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.

Risk of coronary artery disease after adjuvant radiotherapy in 29,662 early breast cancer patients: A population-based Danish Breast Cancer Group study.

PURPOSE: Radiotherapy (RT) for early breast cancer (BC) reduces the risk of recurrence and improves overall survival. However, thoracic RT may cause some incidental RT dose to the heart with subsequent risk of heart disease. During 2000-2010, CT-based RT planning was gradually introduced. The aim of this study was to investigate the risk of cardiac events in left-sided compared with right-sided BC patients treated during a non-CT-based (1999-2007) vs a CT-based period (2008-2016). MATERIAL AND METHODS: Information on BC and cardiac events among Danish women was obtained from population-based medical registers. Patients diagnosed with BC during 1999-2016, were included. A cardiac event was defined as coronary artery disease or severe valvular heart disease. RESULTS: Among 29,662 patients, 22,056 received RT. For those irradiated during the non-CT-based period, the 10-year cumulative risk of cardiac event was 1.7% (95% CI 1.4-2.0) at median follow-up of 11.1 years. The incidence rate ratio (IRR) for cardiac event in left-sided vs right-sided patients was 1.44 (1.07-1.94) and a trend towards worse outcome was seen within the first 10 years after RT and approached statistical significance with longer follow-up. Among patients irradiated during the CT-based period, the 10-year cumulative risk of cardiac event was 2.1% (1.8-2.4) at median 6.8 years follow-up. The IRR for cardiac event in left-sided vs right-sided patients was 0.90 (0.69-1.16) and no trend towards worse outcome within the first 10 years was observed. CONCLUSION: This study confirmed a higher risk of cardiac events in left-sided vs right-sided BC patients irradiated during a non-CT-based period. For patients irradiated during a CT-based period, no increased risk of cardiac events in left-sided vs right-sided patients was observed within the first 10 years after RT, whilst information on cardiac events beyond 10 years after RT was limited.

Delineation of whole heart and substructures in thoracic radiation therapy: National guidelines and contouring atlas by the Danish Multidisciplinary Cancer Groups.

BACKGROUND AND PURPOSE: This study presents Danish consensus guidelines for delineation of the heart and cardiac substructures across relevant Danish Multidisciplinary Cancer Groups. MATERIAL AND METHODS: Consensus guidelines for the heart and cardiac substructures were reached among 15 observers representing the radiotherapy (RT) committees of four Danish Multidisciplinary Cancer Groups. The guidelines were validated on CT scans of 12 patients, each with five independent contour sets. The Sørensen-Dice similarity coefficient (DSC), the distance between the centers of the arteries and the mean surface distance were used to evaluate the inter-observer variation. RESULTS: National guidelines for contouring the heart and cardiac substructures were achieved. The median DSC was 0.78-0.96 for the heart and the four cardiac chambers. For the four substructures of the left ventricle, the median DSC was 0.35-0.57. The coronary arteries were contoured in ten segments, with the best agreement for the left anterior descending coronary artery segments, with a median distance between the arteries ranging from 2.4-4.4 mm. The median variation was 3.7-12.8 mm for the right coronary artery segments and 3.7-6.2 mm for the left circumflex coronary artery segments, with the most pronounced inter-observer variation in the distal segment for all three coronary arteries. CONCLUSION: National guidelines for contouring the heart and cardiac substructures were developed across relevant Danish Multidisciplinary Cancer Groups, where RT dose to the heart is of concern. The inter-observer contour overlap was best for the heart and chambers and decreased for smaller structures.