Prognostic evaluation of retinal reattachment surgery: predictive efficacy analysis of microperfusion indices via contrast-enhanced ultrasound.

We aimed to examine how preoperative contrast-enhanced ultrasound (CEUS) parameters relate to postoperative visual recovery in retinal detachment (RD) patients and to establish new objective markers for clinical prognosis. This single-center prospective observational study enrolled 52 RD patients undergoing pars plana vitrectomy (PPV). Preoperative CEUS parameters (arrival time [AT], time to peak [TTP], peak intensity [PI], time-intensity curve-area under the curve [TIC-AUC], ascending slope [AS]) were collected. Between-group comparisons were performed using parametric or non-parametric tests as appropriate. Ocular covariates including RD duration, macular status, and tamponade type were adjusted in prespecified binary logistic regression models. (Patients were stratified into improvement (n = 39) and non-improvement (n = 13) groups based on 6-month best-corrected visual acuity. Model stability was validated using bootstrap resampling, and post-hoc power analysis was performed to confirm statistical reliability. Receiver operating characteristic (ROC) curves were generated to evaluate predictive performance. The improvement group exhibited significantly higher PI and TIC-AUC, and lower AT (all P < 0.05). TTP showed a borderline significant trend (P = 0.059), whereas AS showed no significant difference (P = 0.100). PI was the only independent CEUS predictor of visual improvement (Receiver operating characteristic-area under the curve [ROC-AUC] = 0.844, P < 0.001), accounting for 34.7% of the variance in visual improvement. Macular status was also an independent predictor. Bootstrap validation confirmed model robustness. Preoperative CEUS parameters, particularly PI, are associated with 6-month visual recovery after RD surgery. PI represents a preliminary candidate prognostic biomarker; caution is needed for clinical translation owing to the single-center design, small sample size, and absence of external validation. Multicenter studies are needed to validate these findings.