File size: 5,331 Bytes
7ca1baf 9ed253c 7ca1baf 3d17d70 ec2329f 28ec6a4 caa599e ec2329f f3fc250 caa599e f3fc250 ec2329f 355b4c2 99b0dd6 355b4c2 7d11710 355b4c2 99b0dd6 7d11710 ec2329f 355b4c2 ec2329f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
---
language:
- en
pipeline_tag: image-segmentation
tags:
- medical
---
# Endometrial cancer segmentation
This repository contains weights and exported learner (encapsulates both the model architecture and its trained parameters) for a deep learning model designed to automate the segmentation of endometrial cancer on MR images.
Our VIBE model utilizes a Residual U-Net architecture, trained on data derived from the study [Automated segmentation of endometrial cancer on MR images using deep learning](https://link.springer.com/content/pdf/10.1038/s41598-020-80068-9.pdf).
The primary objective of this repository is to reproduce the results reported in the study. In addition, we have looked at improving the segmentation performance using multi-sequence MR images (T2w, VIBE, and ADC) as reported in the study [Automatic segmentation of uterine endometrial cancer on multi-sequence MRI using a convolutional neural network](https://www.nature.com/articles/s41598-021-93792-7).
## Requirements
Last checked and validated with fastMONAI version 0.3.4. Please ensure that you have the correct version of fastMONAI installed to guarantee the correct operation of the model.
## Usage
The source code for training the model and running inference on your own data is available at: https://github.com/MMIV-ML/fastMONAI/tree/master/research/endometrial_cancer.
## Results for VIBE
The box plot of the predictions on the validation set:
The results from the validation set are also presented in the table below:
| subject_id | tumor_vol | inter_rater | r1_ml | r2_ml |
|---|---|---|---|---|
| 29 | 4.16 | 0.201835 | 0.806382 | 0.006231 |
| 32 | 8.00 | 0.684142 | 0.293306 | 0.209449 |
| 36 | 19.06 | 0.928750 | 0.793611 | 0.785335 |
| 47 | 11.01 | 0.944209 | 0.905159 | 0.902548 |
| 50 | 6.26 | 0.722867 | 0.619272 | 0.631579 |
| 65 | 13.09 | 0.930613 | 0.879279 | 0.850546 |
| 67 | 3.71 | 0.943498 | 0.887189 | 0.878163 |
| 75 | 7.16 | 0.263539 | 0.774411 | 0.266463 |
| 86 | 7.04 | 0.842577 | 0.821208 | 0.798148 |
| 135 | 8.10 | 0.839964 | 0.758176 | 0.680348 |
| 140 | 19.78 | 0.895506 | 0.936177 | 0.874019 |
| 164 | 16.98 | 0.905008 | 0.923559 | 0.887268 |
| 246 | 6.59 | 0.899448 | 0.907503 | 0.871254 |
| 255 | 36.22 | 0.955784 | 0.927517 | 0.921816 |
| 343 | 0.69 | 0.528261 | 0.840237 | 0.600751 |
| 349 | 2.96 | 0.912664 | 0.828181 | 0.778983 |
| 367 | 1.02 | 0.073485 | 0.392027 | 0.117796 |
| 370 | 10.82 | 0.953443 | 0.917094 | 0.908893 |
| 371 | 3.83 | 0.859781 | 0.685033 | 0.618380 |
| 375 | 11.67 | 0.911141 | 0.921345 | 0.910804 |
| 377 | 4.37 | 0.782994 | 0.712791 | 0.680165 |
| 381 | 7.63 | 0.891990 | 0.245768 | 0.237990 |
| 385 | 2.67 | 0.803215 | 0.641916 | 0.601690 |
| 395 | 0.68 | 0.770738 | 0.204418 | 0.242908 |
| 397 | 5.94 | 0.904544 | 0.882265 | 0.874036 |
| 409 | 11.86 | 0.944934 | 0.900965 | 0.901000 |
| 411 | 5.98 | 0.949977 | 0.933271 | 0.929499 |
| 425 | 0.91 | 0.802867 | 0.602796 | 0.559908 |
| 434 | 94.42 | 0.894601 | 0.590374 | 0.580553 |
| 531 | 22.08 | 0.892250 | 0.555015 | 0.505062 |
| 540 | 8.35 | 0.923702 | 0.895905 | 0.880058 |
<b>Median DSC</b>: 0.8946, 0.8212, 0.779
## Results for multi-sequence (T2, VIBE, and ADC)
The box plot of the predictions on the validation set:
The results from the validation set are also presented in the table below:
Sure, here is your data without the index numbers, in Markdown table format:
| subject_id | tumor_vol | inter_rater | r1_ml | r2_ml |
|------------|-----------|-------------|-------|-------|
| 29 | 4.16 | 0.201835 | 0.859937 | 0.148586 |
| 32 | 8.00 | 0.684142 | 0.662912 | 0.515606 |
| 36 | 19.06 | 0.928750 | 0.903445 | 0.889360 |
| 47 | 11.01 | 0.944209 | 0.907479 | 0.907132 |
| 50 | 6.26 | 0.722867 | 0.587017 | 0.547842 |
| 65 | 13.09 | 0.930613 | 0.889782 | 0.862255 |
| 67 | 3.71 | 0.943498 | 0.872207 | 0.862281 |
| 75 | 7.16 | 0.263539 | 0.750735 | 0.205290 |
| 86 | 7.04 | 0.842577 | 0.872160 | 0.813740 |
| 135 | 8.10 | 0.839964 | 0.808324 | 0.751063 |
| 140 | 19.78 | 0.895506 | 0.907457 | 0.852548 |
| 164 | 16.98 | 0.905008 | 0.925958 | 0.893707 |
| 246 | 6.59 | 0.899448 | 0.906569 | 0.852195 |
| 255 | 36.22 | 0.955784 | 0.942831 | 0.945464 |
| 343 | 0.69 | 0.528261 | 0.875817 | 0.463243 |
| 349 | 2.96 | 0.912664 | 0.853302 | 0.821028 |
| 367 | 1.02 | 0.073485 | 0.391209 | 0.086412 |
| 370 | 10.82 | 0.953443 | 0.916154 | 0.911768 |
| 371 | 3.83 | 0.859781 | 0.665516 | 0.637862 |
| 375 | 11.67 | 0.911141 | 0.898501 | 0.910147 |
| 377 | 4.37 | 0.782994 | 0.714387 | 0.647229 |
| 381 | 7.63 | 0.891990 | 0.437500 | 0.430847 |
| 385 | 2.67 | 0.803215 | 0.688608 | 0.624595 |
| 395 | 0.68 | 0.770738 | 0.400970 | 0.444584 |
| 397 | 5.94 | 0.904544 | 0.868022 | 0.850653 |
| 409 | 11.86 | 0.944934 | 0.888056 | 0.885981 |
| 411 | 5.98 | 0.949977 | 0.884870 | 0.883174 |
| 425 | 0.91 | 0.802867 | 0.573756 | 0.490407 |
| 434 | 94.42 | 0.894601 | 0.618859 | 0.605528 |
| 531 | 22.08 | 0.892250 | 0.348378 | 0.318399 |
| 540 | 8.35 | 0.923702 | 0.894245 | 0.884275 |
<b>Median DSC</b>: 0.8946, 0.868, 0.8137
## Support and Contribution
For any issues related to the model or the source code, please open an issue in the corresponding GitHub repository. Contributions to the code or the model are welcome and should be proposed through a pull request.
|