An open source auto-segmentation algorithm for delineating heart and substructures - Development and validation within a multicenter lung cancer cohort.

BACKGROUND AND PURPOSE: Irradiation of the heart in thoracic cancers raises toxicity concerns. For accurate dose estimation, automated heart and substructure segmentation is potentially useful. In this study, a hybrid automatic segmentation is developed. The accuracy of delineation and dose predictions were evaluated, testing the method's potential within heart toxicity studies. MATERIALS AND METHODS: The hybrid segmentation method delineated the heart, four chambers, three large vessels, and the coronary arteries. The method consisted of a nnU-net heart segmentation and partly atlas- and model-based segmentation of the substructures. The nnU-net training and atlas segmentation was based on lung cancer patients and was validated against a national consensus dataset of 12 patients with breast cancer. The accuracy of dose predictions between manual and auto-segmented heart and substructures was evaluated by transferring the dose distribution of 240 previously treated lung cancer patients to the consensus data set. RESULTS: The hybrid auto-segmentation method performed well with a heart dice similarity coefficient (DSC) of 0.95, with no statistically significant difference between the automatic and manual delineations. The DSC for the chambers varied from 0.78-0.86 for the automatic segmentation and was comparable with the inter-observer variability. Most importantly, the automatic segmentation was as precise as the clinical experts in predicting the dose distribution to the heart and all substructures. CONCLUSION: The hybrid segmentation method performed well in delineating the heart and substructures. The prediction of dose by the automatic segmentation was aligned with the manual delineations, enabling measurement of heart and substructure dose in large cohorts. The delineation algorithm will be available for download.

Consequences of introducing geometric GTV to CTV margin expansion in DAHANCA contouring guidelines for head and neck radiotherapy.

BACKGROUND AND PURPOSE: Defining margins around the Gross Tumour Volume (GTV) to create a Clinical Target Volume (CTV) for head and neck cancer radiotherapy has traditionally been based on presumed knowledge of anatomical routes of spread. However, using a concentric geometric expansion around the GTV may be more reproducible. The purpose of this study was to analyse the inter-observer consistency of geometric CTV delineation with adaptation for anatomical boundaries versus anatomically defined CTVs. MATERIAL AND METHODS: Radiation oncologists at four Danish cancer centres delineated high, intermediate and elective dose CTVs (CTV1, CTV2 and CTV3, respectively) in a patient-case template (stage IV squamous cell carcinoma of the oropharynx), first using mainly anatomical margins (original standard) and then using concentric geometric expansion (new standard). Each centre made a dummy-run radiotherapy plan based on the delineated CTVs. The difference between the CTV contours and the radiotherapy plans was evaluated across the centres. RESULTS: Anatomy-based contours were significantly more heterogenous and showed larger volume differences between centres than geometric margins. Dice similarity coefficient increased by 0.29 and mean surface distance decreased by 4mm for CTV1. Use of consistent CTV volumes resulted in more consistent irradiated volumes between centres. CONCLUSION: Introduction of geometric margins resulted in more uniform CTV1 and CTV2 delineation. Geometric CTV expansion was easier, left less room for misinterpretation, and resulted in more uniform treatment plans with similar irradiated high and intermediate dose volumes across all centres.