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

OBJECTIVES

To evaluate the diagnostic accuracy of infrared thermography (IT) using temperature difference (ΔT) and gray-level co-occurrence matrix (GLCM)-based textural parameters for detecting patellar tendinopathy (PT) in athletes, and to determine the optimal region of interest (ROI) size for analysis.

DESIGN

Cross-sectional study following STARD criteria.

SETTING

Physical therapy center with controlled environmental conditions.

PARTICIPANTS

54 athletes (27 with unilateral PT; 27 healthy controls).

MAIN OUTCOME MEASURES: Temperature differences (ΔT) and GLCM features (energy, homogeneity, contrast, correlation, entropy) from thermal images of tendon and knee ROIs.

RESULTS

Athletes with PT showed higher ΔT in the tendon ROI (0.5 ± 0.36 °C vs. 0.2 ± 0.20 °C; d = 0.77; p = 0.013). Among GLCM parameters, tendon ROI correlation (TCOR) differed significantly between groups (p = 0.042). Diagnostic performance analysis identified ΔT as the best single variable (AUC 0.83; sensitivity 78 %; specificity 78 %). The ΔT + GLCM combination achieved highest accuracy for tendon ROI (AUC 0.93). The patellar tendon-specific ROI provided better discrimination than the knee ROI.

CONCLUSIONS

IRT combined with GLCM enhances PT diagnostic precision, with the tendon ROI optimal for detecting local physiological differences. This accessible multimodal approach could complement clinical diagnosis in athletes.