ARRS 2022 Abstracts

RETURN TO ABSTRACT LISTING


E2162. 3D MRI of the Rotator Cuff: What is the Added Value versus 2D MRI
Authors
  1. Leila Aghaghazvini; Department of Radiology, Shariati Hospital, Tehran University of Medical Sciences
  2. Fardis Vosughi; Department of Orthopedics, Shariati Hospital, Tehran University of Medical Sciences
  3. Mahyar Daskareh; Department of Radiology, Shariati Hospital, Tehran University of Medical Sciences
  4. Mohammadreza Guiti; Department of Orthopedics, Shariati Hospital, Tehran University of Medical Sciences
Objective:
Selecting proper surgical planning for a torn rotator cuff requires accurate preoperative patient information including clinical examination and imaging data. In the current study we aim to use 3D MRI to predict arthroscopic findings on tear shape and compare its accuracy with the conventional two-dimensional MRI.

Materials and Methods:
All patients underwent routine standard 2D protocol MRI examination (3.0 T DiscoveryMR750w, GE Healthcare) plus 3D-MERGE isotropic sequence with specific parameters for evaluating a clinical suspicion of a completely torn rotator cuff. Images were analyzed and post-processed for 3D reconstruction with Advantage Workstation (ADW 4.7-Volume Viewer application, GE Healthcare). We categorized the shape of tearing to be massive, crescentic, and longitudinal. One expert musculoskeletal radiologist (10 years of experience) evaluated the MRI, and an expert shoulder surgeon (25 years of experience) did arthroscopy.

Results:
Data were gathered from 42 consecutive patients (31 men) with a mean age of 57.43 years (SD: 6.32). Arthroscopic examination showed a massive tear in 19 (45.2%), crescent tear in 17 (40.5%), and longitudinal (U or L shaped) tear in six (14.3%) patients. 2D MRI resulted in 16 (38.1%) cases of massive tear, 17 (40.5%) cases of crescent tear, and nine (21.4%) cases of longitudinal tear. 3D MR imaging diagnosed 18 (42.9%) patients with massive tears, 17 (40.5%) patients with crescent tears, and seven (16.7%) patients with longitudinal tears. Sensitivity and specificity values for a crescent tear were 70.5% and 80% for 2D and 88% and 92% for 3D, respectively. For a longitudinal tear, sensitive and specificity were 88% and 89% on 2D and 83% and 94% on 3D, respectively. For a massive tear, sensitivity and specificity were 68% and 87% on 2D and 89% and 96% on 3D, respectively. Our results showed that 3D MRI had the closest prediction of a rotator cuff tear and its shape (p = 0.02)

Conclusion:
3D MRI has the closest prediction of a rotator cuff tear and its shape respective to the arthroscopic findings in all types of complete tears, as compared to 2D MRI.