Semantic Segmentation for Preoperative Planning in Transcatheter Aortic Valve Replacement

2025-07-22

AI Summary

To address the time-consuming and subjective manual measurement of critical anatomical structures—such as the aortic root and coronary ostia—in pre-procedural planning for transcatheter aortic valve replacement (TAVR), this work proposes a clinically deployable semantic segmentation method for CT imaging. We introduce a novel fine-grained pseudo-label generation strategy integrating anatomical prior constraints and uncertainty modeling, and enhance the segmentation loss function to improve boundary accuracy and sensitivity to small structures. Evaluated on a multi-center CT dataset, our model achieves a 1.27% Dice score improvement over baseline methods, demonstrating statistically significant superiority. The high-quality pseudo-label set and corresponding expert annotations are publicly released. This study advances AI-assisted TAVR planning toward clinical utility by enabling precise, interpretable, and verifiable preoperative assessment.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

When preoperative planning for surgeries is conducted on the basis of medical images, artificial intelligence methods can support medical doctors during assessment. In this work, we consider medical guidelines for preoperative planning of the transcatheter aortic valve replacement (TAVR) and identify tasks, that may be supported via semantic segmentation models by making relevant anatomical structures measurable in computed tomography scans. We first derive fine-grained TAVR-relevant pseudo-labels from coarse-grained anatomical information, in order to train segmentation models and quantify how well they are able to find these structures in the scans. Furthermore, we propose an adaptation to the loss function in training these segmentation models and through this achieve a +1.27% Dice increase in performance. Our fine-grained TAVR-relevant pseudo-labels and the computed tomography scans we build upon are available at https://doi.org/10.5281/zenodo.16274176.

Problem

• AI supports TAVR preoperative planning via semantic segmentation
• Deriving fine-grained labels from coarse anatomical data for segmentation
• Improving segmentation accuracy with modified loss function

Innovation

• AI supports TAVR planning via semantic segmentation
• Fine-grained pseudo-labels enhance segmentation accuracy
• Adapted loss function boosts Dice performance by 1.27%