Predicting Reparability of Meniscal Tears Pre-Operatively Using 3-Tesla MRI (111)

OBJECTIVES: The ability to predict meniscus tear reparability based on pre-operative MRI is desirable for pre-operative patient counseling. However, the accuracy of MRI-based predictive methods varies widely within the orthopedic and radiology literature. We hypothesized that modern higher resolution 3-Tesla (T) MRI improves the accuracy of predicting reparability compared to prior investigations using 1.5T MRI assessments. METHODS: We identified 44 patients (age 16 to 40 years) who were known to have undergone arthroscopic meniscal repair at our institution between the dates of January 1, 2013 and June 1, 2019. The MRI characteristics of this meniscus repair group were then compared to 43 age- and sex-matched patients who underwent arthroscopic partial meniscectomy during the same time period. 3T MRI images from the repair (Figure 1A) and the partial meniscectomy (Figure 1B) groups were all obtained pre-operatively at the author’s institution. Images from all 87 patients were independently reviewed by two fellowship-trained musculoskeletal radiologists and one orthopedic surgery fellow specializing in sports medicine. Each examiner was blinded with regard to meniscus tear treatment (repair versus partial meniscectomy). Meniscal tear MRI characteristics were evaluated based on established arthroscopic criteria including tear length greater than 10 mm, tear location within 3 mm of the menisco-synovial junction, tear greater than 50% thickness, and the presence of an intact inner meniscal fragment. We then analyzed the predictive accuracy and interrater reliability of this method. RESULTS: With regards to accurately predicting meniscal reparability using the established criteria, the three MRI examiners accurately predicted repair 58% (orthopedist), 60% (radiologist 1), and 63% (radiologist 2) of the time, with respective positive predictive values of 60%, 62%, and 70%. The three examiners agreed upon tear reparability (i.e., a score of 4 versus not 4) only 41% of the time (κ = 0.173, p = .005). For 2 of the examiners (orthopedist and radiologist), none of the individual criteria were significantly predictive of tear reparability. For the 3(rd) examiner (radiologist), tear location within 3mm of the meniscosynovial junction was the most predictive individual criterion and the only criteria that reached statistical significance (OR = 9.83, p = .04). CONCLUSIONS: Although 3T MRI is higher resolution than 1.5T MRI, 3T MR imaging assessments performed by experienced examiners demonstrated a poor ability to predict the reparability of meniscus tears based on the application of previously established arthroscopic criteria. In addition, inter-observer reliability in this setting was also poor. Arthroscopic inspection remains the gold standard for the determination of meniscus tear reparability.