Work

Age-Related Scattered Hypofluorescent Spots as an Adverse Prognostic Factor for Polypoidal Choroidal Vasculopathy

Year: 2025

Type: article

Abstract: Purpose: Polypoidal choroidal vasculopathy (PCV) demonstrates significant prognostic variability, and the impact of age-related scattered hypofluorescent spots observed in late-phase indocyanine green angiography (ASHS-LIA) on the prognosis of PCV remains under-researched. This study aims to investigate the association between ASHS-LIA in PCV and prognosis using the AdaBoost machine learning model. Design: A cross-sectional study. Participants: The study included patients diagnosed with PCV and treated with anti-VEGF therapy at 2 medical institutions between 2012 and 2021. Methods: We conducted a retrospective analysis of the clinical characteristics, anti-VEGF treatment history, and outcomes of the participants, classifying them based on the presence or absence of ASHS-LIA. An AdaBoost meta-estimator was applied to predict prognosis, including disease stability, injection frequency, and time to first remission, utilizing features selected through principal component analysis. Main Outcome Measures: The prognostic significance of ASHS-LIA was assessed by feature importance, with the mean decrease in impurity serving as the evaluation metric. Results: < 0.001). AdaBoost models confirmed the importance of ASHS-LIA for predicting disease stability, injection demand, and time to first remission, ranking it as the third, seventh, and eighth top contributory factor, respectively. Conclusions: Machine learning analysis identified ASHS-LIA as a negative prognostic factor in PCV, correlating with reduced disease stability, higher recurrence rates, and increased treatment requirements. These findings suggest that ASHS-LIA could serve as a valuable marker for assessing prognosis and guiding treatment strategies in PCV management. Financial Disclosures: The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Source: Ophthalmology Science

DOI: 10.1016/j.xops.2025.100818