Refined Myocardium Segmentation from CT Using a Hybrid-Fusion Transformer

1. Abstract Accurate segmentation of the left ventricle (LV) in cardiac CT images is crucial for assessing ventricular function and diagnosing cardiovascular diseases. Creating a sufficiently large training set with accurate manual labels of LV can be cumbersome. More efficient semi-automatic segmentation, however, often includes unwanted structures, such as papillary muscles, due to low contrast between the LV wall and surrounding tissues. This study introduces a two-input-channel method within a Hybrid-Fusion Transformer deep-learning framework to produce refined LV labels from a combination of CT images and semi-automatic rough labels, effectively removing papillary muscles. By leveraging the efficiency of semi-automatic LV segmentation, we train an automatic refined segmentation model on a small set of images with both refined manual and rough semi-automatic labels. Evaluated through quantitative cross-validation, our method outperformed models that used only either CT images or rough masks as input. Competing Interest Statement The authors have declared no competing interest. Footnotes ↵# J. Woo and I. Aganj are co-senior authors. Funding Information: This work was supported by the Connors Center for Women’s Health and Gender Biology at the Brigham and Women’s Hospital (BWH) and the Office of Chief Academic Officer at Mass General Brigham (MGB) through a Connors BWH-MGB Collaborative IGNITE Award. Additional funding was provided by the National Institutes of Health (R01AG068261). The authors do not declare any conflict of interest relevant to this work. This version has been revised to prepare the manuscript for submission to another journal. Minor textual edits and formatting adjustments have been made. No substantive changes were made to the scientific content.