Internal and external validation of an ESTRO delineation guideline - dependent automated segmentation tool for loco-regional radiation therapy of early breast cancer.

BACKGROUND AND PURPOSE: To internally and externally validate an atlas based automated segmentation (ABAS) in loco-regional radiation therapy of breast cancer. MATERIALS AND METHODS: Structures of 60 patients delineated according to the ESTRO consensus guideline were included in four categorized multi-atlas libraries using MIM Maestro™ software. These libraries were used for auto-segmentation in two different patient groups (50 patients from the local institution and 40 patients from other institutions). Dice Similarity Coefficient, Average Hausdorff Distance, difference in volume and time were computed to compare ABAS before and after correction against a gold standard manual segmentation (MS). RESULTS: ABAS reduced the time of MS before and after correction by 93% and 32%, respectively. ABAS showed high agreement for lung, heart, breast and humeral head, moderate agreement for chest wall and axillary nodal levels and poor agreement for interpectoral, internal mammary nodal regions and LADCA. Correcting ABAS significantly improved all the results. External validation of ABAS showed comparable results. CONCLUSIONS: ABAS is a clinically useful tool for segmenting structures in breast cancer loco-regional radiation therapy in a multi-institutional setting. However, manual correction of some structures is important before clinical use. The ABAS is now available for routine clinical use in Danish patients.