Characterization of mitral valve prolapse with cardiac computed tomography: comparison to echocardiographic and intraoperative findings

Abstract

A single imaging modality that can accurately assess both coronary anatomy and mitral valve (MV) anatomy prior to surgery may be desirable. We sought to determine the diagnostic accuracy of cardiac computed tomography (CT) to detect and characterize mitral valve prolapse (MVP) compared to echocardiography. Consecutive patients referred for 'single-source' cardiac CT for investigation prior to non-coronary cardiac sugery were identified. MV anatomy was assessed for MVP and results were compared to echocardiography and to intra-operative visual assessment of the MV. Comparison between the three modalities was performed at the per-patient, per-leaflet and per-scallop levels. A total of 67 consecutive patients that were referred for Cardiac CT prior to non-coronary cardiac surgery and were prospectively recruited into a Cardiac CT registry. Of these, 65 patients underwent cardiac surgery. 63 patients had echocardiography and 32 patients had intra-operative visual assessment of the mitral valve. Compared to echocardiography, cardiac CT had excellent sensitivity (92.6%) and specificity (97.1%) for the detection of any MVP, but had poor sensitivity (68.5%) for the detection of individual prolapsing scallop. Compared to intra-operative visual assessment of the prolapsing scallop, both cardiac CT and echocardiography had low sensitivity (58.1 and 78.1%, respectively). Cardiac CT was able to identify patients with MVP but had difficulty identifying the prolapsed scallops compared to echocardiography. Single-source CT may not be ready for characterization of individual mitral valve scallops.