,
Kapandjieva N, Marina S, Mateev G, Oparanov B, Rashkov R, Sapundziev L, Todorov
S, Velkova M, Chalem Choueka PS, Jaller Raad JJ, Maldonado Lopez MC, Otero
Escalante WJ, Pinto Peñaranda LF, Velez Sanchez PJ, Lidar M, Mevorach D, Abud
Mendoza C, Aguilar Arreola JE, Aroca Martínez GJ, Barrera Rodriguez AA, Benitez
Cabrera A, Cortes Hernandez M, De la Garza Ramos EH, Enriquez Sosa FE, Garcia de
la Torre I, Guerrero Diaz FI, Miranda Limon JM, Ortiz Jimenez E, Rizo Rodriguez
JC, Romero Diaz J, Xibille Friedmann DX, Arroyo C, Dulos R, Gomez HM, Hao L,
Lanzon A, Lichauco JJ, Perez E, Lorenzo JP, Manapat-Reyes BH, Navarra S,
Ramiterre E, Salido E, Tee M, Adamski Z, Celinska-Lowenhoff M, Dankiewicz-Fares
I, Krajewska-Wlodarczyk M, Narbutt J, Pulka G, Szepietowski J, Arandjelovic S,
Jovanovski A, Petronijevic M, Radunovic G, Zivkovic V, Cho CS, Lee SH, Lim MK,
Park YB, Shim SC, Suh CH, Fang YF, Ho JC, Hsieh SC, Tien YC, Chakkavittumrong P,
Choonhakarn C, Kasitanon N, Rerknimitr P, Siripaitoon B, Anadkat M, Antolini C,
Ayesu K, Barthel R, Berney S, Bunch T, Chindalore V, Chong B, Culton DA, Dore R,
Dvorkina O, Ehrlich A, Elston D, Fernandez B, Ferris L, Firooz N, Fretzin S,
Furie R, George R, Gross M, Helfrich Y, Holdgate N, Huff J, Iglesias N,
Kaffenberger B, Kavanaugh A, Kim W, Kivitz A, Kotha R, Kramer N, Kumar V, Laquer
V, Lee E, Levin R, Mabaquiao A, Mehta C, Merola J, Mesinkovska N, Mishra N, Nahm
W, Nami A, Pattanaik D, Pacheco L, Pariser D, Quinones M, Rangel O, Sheikh S,
Sofen H, Spangenthal S, Stoica G, Tesser J, Torres A, Turner M, Waller P,
Weisman J, Werth V.
Author information:
(1)From the University of Pennsylvania and Corporal Michael J. Crescenz Veterans
Affairs Medical Center - both in Philadelphia (V.P.W.); Northwell Health, Great
Neck, NY (R.A.F.); Instituto Nacional de Ciencias Médicas y Nutrición Salvador
Zubrián, Mexico City (J.R.-D.); the University of Santo Tomas, Manila,
Philippines (S.N.); the University of California, San Diego, La Jolla (K.K.);
Amsterdam University Medical Centers, Amsterdam (R.F.V.); Karolinska University
Hospital, Stockholm (F.N.); Ohio State University, Columbus (B.H.K.); University
of North Carolina at Chapel Hill, Chapel Hill (S.Z.S.); Institute of
Rheumatology, University of Belgrade, Belgrade, Serbia (G.R.); Biogen,
Cambridge, MA (X.H., G.C., H.C., H.N., F.G., A.M., C.M., N.F.); and Biogen,
Baar, Switzerland (C.B.).
Comment in
N Engl J Med. 2022 Oct 20;387(16):1528-1529. doi: 10.1056/NEJMc2211121.
BACKGROUND: Blood dendritic cell antigen 2 (BDCA2) is a receptor that is
exclusively expressed on plasmacytoid dendritic cells, which are implicated in
the pathogenesis of lupus erythematosus. Whether treatment with litifilimab, a
humanized monoclonal antibody against BDCA2, would be efficacious in reducing
disease activity in patients with cutaneous lupus erythematosus has not been
extensively studied.
METHODS: In this phase 2 trial, we randomly assigned adults with histologically
confirmed cutaneous lupus erythematosus with or without systemic manifestations
in a 1:1:1:1 ratio to receive subcutaneous litifilimab (at a dose of 50, 150, or
450 mg) or placebo at weeks 0, 2, 4, 8, and 12. We used a dose-response model to
assess whether there was a response across the four groups on the basis of the
primary end point, which was the percent change from baseline to 16 weeks in the
Cutaneous Lupus Erythematosus Disease Area and Severity Index-Activity score
(CLASI-A; scores range from 0 to 70, with higher scores indicating more
widespread or severe skin involvement). Safety was also assessed.
RESULTS: A total of 132 participants were enrolled; 26 were assigned to the
50-mg litifilimab group, 25 to the 150-mg litifilimab group, 48 to the 450-mg
litifilimab group, and 33 to the placebo group. Mean CLASI-A scores for the
groups at baseline were 15.2, 18.4, 16.5, and 16.5, respectively. The difference
from placebo in the change from baseline in CLASI-A score at week 16 was -24.3
percentage points (95% confidence interval [CI] -43.7 to -4.9) in the 50-mg
litifilimab group, -33.4 percentage points (95% CI, -52.7 to -14.1) in the
150-mg group, and -28.0 percentage points (95% CI, -44.6 to -11.4) in the 450-mg
group. The least squares mean changes were used in the primary analysis of a
best-fitting dose-response model across the three drug-dose levels and placebo,
which showed a significant effect. Most of the secondary end points did not
support the results of the primary analysis. Litifilimab was associated with
three cases each of hypersensitivity and oral herpes infection and one case of
herpes zoster infection. One case of herpes zoster meningitis occurred 4 months
after the participant received the last dose of litifilimab.
CONCLUSIONS: In a phase 2 trial involving participants with cutaneous lupus
erythematosus, treatment with litifilimab was superior to placebo with regard to
a measure of skin disease activity over a period of 16 weeks. Larger and longer
trials are needed to determine the effect and safety of litifilimab for the
treatment of cutaneous lupus erythematosus. (Funded by Biogen; LILAC
ClinicalTrials.gov number, NCT02847598.).
Copyright © 2022 Massachusetts Medical Society.
DOI: 10.1056/NEJMoa2118024
PMID: 35939578 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/36069871 | 1. N Engl J Med. 2022 Sep 8;387(10):894-904. doi: 10.1056/NEJMoa2118025.
Trial of Anti-BDCA2 Antibody Litifilimab for Systemic Lupus Erythematosus.
Furie RA(1), van Vollenhoven RF(1), Kalunian K(1), Navarra S(1), Romero-Diaz
J(1), Werth VP(1), Huang X(1), Clark G(1), Carroll H(1), Meyers A(1), Musselli
C(1), Barbey C(1), Franchimont N(1); LILAC Trial Investigators.
Collaborators: Alvarez A, Borgia A, Crespi G, Carrasco MG, Kerzberg E, Lucero
EDV, Magariños G, Mannucci Walter PA, Moreno JLC, Savio V, Spindler AJ, Tate P,
Velasco Zamora JL, Demerdjieva Z, Goranov I, Gospodinov D, Kapandjieva N, Marina
S, Mateev G, Oparanov B, Rashkov R, Sapundziev L, Todorov S, Velkova M, Aroca
Martínez GJ, Choueka PSC, Jaller Raad JJ, Maldonado Lopez MC, Otero Escalante
WJ, Pinto Peñaranda LF, Velez Sanchez PJ, Lidar M, Mevorach D, Abud Mendoza C,
Aguilar Arreola JE, Barrera Rodriguez AA, Benitez Cabrera A, Cortes Hernandez M,
De la Garza Ramos EH, Enriquez Sosa FE, Garcia de la Torre I, Guerrero Diaz FI,
Miranda Limon JM, Ortiz Jimenez E, Cruz Rizo Rodriguez J, Romero Diaz J, Xibille
Friedman DX, Arroyo C, Dulos R, Gomez HM, Hao L, Lanzon A, Lichauco JJ, Perez E,
Lorenzo JP, Manapat-Reyes BH, Navarra S, Ramiterre E, Salido E, Tee M, Adamski
Z, Celinska-Lowenhoff M, Dankiewicz-Fares I, Krajewska-Wlodarczyk M, Narbutt J,
Pulka G, Szepietowski J, Arandjelovic S, Jovanovski A, Petronijevic M, Radunovic
G, Zivkovic V, Cho CS, Lee SH, Lim MK, Park YB, Shim SC, Suh CH, Fang YF, Ho JC,
Hsieh SC, Tien YC, Chakkavittumrong P, Choonhakarn C, Kasitanon N, Rerknimitr P,
Siripaitoon B, Anadkat M, Antolini C, Ayesu K, Barthel R, Berney S, Bunch T,
Chindalore V, Chong B, Dore R, Dvorkina O, Ehrlich A, Elston D, Fernandez B,
Ferris L, Firooz N, Fretzin S, Furie R, George R, Gross M, Helfrich Y, Holdgate
N, Huff J, Iglesias N, Kaffenberger B, Kavanaugh A, Kim W, Kivitz A, Kotha R,
Kramer N, Kumar V, Laquer V, Lee E, Levin R, Mabaquiao A, Mehta C, Merola J,
Mesinkovska N, Mishra N, Nahm W, Nami A, Pattanaik D, Pacheco L, Pariser D,
Quinones M, Rangel O, Sheikh S, Sofen H, Spangenthal S, Stoica G, Tesser J,
Torres A, Turner M, Waller P, Weisman J, Werth V.
Author information:
(1)From Northwell Health, Great Neck, NY (R.A.F.); Amsterdam University Medical
Centers, Amsterdam (R.F.V.); the University of California San Diego, La Jolla
(K.K.); the University of Santo Tomas, Manila, Philippines (S.N.); Instituto
Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico City
(J.R.-D.); the University of Pennsylvania and Corporal Michael J. Crescenz
Veterans Affairs Medical Center - both in Philadelphia (V.P.W.); Biogen,
Cambridge, MA (X.H., G.C., H.C., A.M., C.M., N.F.); and Biogen, Baar,
Switzerland (C.B.).
Comment in
N Engl J Med. 2022 Sep 8;387(10):939-940. doi: 10.1056/NEJMe2208772.
BACKGROUND: Antibody-binding of blood dendritic cell antigen 2 (BDCA2), which is
expressed exclusively on plasmacytoid dendritic cells, suppresses the production
of type I interferon that is involved in the pathogenesis of systemic lupus
erythematosus (SLE). The safety and efficacy of subcutaneous litifilimab, a
humanized monoclonal antibody that binds to BDCA2, in patients with SLE have not
been extensively studied.
METHODS: We conducted a phase 2 trial of litifilimab involving participants with
SLE. The initial trial design called for randomly assigning participants to
receive litifilimab (at a dose of 50, 150, or 450 mg) or placebo administered
subcutaneously at weeks 0, 2, 4, 8, 12, 16, and 20, with the primary end point
of evaluating cutaneous lupus activity. The trial design was subsequently
modified; adults with SLE, arthritis, and active skin disease were randomly
assigned to receive either litifilimab at a dose of 450 mg or placebo. The
revised primary end point was the change from baseline in the total number of
active joints (defined as the sum of the swollen joints and the tender joints)
at week 24. Secondary end points were changes in cutaneous and global disease
activity. Safety was also assessed.
RESULTS: A total of 334 adults were assessed for eligibility, and 132 underwent
randomization (64 were assigned to receive 450-mg litifilimab, 6 to receive
150-mg litifilimab, 6 to receive 50-mg litifilimab, and 56 to receive placebo).
The primary analysis was conducted in the 102 participants who had received
450-mg litifilimab or placebo and had at least four tender and at least four
swollen joints. The mean (±SD) baseline number of active joints was 19.0±8.4 in
the litifilimab group and 21.6±8.5 in the placebo group. The least-squares mean
(±SE) change from baseline to week 24 in the total number of active joints was
-15.0±1.2 with litifilimab and -11.6±1.3 with placebo (mean difference, -3.4;
95% confidence interval, -6.7 to -0.2; P = 0.04). Most of the secondary end
points did not support the results of the analysis of the primary end point.
Receipt of litifilimab was associated with adverse events, including two cases
of herpes zoster and one case of herpes keratitis.
CONCLUSIONS: In a phase 2 trial involving participants with SLE, litifilimab was
associated with a greater reduction from baseline in the number of swollen and
tender joints than placebo over a period of 24 weeks. Longer and larger trials
are required to determine the safety and efficacy of litifilimab for the
treatment of SLE. (Funded by Biogen; LILAC ClinicalTrials.gov number,
NCT02847598.).
Copyright © 2022 Massachusetts Medical Society.
DOI: 10.1056/NEJMoa2118025
PMID: 36069871 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/23620649 | 1. Acta Myol. 2012 Dec;31(3):179-83.
Risk assessment and genetic counseling in families with Duchenne muscular
dystrophy.
Grimm T(1), Kress W, Meng G, Müller CR.
Author information:
(1)Department of Human Genetics, University Würzburg, Biozentrum, Würzburg,
Germany. [email protected]
The Duchenne Muscular dystrophy (DMD) is the most frequent muscle disorder in
childhood caused by mutations in the Xlinked dystrophin gene (about 65%
deletions, about 7% duplications, about 26% point mutations and about 2% unknown
mutations). The clinically milder Becker muscular dystrophy (BMD) is allelic to
DMD. About 33% of all patients are due to de novo mutations and germ line
mosaicism is frequently observed. While in earlier studies equal mutation rates
in males and females had been reported, a breakdown by mutation types can better
explain the sex ratio of mutations: Point mutations and duplications arise
preferentially during spermatogenesis whereas deletions mostly arise in
oogenesis. With current analytical methods, the underlying mutation can be
identified in the great majority of cases and be used for carrier detection.
However, in families with no mutation carrier available, the genetic model to be
used for counselling of relatives can be quite complex.
PMCID: PMC3631803
PMID: 23620649 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/35610482 | 1. Spinal Cord. 2022 Nov;60(11):990-995. doi: 10.1038/s41393-022-00815-9. Epub
2022 May 24.
Cross-cultural adaptation and validation of the French version of the Spinal
Cord Injury Pain Instrument (SCIPI).
Reynard F(1)(2), Léger B(3), Jordan X(4), Duong HP(3).
Author information:
(1)Department of Physiotherapy, Clinique romande de réadaptation Suva, Sion,
Switzerland. [email protected].
(2)Department of Medical Research, Clinique romande de réadaptation Suva, Sion,
Switzerland. [email protected].
(3)Department of Medical Research, Clinique romande de réadaptation Suva, Sion,
Switzerland.
(4)Department of Paraplegia, Clinique romande de réadaptation Suva, Sion,
Switzerland.
STUDY DESIGN: Cross-sectional.
OBJECTIVES: To assess the reliability and validity of the French version of the
Spinal Cord Injury Pain Instrument (SCIPI) and to determine its performance
versus "Douleur Neuropathique 4 questions" (DN4) in diagnosing neuropathic pain
(NeuP).
SETTING: Clinique romande de réadaptation, spinal cord injury (SCI) center in
the French-speaking part of Switzerland.
METHODS: Backward and forward translation in French of the 4-item SCIPI were
performed by native speakers in both languages. Thirty persons with SCI were
included in the validation study. Internal consistency was measured with the
Kuder-Richardson (KR-20) coefficient. Cohen's kappa coefficients were used to
assess the test-retest reliability and the agreement between SCIPI and DN4.
Clinical assessment was used as the reference standard to diagnose NeuP. The
area under the receiver operator characteristics curve (AUROC) was used to
assess the performance of diagnostic tests.
RESULTS: KR-20 coefficient of internal consistency was 0.50 (95% CI 0.26, 0.74).
Test-retest reliability coefficient was 0.86 (95% CI 0.76, 0.95). The best
cutoff value was 2 points, resulting a sensitivity of 88% (95% CI 69%, 98%) and
a specificity of 92% (95% CI 75%, 99%). SCIPI had an AUROC of 0.90 (95% CI 0.82,
0.98), which was not significantly lower than the AUROC for DN4, 0.92 (95% CI
0.85, 0.99, p = 0.56). Agreement between SCIPI and DN4 was of 0.88 (95% CI 0.77,
1.00).
CONCLUSION: The French version of the SCIPI is a reliable and valid tool that
can identify the presence of NeuP in an individual with SCI.
© 2022. The Author(s), under exclusive licence to International Spinal Cord
Society.
DOI: 10.1038/s41393-022-00815-9
PMID: 35610482 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/35426141 | 1. Pharmacotherapy. 2022 Jun;42(6):460-471. doi: 10.1002/phar.2683. Epub 2022 Apr
20.
Machine learning to predict vasopressin responsiveness in patients with septic
shock.
Scheibner A(1), Betthauser KD(1), Bewley AF(2), Juang P(1)(3), Lizza B(1), Micek
S(3), Lyons PG(2).
Author information:
(1)Department of Pharmacy, Barnes-Jewish Hospital, St. Louis, Missouri, USA.
(2)Division of Pulmonary and Critical Care Medicine, Washington University
School of Medicine, St. Louis, Missouri, USA.
(3)Department of Pharmacy Practice, University of Health Sciences and Pharmacy,
St. Louis, Missouri, USA.
STUDY OBJECTIVES: The objective of this study was to develop and externally
validate a model to predict adjunctive vasopressin response in patients with
septic shock being treated with norepinephrine for bedside use in the intensive
care unit.
DESIGN: This was a retrospective analysis of two adult tertiary intensive care
unit septic shock populations.
SETTING: Barnes-Jewish Hospital (BJH) from 2010 to 2017 and Beth Israel
Deaconess Medical Center (BIDMC) from 2001 to 2012.
PATIENTS: Two septic shock populations (548 BJH patients and 464 BIDMC patients)
that received vasopressin as second-line vasopressor.
INTERVENTION: Patients who were vasopressin responsive were compared with those
who were nonresponsive. Vasopressin response was defined as survival with at
least a 20% decrease in maximum daily norepinephrine requirements by one
calendar day after vasopressin initiation, without a third-line vasopressor.
MEASUREMENTS: Two supervised machine learning models (gradient-boosting machine
[XGBoost] and elastic net penalized logistic regression [EN]) were trained in
1000 bootstrap replications of the BJH data and externally validated in the
BIDMC data to predict vasopressin responsiveness.
MAIN RESULTS: Vasopressin responsiveness was similar among each cohort (BJH 45%
and BIDMC 39%). Mortality was lower for vasopressin responders compared with
nonresponders in the BJH (51% vs. 73%) and BIDMC (45% vs. 83%) cohorts,
respectively. Both models demonstrated modest discrimination in the training
(XGBoost area under receiver operator curve [AUROC] 0.61 [95% confidence
interval (CI) 0.61-0.61], EN 0.59 [95% CI 0.58-0.59]) and external validation
(XGBoost 0.68 [95% CI 0.63-0.73], EN 0.64 [95% CI 0.59-0.69]) datasets.
CONCLUSION: Vasopressin nonresponsiveness is common and associated with
increased mortality. The models' modest performances highlight the complexity of
septic shock and indicate that more research will be required before clinical
decision support tools can aid in anticipating patient-specific responsiveness
to vasopressin.
© 2022 Pharmacotherapy Publications, Inc.
DOI: 10.1002/phar.2683
PMID: 35426141 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/35438215 | 1. J Obstet Gynaecol Res. 2022 Jul;48(7):1775-1785. doi: 10.1111/jog.15266. Epub
2022 Apr 19.
Machine learning prediction models for postpartum depression: A multicenter
study in Japan.
Matsuo S(1), Ushida T(1)(2), Emoto R(3), Moriyama Y(4), Iitani Y(1), Nakamura
N(1), Imai K(1), Nakano-Kobayashi T(1), Yoshida S(5), Yamashita M(5), Matsui
S(3), Kajiyama H(1), Kotani T(1)(2).
Author information:
(1)Department of Obstetrics and Gynecology, Nagoya University Graduate School of
Medicine, Nagoya, Japan.
(2)Division of Perinatology, Center for Maternal-Neonatal Care, Nagoya
University Hospital, Nagoya, Japan.
(3)Department of Biostatistics, Nagoya University Graduate School of Medicine,
Nagoya, Japan.
(4)Department of Obstetrics and Gynecology, Fujita Health University School of
Medicine, Toyoake, Japan.
(5)Kishokai Medical Corporation, Nagoya, Japan.
AIM: Postpartum depression (PPD) and perinatal mental health care are of growing
importance worldwide. Here we aimed to develop and validate machine learning
models for the prediction of PPD, and to evaluate the usefulness of the recently
adopted 2-week postpartum checkup in some parts of Japan for the identification
of women at high risk of PPD.
METHODS: A multicenter retrospective study was conducted using the clinical data
of 10 013 women who delivered at ≥35 weeks of gestation at 12 maternity care
hospitals in Japan. PPD was defined as an Edinburgh Postnatal Depression Scale
score of ≥9 points at 4 weeks postpartum. We developed prediction models using
conventional logistic regression and four machine learning algorithms based on
the information that can be routinely collected in daily clinical practice. The
model performance was evaluated using the area under the receiver operating
characteristic curve (AUROC).
RESULTS: In the machine learning models developed using clinical data before
discharge, the AUROCs were similar to those in the conventional logistic
regression models (AUROC, 0.569-0.630 vs. 0.626). The incorporation of
additional 2-week postpartum checkup data into the model significantly improved
the predictive performance for PPD compared to that without in the Ridge
regression and Elastic net (AUROC, 0.702 vs. 0.630 [p < 0.01] and 0.701 vs.
0.628 [p < 0.01], respectively).
CONCLUSIONS: Our machine learning models did not achieve better predictive
performance for PPD than conventional logistic regression models. However, we
demonstrated the usefulness of the 2-week postpartum checkup for the
identification of women at high risk of PPD.
© 2022 Japan Society of Obstetrics and Gynecology.
DOI: 10.1111/jog.15266
PMID: 35438215 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/31304302 | 1. NPJ Digit Med. 2018 May 8;1:18. doi: 10.1038/s41746-018-0029-1. eCollection
2018.
Scalable and accurate deep learning with electronic health records.
Rajkomar A(#)(1)(2), Oren E(#)(1), Chen K(1), Dai AM(1), Hajaj N(1), Hardt M(1),
Liu PJ(1), Liu X(1), Marcus J(1), Sun M(1), Sundberg P(1), Yee H(1), Zhang K(1),
Zhang Y(1), Flores G(1), Duggan GE(1), Irvine J(1), Le Q(1), Litsch K(1), Mossin
A(1), Tansuwan J(1), Wang D(1), Wexler J(1), Wilson J(1), Ludwig D(2),
Volchenboum SL(3), Chou K(1), Pearson M(1), Madabushi S(1), Shah NH(4), Butte
AJ(2), Howell MD(1), Cui C(1), Corrado GS(1), Dean J(1).
Author information:
(1)1Google Inc, Mountain View, CA USA.
(2)2University of California, San Francisco, San Francisco, CA USA.
(3)3University of Chicago Medicine, Chicago, IL USA.
(4)4Stanford University, Stanford, CA USA.
(#)Contributed equally
Predictive modeling with electronic health record (EHR) data is anticipated to
drive personalized medicine and improve healthcare quality. Constructing
predictive statistical models typically requires extraction of curated predictor
variables from normalized EHR data, a labor-intensive process that discards the
vast majority of information in each patient's record. We propose a
representation of patients' entire raw EHR records based on the Fast Healthcare
Interoperability Resources (FHIR) format. We demonstrate that deep learning
methods using this representation are capable of accurately predicting multiple
medical events from multiple centers without site-specific data harmonization.
We validated our approach using de-identified EHR data from two US academic
medical centers with 216,221 adult patients hospitalized for at least 24 h. In
the sequential format we propose, this volume of EHR data unrolled into a total
of 46,864,534,945 data points, including clinical notes. Deep learning models
achieved high accuracy for tasks such as predicting: in-hospital mortality (area
under the receiver operator curve [AUROC] across sites 0.93-0.94), 30-day
unplanned readmission (AUROC 0.75-0.76), prolonged length of stay (AUROC
0.85-0.86), and all of a patient's final discharge diagnoses (frequency-weighted
AUROC 0.90). These models outperformed traditional, clinically-used predictive
models in all cases. We believe that this approach can be used to create
accurate and scalable predictions for a variety of clinical scenarios. In a case
study of a particular prediction, we demonstrate that neural networks can be
used to identify relevant information from the patient's chart.
DOI: 10.1038/s41746-018-0029-1
PMCID: PMC6550175
PMID: 31304302
Conflict of interest statement: Competing interestsThe authors declare no
competing interests. |
http://www.ncbi.nlm.nih.gov/pubmed/31437918 | 1. Stud Health Technol Inform. 2019 Aug 21;264:223-227. doi: 10.3233/SHTI190216.
Model Performance Metrics in Assessing the Value of Adding Intraoperative Data
for Death Prediction: Applications to Noncardiac Surgery.
Lei VJ(1)(2), Kennedy EH(3), Luong T(4), Chen X(2), Polsky DE(2), Volpp
KG(1)(2)(5), Neuman MD(2)(6), Holmes JH(2)(7), Fleisher LA(2)(6), Navathe
AS(1)(2)(5).
Author information:
(1)Department of Medical Ethics and Health Policy, University of Pennsylvania,
Philadelphia, Pennsylvania, USA.
(2)Leonard Davis Institute of Health Economics, University of Pennsylvania,
Philadelphia, Pennsylvania, USA.
(3)Department of Statistics & Data Science, Carnegie Mellon University,
Pittsburgh, Pennsylvania, USA.
(4)Predictive Healthcare, University of Pennsylvania Health System,
Philadelphia, Pennsylvania, USA.
(5)Veterans Health Administration, Department of Veterans Affairs, Philadelphia,
Pennsylvania, USA.
(6)Department of Anesthesiology, University of Pennsylvania Health System,
Philadelphia, Pennsylvania, USA.
(7)Department of Biostatistics, Epidemiology, and Informatics, University of
Pennsylvania, Philadelphia, Pennsylvania, USA.
We tested the value of adding data from the operating room to models predicting
in-hospital death. We assessed model performance using two metrics, the area
under the receiver operating characteristic curve (AUROC) and the area under the
precision-recall curve (AUPRC), to illustrate the differences in information
they convey in the setting of class imbalance. Data was collected on 74,147
patients who underwent major noncardiac surgery and 112 unique features were
extracted from electronic health records. Sets of features were incrementally
added to models using logistic regression, naïve Bayes, random forest, and
gradient boosted machine methods. AUROC increased as more features were added,
but changes were small for some modeling approaches. In contrast, AUPRC, which
reflects positive predicted value, exhibited improvements across all models.
Using AUPRC highlighted the added value of intraoperative data, not seen
consistently with AUROC, and that with class imbalance AUPRC may serve as the
more clinically relevant criterion.
DOI: 10.3233/SHTI190216
PMID: 31437918 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/32930711 | 1. J Am Med Inform Assoc. 2020 Oct 1;27(10):1593-1599. doi:
10.1093/jamia/ocaa180.
Representation of EHR data for predictive modeling: a comparison between UMLS
and other terminologies.
Rasmy L(1), Tiryaki F(1), Zhou Y(1), Xiang Y(1), Tao C(1), Xu H(1), Zhi D(1).
Author information:
(1)School of Biomedical Informatics University of Texas Health Science Center,
Houston, Texas, USA.
OBJECTIVE: Predictive disease modeling using electronic health record data is a
growing field. Although clinical data in their raw form can be used directly for
predictive modeling, it is a common practice to map data to standard
terminologies to facilitate data aggregation and reuse. There is, however, a
lack of systematic investigation of how different representations could affect
the performance of predictive models, especially in the context of machine
learning and deep learning.
MATERIALS AND METHODS: We projected the input diagnoses data in the Cerner
HealthFacts database to Unified Medical Language System (UMLS) and 5 other
terminologies, including CCS, CCSR, ICD-9, ICD-10, and PheWAS, and evaluated the
prediction performances of these terminologies on 2 different tasks: the risk
prediction of heart failure in diabetes patients and the risk prediction of
pancreatic cancer. Two popular models were evaluated: logistic regression and a
recurrent neural network.
RESULTS: For logistic regression, using UMLS delivered the optimal area under
the receiver operating characteristics (AUROC) results in both dengue
hemorrhagic fever (81.15%) and pancreatic cancer (80.53%) tasks. For recurrent
neural network, UMLS worked best for pancreatic cancer prediction (AUROC
82.24%), second only (AUROC 85.55%) to PheWAS (AUROC 85.87%) for dengue
hemorrhagic fever prediction.
DISCUSSION/CONCLUSION: In our experiments, terminologies with larger
vocabularies and finer-grained representations were associated with better
prediction performances. In particular, UMLS is consistently 1 of the
best-performing ones. We believe that our work may help to inform better designs
of predictive models, although further investigation is warranted.
© The Author(s) 2020. Published by Oxford University Press on behalf of the
American Medical Informatics Association. All rights reserved. For permissions,
please email: [email protected].
DOI: 10.1093/jamia/ocaa180
PMCID: PMC7647355
PMID: 32930711 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/25958031 | 1. J Pediatr Urol. 2015 Aug;11(4):176.e1-7. doi: 10.1016/j.jpurol.2015.03.006.
Epub 2015 Apr 16.
Predictive value of specific ultrasound findings when used as a screening test
for abnormalities on VCUG.
Logvinenko T(1), Chow JS(2), Nelson CP(3).
Author information:
(1)Department of Urology, Boston Children's Hospital, Boston, MA, USA; Clinical
Research Center, Boston Children's Hospital, Boston, MA, USA. Electronic
address: [email protected].
(2)Department of Urology, Boston Children's Hospital, Boston, MA, USA;
Department of Radiology, Boston Children's Hospital, Boston, MA, USA; Department
of Radiology, Harvard Medical School, Boston, MA, USA. Electronic address:
[email protected].
(3)Department of Urology, Boston Children's Hospital, Boston, MA, USA;
Department of Radiology, Boston Children's Hospital, Boston, MA, USA; Department
of Surgery, Harvard Medical School, Boston, MA, USA. Electronic address:
[email protected].
Comment in
J Urol. 2016 May;195(5):1575-1576. doi: 10.1016/j.juro.2016.02.032.
BACKGROUND: Renal and bladder ultrasound (RBUS) is often used as an initial
screening test for children after urinary tract infection (UTI), and the 2011
AAP guidelines specifically recommend RBUS be performed first, with voiding
cystourethrogram (VCUG) to be performed only if the ultrasound is abnormal. It
is uncertain whether specific RBUS findings, alone or in combination, might make
RBUS more useful as a predictor of VCUG abnormalities.
AIMS: To evaluate the association of specific RBUS with VCUG findings, and
determine whether predictive models that accurately predict patients at high
risk of VCUG abnormalities, based on RBUS findings, can be constructed.
METHODS: and study sample: A total of 3995 patients were identified with VCUG
and RBUS performed on the same day. The RBUS and VCUG reports were reviewed and
the findings were classified. Analysis was limited to patients aged 0-60 months
with no prior postnatal genitourinary imaging and no history of prenatal
hydronephrosis.
ANALYSIS: The associations between large numbers of specific RBUS findings with
abnormalities seen on VCUG were investigated. Both multivariate logistic models
and a neural network machine learning algorithms were constructed to evaluate
the predictive power of RBUS for VCUG abnormalities (including VUR or
bladder/urethral findings). Sensitivity, specificity, predictive values and area
under receiving operating curves (AUROC) of RBUS for VCUG abnormalities were
determined.
RESULTS: A total of 2259 patients with UTI as the indication for imaging were
identified. The RBUS was reported as "normal" in 75.0%. On VCUG, any VUR was
identified in 41.7%, VUR grade > II in 20.9%, and VUR grade > III in 2.8%. Many
individual RBUS findings were significantly associated with VUR on VCUG. Despite
these strong univariate associations, multivariate modeling didn't result in a
predictive model that was highly accurate. Multivariate logistic regression
built via stepwise selection had: AUROC = 0.57, sensitivity = 86% and
specificity = 25% for any VUR; AUROC = 0.60, sensitivity = 5% and specificity =
99% for VUR grade > II; and AUROC = 0.67, sensitivity = 6% and specificity = 99%
for VUR grade > III. The best predictive model constructed via neural networks
had: AUROC = 0.69, sensitivity = 64% and specificity = 60% for any VUR; AUROC =
0.67, sensitivity = 18% and specificity = 98% for VUR grade > II; and AUROC =
0.79, sensitivity = 32% and specificity = 100% for VUR grade > III.
CONCLUSIONS: Even with the state-of-the-art predictive models, abnormal findings
on RBUS provide a poor screening test for genitourinary abnormalities. Renal
bladder ultrasound and VCUG should be considered complementary, as they provide
important, but different, information.
Copyright © 2015 Journal of Pediatric Urology Company. Published by Elsevier
Ltd. All rights reserved.
DOI: 10.1016/j.jpurol.2015.03.006
PMCID: PMC4540607
PMID: 25958031 [Indexed for MEDLINE]
Conflict of interest statement: Conflict of interest statement None of the
authors have any financial and personal relationships with other people or
organizations that could inappropriately influence (bias) their work. |
http://www.ncbi.nlm.nih.gov/pubmed/33203588 | 1. J Neurol Sci. 2021 Jan 15;420:117184. doi: 10.1016/j.jns.2020.117184. Epub
2020 Nov 2.
Systematic review and evaluation of predictive modeling algorithms in spinal
surgeries.
Romiyo P(1), Ding K(1), Dejam D(1), Franks A(1), Ng E(1), Preet K(1), Tucker
AM(1), Niu T(1), Nagasawa DT(1), Rahman S(2), Yang I(3).
Author information:
(1)Departments of Neurosurgery, Ronald Reagan UCLA Medical Center, 757 Westwood
Plaza, Los Angeles, CA 90095, United States.
(2)Kaiser Permanente, United States.
(3)Departments of Neurosurgery, Ronald Reagan UCLA Medical Center, 757 Westwood
Plaza, Los Angeles, CA 90095, United States; Departments of Office of the
Patient Experience, Ronald Reagan UCLA Medical Center, 757 Westwood Plaza, Los
Angeles, CA 90095, United States; Departments of Radiation Oncology, Ronald
Reagan UCLA Medical Center, 757 Westwood Plaza, Los Angeles, CA 90095, United
States; Departments of Head and Neck Surgery, Ronald Reagan UCLA Medical Center,
757 Westwood Plaza, Los Angeles, CA 90095, United States; Departments of UCLA
Jonsson Comprehensive Cancer Center, Ronald Reagan UCLA Medical Center, 757
Westwood Plaza, Los Angeles, CA 90095, United States; Department of
Neurosurgery, Harbor-UCLA Medical Center, 1000 W Carson St, Torrance, CA 90509,
United States; Department of Los Angeles Biomedical Research Institute (LA
BioMed), Harbor-UCLA Medical Center, 1000 W Carson St, Torrance, CA 90509,
United States. Electronic address: [email protected].
In order to better educate patients, predictive models have been implemented to
stratify surgical risk, thereby instituting greater uniformity across surgical
practices and prioritizing the safety and outcomes of patients. The purpose of
this study is to conduct a systematic review summarizing the major predictive
models used to evaluate patients as candidates for spinal surgery. A search was
conducted for articles related to predictive modeling in spinal surgeries using
PubMed, MEDLINE, and Scopus databases. Papers with area under the receiver
operating curve (AUROC) scores reported were included in the analysis. Models
not relevant to spinal procedures were excluded. Comparison between models was
only attainable for those that reported AUROCs for individual procedures. Based
on a combination of AUROC scores and demonstrated applicability to spinal
procedures, the models by Scheer et al. (0.89), Ratliff et al. (0.70), the
Seattle Spine Score (0.712), Risk Assessment Tool (0.67-0.7), and the Spine Sage
calculator (0.81-0.85) were determined to be ideal for predictive modeling in
spinal surgeries and were subsequently broken down into their individual inputs
and outputs to determine what elements a theoretical model should assimilate.
Alongside the model by Scheer et al., the Spine Sage calculator, Seattle Spine
Score, Risk Assessment Tool, and a model by Ratliff et al. showed the most
promise for patients undergoing spinal procedures. Using the first model as a
springboard, new spinal predictive models can be optimized through use of larger
prospective databases, with longer follow-up times, and greater inclusion of
reliable high impact variables.
Copyright © 2020 Elsevier B.V. All rights reserved.
DOI: 10.1016/j.jns.2020.117184
PMID: 33203588 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/30842277 | 1. Proc Natl Acad Sci U S A. 2019 Mar 19;116(12):5542-5549. doi:
10.1073/pnas.1814551116. Epub 2019 Mar 6.
Evolutionarily informed deep learning methods for predicting relative transcript
abundance from DNA sequence.
Washburn JD(1), Mejia-Guerra MK(1), Ramstein G(1), Kremling KA(1), Valluru R(1),
Buckler ES(2)(3), Wang H(4)(1).
Author information:
(1)Institute for Genomic Diversity, Cornell University, Ithaca, NY 14853.
(2)Institute for Genomic Diversity, Cornell University, Ithaca, NY 14853;
[email protected] [email protected].
(3)Agricultural Research Service, United States Department of Agriculture,
Ithaca, NY 14850.
(4)Biotechnology Research Institute, Chinese Academy of Agricultural Sciences,
100081 Beijing, China [email protected] [email protected].
Deep learning methodologies have revolutionized prediction in many fields and
show potential to do the same in molecular biology and genetics. However,
applying these methods in their current forms ignores evolutionary dependencies
within biological systems and can result in false positives and spurious
conclusions. We developed two approaches that account for evolutionary
relatedness in machine learning models: (i) gene-family-guided splitting and
(ii) ortholog contrasts. The first approach accounts for evolution by
constraining model training and testing sets to include different gene families.
The second approach uses evolutionarily informed comparisons between orthologous
genes to both control for and leverage evolutionary divergence during the
training process. The two approaches were explored and validated within the
context of mRNA expression level prediction and have the area under the ROC
curve (auROC) values ranging from 0.75 to 0.94. Model weight inspections showed
biologically interpretable patterns, resulting in the hypothesis that the 3' UTR
is more important for fine-tuning mRNA abundance levels while the 5' UTR is more
important for large-scale changes.
DOI: 10.1073/pnas.1814551116
PMCID: PMC6431157
PMID: 30842277 [Indexed for MEDLINE]
Conflict of interest statement: The authors declare no conflict of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/34908548 | 1. Anesth Analg. 2022 Jan 1;134(1):102-113. doi: 10.1213/ANE.0000000000005694.
Impact of Intraoperative Data on Risk Prediction for Mortality After
Intra-Abdominal Surgery.
Yan X(1), Goldsmith J(1), Mohan S(2)(3), Turnbull ZA(4), Freundlich RE(5),
Billings FT 4th(5), Kiran RP(3)(6), Li G(3)(7), Kim M(3)(7).
Author information:
(1)From the Department of Biostatistics, Mailman School of Public Health,
Columbia University, New York, New York.
(2)Department of Medicine, Division of Nephrology, Columbia University Medical
Center, New York, New York.
(3)Department of Epidemiology, Mailman School of Public Health, Columbia
University, New York, New York.
(4)Department of Anesthesiology, Weill Cornell Medicine, New York, New York.
(5)Department of Anesthesiology, Vanderbilt University Medical Center,
Nashville, Tennessee.
(6)Department of Surgery, Division of Colorectal Surgery.
(7)Department of Anesthesiology, Columbia University Medical Center, New York,
New York.
Comment in
Anesth Analg. 2022 May 1;134(5):e29. doi: 10.1213/ANE.0000000000005977.
Anesth Analg. 2022 May 1;134(5):e29-e30. doi: 10.1213/ANE.0000000000005978.
BACKGROUND: Risk prediction models for postoperative mortality after
intra-abdominal surgery have typically been developed using preoperative
variables. It is unclear if intraoperative data add significant value to these
risk prediction models.
METHODS: With IRB approval, an institutional retrospective cohort of
intra-abdominal surgery patients in the 2005 to 2015 American College of
Surgeons National Surgical Quality Improvement Program was identified.
Intraoperative data were obtained from the electronic health record. The primary
outcome was 30-day mortality. We evaluated the performance of machine learning
algorithms to predict 30-day mortality using: 1) baseline variables and 2)
baseline + intraoperative variables. Algorithms evaluated were: 1) logistic
regression with elastic net selection, 2) random forest (RF), 3) gradient
boosting machine (GBM), 4) support vector machine (SVM), and 5) convolutional
neural networks (CNNs). Model performance was evaluated using the area under the
receiver operator characteristic curve (AUROC). The sample was randomly divided
into a training/testing split with 80%/20% probabilities. Repeated 10-fold
cross-validation identified the optimal model hyperparameters in the training
dataset for each model, which were then applied to the entire training dataset
to train the model. Trained models were applied to the test cohort to evaluate
model performance. Statistical significance was evaluated using P < .05.
RESULTS: The training and testing cohorts contained 4322 and 1079 patients,
respectively, with 62 (1.4%) and 15 (1.4%) experiencing 30-day mortality,
respectively. When using only baseline variables to predict mortality, all
algorithms except SVM (area under the receiver operator characteristic curve
[AUROC], 0.83 [95% confidence interval {CI}, 0.69-0.97]) had AUROC >0.9: GBM
(AUROC, 0.96 [0.94-1.0]), RF (AUROC, 0.96 [0.92-1.0]), CNN (AUROC, 0.96
[0.92-0.99]), and logistic regression (AUROC, 0.95 [0.91-0.99]). AUROC
significantly increased with intraoperative variables with CNN (AUROC, 0.97
[0.96-0.99]; P = .047 versus baseline), but there was no improvement with GBM
(AUROC, 0.97 [0.95-0.99]; P = .3 versus baseline), RF (AUROC, 0.96 [0.93-1.0]; P
= .5 versus baseline), and logistic regression (AUROC, 0.94 [0.90-0.99]; P = .6
versus baseline).
CONCLUSIONS: Postoperative mortality is predicted with excellent discrimination
in intra-abdominal surgery patients using only preoperative variables in various
machine learning algorithms. The addition of intraoperative data to preoperative
data also resulted in models with excellent discrimination, but model
performance did not improve.
Copyright © 2021 International Anesthesia Research Society.
DOI: 10.1213/ANE.0000000000005694
PMCID: PMC8682663
PMID: 34908548 [Indexed for MEDLINE]
Conflict of interest statement: The authors declare no conflicts of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/34865209 | 1. Thromb Haemost. 2022 Jun;122(6):913-925. doi: 10.1055/s-0041-1739514. Epub
2021 Dec 5.
Predictive Modeling Identifies Total Bleeds at 12-Weeks Postswitch to N8-GP
Prophylaxis as a Predictor of Treatment Response.
Chowdary P(1), Hampton K(2), Jiménez-Yuste V(3), Young G(4), Benchikh El Fegoun
S(5), Cooper A(6), Scalfaro E(7), Tiede A(8).
Author information:
(1)Katharine Dormandy Haemophilia and Thrombosis Centre, Royal Free Hospital,
London, United Kingdom.
(2)Department of Cardiovascular Science, University of Sheffield, Sheffield,
United Kingdom.
(3)Department of Hematology, La Paz University Hospital-IdiPaz, Autónoma
University, Madrid, Spain.
(4)Hemostasis and Thrombosis Center, Cancer and Blood Disorders Institute,
Children's Hospital Los Angeles, University of Southern California Keck School
of Medicine, Los Angeles, California, United Sates.
(5)Global Medical Affairs Biopharm, Novo Nordisk Health Care AG, Zürich,
Switzerland.
(6)Predictive Analytics, Real World Solutions, IQVIA, London, United Kingdom.
(7)Real World Insights, IQVIA, Basel, Switzerland.
(8)Department of Hematology, Hemostasis, Oncology and Stem Cell Transplantation,
Hannover Medical School, Hanover, Germany.
BACKGROUND: Predicting annualized bleeding rate (ABR) during factor VIII
(FVIII) prophylaxis for severe hemophilia A (SHA) is important for long-term
outcomes. This study used supervised machine learning-based predictive modeling
to identify predictors of long-term ABR during prophylaxis with an extended
half-life FVIII.
METHODS: Data were from 166 SHA patients who received N8-GP prophylaxis
(50 IU/kg every 4 days) in the pathfinder 2 study. Predictive models were
developed to identify variables associated with an ABR of ≤1 versus >1 during
the trial's main phase (median follow-up of 469 days). Model performance was
assessed using area under the receiver operator characteristic curve (AUROC).
Pre-N8-GP prophylaxis models learned from data collected at baseline; post-N8-GP
prophylaxis models learned from data collected up to 12-weeks postswitch to
N8-GP, and predicted ABR at the end of the outcome period (final year of
treatment in the main phase).
RESULTS: The predictive model using baseline variables had moderate performance
(AUROC = 0.64) for predicting observed ABR. The most performant model used data
collected at 12-weeks postswitch (AUROC = 0.79) with cumulative bleed count up
to 12 weeks as the most informative variable, followed by baseline von
Willebrand factor and mean FVIII at 30 minutes postdose. Univariate cumulative
bleed count at 12 weeks performed equally well to the 12-weeks postswitch model
(AUROC = 0.75). Pharmacokinetic measures were indicative, but not essential, to
predict ABR.
CONCLUSION: Cumulative bleed count up to 12-weeks postswitch was as informative
as the 12-week post-switch predictive model for predicting long-term ABR,
supporting alterations in prophylaxis based on treatment response.
The Author(s). This is an open access article published by Thieme under the
terms of the Creative Commons Attribution License, permitting unrestricted use,
distribution, and reproduction so long as the original work is properly cited.
(https://creativecommons.org/licenses/by/4.0/).
DOI: 10.1055/s-0041-1739514
PMCID: PMC9251711
PMID: 34865209 [Indexed for MEDLINE]
Conflict of interest statement: P.C. has received grant/research support from
Alnylam, Bayer, Biogen, Cangene, CSL Behring, Freeline Therapeutics, Novo
Nordisk, Pfizer, Sobi, and Takeda (Shire); and consultancy fees from Bayer,
Biogen, Cangene, CSL Behring, Chugai, Freeline Therapeutics, Novo Nordisk,
Pfizer, Roche, Spark Therapeutics, Sobi, and Takeda (Shire). K.H. has received
consultancy fees from Novo Nordisk. V.J-.Y. has received reimbursement for
attending symposia/congresses and/or honoraria for speaking and/or honoraria for
consulting, and/or funds for research from Bayer, CSL Behring, Grifols, Novo
Nordisk, Octapharma, Sobi, and Takeda. G.Y. has received consultancy fees from
Novo Nordisk. S.B.F. is a shareholder and employee of Novo Nordisk. A.C. is an
employee of IQVIA UK, which received funding from Novo Nordisk to design,
conduct, and interpret the reported analyses. E.S. is an employee of IQVIA AG,
which received funding from Novo Nordisk to design, conduct, and interpret the
reported analyses. A.T. has received grants for research/study support from
Bayer, Biotest, Chugai, Novo Nordisk, Octapharma, Pfizer, Roche, SOBI, and
Takeda; and received honoraria for lectures/consultancy from Bayer, Biotest,
Chugai, CSL Behring, Novo Nordisk, Octapharma, Pfizer, Roche, SOBI, and Takeda. |
http://www.ncbi.nlm.nih.gov/pubmed/35509018 | 1. BMC Pregnancy Childbirth. 2022 May 4;22(1):388. doi:
10.1186/s12884-022-04699-8.
Application of machine learning methods for predicting infant mortality in
Rwanda: analysis of Rwanda demographic health survey 2014-15 dataset.
Mfateneza E(1), Rutayisire PC(2), Biracyaza E(3), Musafiri S(4), Mpabuka WG(5).
Author information:
(1)African Centre of Excellence in Data Science, University of Rwanda, Kigali,
Rwanda. [email protected].
(2)Applied Statistics Department, University of Rwanda, Kigali, Rwanda.
(3)Prison Fellowship Rwanda, Kigali, Rwanda.
(4)Clinical Department of Internal Medicine, University of Rwanda, Kigali,
Rwanda.
(5)Transparency International Rwanda, Kigali, Rwanda.
BACKGROUND: Extensive research on infant mortality (IM) exists in developing
countries; however, most of the methods applied thus far relied on conventional
regression analyses with limited prediction capability. Advanced of Machine
Learning (AML) methods provide accurate prediction of IM; however, there is no
study conducted using ML methods in Rwanda. This study, therefore, applied
Machine Learning Methods for predicting infant mortality in Rwanda. METHODS: A
cross-sectional study design was conducted using the 2014-15 Rwanda Demographic
and Health Survey. Python software version 3.8 was employed to test and apply ML
methods through Random Forest (RF), Decision Tree, Support Vector Machine and
Logistic regression. STATA version 13 was used for analysing conventional
methods. Evaluation metrics methods specifically confusion matrix, accuracy,
precision, recall, F1 score, and Area under the Receiver Operating
Characteristics (AUROC) were used to evaluate the performance of predictive
models.
RESULTS: Ability of prediction was between 68.6% and 61.5% for AML. We preferred
with the RF model (61.5%) presenting the best performance. The RF model was the
best predictive model of IM with accuracy (84.3%), recall (91.3%), precision
(80.3%), F1 score (85.5%), and AUROC (84.2%); followed by decision tree model
with model accuracy (83%), recall (91%), precision (79%), F1 score (84.67%) and
AUROC(82.9%), followed by support vector machine with model accuracy (68.6%),
recall (74.9%), precision(67%), F1 score (70.73%) and AUROC (68.6%) and last was
a logistic regression with the low accuracy of prediction (61.5%), recall
(61.1%), precision (62.2%), F1 score (61.6%) and AUROC (61.5%) compared to other
predictive models. Our predictive models showed that marital status, children
ever born, birth order and wealth index are the 4 top predictors of IM.
CONCLUSIONS: In developing a predictive model, ML methods are used to classify
certain hidden information that could not be detected by traditional statistical
methods. Random Forest was classified as the best classifier to be used for the
predictive models of IM.
© 2022. The Author(s).
DOI: 10.1186/s12884-022-04699-8
PMCID: PMC9066935
PMID: 35509018 [Indexed for MEDLINE]
Conflict of interest statement: The authors declare that they have no competing
interests. |
http://www.ncbi.nlm.nih.gov/pubmed/31093546 | 1. Diagn Progn Res. 2017 Nov 15;1:17. doi: 10.1186/s41512-017-0017-y. eCollection
2017.
A novel method for interrogating receiver operating characteristic curves for
assessing prognostic tests.
Thomas G(1), Kenny LC(2)(3), Baker PN(4), Tuytten R(5).
Author information:
(1)SQU4RE, 8800 Roeselare, Belgium.
(2)2Department of Obstetrics and Gynaecology, University College Cork, Cork,
Ireland.
(3)Irish Centre for Fetal and Neonatal Translational Research (INFANT), 5th
Floor, Cork University Maternity Hospital, Cork, Ireland.
(4)3College of Medicine, Biological Sciences and Psychology, University of
Leicester, University Road, Leicester, LE1 7RH UK.
(5)Metabolomic Diagnostics, Hoffmann Park, Little Island, Cork, Ireland.
BACKGROUND: Disease prevalence is rarely explicitly considered in the early
stages of the development of novel prognostic tests. Rather, researchers use the
area under the receiver operating characteristic (AUROC) as the key metric to
gauge and report predictive performance ability. Because this statistic does not
account for disease prevalence, proposed tests may not appropriately address
clinical requirements. This ultimately impedes the translation of prognostic
tests into clinical practice.
METHODS: A method to express positive- and/or negative predictive value criteria
(PPV, NPV) within the ROC space is presented. Equations are derived for
so-called equi-PPV (and equi-NPV) lines. Herewith it is possible, for any given
prevalence, to plot a series of sensitivity-specificity pairs which meet a
specified PPV (or NPV) criterion onto the ROC space.This concept is introduced
by firstly reviewing the well-established "mechanics", strengths and limitations
of the ROC analysis in the context of developing prognostic models. Then, the
use of PPV (and/or) NPV criteria to augment the ROC analysis is
elaborated.Additionally, an interactive web tool was also created to enable
people to explore the dynamics of lines of equi-predictive value in function of
prevalence. The web tool also allows to gauge what ROC curve shapes best meet
specific positive and/or negative predictive value criteria
(http://d4ta.link/ppvnpv/).
RESULTS: To illustrate the merits and implications of this concept, an example
on the prediction of pre-eclampsia risk in low-risk nulliparous pregnancies is
elaborated.
CONCLUSIONS: In risk stratification, the clinical usefulness of a prognostic
test can be expressed in positive- and negative predictive value criteria; the
development of novel prognostic tests will be facilitated by the possibility to
co-visualise such criteria together with ROC curves. To achieve clinically
meaningful risk stratification, the development of separate tests to meet either
a pre-specified positive value (rule-in) or a negative predictive value
(rule-out) criteria should be considered: the characteristics of successful
rule-in and rule-out tests may markedly differ.
DOI: 10.1186/s41512-017-0017-y
PMCID: PMC6460848
PMID: 31093546
Conflict of interest statement: In addition of being practising clinicians, PNB
and LCK have a long-standing record in both basic and translational preeclampsia
research. PNB and LCK are well-recognised pioneers in the use of “omics” for the
discovery of novel biomarkers to predict pre-eclampsia. RT and GT have been
involved in “omics” biomarker discovery and biomarker translational research for
over a decade.Not applicable.Not applicable.GT is the owner of SQU4RE, an
independent Statistics and Data mining provider. LCK is a minority shareholder
in Metabolomic Diagnostics, a company that has licenced technology concerning
the use of metabolomics biomarkers in the prediction of pre-eclampsia. LCK has
also received consultancy fees and honoraria payments from Alere relating to the
Triage PlGF test for the prediction of complications in women with suspected
pre-eclampsia. LCK is also Director of INFANT which is funded in part by a range
of industry partnerships. Full details can be found at www.infantcentre.ie. PNB
is a minority shareholder in Metabolomic Diagnostics, a company that has
licenced technology concerning the use of metabolomics biomarkers in the
prediction of pre-eclampsia. RT is employed by Metabolomic Diagnostics, which is
developing metabolomics-based prognostic tests for adverse pregnancy
outcomes.Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations. |
http://www.ncbi.nlm.nih.gov/pubmed/35505048 | 1. Sci Rep. 2022 May 3;12(1):7180. doi: 10.1038/s41598-022-11226-4.
Mortality prediction of patients in intensive care units using machine learning
algorithms based on electronic health records.
Choi MH(1), Kim D(1), Choi EJ(2), Jung YJ(2), Choi YJ(3), Cho JH(3), Jeong
SH(4).
Author information:
(1)Department of Laboratory Medicine and Research Institute of Bacterial
Resistance, Gangnam Severance Hospital, Yonsei University College of Medicine,
211 Eonju-ro, Gangnam-gu, Seoul, 06273, South Korea.
(2)Department of Statistics and Data Science, Yonsei University, Seoul, South
Korea.
(3)Department of Internal Medicine, Gangnam Severance Hospital, Yonsei
University College of Medicine, Seoul, South Korea.
(4)Department of Laboratory Medicine and Research Institute of Bacterial
Resistance, Gangnam Severance Hospital, Yonsei University College of Medicine,
211 Eonju-ro, Gangnam-gu, Seoul, 06273, South Korea. [email protected].
Improving predictive models for intensive care unit (ICU) inpatients requires a
new strategy that periodically includes the latest clinical data and can be
updated to reflect local characteristics. We extracted data from all adult
patients admitted to the ICUs of two university hospitals with different
characteristics from 2006 to 2020, and a total of 85,146 patients were included
in this study. Machine learning algorithms were trained to predict in-hospital
mortality. The predictive performance of conventional scoring models and machine
learning algorithms was assessed by the area under the receiver operating
characteristic curve (AUROC). The conventional scoring models had various
predictive powers, with the SAPS III (AUROC 0.773 [0.766-0.779] for hospital S)
and APACHE III (AUROC 0.803 [0.795-0.810] for hospital G) showing the highest
AUROC among them. The best performing machine learning models achieved an AUROC
of 0.977 (0.973-0.980) in hospital S and 0.955 (0.950-0.961) in hospital G. The
use of ML models in conjunction with conventional scoring systems can provide
more useful information for predicting the prognosis of critically ill patients.
In this study, we suggest that the predictive model can be made more robust by
training with the individual data of each hospital.
© 2022. The Author(s).
DOI: 10.1038/s41598-022-11226-4
PMCID: PMC9065110
PMID: 35505048 [Indexed for MEDLINE]
Conflict of interest statement: The authors declare no competing interests. |
http://www.ncbi.nlm.nih.gov/pubmed/31807867 | 1. J Cancer Res Clin Oncol. 2020 Mar;146(3):767-775. doi:
10.1007/s00432-019-03103-x. Epub 2019 Dec 5.
Predictive models for patients with lung carcinomas to identify EGFR mutation
status via an artificial neural network based on multiple clinical information.
Qin X(1), Wang H(2), Hu X(3), Gu X(4), Zhou W(5).
Author information:
(1)Department of Hematology, The First Affiliated Hospital of Wenzhou Medical
University, Wenzhou, Zhejiang, China.
(2)Wenzhou Medical University, Wenzhou, Zhejiang, China.
(3)Department of Endocrinology, The First Affiliated Hospital of Wenzhou Medical
University, Wenzhou, Zhejiang, China.
(4)Department of Pneumology, Ningbo Yinzhou NO.2 Hospital, Ningbo, Zhejiang,
China.
(5)Department of Intensive Care Unit, The First Affiliated Hospital of Wenzhou
Medical University, Nan Bai Xiang Street, Ouhai District, Wenzhou, Zhejiang,
325000, China. [email protected].
PURPOSE: Epidermal growth factor receptor (EGFR) mutation testing has several
limitations. Therefore, we built predictive models to determine the EGFR
mutation status of patients and guide therapeutic decision-making.
METHODS: We collected data from 320 patients with lung carcinoma, including sex,
age, smoking history, serum tumour marker levels, maximum standardized uptake
value, pathological results, computed tomography images, and EGFR mutation
status. Artificial neural network (ANN) models based on multiple clinical
characteristics were proposed to predict EGFR mutation status.
RESULTS: A training set (n = 200) was used to develop predictive models of the
EGFR mutation status (Model 1: area under the receiver operating characteristic
curve [AUROC] = 0.910, 95% CI 0.861-0.945; Model 2: AUROC = 0.859, 95% CI
0.803-0.904; Model 3: AUROC = 0.711, 95% CI 0.643-0.773). A testing set (n = 50)
and temporal validation data set (n = 70) were used to evaluate the
generalisation performance of the established models (testing set: Model 1,
AUROC = 0.845, 95% CI 0.715-0.932; Model 2, AUROC = 0.882, 95% CI 0.759-0.956;
Model 3, AUROC = 0.817, 95% CI 0.682-0.912; temporal validation dataset: Model
1, AUROC = 0.909, 95% CI 0.816-0.964; Model 2, AUROC = 0.855, 95% CI
0.751-0.928; Model 3, AUROC = 0.831, 95% CI 0.723-0.910). The predictive
abilities of the three ANN models were superior to that of a previous logistic
regression model (P < 0.001, 0.027, and 0.050, respectively).
CONCLUSIONS: ANN models provide a non-invasive and readily available method for
EGFR mutation status prediction.
DOI: 10.1007/s00432-019-03103-x
PMID: 31807867 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/31932807 | 1. Nat Med. 2020 Jan;26(1):71-76. doi: 10.1038/s41591-019-0724-8. Epub 2020 Jan
13.
Prediction of gestational diabetes based on nationwide electronic health
records.
Artzi NS(#)(1)(2), Shilo S(#)(1)(2)(3), Hadar E(#)(4)(5), Rossman H(1)(2),
Barbash-Hazan S(4), Ben-Haroush A(4)(5), Balicer RD(6)(7), Feldman B(6),
Wiznitzer A(8)(9), Segal E(10)(11).
Author information:
(1)Department of Computer Science and Applied Mathematics, Weizmann Institute of
Science, Rehovot, Israel.
(2)Department of Molecular Cell Biology, Weizmann Institute of Science, Rehovot,
Israel.
(3)Pediatric Diabetes Unit, Ruth Rappaport Children's Hospital, Rambam
Healthcare Campus, Haifa, Israel.
(4)Helen Schneider Hospital for Women, Rabin Medical Center, Petach Tikva,
Israel.
(5)Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel.
(6)Clalit Research Institute, Clalit Health Services, Tel Aviv, Israel.
(7)Department of Public Health, Faculty of Health Sciences, Ben-Gurion
University, Beer-Sheva, Israel.
(8)Helen Schneider Hospital for Women, Rabin Medical Center, Petach Tikva,
Israel. [email protected].
(9)Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel.
[email protected].
(10)Department of Computer Science and Applied Mathematics, Weizmann Institute
of Science, Rehovot, Israel. [email protected].
(11)Department of Molecular Cell Biology, Weizmann Institute of Science,
Rehovot, Israel. [email protected].
(#)Contributed equally
Comment in
Nat Rev Endocrinol. 2020 Mar;16(3):130. doi: 10.1038/s41574-020-0326-z.
Trends Pharmacol Sci. 2020 May;41(5):301-304. doi:
10.1016/j.tips.2020.03.003.
Gestational diabetes mellitus (GDM) poses increased risk of short- and long-term
complications for mother and offspring1-4. GDM is typically diagnosed at
24-28 weeks of gestation, but earlier detection is desirable as this may prevent
or considerably reduce the risk of adverse pregnancy outcomes5,6. Here we used a
machine-learning approach to predict GDM on retrospective data of
588,622 pregnancies in Israel for which comprehensive electronic health records
were available. Our models predict GDM with high accuracy even at pregnancy
initiation (area under the receiver operating curve (auROC) = 0.85),
substantially outperforming a baseline risk score (auROC = 0.68). We validated
our results on both a future validation set and a geographical validation set
from the most populated city in Israel, Jerusalem, thereby emulating real-world
performance. Interrogating our model, we uncovered previously unreported risk
factors, including results of previous pregnancy glucose challenge tests.
Finally, we devised a simpler model based on just nine questions that a patient
could answer, with only a modest reduction in accuracy (auROC = 0.80). Overall,
our models may allow early-stage intervention in high-risk women, as well as a
cost-effective screening approach that could avoid the need for glucose
tolerance tests by identifying low-risk women. Future prospective studies and
studies on additional populations are needed to assess the real-world clinical
utility of the model.
DOI: 10.1038/s41591-019-0724-8
PMID: 31932807 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/21958955 | 1. J Clin Densitom. 2011 Oct-Dec;14(4):407-15. doi: 10.1016/j.jocd.2011.06.006.
Epub 2011 Oct 1.
Methods for assessing fracture risk prediction models: experience with FRAX in a
large integrated health care delivery system.
Pressman AR(1), Lo JC, Chandra M, Ettinger B.
Author information:
(1)Division of Research, Kaiser Permanente Northern California, Oakland, CA,
USA. [email protected]
Area under the receiver operating characteristics (AUROC) curve is often used to
evaluate risk models. However, reclassification tests provide an alternative
assessment of model performance. We performed both evaluations on results from
FRAX (World Health Organization Collaborating Centre for Metabolic Bone
Diseases, University of Sheffield, UK), a fracture risk tool, using Kaiser
Permanente Northern California women older than 50yr with bone mineral density
(BMD) measured during 1997-2003. We compared FRAX performance with and without
BMD in the model. Among 94,489 women with mean follow-up of 6.6yr, 1579 (1.7%)
sustained a hip fracture. Overall, AUROCs were 0.83 and 0.84 for FRAX without
and with BMD, suggesting that BMD did not contribute to model performance. AUROC
decreased with increasing age, and BMD contributed significantly to higher AUROC
among those aged 70yr and older. Using an 81% sensitivity threshold (optimum
level from receiver operating characteristic curve, corresponding to 1.2%
cutoff), 35% of those categorized above were reassigned below when BMD was
added. In contrast, only 10% of those categorized below were reassigned to the
higher risk category when BMD was added. The net reclassification improvement
was 5.5% (p<0.01). Two versions of this risk tool have similar AUROCs, but
alternative assessments indicate that addition of BMD improves performance.
Multiple methods should be used to evaluate risk tool performance with less
reliance on AUROC alone.
Copyright © 2011 The International Society for Clinical Densitometry. Published
by Elsevier Inc. All rights reserved.
DOI: 10.1016/j.jocd.2011.06.006
PMID: 21958955 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/17309939 | 1. Circulation. 2007 Feb 20;115(7):928-35. doi:
10.1161/CIRCULATIONAHA.106.672402.
Use and misuse of the receiver operating characteristic curve in risk
prediction.
Cook NR(1).
Author information:
(1)Division of Preventive Medicine, Brigham and Women's Hospital, 900
Commonwealth Ave East, Boston, MA 02215, USA. [email protected]
Comment in
Circulation. 2007 Aug 7;116(6):e132; author reply e134. doi:
10.1161/CIRCULATIONAHA.107.709253.
Circulation. 2007 Aug 7;116(6):e133; author reply e134. doi:
10.1161/CIRCULATIONAHA.107.714360.
The c statistic, or area under the receiver operating characteristic (ROC)
curve, achieved popularity in diagnostic testing, in which the test
characteristics of sensitivity and specificity are relevant to discriminating
diseased versus nondiseased patients. The c statistic, however, may not be
optimal in assessing models that predict future risk or stratify individuals
into risk categories. In this setting, calibration is as important to the
accurate assessment of risk. For example, a biomarker with an odds ratio of 3
may have little effect on the c statistic, yet an increased level could shift
estimated 10-year cardiovascular risk for an individual patient from 8% to 24%,
which would lead to different treatment recommendations under current Adult
Treatment Panel III guidelines. Accepted risk factors such as lipids,
hypertension, and smoking have only marginal impact on the c statistic
individually yet lead to more accurate reclassification of large proportions of
patients into higher-risk or lower-risk categories. Perfectly calibrated models
for complex disease can, in fact, only achieve values for the c statistic well
below the theoretical maximum of 1. Use of the c statistic for model selection
could thus naively eliminate established risk factors from cardiovascular risk
prediction scores. As novel risk factors are discovered, sole reliance on the c
statistic to evaluate their utility as risk predictors thus seems ill-advised.
DOI: 10.1161/CIRCULATIONAHA.106.672402
PMID: 17309939 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/26592354 | 1. J Hepatol. 2016 Feb;64(2):390-398. doi: 10.1016/j.jhep.2015.11.008. Epub 2015
Dec 1.
FibroGENE: A gene-based model for staging liver fibrosis.
Eslam M(1), Hashem AM(2), Romero-Gomez M(3), Berg T(4), Dore GJ(5), Mangia A(6),
Chan HLY(7), Irving WL(8), Sheridan D(9), Abate ML(10), Adams LA(11), Weltman
M(12), Bugianesi E(10), Spengler U(13), Shaker O(14), Fischer J(15), Mollison
L(16), Cheng W(17), Nattermann J(13), Riordan S(18), Miele L(19), Kelaeng KS(1),
Ampuero J(3), Ahlenstiel G(1), McLeod D(20), Powell E(21), Liddle C(1), Douglas
MW(1), Booth DR(22), George J(23); International Liver Disease Genetics
Consortium (ILDGC).
Author information:
(1)Storr Liver Centre, The Westmead Millennium Institute for Medical Research
and Westmead Hospital, The University of Sydney, NSW, Australia.
(2)Department of Systems and Biomedical Engineering, Faculty of Engineering,
Minia University, Minia, Egypt.
(3)Unit for The Clinical Management of Digestive Diseases and CIBERehd, Hospital
Universitario de Valme, Sevilla, Spain.
(4)Medizinische Klinik m.S. Hepatologie und Gastroenterologie, Charite, Campus
Virchow-Klinikum, Universitätsmedizin Berlin, Germany; Department of Hepatology,
Clinic for Gastroenterology and Rheumatology, University Clinic Leipzig,
Leipzig, Germany.
(5)Kirby Institute, The University of New South Wales, Sydney, NSW, Australia;
St Vincent's Hospital, Sydney, NSW, Australia.
(6)Division of Hepatology, Ospedale Casa Sollievo della Sofferenza, IRCCS, San
Giovanni Rotondo, Italy.
(7)Department of Medicine and Therapeutics, Prince of Wales Hospital, The
Chinese University of Hong Kong, Hong Kong, China.
(8)NIHR Biomedical Research Unit in Gastroenterology and the Liver, University
of Nottingham, Nottingham, United Kingdom.
(9)Liver Research Group, Institute of Cellular Medicine, Medical School,
Newcastle University, Newcastle upon Tyne, United Kingdom; Institute of
Translational and Stratified Medicine, Plymouth University, United Kingdom.
(10)Division of Gastroenterology and Hepatology, Department of Medical Science,
University of Turin, Turin, Italy.
(11)School of Medicine and Pharmacology, Sir Charles Gairdner Hospital Unit,
University of Western Australia, Nedlands, WA, Australia.
(12)Department of Gastroenterology and Hepatology, Nepean Hospital, Sydney, NSW,
Australia.
(13)Department of Internal Medicine I, University of Bonn, Bonn, Germany.
(14)Medical Biochemistry and Molecular Biology Department, Faculty of Medicine,
Cairo University, Cairo, Egypt.
(15)Medizinische Klinik m.S. Hepatologie und Gastroenterologie, Charite, Campus
Virchow-Klinikum, Universitätsmedizin Berlin, Germany.
(16)School of Medicine and Pharmacology, Fremantle Hospital, UWA, Fremantle, WA,
Australia.
(17)Department of Gastroenterology & Hepatology, Royal Perth Hospital, WA,
Australia.
(18)Gastrointestinal and Liver Unit, Prince of Wales Hospital and University of
New South Wales, Sydney, NSW, Australia.
(19)Department of Internal Medicine, Catholic University of the Sacred Heart,
Rome, Italy.
(20)Department of Anatomical Pathology, Institute of Clinical Pathology and
Medical Research (ICPMR), Westmead Hospital, Sydney, Australia.
(21)Princess Alexandra Hospital, Department of Gastroenterology and Hepatology,
Woolloongabba, QLD, Australia; The University of Queensland, School of Medicine,
Princess Alexandra Hospital, Woolloongabba, QLD, Australia.
(22)Institute of Immunology and Allergy Research, Westmead Hospital and Westmead
Millennium Institute, University of Sydney, NSW, Australia.
(23)Storr Liver Centre, The Westmead Millennium Institute for Medical Research
and Westmead Hospital, The University of Sydney, NSW, Australia. Electronic
address: [email protected].
BACKGROUND & AIMS: The extent of liver fibrosis predicts long-term outcomes, and
hence impacts management and therapy. We developed a non-invasive algorithm to
stage fibrosis using non-parametric, machine learning methods designed for
predictive modeling, and incorporated an invariant genetic marker of liver
fibrosis risk.
METHODS: Of 4277 patients with chronic liver disease, 1992 with chronic
hepatitis C (derivation cohort) were analyzed to develop the model, and
subsequently validated in an independent cohort of 1242 patients. The model was
assessed in cohorts with chronic hepatitis B (CHB) (n=555) and non-alcoholic
fatty liver disease (NAFLD) (n=488). Model performance was compared to FIB-4 and
APRI, and also to the NAFLD fibrosis score (NFS) and Forns' index, in those with
NAFLD.
RESULTS: Significant fibrosis (⩾F2) was similar in the derivation (48.4%) and
validation (47.4%) cohorts. The FibroGENE-DT yielded the area under the receiver
operating characteristic curve (AUROCs) of 0.87, 0.85 and 0.804 for the
prediction of fast fibrosis progression, cirrhosis and significant fibrosis
risk, respectively, with comparable results in the validation cohort. The model
performed well in NAFLD and CHB with AUROCs of 0.791, and 0.726, respectively.
The negative predictive value to exclude cirrhosis was>0.96 in all three liver
diseases. The AUROC of the FibroGENE-DT performed better than FIB-4, APRI, and
NFS and Forns' index in most comparisons.
CONCLUSION: A non-invasive decision tree model can predict liver fibrosis risk
and aid decision making.
Copyright © 2015 European Association for the Study of the Liver. Published by
Elsevier B.V. All rights reserved.
DOI: 10.1016/j.jhep.2015.11.008
PMID: 26592354 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/28860560 | 1. Sci Rep. 2017 Aug 31;7(1):10155. doi: 10.1038/s41598-017-10408-9.
Effect of non-linearity of a predictor on the shape and magnitude of its
receiver-operating-characteristic curve in predicting a binary outcome.
Ho KM(1)(2)(3).
Author information:
(1)Department of Intensive Care Medicine, Royal Perth Hospital, Perth,
Australia. [email protected].
(2)School of Population Health, University of Western Australia, Perth,
Australia. [email protected].
(3)School of Veterinary & Life Science, Murdoch University, Perth, Australia.
[email protected].
Area under a receiver-operating-characteristic (AUROC) curve is widely used in
medicine to summarize the ability of a continuous predictive marker to predict a
binary outcome. This study illustrated how a U-shaped or inverted U-shaped
continuous predictor would affect the shape and magnitude of its AUROC curve in
predicting a binary outcome by comparing the ROC curves of the worst first
24-hour arterial pH values of 9549 consecutive critically ill patients in
predicting hospital mortality before and after centering the predictor by its
mean or median. A simulation dataset with an inverted U-shaped predictor was
used to assess how this would affect the shape and magnitude of the AUROC curve.
An asymmetrical U-shaped relationship between pH and hospital mortality,
resulting in an inverse-sigmoidal ROC curve, was observed. The AUROC
substantially increased after centering the predictor by its mean (0.611 vs
0.722, difference = 0.111, 95% confidence interval [CI] 0.087-0.135), and was
further improved after centering by its median (0.611 vs 0.745,
difference = 0.133, 95%CI 0.110-0.157). A sigmoidal-shaped ROC curve was
observed for an inverted U-shaped predictor. In summary, a non-linear predictor
can result in a biphasic-shaped ROC curve; and centering the predictor can
reduce its bias towards null predictive ability.
DOI: 10.1038/s41598-017-10408-9
PMCID: PMC5578972
PMID: 28860560
Conflict of interest statement: The authors declare that they have no competing
interests. |
http://www.ncbi.nlm.nih.gov/pubmed/31094792 | 1. Anesth Analg. 2019 Jun;128(6):1225-1233. doi: 10.1213/ANE.0000000000003701.
Factors Associated With Recovery Room Intravenous Opiate Requirement After
Pediatric Outpatient Operations.
Nafiu OO(1), Thompson A, Chiravuri SD, Cloyd B, Reynolds PI.
Author information:
(1)From the Department of Anesthesiology, Section of Pediatric Anesthesiology,
University of Michigan, Ann Arbor, Michigan.
BACKGROUND: Many children recovering from anesthesia experience pain that is
severe enough to warrant intravenous (IV) opioid treatment within moments of
admission to the postanesthesia care unit (PACU). Postoperative pain has several
negative consequences; therefore, preventing significant PACU pain in children
is both a major clinical goal and a moral/ethical imperative. This requires
identifying patient-level and perioperative factors that may be used to predict
PACU IV opioid requirement. This should allow for the development of
personalized care protocols to prevent clinically significant PACU pain in
children. Our objective was to develop prediction models enabling practitioners
to identify children at risk for PACU IV opioid requirement after various
painful ambulatory surgical procedures.
METHODS: After Institutional Review Board approval, clinical, demographic, and
anthropometric data were prospectively collected on 1256 children 4-17 years of
age scheduled for painful ambulatory surgery (defined as intraoperative
administration of analgesia or local anesthetic infiltration). Three
multivariable logistic regression models to determine possible predictors of
PACU IV opioid requirement were constructed based on (1) preoperative history;
(2) history + intraoperative variables; and (3) history + intraoperative
variables + PACU variables. Candidate predictors were chosen from readily
obtainable parameters routinely collected during the surgical visit. Predictive
performance of each model was assessed by calculating the area under the
respective receiver operating characteristic curves.
RESULTS: Overall, 29.5% of patients required a PACU IV opioid, while total PACU
analgesia requirement (oral or IV) was 41.1%. Independent predictors using
history alone were female sex, decreasing age, surgical history, and
non-Caucasian ethnicity (model area under the receiver operating characteristic
curve [AUROC], 0.59 [95% confidence interval {CI}, 0.55-0.63]). Adding a few
intraoperative variables improved the discriminant ability of the model (AUROC
for the history + intraoperative variables model, 0.71 [95% CI, 0.67-0.74]).
Addition of first-documented PACU pain score produced a substantially improved
model (AUROC, 0.85 [95% CI, 0.82-0.87]).
CONCLUSIONS: Postoperative pain requiring PACU IV opioid in children may be
determined using a small set of easily obtainable perioperative variables. Our
models require validation in other settings to determine their clinical
usefulness.
DOI: 10.1213/ANE.0000000000003701
PMID: 31094792 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/33088899 | 1. Int J Cardiol Heart Vasc. 2020 Oct 12;31:100641. doi:
10.1016/j.ijcha.2020.100641. eCollection 2020 Dec.
Differential prognostic accuracy of right ventricular dysfunction, the Seattle
heart failure model and the MAGGIC score in patients with severe mitral
regurgitation undergoing the MitraClip® procedure.
Heyl S(1), Luu B(1), Wieszner M(1), Nikkhoo A(1), Seeger F(1), Hemmann K(1),
Assmus B(1), Kaess B(1), Zeiher AM(1), Walther C(1), Fichtlscherer S(1), Honold
J(1).
Author information:
(1)University Hospital Frankfurt, Department of Cardiology, Germany.
BACKGROUND: MitraClip ® (MC) is an established procedure for severe mitral
regurgitation (MR) in patients deemed unsuitable for surgery.Right ventricular
dysfunction (RVD) is associated with a higher mortality risk. The prognostic
accuracy of heart failure risk scores like the Seattle heart failure model
(SHFM) and Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) score in
pts undergoing MC with or without RVD has not been investigated so far.
METHODS: SHFM and MAGGIC score were calculated retrospectively. RVD was
determined as tricuspid annular plane systolic excursion (TAPSE) ≤15 mm. Area
under receiver operating curves (AUROC) of SHFM and MAGGIC were performed for
one-year all-cause mortality after MC.
RESULTS: N = 103 pts with MR III° (73 ± 11 years, LVEF 37 ± 17%) underwent MC
with a reduction of at least I° MR. One-year mortality was 28.2%.In Kaplan-Meier
analysis, one- year mortality was significantly higher in RVD-pts (34.8% vs
2.8%, p = 0.009).Area under the Receiver Operating Characteristic (AUROC) for
SHFM and MAGGIC were comparable for both scores (SHFM: 0.704, MAGGIC: 0.692). In
pts without RVD, SHFM displayed a higher AUROC and therefore better diagnostic
accuracy (SHFM: 0.776; MAGGIC: 0.551, p < 0.05). In pts with RVD, MAGGIC and
SHFM displayed comparable AUROCs.
CONCLUSION: RVD is an important prognostic marker in pts undergoing MC. SHFM and
MAGGIC displayed adequate over-all prognostic power in these pts. Accuracy
differed in pts with and without RVD, indicating higher predictive power of the
SHFM score in pts without RVD.
© 2020 Published by Elsevier B.V.
DOI: 10.1016/j.ijcha.2020.100641
PMCID: PMC7566949
PMID: 33088899
Conflict of interest statement: The authors report no relationships that could
be construed as a conflict of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/23620757 | 1. PLoS One. 2013 Apr 19;8(4):e61468. doi: 10.1371/journal.pone.0061468. Print
2013.
Pharmacointeraction network models predict unknown drug-drug interactions.
Cami A(1), Manzi S, Arnold A, Reis BY.
Author information:
(1)Division of Emergency Medicine, Boston Children's Hospital, Boston,
Massachusetts, USA. [email protected]
Drug-drug interactions (DDIs) can lead to serious and potentially lethal adverse
events. In recent years, several drugs have been withdrawn from the market due
to interaction-related adverse events (AEs). Current methods for detecting DDIs
rely on the accumulation of sufficient clinical evidence in the post-market
stage - a lengthy process that often takes years, during which time numerous
patients may suffer from the adverse effects of the DDI. Detection methods are
further hindered by the extremely large combinatoric space of possible
drug-drug-AE combinations. There is therefore a practical need for predictive
tools that can identify potential DDIs years in advance, enabling drug safety
professionals to better prioritize their limited investigative resources and
take appropriate regulatory action. To meet this need, we describe Predictive
Pharmacointeraction Networks (PPINs) - a novel approach that predicts unknown
DDIs by exploiting the network structure of all known DDIs, together with other
intrinsic and taxonomic properties of drugs and AEs. We constructed an 856-drug
DDI network from a 2009 snapshot of a widely-used drug safety database, and used
it to develop PPIN models for predicting future DDIs. We compared the DDIs
predicted based solely on these 2009 data, with newly reported DDIs that
appeared in a 2012 snapshot of the same database. Using a standard multivariate
approach to combine predictors, the PPIN model achieved an AUROC (area under the
receiver operating characteristic curve) of 0.81 with a sensitivity of 48% given
a specificity of 90%. An analysis of DDIs by severity level revealed that the
model was most effective for predicting "contraindicated" DDIs (AUROC = 0.92)
and less effective for "minor" DDIs (AUROC = 0.63). These results indicate that
network based methods can be useful for predicting unknown drug-drug
interactions.
DOI: 10.1371/journal.pone.0061468
PMCID: PMC3631217
PMID: 23620757 [Indexed for MEDLINE]
Conflict of interest statement: Competing Interests: The authors have declared
that no competing interests exist. |
http://www.ncbi.nlm.nih.gov/pubmed/34154565 | 1. BMC Pediatr. 2021 Jun 21;21(1):287. doi: 10.1186/s12887-021-02732-x.
Risk of Mycoplasma pneumoniae-related hepatitis in MP pneumonia pediatric
patients: a predictive model construction and assessment.
Bi Y(1), Ma Y(2), Zhuo J(3), Zhang L(4), Yin L(2), Sheng H(2), Luan J(2), Li
T(5).
Author information:
(1)Department of Pediatrics, Shandong Provincial Maternal and Child Health Care
Hospital, Cheeloo College of Medicine, Shandong University, 238#, Jing 10 East
Road, Jinan, Shandong, China.
(2)Department of Pediatrics, Cheeloo College of Medicine, Shandong Provincial
Third Hospital, Shandong University, 12#, Wuyingshan Middle Road, Jinan, 250014,
Shandong, China.
(3)Department of Clinical Laboratory, Cheeloo College of Medicine, Shandong
Provincial Third Hospital, Shandong University, 12#, Wuyingshan Middle Road,
Jinan, 250014, Shandong, China.
(4)Department of Pediatrics, Affiliated Hospital of Yangzhou University, 45#,
Taizhou Road, Yangzhou, Jiangsu, China.
(5)Department of Infectious Diseases, Shandong Provincial Hospital Affiliated to
Shandong First Medical University, 324#, Jing 5 Road, 250021, Jinan, China.
[email protected].
BACKGROUND: A predictive model for risk of Mycoplasma pneumoniae (MP)-related
hepatitis in MP pneumonia pediatric patients can improve treatment selection and
therapeutic effect. However, currently, no predictive model is available.
METHODS: Three hundred seventy-four pneumonia pediatric patients with/without
serologically-confirmed MP infection and ninety-three health controls were
enrolled. Logistic regressions were performed to identify the determinant
variables and develop predictive model. Predictive performance and optimal
diagnostic threshold were evaluated using area under the receiver operating
characteristic curve (AUROC). Stratification analysis by age and MP-IgM titer
was used to optimize model's clinical utility. An external validation set,
including 84 MP pneumonia pediatric patients, was used to verify the predictive
efficiency. After univariate analysis to screen significant variables, monocyte
count (MO), erythrocyte distribution width (RDW) and platelet count (PLT) were
identified as independent predictors in multivariate analysis.
RESULTS: We constructed MRP model: MO [^109/L] × 4 + RDW [%] - PLT [^109/L] ×
0.01. MRP achieved an AUROC of 0.754 and the sensitivity and specificity at
cut-off value 10.44 were 71.72 and 61.00 %, respectively in predicting
MP-related hepatitis from MP pneumonia. These results were verified by the
external validation set, whereas it merely achieved an AUROC of 0.540 in
pneumonia without MP infection. The AUROC of MRP was 0.812 and 0.787 in infants
and toddlers (0-36 months) and low MP-IgM titer subgroup (1:160-1:320),
respectively. It can achieve an AUROC of 0.804 in infants and toddler with low
MP-IgM titer subgroup.
CONCLUSIONS: MRP is an effective predictive model for risk of MP-related
hepatitis in MP pneumonia pediatric patients, especially infants and toddlers
with low MP-IgM titer.
DOI: 10.1186/s12887-021-02732-x
PMCID: PMC8218438
PMID: 34154565 [Indexed for MEDLINE]
Conflict of interest statement: The authors declare that they have no competing
interests. |
http://www.ncbi.nlm.nih.gov/pubmed/28323524 | 1. J Neurotrauma. 2017 Jul 15;34(14):2235-2242. doi: 10.1089/neu.2016.4606. Epub
2017 Apr 26.
Comparison of Two Predictive Models for Short-Term Mortality in Patients after
Severe Traumatic Brain Injury.
Kesmarky K(1), Delhumeau C(1), Zenobi M(1), Walder B(1).
Author information:
(1)Department of Anesthesiology, Intensive Care and Clinical Pharmacology,
University Hospitals of Geneva , Geneva, Switzerland .
The Glasgow Coma Scale (GCS) and the Abbreviated Injury Score of the head region
(HAIS) are validated prognostic factors in traumatic brain injury (TBI). The aim
of this study was to compare the prognostic performance of an alternative
predictive model including motor GCS, pupillary reactivity, age, HAIS, and
presence of multi-trauma for short-term mortality with a reference predictive
model including motor GCS, pupil reaction, and age (IMPACT core model). A
secondary analysis of a prospective epidemiological cohort study in Switzerland
including patients after severe TBI (HAIS >3) with the outcome death at 14 days
was performed. Performance of prediction, accuracy of discrimination (area under
the receiver operating characteristic curve [AUROC]), calibration, and validity
of the two predictive models were investigated. The cohort included 808 patients
(median age, 56; interquartile range, 33-71), median GCS at hospital admission 3
(3-14), abnormal pupil reaction 29%, with a death rate of 29.7% at 14 days. The
alternative predictive model had a higher accuracy of discrimination to predict
death at 14 days than the reference predictive model (AUROC 0.852, 95%
confidence interval [CI] 0.824-0.880 vs. AUROC 0.826, 95% CI 0.795-0.857;
p < 0.0001). The alternative predictive model had an equivalent calibration,
compared with the reference predictive model Hosmer-Lemeshow p values (Chi2
8.52, Hosmer-Lemeshow p = 0.345 vs. Chi2 8.66, Hosmer-Lemeshow p = 0.372). The
optimism-corrected value of AUROC for the alternative predictive model was
0.845. After severe TBI, a higher performance of prediction for short-term
mortality was observed with the alternative predictive model, compared with the
reference predictive model.
DOI: 10.1089/neu.2016.4606
PMID: 28323524 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/36329470 | 1. BMC Endocr Disord. 2022 Nov 4;22(1):269. doi: 10.1186/s12902-022-01186-1.
Exploring risk factors for cervical lymph node metastasis in papillary thyroid
microcarcinoma: construction of a novel population-based predictive model.
Huang Y(#)(1)(2), Mao Y(#)(1)(3), Xu L(#)(1), Wen J(1)(4), Chen G(5)(6).
Author information:
(1)Shengli Clinical Medical College of Fujian Medical University, Fuzhou, China.
(2)Department of Endocrinology, Zhongshan Hospital Xiamen University, Xiamen,
China.
(3)Department of Internal Medicine, Fujian Provincial Hospital Jinshan Branch,
Shengli Clinical Medical College of Fujian Medical University, Fuzhou, China.
(4)Department of Endocrinology, Fujian Provincial Hospital, Shengli Clinical
Medical College of Fujian Medical University, Fuzhou, 350001, China.
(5)Shengli Clinical Medical College of Fujian Medical University, Fuzhou, China.
[email protected].
(6)Department of Endocrinology, Fujian Provincial Hospital, Shengli Clinical
Medical College of Fujian Medical University, Fuzhou, 350001, China.
[email protected].
(#)Contributed equally
BACKGROUND: Machine learning was a highly effective tool in model construction.
We aim to establish a machine learning-based predictive model for predicting the
cervical lymph node metastasis (LNM) in papillary thyroid microcarcinoma (PTMC).
METHODS: We obtained data on PTMC from the SEER database, including 10
demographic and clinicopathological characteristics. Univariate and multivariate
logistic regression (LR) analyses were applied to screen the risk factors for
cervical LNM in PTMC. Risk factors with P < 0.05 in multivariate LR analysis
were used as modeling variables. Five different machine learning (ML) algorithms
including extreme gradient boosting (XGBoost), random forest (RF), adaptive
boosting (AdaBoost), gaussian naive bayes (GNB) and multi-layer perceptron (MLP)
and traditional regression analysis were used to construct the prediction model.
Finally, the area under the receiver operating characteristic (AUROC) curve was
used to compare the model performance.
RESULTS: Through univariate and multivariate LR analysis, we screened out 9
independent risk factors most closely associated with cervical LNM in PTMC,
including age, sex, race, marital status, region, histology, tumor size, and
extrathyroidal extension (ETE) and multifocality. We used these risk factors to
build an ML prediction model, in which the AUROC value of the XGBoost algorithm
was higher than the other 4 ML algorithms and was the best ML model. We
optimized the XGBoost algorithm through 10-fold cross-validation, and its best
performance on the training set (AUROC: 0.809, 95%CI 0.800-0.818) was better
than traditional LR analysis (AUROC: 0.780, 95%CI 0.772-0.787).
CONCLUSIONS: ML algorithms have good predictive performance, especially the
XGBoost algorithm. With the continuous development of artificial intelligence,
ML algorithms have broad prospects in clinical prognosis prediction.
© 2022. The Author(s).
DOI: 10.1186/s12902-022-01186-1
PMCID: PMC9635156
PMID: 36329470 [Indexed for MEDLINE]
Conflict of interest statement: The authors declare that they have no competing
interests. |
http://www.ncbi.nlm.nih.gov/pubmed/31509205 | 1. JAMA Netw Open. 2019 Sep 4;2(9):e1910967. doi:
10.1001/jamanetworkopen.2019.10967.
Prevalence and Predictability of Low-Yield Inpatient Laboratory Diagnostic
Tests.
Xu S(1), Hom J(2), Balasubramanian S(1), Schroeder LF(3), Najafi N(4), Roy S(5),
Chen JH(1)(2).
Author information:
(1)Center for Biomedical Informatics Research, Department of Medicine, Stanford
University, Stanford, California.
(2)Division of Hospital Medicine, Department of Medicine, Stanford University,
Stanford, California.
(3)Department of Pathology, University of Michigan School of Medicine, Ann
Arbor.
(4)Department of Medicine, University of California, San Francisco.
(5)Department of Computer Science, Stanford University, Stanford, California.
Erratum in
JAMA Netw Open. 2019 Oct 2;2(10):e1914190. doi:
10.1001/jamanetworkopen.2019.14190.
IMPORTANCE: Laboratory testing is an important target for high-value care
initiatives, constituting the highest volume of medical procedures. Prior
studies have found that up to half of all inpatient laboratory tests may be
medically unnecessary, but a systematic method to identify these unnecessary
tests in individual cases is lacking.
OBJECTIVE: To systematically identify low-yield inpatient laboratory testing
through personalized predictions.
DESIGN, SETTING, AND PARTICIPANTS: In this retrospective diagnostic study with
multivariable prediction models, 116 637 inpatients treated at Stanford
University Hospital from January 1, 2008, to December 31, 2017, a total of
60 929 inpatients treated at University of Michigan from January 1, 2015, to
December 31, 2018, and 13 940 inpatients treated at the University of
California, San Francisco from January 1 to December 31, 2018, were assessed.
MAIN OUTCOMES AND MEASURES: Diagnostic accuracy measures, including sensitivity,
specificity, negative predictive values (NPVs), positive predictive values
(PPVs), and area under the receiver operating characteristic curve (AUROC), of
machine learning models when predicting whether inpatient laboratory tests yield
a normal result as defined by local laboratory reference ranges.
RESULTS: In the recent data sets (July 1, 2014, to June 30, 2017) from Stanford
University Hospital (including 22 664 female inpatients with a mean [SD] age of
58.8 [19.0] years and 22 016 male inpatients with a mean [SD] age of 59.0 [18.1]
years), among the top 20 highest-volume tests, 792 397 were repeats of orders
within 24 hours, including tests that are physiologically unlikely to yield new
information that quickly (eg, white blood cell differential, glycated
hemoglobin, and serum albumin level). The best-performing machine learning
models predicted normal results with an AUROC of 0.90 or greater for 12
stand-alone laboratory tests (eg, sodium AUROC, 0.92 [95% CI, 0.91-0.93];
sensitivity, 98%; specificity, 35%; PPV, 66%; NPV, 93%; lactate dehydrogenase
AUROC, 0.93 [95% CI, 0.93-0.94]; sensitivity, 96%; specificity, 65%; PPV, 71%;
NPV, 95%; and troponin I AUROC, 0.92 [95% CI, 0.91-0.93]; sensitivity, 88%;
specificity, 79%; PPV, 67%; NPV, 93%) and 10 common laboratory test components
(eg, hemoglobin AUROC, 0.94 [95% CI, 0.92-0.95]; sensitivity, 99%; specificity,
17%; PPV, 90%; NPV, 81%; creatinine AUROC, 0.96 [95% CI, 0.96-0.97];
sensitivity, 93%; specificity, 83%; PPV, 79%; NPV, 94%; and urea nitrogen AUROC,
0.95 [95% CI, 0.94, 0.96]; sensitivity, 87%; specificity, 89%; PPV, 77%; NPV
94%).
CONCLUSIONS AND RELEVANCE: The findings suggest that low-yield diagnostic
testing is common and can be systematically identified through data-driven
methods and patient context-aware predictions. Implementing machine learning
models appear to be able to quantify the level of uncertainty and expected
information gained from diagnostic tests explicitly, with the potential to
encourage useful testing and discourage low-value testing that incurs direct
costs and indirect harms.
DOI: 10.1001/jamanetworkopen.2019.10967
PMCID: PMC6739729
PMID: 31509205 [Indexed for MEDLINE]
Conflict of interest statement: Conflict of Interest Disclosures: Dr Chen
reported receiving grants from the National Institute of Environmental Health
Sciences and the Gordon and Betty Moore Foundation during the conduct of the
study and having co-ownership of Reaction Explorer LLC (chemistry education
software company). No other disclosures were reported. |
http://www.ncbi.nlm.nih.gov/pubmed/25885260 | 1. PLoS One. 2015 Apr 17;10(4):e0122929. doi: 10.1371/journal.pone.0122929.
eCollection 2015.
Predicting outcome on admission and post-admission for acetaminophen-induced
acute liver failure using classification and regression tree models.
Speiser JL(1), Lee WM(2), Karvellas CJ(3); US Acute Liver Failure Study Group.
Collaborators: Lee WM, Larson AM, Liou I, Davern T, Fix O, Schilsky M,
McCashland T, Hay J, Murray N, Shaikh A, Blei A, Ganger D, Zaman A, Han SH,
Fontana R, McGuire B, Chung RT, Smith A, Brown R, Crippin J, Harrison E, Reuben
A, Munoz S, Reddy R, Stravitz R, Rossaro L, Satyanarayana R, Hassanein T, Hanje
J, Olson J, Subramanian R, Karvellas CJ, Samuel G, Lalani E, Pezzia C, Sanders
C, Attar N, Hynan LS, Durkalski V, Zhao W, Speiser J, Dillon C, Battenhouse H,
Gottfried M.
Author information:
(1)Department of Public Health Sciences, Medical University of South Carolina,
Charleston, South Carolina, United States of America.
(2)Division of Digestive and Liver Diseases, Department of Internal Medicine,
University of Texas Southwestern Medical Center, Dallas, Texas, United States of
America.
(3)Divisions of Hepatology and Critical Care Medicine, University of Alberta,
Edmonton, Canada.
BACKGROUND/AIM: Assessing prognosis for acetaminophen-induced acute liver
failure (APAP-ALF) patients often presents significant challenges. King's
College (KCC) has been validated on hospital admission, but little has been
published on later phases of illness. We aimed to improve determinations of
prognosis both at the time of and following admission for APAP-ALF using
Classification and Regression Tree (CART) models.
METHODS: CART models were applied to US ALFSG registry data to predict 21-day
death or liver transplant early (on admission) and post-admission (days 3-7) for
803 APAP-ALF patients enrolled 01/1998-09/2013. Accuracy in prediction of
outcome (AC), sensitivity (SN), specificity (SP), and area under
receiver-operating curve (AUROC) were compared between 3 models: KCC (INR,
creatinine, coma grade, pH), CART analysis using only KCC variables (KCC-CART)
and a CART model using new variables (NEW-CART).
RESULTS: Traditional KCC yielded 69% AC, 90% SP, 27% SN, and 0.58 AUROC on
admission, with similar performance post-admission. KCC-CART at admission
offered predictive 66% AC, 65% SP, 67% SN, and 0.74 AUROC. Post-admission,
KCC-CART had predictive 82% AC, 86% SP, 46% SN and 0.81 AUROC. NEW-CART models
using MELD (Model for end stage liver disease), lactate and mechanical
ventilation on admission yielded predictive 72% AC, 71% SP, 77% SN and AUROC
0.79. For later stages, NEW-CART (MELD, lactate, coma grade) offered predictive
AC 86%, SP 91%, SN 46%, AUROC 0.73.
CONCLUSION: CARTs offer simple prognostic models for APAP-ALF patients, which
have higher AUROC and SN than KCC, with similar AC and negligibly worse SP.
Admission and post-admission predictions were developed.
KEY POINTS: • Prognostication in acetaminophen-induced acute liver failure
(APAP-ALF) is challenging beyond admission • Little has been published regarding
the use of King's College Criteria (KCC) beyond admission and KCC has shown
limited sensitivity in subsequent studies • Classification and Regression Tree
(CART) methodology allows the development of predictive models using binary
splits and offers an intuitive method for predicting outcome, using processes
familiar to clinicians • Data from the ALFSG registry suggested that CART
prognosis models for the APAP population offer improved sensitivity and model
performance over traditional regression-based KCC, while maintaining similar
accuracy and negligibly worse specificity • KCC-CART models offered modest
improvement over traditional KCC, with NEW-CART models performing better than
KCC-CART particularly at late time points.
DOI: 10.1371/journal.pone.0122929
PMCID: PMC4401567
PMID: 25885260 [Indexed for MEDLINE]
Conflict of interest statement: Competing Interests: The authors have declared
that no competing interests exist. |
http://www.ncbi.nlm.nih.gov/pubmed/31967640 | 1. Int J Epidemiol. 2020 Aug 1;49(4):1397-1403. doi: 10.1093/ije/dyz274.
Reflection on modern methods: Revisiting the area under the ROC Curve.
Janssens ACJW(1)(2), Martens FK(2).
Author information:
(1)Department of Epidemiology, Rollins School of Public Health, Emory
University, Atlanta, GA, USA.
(2)Department of Clinical Genetics, Amsterdam Public Health Research Institute,
Amsterdam UMC, Amsterdam, The Netherlands.
The area under the receiver operating characteristic (ROC) curve (AUC) is
commonly used for assessing the discriminative ability of prediction models even
though the measure is criticized for being clinically irrelevant and lacking an
intuitive interpretation. Every tutorial explains how the coordinates of the ROC
curve are obtained from the risk distributions of diseased and non-diseased
individuals, but it has not become common sense that therewith the ROC plot is
just another way of presenting these risk distributions. We show how the ROC
curve is an alternative way to present risk distributions of diseased and
non-diseased individuals and how the shape of the ROC curve informs about the
overlap of the risk distributions. For example, ROC curves are rounded when the
prediction model included variables with similar effect on disease risk and have
an angle when, for example, one binary risk factor has a stronger effect; and
ROC curves are stepped rather than smooth when the sample size or incidence is
low, when the prediction model is based on a relatively small set of categorical
predictors. This alternative perspective on the ROC plot invalidates most
purported limitations of the AUC and attributes others to the underlying risk
distributions. AUC is a measure of the discriminative ability of prediction
models. The assessment of prediction models should be supplemented with other
metrics to assess their clinical utility.
© The Author(s) 2020; all rights reserved. Published by Oxford University Press
on behalf of the International Epidemiological Association.
DOI: 10.1093/ije/dyz274
PMID: 31967640 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/35879562 | 1. J Neurol. 2022 Dec;269(12):6377-6385. doi: 10.1007/s00415-022-11287-5. Epub
2022 Jul 25.
Plasma biomarkers inclusive of α-synuclein/amyloid-beta40 ratio strongly
correlate with Mini-Mental State Examination score in Parkinson's disease and
predict cognitive impairment.
Chan DKY(1)(2)(3), Chen J(4), Chen RF(5), Parikh J(6), Xu YH(4)(7)(6), Silburn
PA(8), Mellick GD(9).
Author information:
(1)University of New South Wales, Sydney, Australia. [email protected].
(2)NICM Health Research Institute, Western Sydney University, Sydney, Australia.
[email protected].
(3)Bankstown-Lidcombe Hospital, Eldridge Rd,, Bankstown, NSW, 2200, Australia.
[email protected].
(4)University of New South Wales, Sydney, Australia.
(5)Central Sydney Immunology Laboratory at Royal Prince Alfred Hospital, Sydney,
NSW, Australia.
(6)Bankstown-Lidcombe Hospital, Eldridge Rd,, Bankstown, NSW, 2200, Australia.
(7)NICM Health Research Institute, Western Sydney University, Sydney, Australia.
(8)Queensland Brain Institute, University of Queensland, Brisbane, QLD,
Australia.
(9)Griffith University, Brisbane, QLD, Australia.
Plasma biomarkers for Parkinson's disease (PD) diagnosis that carry predictive
value for cognitive impairment are valuable. We explored the relationship of
Mini-Mental State Examination (MMSE) score with plasma biomarkers in PD patients
and compared results to vascular dementia (VaD) and normal controls. The
predictive accuracy of an individual biomarker on cognitive impairment was
evaluated using area under the receiver operating characteristic curve (AUROC),
and multivariate logistic regression was applied to evaluate predictive accuracy
of biomarkers on cognitive impairment; 178 subjects (41 PD, 31 VaD and 106
normal controls) were included. In multiple linear regression analysis of PD
patients, α-synuclein, anti-α-synuclein, α-synuclein/Aβ40 and
anti-α-synuclein/Aβ40 were highly predictive of MMSE score in both full model
and parsimonious model (R2 = 0.838 and 0.835, respectively) compared to
non-significant results in VaD group (R2 = 0.149) and in normal controls
(R2 = 0.056). Α-synuclein and anti-α-synuclein/Aβ40 were positively associated
with MMSE score, and anti-α-synuclein, α-synuclein/Aβ40 were negatively
associated with the MMSE score among PD patients (all Ps < 0.005). In the AUROC
analysis, anti-α-synuclein (AUROC = 0.788) and anti-α-synuclein/Aβ40
(AUROC = 0.749) were significant individual predictors of cognitive impairment.
In multivariate logistic regression, full model of combined biomarkers showed
high accuracy in predicting cognitive impairment (AUROC = 0.890; 95%CI
0.796-0.984) for PD versus controls, as was parsimonious model (AUROC = 0.866;
95%CI 0.764-0.968). In conclusion, simple combination of biomarkers inclusive of
α-synuclein/Aβ40 strongly correlates with MMSE score in PD patients versus
controls and is highly predictive of cognitive impairment.
© 2022. The Author(s).
DOI: 10.1007/s00415-022-11287-5
PMCID: PMC9618522
PMID: 35879562 [Indexed for MEDLINE]
Conflict of interest statement: The authors declare that they have no conflict
of interest. |
http://www.ncbi.nlm.nih.gov/pubmed/34348142 | 1. Cell Rep. 2021 Aug 3;36(5):109464. doi: 10.1016/j.celrep.2021.109464.
Spliceosomal component PRP-40 is a central regulator of microexon splicing.
Choudhary B(1), Marx O(1), Norris AD(2).
Author information:
(1)Biological Sciences, Southern Methodist University, Dallas, TX 75275, USA.
(2)Biological Sciences, Southern Methodist University, Dallas, TX 75275, USA.
Electronic address: [email protected].
Microexons (≤27 nt) play critical roles in nervous system development and
function but create unique challenges for the splicing machinery. The mechanisms
of microexon regulation are therefore of great interest. We performed a genetic
screen for alternative splicing regulators in the C. elegans nervous system and
identify PRP-40, a core component of the U1 snRNP. RNA-seq reveals that PRP-40
is required for inclusion of alternatively spliced, but not constitutively
spliced, exons. PRP-40 is particularly required for inclusion of neuronal
microexons, and our data indicate that PRP-40 is a central regulator of
microexon splicing. Microexons can be relieved from PRP-40 dependence by
artificially increasing exon size or reducing flanking intron size, indicating
that PRP-40 is specifically required for microexons surrounded by conventionally
sized introns. Knockdown of the orthologous PRPF40A in mouse neuroblastoma cells
causes widespread dysregulation of microexons but not conventionally sized
exons. PRP-40 regulation of neuronal microexons is therefore a widely conserved
phenomenon.
Copyright © 2021 The Author(s). Published by Elsevier Inc. All rights reserved.
DOI: 10.1016/j.celrep.2021.109464
PMCID: PMC8378409
PMID: 34348142 [Indexed for MEDLINE]
Conflict of interest statement: Declaration of interests The authors declare no
competing interests. |
http://www.ncbi.nlm.nih.gov/pubmed/33861387 | 1. Drugs. 2021 May;81(7):875-879. doi: 10.1007/s40265-021-01512-2. Epub 2021 Apr
16.
Casimersen: First Approval.
Shirley M(1).
Author information:
(1)Springer Nature, Mairangi Bay, Private Bag 65901, Auckland, 0754, New
Zealand. [email protected].
Casimersen (Amondys 45™) is an antisense oligonucleotide of the
phosphorodiamidate morpholino oligomer subclass developed by Sarepta
Therapeutics for the treatment of Duchenne muscular dystrophy (DMD) in patients
who have a mutation in the DMD gene that is amenable to exon 45 skipping.
Administered by intravenous infusion, casimersen is designed to bind to exon 45
of the DMD gene pre-mRNA, resulting in skipping of this exon during mRNA
processing, intended to allow for production of an internally truncated but
functional dystrophin protein in patients with DMD. Casimersen received its
first approval on 25 February 2021, in the USA, for the treatment of DMD in
patients who have a confirmed mutation of the DMD gene that is amenable to
exon 45 skipping. The approval, granted under the US FDA Accelerated Approval
Program, was based on an observed increase in dystrophin production in skeletal
muscle in patients treated with casimersen. Casimersen is continuing in
phase III development for the treatment of DMD in several other countries
worldwide. This article summarises the milestones in the development of
casimersen leading to this first approval for DMD. As with other approvals under
the Accelerated Approval Program, continued approval for this indication may be
contingent upon verification of a clinical benefit in confirmatory trials.
DOI: 10.1007/s40265-021-01512-2
PMID: 33861387 [Indexed for MEDLINE] |
http://www.ncbi.nlm.nih.gov/pubmed/35651477 | 1. Drug Des Devel Ther. 2022 May 25;16:1547-1559. doi: 10.2147/DDDT.S358989.
eCollection 2022.
Designing a Dual GLP-1R/GIPR Agonist from Tirzepatide: Comparing Residues
Between Tirzepatide, GLP-1, and GIP.
Wang L(1).
Author information:
(1)College of Life Sciences and Technology, China Pharmaceutical University,
Nanjing, Jiangsu, People's Republic of China.
Improving type 2 diabetes using incretin analogues is becoming increasingly
plausible. Currently, tirzepatide is the most promising listed incretin
analogue. Here, I briefly explain the evolution of drugs of this kind, analyze
the residue discrepancies between tirzepatide and endogenous incretins,
summarize some existing strategies for prolonging half-life, and present
suggestions for future research, mainly involving biased functions. This review
aims to present some useful information for designing a dual glucagon like
peptide-1 receptor/glucose-dependent insulinotropic polypeptide receptor
agonist.
© 2022 Wang.
DOI: 10.2147/DDDT.S358989
PMCID: PMC9149770
PMID: 35651477 [Indexed for MEDLINE]
Conflict of interest statement: The author reports no conflicts of interest in
this work. |
End of preview. Expand
in Dataset Viewer.
README.md exists but content is empty.
Use the Edit dataset card button to edit it.
- Downloads last month
- 49