by Peiyan Wu, Yan Jiang, Hanshuo Xing, Wenbo Song, Xin-wu Cui, Xinglong Wu and Guoping Xu. This is a code repo of the paper.

Methods: The clinical and ultrasound imaging data, including brightness mode (B-mode) and color Doppler flow imaging (CDFI), of 611 breast cancer patients from multiple hospitals in China were retrospectively analyzed. Patients were divided into one primary cohort (PC), one validation cohort (VC) and two test cohorts (TC1 and TC2). A multimodality deep learning radiomics nomogram (DLRN) was constructed for predicting the MBC. The performance of the proposed DLRN was comprehensively assessed and compared with three unimodal models via the calibration curve, the area under the curve (AUC) of receiver operating characteristics (ROC) and the decision curve analysis (DCA).

Calibration ROC DCA

nomogram points and confusion matrix

nomogram

Ubuntu 18.04
Conda Environment Setup
Create your own conda environment

conda create -n MDLRN python=3.8
conda activate MDLRN

Install Pytorch == 2.0.0(depends on your NVIDIA driver and you can see your compatible CUDA version at the right hand corner in nvidia-smi)

conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia

Install monai == 1.1.0

pip install monai[all]

other python required packages

pip install -r requirements.txt

Install R 4.3.0 and r packages in R_requirements.txt

Input Folder Format

├── MDLRN
│   ├── code
│   │   ├── deeplearning
│   │   ├── form
│   │   │   ├── clinical
│   │   │   │   ├── pc.csv
│   │   │   │   ├── vc.csv
│   │   │   │   ├── tc1.csv
│   │   │   │   ├── tc2.csv
│   │   │   ├── data_split
│   │   │   │   ├── pc.csv
│   │   │   │   ├── vc.csv
│   │   │   │   ├── tc1.csv
│   │   │   │   ├── tc2.csv
│   │   │   ├── ...
│   │   ├── config.yaml
│   │   ├── MDLRN.ipynb
│   │   ├── nomogram.ipynb
│   ├── data
│   │   ├── origin
│   │   │   ├── PC
│   │   │   │   ├── bus
│   │   │   │   ├── ├── ne_xxxxxx.png(ne means negative)
│   │   │   │   ├── ├── po_xxxxxx.png(po means positive)
│   │   │   │   ├── ├── ...
│   │   │   │   ├── cdfi
│   │   │   │   ├── roi
│   │   │   ├── VC
│   │   │   ├── TC1