Transcatheter Aortic Valve Replacement Planning Pseudo-label CT dataset (TAVRP-PL)

This dataset contains volumetric segmentation pseudo-labels of relevant anatomical structures for Transcatheter Aortic Valve Replacement (TAVR) Planning and their associated computed tomography scans.

It is based on the TotalSegmentator dataset [1], from which a subset of labels are used to serve as basis for deriving additional anatomical TAVR-relevant structures, namely:

• Class 1: Aorta
• Class 2: Aortic Root
• Class 3: Valve
• Class 4: Left Ventricle
• Class 5: Annulus
• Class 6: Iliac Artery Left
• Class 7: Iliac Artery Right

This datasets contains a subset of 578 CT scans from the TotalSegmentator dataset and it is structured in train, validation and test splits, for training and evaluating segmentation models.

More information about this pseudo-label TAVR dataset can be found in an accompaying paper [2].

Details on the TotalSegmentator dataset, which we base our additional labels on can be found in their paper or dataset webpage.

[1] Jakob Wasserthal. (2023). Dataset with segmentations of 117 important anatomical structures in 1228 CT images (2.0.1) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.10047292

[2] Zöllner, C., Reiß, S., Jaus, A., Sholi, A., Sodian, R., and Stiefelhagen, R., “Semantic Segmentation for Preoperative Planning in Transcatheter Aortic Valve Replacement”, https://arxiv.org/abs/2507.16573