ClinicalDatasets

Overview of public clinical datasets

	Dataset	Input type	Task	Sample size
	BreakHis	images	binary classification	700 x 460 x 3
OpenML	dermatology pbcseq	tabular data	classification	1 x 34 1 x 19
FLamby	Camelyon16 TCGA-BRCA ISIC2019 Heart-Disease LIDC-IDRI IXI Tiny KiTS2019	images tabular data images tabular data 3D CT scans 3D MRI images 3D CT scans	cancer detection survival prediction melanoma classification heart disease detection segmentation segmentation segmentation	10,000 x 2,048 1 x 39 200 x 200 x 3 1 x 13 128 x 128 x 128 83 x 44 x 55 64 x 192 x 192
FLOP	COVIDx Kvasir	images images	classification Gastrointestinal disease classification	480 x 480 x 3 from 720 x 576 up to 1920 x 1072
	RSNA Breast Cancer	images	breast cancer detection	from 2,776 x 2,082 to 3,328 x 2,560
	MIMIC-III	mixed (mostly tabular)	?
	NHANES	tabular data	?

Additional dataset information:

Joshi et al.

Joshi et al. Federated Learning for Healthcare Domain - Pipeline, Applications and Challenges. 2022.

Tables 1 and 2 list studies on FL and ML in healthcare. Used datasets of these studies need to be investigated.

UK Biobank

https://www.ukbiobank.ac.uk/

PhysioNet

https://physionet.org/

The PhysioNet dataset is a collection of physiological signals and clinical data for use in research and education. It includes data on electrocardiograms (ECGs), electroencephalograms (EEGs), blood pressure, and other physiological signals.

Global Health Observatory (GHO) data

https://www.who.int/data/gho

The GHO data from the World Health Organization (WHO) includes health-related statistics for countries around the world. The dataset includes information on mortality rates, disease incidence, health expenditures, and other health indicators.

Todo

Diabetes Dataset (1988)

This dataset contains data related to diabetes patients and includes information such as age, gender, body mass index, blood pressure, and various blood serum measurements.

Kaggle dataset

• 768 patients (females at least 21 years old of Pima Indian heritage)
• features: pregnancies, glucose, blood pressure, skin thickness, insulin, bmi, diabetes pedigree function, age
• 500 without diabetes, 268 with diabetes

TCGA-LUAD

The Cancer Genome Atlas Lung Adenocarcinoma Collection

https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=6881474

OASIS

OASIS: Brain MRIs, 400+, 256 x 256

https://oasis-brains.org/

ADNI

ADNI: Alzheimer's Disease Neuroimaging Initiative, MRI

https://adni.loni.usc.edu/

NLST

Cancer Imaging Archive: 21M images (11 TB), only cancer images??

https://wiki.cancerimagingarchive.net/display/NLST/National+Lung+Screening+Trial

Covid-19

chest x-ray (normal, pneumonia, covid), 100-200 per class

https://www.kaggle.com/datasets/pranavraikokte/covid19-image-dataset