AI in Oncology · AI Models

A researcher-grade catalog of AI models, datasets and open data needs in oncology — every card structured, sourced and dated.

Datasets, registries and benchmarks you can train or evaluate oncology models on — with access conditions, licences, sizes, annotations and the models already using them. { } export JSON

Data type Task Cancer Access Kind Search × Clear filters

BraTS — Brain Tumor Segmentation Challenge

BraTS organisers (University of Pennsylvania, Indiana University, MICCAI)

Benchmark / challenge Free registration challenge terms (registration on Synapse)

The long-running benchmark for brain tumour segmentation from MRI; nnU-Net-based methods have dominated its leaderboards.

1 250 patients ≥1,250 glioma cases since 2021; later editions add paediatric, metastasis, meningioma and Sub-Saharan Africa tracks

MRI Segmentation NIfTI

Adult-Type Diffuse Gliomas

Dataset site used by models: 1 Open card →

CAMELYON16 / CAMELYON17

Radboud University Medical Center and partners (grand-challenge.org)

Benchmark / challenge Open download CC0

The canonical whole-slide benchmark for breast cancer lymph-node metastasis detection; still the first sanity check for any new pathology encoder.

1 399 WSIs CAMELYON16: 399 slides (270 train / 129 test) with pixel-level metastasis annotations; CAMELYON17: 1,000 slides from 5 centres with patient-level pN stage

Histopathology (WSI) Detection Classification Segmentation TIFFCSV

Invasive Breast Carcinoma

Dataset site used by models: 2 Open card →

ISIC Archive — International Skin Imaging Collaboration

ISIC (Memorial Sloan Kettering and partners)

Registry / portal Open download CC-0 / CC BY-NC per contributor

The public backbone of skin-cancer AI; strongly skewed towards light skin tones, which every model card built on it should say.

70 000 dermoscopic images tens of thousands of images with diagnosis; annual challenge subsets (e.g. 2020: 33,126 images)

Dermoscopy Classification Detection Segmentation JPEGCSV

Malignant Melanoma Non-Melanoma Skin Cancer (NMSC)

Dataset site Open card →

LIDC-IDRI — Lung Image Database Consortium

NCI / TCIA

Dataset Open download CC BY 3.0

The standard open CT dataset for lung nodule detection and characterisation; basis of the LUNA16 challenge.

1 010 patients 1 018 CT scans nodule annotations by four thoracic radiologists in a two-phase reading

CT Detection Segmentation Classification DICOMJSON

Lung & Bronchus

Dataset site Open card →

TCIA — The Cancer Imaging Archive

NCI Cancer Imaging Program; hosted by the University of Arkansas for Medical Sciences

Registry / portal Open download mostly CC BY 3.0/4.0 per collection; some restricted collections

The main public archive of de-identified cancer imaging, organised into collections by disease and modality; the source of most public radiology training data in oncology.

200 collections hundreds of collections, tens of thousands of patients; DICOM with linked clinical and sometimes genomic data

CT MRI PET Mammography Histopathology (WSI)X-ray Segmentation Detection Classification Prognosis DICOMNIfTICSV

Pan-cancer

Dataset site used by models: 1 Open card →

Cards follow the cancer3.ai model-card standard: AI_MODEL_CARD_STANDARD.md. Corrections and new entries: contact the editorial team; every fact needs a public source.

This page is educational — it is not medical advice and does not replace consultation with an oncologist. Diagnostic and treatment decisions are made solely by specialist physicians.