Accuracy of predicted orthogonal projection angles for valve deployment during transcatheter aortic valve replacement

Background: Multi-detector computed tomography (MDCT) predicted orthogonal projection angles have been introduced to guide valve deployment during transcatheter aortic valve replacement (TAVR). Our aim was to investigate the accuracy of MDCT prediction methods versus actual angiographic deployment angles. Methods: Retrospective analysis of 2 currently used MDCT methods: manual multiplanar reformations (MR) and the semiautomatic optimal angle graph (OAG). Paired analysis was used to compare the 2-dimensional distributions and means. Results: We included 101 patients with a mean (±SD) age of 81 ± 9 years. The MR and OAG methods were used in 46 and 55 patients, respectively. A ≥5% change from the predicted MDCT range in left anterior oblique/right anterior oblique (LAO/RAO) and the cranial/caudal (CRA/CAU) angle occurred in 42% and 58% of patients, respectively. The mean predicted versus actual deployment angles were significantly different (CRA/CAU: -2.6 ± 11.5 vs. -7.6 ± 10.7, p < 0.001; RAO/LAO 8.1 ± 10.9 vs. 9.5 ± 10.6, p = 0.048; respectively). The MR method resulted in a more accurate CRA/CAU angle (CRA/CAU: -4.6 ± 11.1 vs. -6.5 ± 11.8, p = 0.139; RAO/LAO 7.4 ± 11.2 vs. 10.4 ± 11.2, p = 0.008; respectively), whereas the use of the OAG resulted in a more accurate RAO/LAO angle (CRA/CAU: -0.9 ± 10.8 vs. -9±11.2, p < 0.001; RAO/LAO 9.05 ± 10.6 vs. 8.5 ± 9.9, p = 0.458; respectively). For the entire cohort, the 2-dimensional distributions and means of the predicted versus the actual angles were significantly different from each other (p < 0.001). We repeated our analysis using both MDCT methods and demonstrated similar results with each method. Conclusions: Currently used MDCT methods for TAVR implantation angles are significantly modified before actual valve deployment. Thus, further refinement of these prediction methods is required.