Artificial intelligence in predicting macular hole surgery outcomes: a focus on optical coherence tomography...

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Macular hole (MH) surgery aims for anatomical success, influenced by preoperative metrics.
• Optical coherence tomography (OCT) provides key measurements for evaluating MH characteristics.

Purpose of the Study:

• To compare the predictive performance of OCT-based indices and a GPT artificial intelligence (AI) model against logistic regression for forecasting anatomical success in MH surgery.

Main Methods:

• Retrospective analysis of 51 eyes undergoing pars plana vitrectomy for idiopathic MH.
• Utilized preoperative OCT measurements: macular hole index (MHI), traction hole index (THI), hole form factor (HFF), basal hole diameter (BHD), and minimum hole diameter (MHD).
• Developed GPT-AI and logistic regression models; compared predictive accuracy, AUC, POPV, NPV, and Kappa statistics.

Main Results:

• Anatomical success rate was 72.5%.
• MHI, THI, and HFF were significant predictors of success (p < 0.0001).
• Logistic regression (84.3% accuracy, 0.759 AUC, 0.800 NPV) outperformed GPT (77.0% accuracy, 0.770 AUC, 0.452 NPV). BHD and MHD had low predictive power (0.291 AUC).

Conclusions:

• OCT-derived indices (MHI, THI, HFF) are valuable for predicting MH surgery outcomes.
• Logistic regression demonstrated superior predictive performance over the GPT model in this cohort.
• AI models show promise but require further development and validation for clinical use in MH surgery prediction.