Artificial Intelligence Unravels Interpretable Malignancy Grades of Prostate Cancer on Histology Images
-
You need a source-code editor (e.g., Visual Studio Code from https://code.visualstudio.com)
-
Python ≥3.8 (Either built-in or from https://www.python.org)
-
Jupyter notebook (How to install: https://jupyter.org/install)
-
Packages installation:
https://github.com/oeminaga/AI_PCA_GRADE/blob/main/install_packages.sh -
Extraction of TMA core images
- Download QuPath
https://qupath.github.io - The instruction given by Andrew Janowczyk
http://www.andrewjanowczyk.com/de-array-a-tissue-microarray-tma-using-qupath-and-python/
- Download QuPath
-
Neural Architecture Search:
https://github.com/oeminaga/AI_PCA_GRADE/blob/main/NAS_PlexusNet.ipynb -
Model development and inference on TMA images
https://github.com/oeminaga/AI_PCA_GRADE/blob/main/main_notebook_for_model_development.ipynb -
Inference on whole-slide images with support for parallel computing-processing multiple whole slide images at the same time- (the ImageScope's xml annotation file including the demarcation of cancer lesions is required)
- Aperio ImageScope (Image Viewer and Annotion Tool):
https://www.leicabiosystems.com/us/digital-pathology/manage/aperio-imagescope/ - SCN files:
https://github.com/oeminaga/AI_PCA_GRADE/tree/main/PROCESS_SCN_WSI - SVS files:
https://github.com/oeminaga/AI_PCA_GRADE/tree/main/PROCESS_SVS_WSI - TIFF and other formats: ⚙️🚧
- Aperio ImageScope (Image Viewer and Annotion Tool):
-
Statistical analyses -abstract version for clarity-
- Survival analyses and comparsion of cox models:
https://github.com/oeminaga/AI_PCA_GRADE/blob/main/Analysis/evaluation.R
- Please use R Studio to open this file. To install, please follow the instruction given in the link [https://posit.co/download/rstudio-desktop/]
- Distribution of 64 representation features for our novel model:
https://github.com/oeminaga/AI_PCA_GRADE/blob/main/Analysis/FeatureDistributionAnalysis.ipynb
- Survival analyses and comparsion of cox models:
Should you have issues, please open a thread in the issue section.
Please cite the following paper when you use these scripts [https://www.nature.com/articles/s44303-023-00005-z?utm_source=rct_congratemailt&utm_medium=email&utm_campaign=oa_20240306&utm_content=10.1038/s44303-023-00005-z]