PO-1177 A heart valves contouring atlas on average intensity projection 4D-CT for lung cancer radiotherapy

A pulmonary vein atlas for radiotherapy planning

“…It would not be possible to encompass all of these factors in a user-friendly atlas, but the guidance presented provides basic principles. Observer accuracy metrics compared favourably to the other atlases in the cardiac substructure field (16)(17)(18)(19)(20). For example, the mean DSCs reported by Socha et al for a cardiac valve atlas ranged between 0.45 and 0.69, and HDMs were > 2 mm (20), whereas in this study, mean DSCs ranged between 0.78 and 0.83, and HDMs 1.1 mm.…”

“…Observer accuracy metrics compared favourably to the other atlases in the cardiac substructure field (16)(17)(18)(19)(20). For example, the mean DSCs reported by Socha et al for a cardiac valve atlas ranged between 0.45 and 0.69, and HDMs were > 2 mm (20), whereas in this study, mean DSCs ranged between 0.78 and 0.83, and HDMs 1.1 mm. As such atlas validation studies typically demonstrate lower concordance for small structures (16,17), the results presented are better than anticipated.…”

“…As radiotherapy is known to cause fibrosis of the cardiac structures, it is plausible that incidental dose deposited in the myocardial sleeves of the PVs during radical thoracic treatment may explain high rates of new arrhythmia during follow-up (15). A radiation dose response not been established for the PVs however, and these vessels have not been included in published cardiac atlas' to date (16)(17)(18)(19)(20). A standardised definition for delineation of the PVs will be important for investigators seeking to establish safe dose constraints for these structures.…”

Validation of an established deep learning auto-segmentation tool for cardiac substructures in 4D radiotherapy planning scans

“…Beyond cardiac motion, other unresolved issues in cardiac substructure segmentation include the lack of standardised great vessel definitions, as these are not defined in the published cardiac atlases [23] , [42] , [49] , [50] , [51] . Current atlases also do not recommend subtraction of the cardiac chamber blood pool, which is likely to be confounding how the dose metrics of the cardiac chamber muscles are interpreted [37] , [46] .…”