Lucas Busnardo Ramadan1, Eduardo Baptista2, Felipe Ferreira de Souza3, Mauro Emilio Conforto Gracitelli4, Jorge Henrique Assunção1, Fernando Brandao Andrade-Silva4, Arnaldo Amado Ferreira-Neto5, Eduardo Angeli Malavolta4. 1. MD. Attending Orthopedic Surgeon, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Sao Paulo, SP, BR. 2. MD, PhD. Attending Musculoskeletal Radiologist, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Sao Paulo, SP, BR; Attending Musculoskeletal Radiologist, Imaging Department, Musculoskeletal Radiology Division, Hospital Israelita Albert Einstein (HIAE), São Paulo, Brazil. 3. MD. Attending Musculoskeletal Radiologist, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Sao Paulo, SP, BR. 4. MD, PhD. Attending Orthopedic Surgeon, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Sao Paulo, SP, BR. 5. MD, PhD. Head of the Shoulder and Elbow Group, Orthopedic Department, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Sao Paulo, SP, BR.
Abstract
BACKGROUND: The accuracy of magnetic resonance imaging (MRI) for making the diagnosis of subscapularis tears presents wide variation in the literature and there are few prospective studies. OBJECTIVE: To compare the findings from MRI and arthroscopy for diagnosing subscapularis tears. DESIGN AND SETTING: Diagnostic test study performed in a tertiary care hospital. METHODS: We included patients who underwent arthroscopic rotator cuff repair and who had firstly undergone high magnetic field MRI without contrast. The images were independently evaluated by a shoulder surgeon and two musculoskeletal radiologists. Sensitivity, specificity, positive and negative predictive values, accuracy and inter and intra-observer agreement were calculated. RESULTS: MRIs on 200 shoulders were evaluated. The incidence of subscapularis tears was 69.5% (41.5% partial and 28.0% full-thickness). The inter and intra-observer agreement was moderate for detection of subscapularis tears. The shoulder surgeon presented sensitivity of 51.1% to 59.0% and specificity of 91.7% to 94.4%. The radiologists showed sensitivity of 83.5% to 87.1% and specificity of 41% to 45.9%. Accuracy ranged from 60.5% to 73.0%. CONCLUSION: The 1.5-T MRIs without contrast showed mean sensitivity of 70.2% and mean specificity of 61.9% for detection of subscapularis tears. Sensitivity was higher for the musculoskeletal radiologists, while specificity was higher for the shoulder surgeon. The mean accuracy was 67.6%, i.e. lower than that of rotator cuff tears overall.
BACKGROUND: The accuracy of magnetic resonance imaging (MRI) for making the diagnosis of subscapularis tears presents wide variation in the literature and there are few prospective studies. OBJECTIVE: To compare the findings from MRI and arthroscopy for diagnosing subscapularis tears. DESIGN AND SETTING: Diagnostic test study performed in a tertiary care hospital. METHODS: We included patients who underwent arthroscopic rotator cuff repair and who had firstly undergone high magnetic field MRI without contrast. The images were independently evaluated by a shoulder surgeon and two musculoskeletal radiologists. Sensitivity, specificity, positive and negative predictive values, accuracy and inter and intra-observer agreement were calculated. RESULTS: MRIs on 200 shoulders were evaluated. The incidence of subscapularis tears was 69.5% (41.5% partial and 28.0% full-thickness). The inter and intra-observer agreement was moderate for detection of subscapularis tears. The shoulder surgeon presented sensitivity of 51.1% to 59.0% and specificity of 91.7% to 94.4%. The radiologists showed sensitivity of 83.5% to 87.1% and specificity of 41% to 45.9%. Accuracy ranged from 60.5% to 73.0%. CONCLUSION: The 1.5-T MRIs without contrast showed mean sensitivity of 70.2% and mean specificity of 61.9% for detection of subscapularis tears. Sensitivity was higher for the musculoskeletal radiologists, while specificity was higher for the shoulder surgeon. The mean accuracy was 67.6%, i.e. lower than that of rotator cuff tears overall.