Data from: Diagnostic accuracy of infrared thermography and textural features in patellar tendinopathy: optimizing ROI selection in athletes.

The aim of this study was to evaluate the diagnostic accuracy of infrared thermography (IT) and grey-level co-occurrence matrix (GLCM) textural features for detecting patellar tendinopathy (PT) in athletes, optimizing region of interest (ROI) selection. Fifty-four athletes (27 with unilateral PT, 27 healthy controls) were assessed using IT and the Victorian Institute of Sport Assessment-Patella (VISA-P) questionnaire. Thermal differences (ΔT) and GLCM features (energy, homogeneity, contrast, correlation, entropy) were analyzed for two ROIs: patellar tendon and anterior knee. Bayesian methods were used, reporting Bayes Factor (BF10) and 95% credibility intervals (CrI). The tendon ROI showed higher ΔT in PT patients (0.5 ± 0.36°C) than controls (0.2 ± 0.20°C; BF10 = 19, effect = 0.77), with GLCM textural correlation differing significantly (BF10 = 14, effect = 0.69). Combining ΔT and GLCM features enhanced diagnostic accuracy (AUC = 0.93, sensitivity = 78%, specificity = 78%) for the tendon ROI, outperforming the knee ROI. Precise ROI selection improved detection of subtle pathological changes. These findings highlight IT and GLCM as cost-effective, non-invasive tools for PT diagnosis, warranting further validation in diverse clinical contexts.