Diagnostic Comparison of 1.5 Tesla and 3.0 Tesla Preoperative MRI of the Wrist in Patients With Ulnar-Sided Wrist Pain

Abstract

The diagnostic sensitivity, specificity, and accuracy of 1.5 Tesla (T) and of 3.0T magnetic resonance imaging (MRI) are correlated with wrist arthroscopy findings in patients presenting with ulnar-sided wrist pain. The records and diagnostic MRI scans of 102 patients who presented between 1997 and 2006 with ulnar-sided wrist pain were evaluated. Preoperative MRI scans at 1.5T (n = 70) and 3.0T (n = 32) were evaluated by 2 experienced musculoskeletal radiologists with different levels of experience who were blinded to the arthroscopic findings. Preoperative MRI findings for the triangular fibrocartilage complex (TFCC), scapholunate, ulnotriquetral, and lunotriquetral ligaments were recorded and compared with findings at diagnostic arthroscopy. The sensitivity, specificity, and accuracy were calculated for both the 1.5T and 3.0T preoperative MRI scans. Statistical comparisons were made using chi-square test and JMP 6.0 software. A tear of the TFCC was identified retrospectively on 1.5T images in 49 of 58 patients and on 3.0T images in 15 of 16 patients. Compared with the gold standard of arthroscopy, 1.5T wrist MRI in this patient population had a sensitivity of 85%, a specificity of 75%, and an accuracy of 83% for reader 1 for the detection of a tear of the TFCC. In the same patient population, 3.0T wrist MRI had a sensitivity of 94%, a specificity of 88%, and an accuracy of 91% for reader 1. For reader 2, the improvement in sensitivity for the lunotriquetral ligament between the 1.5T and 3.0T images was statistically significant. The sensitivity, specificity, and accuracy of 3.0T wrist MRI for the TFCC is consistently higher compared with those of 1.5T wrist MRI. The trend suggests that 3.0T wrist MRI provides improved capability for detection of TFCC injuries. Given the available sample size, however, the confidence intervals around the point estimates are wide and overlapping. Further studies are needed to confirm or refute our results of the estimated sensitivity, specificity, and accuracy parameters.