T2* mapping values across MRI-defined normal, partial-tear, and complete-tear anterior cruciate ligament categories: a prospective comparative study

Anterior cruciate ligament injury is commonly evaluated using conventional Magnetic Resonance Imaging; however, morphology-based assessment may be challenging, particularly in partial ACL tears. T2* mapping is a quantitative MRI technique that may provide numerical information related to ligament tissue characteristics. This study aimed to compare ACL T2* values among MRI-defined normal ACL, partial ACL tear and complete ACL tear groups. This prospective comparative observational study included 90 knee MRI examinations from 78 participants; 66 participants contributed one knee examination and 12 contributed bilateral examinations. The examinations were allocated to three MRI-defined categories, with 30 examinations each in the normal ACL, partial ACL tear, and complete ACL tear groups. The knee MRI examination was the unit of analysis. MRI was performed on a 1.5 T Siemens MAGNETOM Sempra system using a dedicated knee coil. Scanner-generated sagittal multi-echo gradient-echo T2* maps were analysed by two observers using proximal/femoral and distal/tibial circular regions of interest. The two regional values were averaged for each observer, and the observer-specific means were then averaged to obtain the final ACL T2* value. Group comparisons used the Kruskal–Wallis test with Dunn–Bonferroni correction. Adjusted multivariable regression with participant-clustered robust standard errors assessed the association between MRI-defined ACL category and final mean T2* value after adjustment for age, sex, and knee side. The final mean ACL T2* value was lowest in the normal ACL group, higher in the partial ACL tear group and highest in the complete ACL tear group. Median (IQR) final ACL T2* values were 11.69 (11.20–12.04) ms, 14.83 (13.46–15.45) ms, and 17.52 (16.84–22.12) ms in the normal ACL, partial ACL tear, and complete ACL tear groups, respectively. A statistically significant difference was observed among the three groups using the Kruskal–Wallis test (H = 59.948, p < 0.001), with a large effect size (epsilon-squared = 0.666). Dunn–Bonferroni post hoc analysis showed significant differences between all pairwise group comparisons. Interobserver reliability was excellent, with an overall ICC (2, 1) of 0.998 (95% CI, 0.997–0.998). Adjusted robust multivariable regression analysis showed that ACL category remained significantly associated with final mean ACL T2* value after adjustment for age, sex, and knee side. Clinically available T2* mapping showed significant quantitative differences among MRI-defined normal ACL, partial ACL tear, and complete ACL tear categories. However, the study did not evaluate whether T2* mapping improves the diagnostic accuracy of conventional MRI, provides incremental diagnostic value, or establishes diagnostic thresholds. Further prospective diagnostic-accuracy studies using an independent reference standard, such as arthroscopy where clinically indicated, are required.