Clinical Image Data Sources: An Overview

by | Nov 17, 2017 | Healthcare, Research

This article contains an overview of Clinical Image Data Sources. Due to divers angels of clinical studies the data collected in these studies have various formats. This data variety provides different solutions for a lot of complex biological questions. However, the downside of this data variety is the complex data storage, processing and comparing. For every different data set new ways of storing and processing must be dealt with. This is a problem that can be hard to deal with, especially for intensive processing or deep learning tasks. In order to tackle these problems in the Genematics Cloud imaging project, we have set up a list of clinical data sources deemed useful or relevant to this project. For each source the license and availability has been checked. Moreover, every source on this list can be used for other projects, studies or data processing tasks, although individual restrictions may apply at specific sources.

Although, each source varies in data type, origin, size, focus of data and accessibility, the following topics have been reviewed for each data source:

Source: The name of the database, dataset or publisher.

Data Types: Types of images found in this source.

  • XR: Plain Xray
  • CT: CT scan
  • MR: MRI scan
  • US: Ultrasound
  • PE: PET scan

Data Focus: The main focus of the dataset images in terms of body location or function.

Images: The amount of images in this data source.

Patients: The amount of patients that contributed to this data source.

Availability: The requirements for accessing this data source.

License: The license and restrictions for this data source.

Link: The Direct URL to the data source, apply form or website closest to the data source. Unknown values are annotated as X.


SourceData TypeData FocusImagesPatientsAvailabilityLicenseLink
Alzheimer’s Disease Neuroimaging Initiative (ADNI)MR, PEBrainXXRequest on application formNon-commercial, non-modify, non-distribute
AMRG AtlasMRHeart781Request on application formNon-commercial, non-modify, non-distribute
Autism Brain Imaging Data Exchange (ABIDE)MRBrainX539Registration Non-commercial
Belarus Tuberculosis PortalXR, CTLungsX500+Open Non-commercial
CHD AtlasMRHeartXXRequest on application formNon-commercial, non-modify, non-distribute
DETERMINEMRHeartXXRequest on application formNon-commercial, non-modify, non-distribute
Digital Database for Screening Mammography (DDSM)XRAllX2620OpenPrimarily research and algorithm development
Digital Retinal Images for Vessel Extraction (DRIVE)MRHeartX400RegistrationNon-commercial, non-modify, non-distribute
Initiative for Collaborative Computer Vision Benchmarking (Prostate)MRProstateXXOpenMIT license
Japanese Society of Radiological Technology Database (JSRT)CTLungs250+250+RegistrationNon-commercial, non-distribute
MedPixXR, CT, MR, US, PEAll53,000+13,000+RegistrationPersonal use only, including local private distribution
 Multi-Ethnic Study of AtherosclerosisMRHeartX6,500+RegistrationNon-commercial, non-modify, non-distribute
Montgomery County X-ray SetMRLungs138XOpenMIT license
Open Access Series of Imaging Series (OASIS)MRBrain373186RegistrationModify and distribute on creator agreement
Osirix DICOM LibraryCT, MR, PEAllXXMonthly subscriptionNon-commercial, non-distribute
SCMR Consensus DataMRAll193¹15Request on application formNon-commercial, non-modify, non-distribute
Shenzhen Chest X-ray SetCTLungs662XOpenMIT license
Sunnybrook Cardiac Data (SCD)MRHeartX46Request on application formNon-commercial, non-modify, non-distribute
The cancer Imaging Archive (TCIA)XR, CT, MR, US, PEAllX40,000+OpenNon-modify, additional restrictions on specific sub datasets

¹ Number of MRI slices

Feature image credit:  Marcin Sadlowski. Licensed via Adobe Stock.