This repository contains code for running a microscopic tumor presence (MTP) model based on the publication Probabilistic Clinical Target Definition with Nearest Neighbor Correlation.
The simulations can be executed using CUDA or C++ implementations.
To run with CUDA, make sure you have the CUDA Toolkit installed on your system.
If you prefer to run the C++ implementation, compile the code as follows:
mkdir build
cd build
cmake ..
cmake --build .To run a simple 2D simulation using a circular GTV, execute:
python run_CMP_model_2D.py
To run the CMP model in 3D:
Download the GTV and prostate data files from [https://drive.google.com/drive/folders/1JS4AE_O_Iynoq18s2h4MehCWmYSgOFgl?usp=sharing].
Save them in the following directory:
data/prostate_data/
The prostate dataset corresponds to the third patient from the publicly available QIN PROSTATE dataset. In this repository, the RTStructure files were converted to .nrrd format.
You can download the original dataset here: 👉 https://www.cancerimagingarchive.net/collection/qin-prostate/
To run the VMP model in 3D:
Download the GTV and breast data files from [https://drive.google.com/drive/folders/1UgMTl9n0jqNiXJo2XwcAou2FlqlGq5U0?usp=sharing].
Save them in the following directory:
data/breast_data/
The breast dataset corresponds to the fifth patient from the Advanced MRI Breast Lesions dataset. The breast was manually segmented, and the GTV was extracted from the structure provided by the dataset.
You can download the original dataset here: 👉 https://www.cancerimagingarchive.net/collection/advanced-mri-breast-lesions/