Purpose To evaluate the intra/inter-rater and diagnostic reliability of the sagittal plane adjusted patellar instability ratios (PIRs) compared to tibial tubercle-trochlear groove (TT-TG) distance alone while employing a matched case-control analysis for age and sex to minimize a potential confounding efect. Methods A retrospective case-control study was performed of all knee MRI studies of patients diagnosed with patellar instability, between 2005 and 2020 at a regional tertiary medical centre. Using a 1:1 case-control matching of sex and age at the time of the diagnosis, one control subject was assigned to each case of patellar instability. Measurements of TT-TG distance, sagittal patellar length (PL), sagittal patellar tendon length (PTL), TT-TG/PL ratio, and TT-TG/PTL ratio were conducted. Two orthopaedic surgery residents and a senior musculoskeletal radiologist were assigned to assess the intraand inter-rater reliability. Inter-class coeicients were calculated (ICC). The receiver operating characteristic (ROC) curve and area under curve (AUC) for each parameter were compared to evaluate for diagnostic reliability. Odds ratios (OR) and their 95% conidence intervals (CI) were calculated and a multivariable logistic regression model was performed to control for possible confounders.
ResultsThe study included 324 individuals (162 case-control matched pairs). In terms of intra-and inter-rater reliability, TT-TG/PL and TT-TG/PTL ratios showed an excellent correlation within and between readers (TT-TG/PL; intra-rater ICC 0.94 and inter-rater ICC 0.92, TT-TG/PTL; intra-rater ICC 0.91 and inter-rater ICC 0.88). The ROC curve showed a slightly greater AUC of the TT-TG/PL ratio compared to TT-TG distance alone (0.75 vs 0.73, p < 0.001). When applying the pathologic cutof of TT-TG ≥ 20 mm and TT-TG/PL ≥ 0.5; the calculated odds ratios for the above cutof were as follows; TT-TG distance alone had an OR of 14 (95% CI 1.8-106.5, p = 0.011) and OR for TT-TG/PL ratio was 23 (95% CI 3.1-170.3, p = 0.002). In the multivariable analysis, while controlling for height and weight, only the association between TT-TG/PL ratio and patellar dislocation remained statistically signiicant with an adjusted OR of 2.7 (CI 1.3-5.4, p = 0.006), compared to TTTG distance alone (OR = 1.9, n.s.). Conclusions Patellar instability ratios are signiicantly more reliable compared to TT-TG distance alone for the evaluation of patellar instability. Patellar instability ratios present superior diagnostic reliability, sensitivity and speciicity, and intra\ inter rater reliability. Thus, patellar instability ratios could function as a valuable diagnostic tool for the evaluation of patellar instability. Level of evidence III.