Healthcare datasets for medical AI development and evaluation. Review the modality, data origin, access conditions and technical profile of each offering before requesting a sample.
Search by product, modality or specialty, then refine the results with the filters below.
Images illustrate dataset categories. Technical samples are available during evaluation.
300 echocardiography studies: 200 pathology and 100 normal; coverage includes EF, valves, wall motion, hypertrophy and congenital findings.
Coronary CT angiography collection with 10,000 abnormal CCTA studies represented.
Interventional cardiology collection covering PTCA / cath-lab / angioplasty, with 15,000 abnormal studies represented.
Mixed-modality oncology imaging collection.
Plain and contrast brain MRI.
Brain MR angiography collection.
Cervical spine MRI collection.
Lumbar spine MRI collection.
Breast MRI collection with 10,000 studies represented; evaluation sample available.
Plain and contrast CT brain collection.
Cervical spine CT collection.
Abdominal and pelvic CT collection.
CT pulmonary angiography collection.
Contrast-enhanced chest CT collection.
Chest radiography collection.
Cervical spine radiography collection.
Lumbar spine radiography collection.
Abdominal and pelvic radiography collection.
Knee radiography collection.
Pelvis and bilateral hip radiography collection.
Mammography imaging collection.
United States four-view mammography DICOM offering with L-CC, R-CC, L-MLO and R-MLO images and linked CSV metadata. 5,000 records are currently available, with a proposed total volume of up to 50,000.
Structured radiology report collection for report understanding, structuring, classification and research workflows.
Mixed-modality medical imaging collection.
Structured pathology-oriented profiles; does not represent whole-slide pathology images or spatial-omics data.
Structured clinical case collection spanning multiple specialties. Longitudinal depth: case-based.
Approx. 5 years of longitudinal history with an average of 2–3 OPD/IPD interactions per year. Represented fields include demographics, encounter date, OPD/IPD, diagnosis, complaints, prescriptions, clinical notes, follow-up…
USA longitudinal patient-record packages with 2+ years of history; PDF delivery.
Brain MRI and longitudinal clinical information linked at the same-patient level.
Patient-linked imaging, radiology reports and longitudinal EMR; DICOM + EMR delivery represented.
Same-patient lung adenocarcinoma evaluation case with clinical history, pathology confirmation, treatment/follow-up information and chest CT image exports.
Two knee MRI samples and two knee X-ray samples are available. Intended scope includes matched CT/MRI for elbow, shoulder and hip, and matched XR/CT/MRI for knee; matching and CT coverage are not yet confirmed.
Panoramic dental X-ray collection.
Endoscopy and surgical procedure video collection; 100 videos represented as available.
CCTA-FFR studies with imaging, fractional-flow-reserve outputs and corresponding report.
A doctor–patient conversation sample has been reviewed. The reviewed 2-channel sample is effectively dual-mono.
US English audio files with a JSON metadata file represented for each audio file.
Longitudinal Parkinson’s dataset comprising 100,000 synthetic patient journeys.
Synthetic patient-population product designed for healthcare analytics, modeling and pipeline development without real patient records.
Synthetic longitudinal EHR data combining encounter-level records with FHIR R4 resources and standard clinical coding structures.
Synthetic clinical-trial product covering continuous, binary and time-to-event endpoint structures for method development and data-engineering workflows.
Synthetic longitudinal disease-progression data for model development, survival analysis and outcome-oriented evaluation.
Synthetic hospital-admission product combining patient-level admission records with operational bed-utilization context.
Synthetic medical-imaging metadata, COCO-style annotations, DICOM-tag fields and structured radiology reports across multiple modalities. No pixel data is represented in this product.
Synthetic patient-treatment data designed for drug-response, pharmacogenomic and pharmacokinetic modeling workflows.
Synthetic healthcare-claims product for reimbursement analytics, payer workflows and controlled model-development use cases.
Synthetic continuous vital-sign and alarm data for ICU/monitoring analytics and time-series model-development workflows.
Synthetic hospital-resource utilization data for operational modeling, forecasting and workflow evaluation.
Fully synthetic oncology data product focused on breast cancer for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on lung cancer for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on prostate cancer for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on colorectal cancer for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on pancreatic cancer for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on liver / hcc for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on leukemia for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on lymphoma for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on melanoma for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on ovarian cancer for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on multi-cancer tumor progression for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on chemotherapy response for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on immunotherapy / checkpoint inhibitor response for model development, analytics, evaluation and controlled research workflows.
Fully synthetic oncology data product focused on metastatic cancer for model development, analytics, evaluation and controlled research workflows.
Fully synthetic cohort spanning cardiovascular risk factors, atrial-fibrillation phenotyping, vascular/cardiac/neuroimaging fields, validated risk-score structures and outcome variables.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Annotated studies with clinical data, specimen data, tissue images and pathology reports.
Try adjusting your search or clearing the filters.