CATE (Coronary Artery Tortuosity Evaluator): A Semi-Automatic Tool for Quantitative Assessment of Coronary Artery Tortuosity from CT Angiography

In recent years, there has been a growing interest in evaluating coronary artery tortuosity and its role in cardiovascular risk stratification. Coronary Computed Tomography Angiography (CCTA) offers a non-invasive approach that provides a comprehensive three-dimensional view of the arteries which is ideal for assessing the presence of anomalous twists and bending in the vessel's course. However, coronary tortuosity is often overlooked as the evidence of its association with acquired or inherited cardiovascular diseases is still limited. Additionally, the absence of a universally accepted tortuosity assessment method poses challenges. To address this gap, our work focuses on developing a pipeline for CCTA image analysis implemented in 3D Slicer. The resulting 3D Slicer extension optimizes image management, guiding the user from visualization to the extraction of tortuosity measures while minimizing the need for manual interventions. Here we present the preliminary results regarding the 3D Slicer implementation of eight tortuosity measures and the validation of a new tortuosity classification method based on these indices. The validation phase was carried out by comparing the proposed classification method with the qualitative assessment employed by radiologists in clinical practice. The agreement between CATE and the evaluations provided by three independent readers has been assessed employing the Cohen's Kappa index. The Inflection Count Metric exhibited the weakest performance (Kappa = −0.08, no agreement), while the Tortuosity Index (ratio between the length of the vessel and the length of the cord between the two vessel endpoints), despite its widespread use in assessing arterial tortuosity, showed only minimal agreement (Kappa = 0.38). Curvature-based indices have emerged as the most promising tools for categorizing tortuosity patterns and enhancing the descriptive abilities of radiologists with the Average Absolute curvature index (Kappa = 0.86) and the Root Mean Square curvature index (Kappa = 0.84) reaching a strong agreement.