Accuracy of clinical evaluation and MRI reporting compared to knee arthroscopic findings

Abstract

Background: Intra-articular knee injuries are diagnosed using multiple modalities including Magnetic Resonance Imaging (MRI) or arthrography and various clinical tests. However, there is a high incidence of false negative and false positive findings in MRI assessment. The deeper intra articular structures such as the medial meniscus brings about the high incidence of false positives and false negatives in MRI assessment. The clinical tests also have benefits in diagnosing these injuries but its sensitivity and specificity is also under question. Arthroscopy has quickly become the “gold standard” but diagnostic arthroscopy on its own is rarely done and is only done before a treatment procedure. No regional studies have been done and only a few handful studies done worldwide especially after recent advances in 3 tesla MRI apparatus.

Objectives: The aim of this study was to assess the diagnostic accuracy based on various clinical tests and validity of MRI in our populace in association with arthroscopic findings.

Design: This was a multi-center prospective cross-sectional study conducted at Aga Khan University Hospital Nairobi (AKUHN), Kikuyu Hospital, M.P. Shah Hospital and Kijabe Hospital over a six-month period.

Methodology: Preoperative clinical diagnosis based on various specific tests and preoperative MRI was compared with operative diagnostic arthroscopic outcomes separately to ascertain the accuracy of each using arthroscopy as the “gold standard”. The study population categorical variables was described and summarized as percentages and counts by use of frequency tables and continuous variables was summarized using central tendency and dispersion (minimum, maximum, range, IQR, mean, standard deviation, median). Sensitivity, specificity, positive and negative predictive values, and finally accuracy was then tabulated.

Results: One hundred and forty patients were recruited with 85 males and 55 females undergoing knee arthroscopy. One hundred and forty six knees were evaluated with the right knee (53.1%) more injured than the left (49%). Lateral meniscus (53.1%) was found to be more injured than the medial meniscus (46.8%) with the posterior horn (medial-73.3%; lateral-60%) of the meniscus being the commonest part of the meniscus injured. Anterior Cruciate Ligament (ACL) injury (39 patients) was more common than PCL injury (9 patients). Among the clinical tests, joint-line tenderness was the most sensitive test (medial-69.1%; lateral-64.4%) and the McMurrays test most specific (medial-92.2%; lateral-91.9%). Thessalys test was not found to be more sensitive (medial-58.8%; lateral-50.8%) or specific (medial-89.6%; lateral-88.4%) as reported previously. For ACL tears, Lachman test was specific (96.4%) and the anterior drawer test was more sensitive (81.8%). Accuracy for medial meniscus tear (47%) on MRI was more than lateral meniscus (41%). MRI did poorly in diagnosing ACL tears due to over reporting of partial ACL tears (complete ACL tear-23%; partial ACL tear-4%).

Conclusion: The diagnosis of intra-articular injuries is still a challenge and both clinical examination and MRI has its limitations. MRI reporting is no better than clinical examination in diagnosing meniscal tears and clinical examination is better at diagnosing cruciate ligament injuries than MRI reporting. Overall the accuracy of MRI reporting is poor in our set-up.

Keywords: Knee Arthroscopy, Knee MRI, 3 Tesla, Accuracy, Knee examination