Retrospective evaluation of visually monitored deep inspiration breath hold for breast cancer patients using edge detection

Abstract

Purpose: Deep inspiration breath hold (DIBH) can reduce cardiac dose during left-sided breast cancer radiotherapy. This study uses cine imaging with edge detection to evaluate a visually-monitored DIBH technique (VM-DIBH). Methods: Cine images were acquired weekly during the medial tangent field of patients treated with VM-DIBH. Edge detection was used to identify the field borders and chest wall edges in digitally reconstructed radiographs (DRRs) and cine images of 15 patients. The distance between the field border and chest wall was measured at the geometric center of the field and used as a surrogate for patient position during breath hold. Setup uncertainties were found by comparing DRR measurements to the first cine image measurement for each fraction. Intra-beam motion during individual breath holds was assessed by comparing the first cine image measurement to all subsequent cine image measurements for each fraction. Results: The mean setup uncertainty (M) was 1.2 mm; random (ct) and systematic (E) setup errors were both 2.0 mm. The chest wall position was within 5 mm of the DRR position in 92% of cine images. Intra-beam motion was within +/- 2 mm in 98% of images, and was slightly skewed in the posterior direction, indicating that patients tend to relax or exhale during breath hold. Conclusion: Edge detection of field borders and the chest wall in cine images and DRRs was successfully used to evaluate inter-fraction and intra-beam uncertainties for a VM-DIBH technique. Setup uncertainties and chest wall position measurements indicated adequate breath hold setup reproducibility for the majority of patients. Intra-beam motion measurements showed excellent stability of breath hold during treatment.