Skip to content

Predicting gastric cancer outcome from resected lymph node histopathology images using deep learning

License

Notifications You must be signed in to change notification settings

MHMAILab/auto_lymph

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Predicting gastric cancer outcome from resected lymph node histopathology images using deep learning

Link to paper

Abstract

N-staging is a determining factor for prognostic assessment and decision-making for stage-based cancer therapeutic strategies. Visual inspection of whole-slides of intact lymph nodes is currently the main method used by pathologists to calculate the number of metastatic lymph nodes (MLNs). Moreover, even at the same N stage, the outcome of patients varies dramatically. Here, we propose a deep-learning framework for analyzing lymph node whole-slide images (WSIs) to identify lymph nodes and tumor regions, and then to uncover tumor-area-to-MLN-area ratio (T/MLN). After training, our model’s tumor detection performance was comparable to that of experienced pathologists and achieved similar performance on two independent gastric cancer validation cohorts. Further, we demonstrate that T/MLN is an interpretable independent prognostic factor. These !ndings indicate that deep-learning models could assist not only pathologists in detecting lymph nodes with metastases but also oncologists in exploring new prognostic factors, especially those that are dif!cult to calculate manually.

System requirements

Hardware Requirements

At least NVIDIA GTX 2080Ti

OS Requirements

This package is supported for Linux. The package has been tested on the following systems:

Linux: Ubuntu 16.04

Software Prerequisites

Python 3.7+
Numpy 1.17.2
Scipy 1.3.0
Pytorch 1.3.0+/CUDA 10.1
torchvision 0.4.1+
Pillow 6.0.0
opencv-python 4.1.0.25
openslide-python 1.1.1
Scikit-learn 0.21

Installation guide

It is recommended to install the environment in the Ubuntu 16.04 system.

  • First install Anconda3.

  • Then install CUDA 10.x and cudnn.

  • Finall intall these dependent python software library.

The installation is estimated to take 1 hour, depending on the network environment.

Demo

Train Segmentation model

train model
python ./segmentation/bin/train.py 
test model
python ./segmentation/bin/test.py 

Train Classification model

train model
python ./classification/bin/train.py 
test model
python ./classification/bin/test.py 

Get T/MLN

python ./test/bin/get_T_MLN.py --wsi_path './tiff/test_patients/'

where --wsi_path is the path to all the WSI tiff of the patient you are interested.

References

Appreciate the great work from the following repositories:

Citation

@article{wang_predicting_2021,
	title = {Predicting gastric cancer outcome from resected lymph node histopathology images using deep learning},
	volume = {12},
	issn = {2041-1723},
	url = {https://doi.org/10.1038/s41467-021-21674-7},
	doi = {10.1038/s41467-021-21674-7},
	number = {1},
	journal = {Nature Communications},
	author = {Wang, Xiaodong and Chen, Ying and Gao, Yunshu and Zhang, Huiqing and Guan, Zehui and Dong, Zhou and Zheng, Yuxuan and Jiang, Jiarui and Yang, Haoqing and Wang, Liming and Huang, Xianming and Ai, Lirong and Yu, Wenlong and Li, Hongwei and Dong, Changsheng and Zhou, Zhou and Liu, Xiyang and Yu, Guanzhen},
	year = {2021},
	pages = {1637}
}

License

This project is covered under the Apache 2.0 License.

About

Predicting gastric cancer outcome from resected lymph node histopathology images using deep learning

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages