Carpal angles as measured on CT and MRI: can we simply translate radiographic measurements?

Abstract

To determine the reliability of carpal angles measured on CT and MRI compared to radiography and assess if these measurements are interchangeable. Our institutional ethic research committee approved this study. For this retrospective study, two independent observers measured the scapholunate (SL), capitolunate (CL), radiolunate (RL), and radioscaphoid (RS) angles on 21 sets of exams, with each set including a radiograph, CT, and MRI of the same wrist. Inter- and intra-observer agreements were evaluated with the intraclass correlation coefficient (ICC). Linear mixed models and two-way contingency tables were used to determine if the angles measured on cross-sectional modalities were significantly different from those obtained on radiography. Inter-observer agreement was strong (ICC >0.8) for all angles, except for the RL angle measured on MRI (ICC 0.68). Intra-observer agreement was also strong for all angles, except for the CL angle measured on CT (ICC 0.66). SL angles measured on CT and MRI were not statistically different from those measured on radiographs (p = 0.37 and 0.36, respectively), unlike CL, RL, and RS angles (p < 0.05). Accuracy between modalities varied between 76 and 86% for the SL angle and ranged between 43 and 76% for the other angles. CL, RL, and RS angles showed large intermodality variability. Therefore, their measurements on CT or MRI could potentially lead to miscategorization. Conversely, our data showing no significant difference between modalities, SL angle could be measured on CT and MRI to assess wrist instability with a lower risk of error.