English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Planning the optimal time of intervention for pulmonary valve replacementsurgery in patients with the congenital heart disease Tetralogy of Fallot (TOF)is mainly based on ventricular volume and function according to currentguidelines. Both of these two biomarkers are most reliably assessed bysegmentation of 3D cardiac magnetic resonance (CMR) images. In several grandchallenges in the last years, U-Net architectures have shown impressive resultson the provided data. However, in clinical practice, data sets are more diverseconsidering individual pathologies and image properties derived from differentscanner properties. Additionally, specific training data for complex rarediseases like TOF is scarce.  For this work, 1) we assessed the accuracy gap when using a publiclyavailable labelled data set (the Automatic Cardiac Diagnosis Challenge (ACDC)data set) for training and subsequent applying it to CMR data of TOF patientsand vice versa and 2) whether we can achieve similar results when applying themodel to a more heterogeneous data base.  Multiple deep learning models were trained with four-fold cross validation.Afterwards they were evaluated on the respective unseen CMR images from theother collection. Our results confirm that current deep learning models canachieve excellent results (left ventricle dice of$0.951\pm{0.003}$/$0.941\pm{0.007}$ train/validation) within a single datacollection. But once they are applied to other pathologies, it becomes apparenthow much they overfit to the training pathologies (dice score drops between$0.072\pm{0.001}$ for the left and $0.165\pm{0.001}$ for the right ventricle).

How well do U-Net-based segmentation trained on adult cardiac magnetic  resonance imaging data generalise to rare congenital heart diseases for  surgical planning?