Valid MR imaging predictors of prior knee arthroscopy

Abstract

To determine whether fibrosis of the medial patellar reticulum (MPR), lateral patellar reticulum (LPR), deep medial aspect of Hoffa's fat pad (MDH), or deep lateral aspect of Hoffa's fat pad (LDH) is a valid predictor of prior knee arthroscopy. Institutional review board approval and waiver of informed consent were obtained for this HIPPA-compliant study. Initially, fibrosis of the MPR, LPR, MDH, or LDH in MR imaging studies of 50 patients with prior knee arthroscopy and 100 patients without was recorded. Subsequently, two additional radiologists, blinded to clinical data, retrospectively and independently recorded the presence of fibrosis of the MPR in 50 patients with prior knee arthroscopy and 50 without. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy for detecting the presence of fibrosis in the MPR were calculated. κ statistics were used to analyze inter-observer agreement. Fibrosis of each of the regions examined during the first portion of the study showed a significant association with prior knee arthroscopy (p < 0.005 for each). A patient with fibrosis of the MPR, LDH, or LPR was 45.5, 9, or 3.7 times more likely, respectively, to have had a prior knee arthroscopy. Logistic regression analysis indicated that fibrosis of the MPR supplanted the diagnostic utility of identifying fibrosis of the LPR, LDH, or MDH, or combinations of these (p ≥ 0.09 for all combinations). In the second portion of the study, fibrosis of the MPR demonstrated a mean sensitivity of 82%, specificity of 72%, PPV of 75%, NPV of 81%, and accuracy of 77% for predicting prior knee arthroscopy. Analysis of MR images can be used to determine if a patient has had prior knee arthroscopy by identifying fibrosis of the MPR, LPR, MDH, or LDH. Fibrosis of the MPR was the strongest predictor of prior knee arthroscopy.