Patient Data for Analysis: Inherited Retinal Disease Images and Genetic Tests

409 patient studies from Hadassah University Medical Center, including 330 patients with inherited retinal diseases (IRDs) and 79 control patients. The dataset contains wide-field fundus autofluorescence (FAF) and pseudocolor fundus (pCF) images paired with genetic test results for IRD. It was created by Leo Joskowicz and last updated in May 2026 to evaluate a multimodal deep learning method for automated disease identification.

Use Cases

• Training binary classifiers to distinguish IRD from non-IRD cases based on fundus images.
• Developing multimodal deep learning models using both FAF and pCF image types.
• Exploring the feasibility of predicting causative gene groups from retinal imaging data.
• Benchmarking model performance using metrics like accuracy, precision, F1 scores, and AUC.

Strengths

• Contains 409 patient records with paired imaging and genetic data.
• Focuses on 25 of the most commonly affected genes in the study population.
• Includes both FAF and pCF image modalities for multimodal analysis.
• Model performance metrics are reported, with binary classification accuracy reaching 0.95.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Data may reflect geographic and institutional bias inherent to a single medical center.

Provenance

Source

Hadassah University Medical Center

Collection Method

Retrospective collection of patient studies containing FAF and pCF images and genetic tests.

Freshness

Last updated 2026-05-11 17:27:44; freshness should be verified.