GPT-4o Cataract Surgery IOL Recommendations for 74 Patients

A retrospective clinical study of 74 cataract patients (146 eyes) who underwent surgery. The dataset contains preoperative clinical findings, biometric parameters, and lifestyle questionnaires used to evaluate GPT-4o's intraocular lens (IOL) recommendations against expert consensus and human surgeons. The data was published by author Yan Liu on figshare in May 2026.

Use Cases

• Benchmarking AI model performance for routine vs. complex IOL selection based on patient data.
• Analyzing the impact of ocular and systemic comorbidities on AI-assisted clinical decisions.
• Studying the consistency and stability of LLM outputs in repeated clinical simulations.
• Comparing AI recommendations to decisions made by high- and low-volume human surgeons.

Strengths

• Includes data from 74 real-world cataract surgery patients (146 eyes).
• Model outputs were evaluated against an adjudicated expert consensus and human surgeons.
• Analysis includes specific metrics: appropriateness rates, Kappa values, and intraclass correlation coefficients (ICC).

Limitations

• Row count and column-level documentation are unknown, limiting suitability assessment.
• The dataset is small (5.2 MB), which may restrict the scope of machine learning applications.
• Data freshness should be verified, as the last update is listed as 2026-05-29.

Provenance

Source

figshare, author Yan Liu.

Collection Method

Retrospective clinical observational study; preoperative datasets were processed via GPT-4o.

Freshness

Last updated 2026-05-29 06:20:26.