Assessment of a new 3D visualisation technique in CT Colonography, the Funnel view, compared to standard interpretation strategies and other recent 3D rendering technologies

F. Chandelier, L. Stein, T. Cabrera; ECR 2012 Book of Abstracts (B-0142)
Purpose: The study objective is to assess the performance of a new 3D rendering technique applied to CT colonography (CTC), namely the Funnel view (FV). It aims at addressing current reviewing short-comings during uni- and bi-directional standard reviews (USV & BSV) such as limited colonic surface coverage despite long reviewing-time [EurRadiol;(2007)61:388-399], and the higher degree of distortions and reduced "polyp display time" of new rendering techniques [EurRadiol;(2011)21:653-662].

Methods and Materials: We designed a unique 3D virtual lens, arranging rays transmission based on the best ray-casting strategy optimising the respective field of view. It allows for distortion-limited visualisation of areas behind folds within a traditional perspective rendering. 20 endoscopy-validated CTC datasets with good colonic distensions were involved in this retrospective study, to compare USV, BSV and FV; 3D manual fly-through was deemed reference standard by manually assessing the complete colonic surface visible. Furthermore, FV will be involved in an academic protocol (McGill University, Quebec, Canada) to prospectively investigate its contribution to CTC reviews.

Results: A single fly-through with FV rendered 100% of the visible colonic mucosa, compared to #73% and #92%, respectively, for the USV and BDV. In addition, it significantly increased the display-time closer to the display periphery, required for quality assessment of potential lesions without inducing major distortion.

Conclusion: Funnel view for CTC provides a time-efficient interpretation strategy to assess the entire colonic mucosa with a single automatic fly-through, increasing potential lesion conspicuity and display-time. It has the potential to significantly reduce reviewing time, while maintaining sensitivity and specificity.