Abstract:AIM: To establish and validate a multidimensional predictive model of postoperative visual recovery after idiopathic macular hole (IMH) surgery. METHODS: Retrospective cohort study. Examinations within a three-month period, both pre- and postoperative, included assessments of best corrected visual acuity (BCVA), intraocular pressure (IOP), and morphological parameters of IMH with optical coherence tomography (OCT). Then, a series of indices were derived, including the IMH index (MHI), diameter hole index (DHI), macular hole closure index (MHCI), hole form factor (HFF), and tractional hole index (THI). Subfoveal anatomical damage (macular hole inferior volume, Vi) was calculated based on the basal diameter (BD), minimum diameter (MD), and height at the narrowest point (HMD) of IMH. Pearson correlation analysis was utilized to discern significant correlations between postoperative BCVA and the multiple indices examined. A subsequent linear correlation analysis was performed. RESULTS: The study involved 51 eyes from 51 patients (mean age 66.90±6.07y) diagnosed with IMH. Preoperative BCVA was 1.22±0.76 logMAR and improved to 0.88±0.38 logMAR after surgery (P<0.001). The correlation analysis results showed significant correlations between postoperative BCVA and preoperative BCVA (P<0.001), BD (P=0.042), MD (P=0.001), MHI (P=0.047), THI (P=0.004), and Vi (P=0.007). The multidimensional model integrating THI, Vi, and preoperative BCVA significantly outperformed traditional predictors (MD, BD, and height) in terms of postoperative visual recovery prediction. THI, reflecting posterior vitreous traction mechanics, independently predicted anatomical reset potential (β=-0.06, P=0.022), while Vi, quantifying subfoveal photoreceptor disruption, was correlated with structural-functional recovery (β=0.01, P=0.046). Preoperative visual acuity served as a critical surrogate for retinal functional reserve (β=0.15, P=0.020). Redundant morphometric parameters (MHI, DHI, MHCI, HFF) were excluded, as their predictive contributions were subsumed by THI/Vi or mediated by preoperative vision. CONCLUSION: The combination of biomechanical traction (THI), subfoveal anatomical damage (Vi), and preoperative BCVA represents a clinically applicable framework for predicting postoperative visual recovery after IMH surgery. This model can be used as a practical tool to guide surgical planning, facilitating the identification of high-risk patients who may benefit from additional techniques (such as an internal limiting membrane flap) while optimizing resource allocation for standard cases.
