Cancer-Net Open Initiative

Cancer-Net is a global open source, open access initiative dedicated to accelerating advancement in machine learning to aid clinicians in the fight against cancer. Towards this goal, our global multi-disciplinary team of researchers, developers, and clinicians have made publicly available a suite of tailored deep neural network models for tackling different challenges ranging from screening to risk stratification to treatment planning for patients with various forms of cancer. Furthermore, we have made available fully curated, open access benchmark datasets comprised of the largest, most diverse patient cohorts from around the world for correlated diffusion imaging, a medical imaging modality tailored for cancer imaging. We hope the open-source, open-access release of Cancer-Net deep learning models and associated large-scale benchmark datasets will motivate and enable researchers, clinicians, and citizen data scientists alike from around the world to build upon them and accelerate progress in this field. We continue to regularly release new models and benchmark datasets to keep up with the lastest developments in cancer research. Cancer-Net is a joint initiative with the COVID-Net initiative and the GenAI4Good initiative.

Updates

• February 14, 2024: Major release of Cancer-Net SCa-DC-AC, a tailored Double-Condensing Attention Condenser deep neural network for detection of skin cancer from dermoscopy images (Paper)(Models)
• September 28, 2023: Major release of TRUDLMIA, a deep learning framework for building trustworthy models for medical image analysis (Paper)(Datasets)
• May 17, 2023: Major release of Cancer-Net PCa-Data, an open-source benchmark dataset for prostate cancer clinical decision support using correlated diffusion imaging data (Datasets)
• April 12, 2023: Major release of Cancer-Net BCa-S, a volumetric convolutional neural network to learn volumetric deep radiomic features for predicting grading for breast cancer using correlated diffusion imaging (Paper)(Models)(Datasets)
• April 12, 2023: Major release of Cancer-Net BCa-Data, a multi-Institutional open-source benchmark dataset for breast cancer clinical decision support using correlated diffusion imaging data (Paper)(Datasets)
• November 26, 2022: Major release of Cancer-Net BCa, a volumetric convolutional neural network to learn volumetric deep radiomic features for predicting the post-treatment response for breast cancer using correlated diffusion imaging, along with a new benchmark dataset of volumetric correlated diffusion imaging data from 253 patient cases (Paper)(Models)(Datasets)
• March 1, 2022: Major release of Cancer-Net PCa-CDIs, Synthetic correlated diffusion imaging hyperintensity delineates clinically significant prostate cancer (Paper)
• January 19, 2021: Major release of Cancer-Net SCa, tailored deep convolutional neural networks for detection of skin cancer from dermoscopy images (Paper)(Models)
  

Benchmark Dataset Status:

• Correlated diffusion imaging data (BCa): 253 patients Click here
• Correlated diffusion imaging data (PCa): 200 patients Click here

Resources

• Cancer-Net BCa-S: Breast Cancer Grade Prediction using Volumetric Deep Radiomic Features from Synthetic Correlated Diffusion Imaging:
  • Repo: Click here
  • Models: Click here
  • Dataset: Click here
  • Paper: Click here
    
    
• Cancer-Net BCa-Data: A Multi-Institutional Open-Source Benchmark Dataset for Breast Cancer Clinical Decision Support using Synthetic Correlated Diffusion Imaging Data:
  • Dataset: Click here
  • Paper: Click here
    
    
• Cancer-Net BCa: Breast Cancer Pathologic Complete Response Prediction using Volumetric Deep Radiomic Features from Synthetic Correlated Diffusion Imaging:
  • Repo: Click here
  • Models: Click here
  • Dataset: Click here
  • Paper: Click here
    
    
• Cancer-Net PCa-Data: An Open-Source Benchmark Dataset for Prostate Cancer Clinical Decision Support using Correlated Diffusion Imaging Data:
  • Dataset: Click here
    
    
• Cancer-Net PCa-CDIs: Synthetic correlated diffusion imaging hyperintensity delineates clinically significant prostate cancer:
  • Paper: Click here
    
    
• Cancer-Net SCa-DC-AC: tailored Double-Condensing Attention Condenser deep neural networks for detection of skin cancer from dermoscopy images:
  • Repo: Click here
  • Models: Click here
  • Paper: Click here
    
• Cancer-Net SCa: tailored deep convolutional neural networks for detection of skin cancer from dermoscopy images:
  • Repo: Click here
  • Models: Click here
  • Paper: Click here
    
• TRUDLMIA: a deep learning framework for building trustworthy models for medical image analysis:
  • Paper: Click here
  • Benchmark dataset: Click here
    

Core Cancer-Net Team

Project Lead: Alexander Wong (a28wong@uwaterloo.ca)

• Vision and Image Processing Research Group, University of Waterloo, Canada
  • Amy Tai
  • Hayden Gunraj
  • Elizabeth Janes
• DarwinAI Corp., Canada and Vision and Image Processing Research Group, University of Waterloo, Canada
  • James Lee
• Ashkan Ebadi and Pengcheng Xi (National Research Council Canada)
• Ali Sabri (Niagara Health, McMaster University, Canada)
• Amer Alaref (Thunder Bay Regional Health Sciences Centre, Northern Ontario School of Medicine, Canada)
• Adrian Florea (CIUSSS de l’Ouest-de-l’Île-de-Montréal)