MO‐A‐137‐07: Assessment of Metrics to Identify Which Left‐Sided Breast Cancer Patients Will Benefit From a Moderate Deep Inspiration Breath‐Hold Technique

Abstract

Purpose: To derive and examine the robustness of metrics for prediction of dosimetric benefits for left‐sided breast cancer patients from a moderate deep inspiration breath‐hold (mDIBH) technique. Methods: Radiotherapy treatment plans using a field‐in‐field tangent approach were retrospectively generated on both free‐breathing (FB) and mDIBH CT scans acquired for fourteen consecutive left‐sided breast cancer patients. Breast PTV‐eval, heart and left anterior descending coronary artery (LAD) were defined as per RTOG and the heart atlas (Feng et al. IJROBP 79:10‐18, 2011). The same PTV coverage was maintained in both FB and mDIBH plans. Volumetric metrics, irradiated heart and lung volume, and distance metrics from representative images were defined. Correlations with dosimetric parameters and robustness were examined. Results for single planner retrospective plans were compared with results from treated plans prepared by a group of dosimetrists for both the cohort of 14 mDIBH patients as well as a cohort of 19 previously treated FB left‐sided breast cancer patients. Results: FB cardiac dosimetric parameters for our cohort appeared to follow exponential distributions, with a significant reduction of sigma shown for mDIBH. The extent of dosimetric benefit is patient specific. Linear correlations were found between cardiac and pulmonary dosimetric parameters and volumetric metrics using irradiated volumes, as well as for distance metrics if volumetric corrections were applied. Linear correlations were preserved (R > 0.9) regardless of whether determined from a single planner or a group of dosimetrists. Fitted parameters were slightly higher for data derived from the group of planners, however, the difference was not found to be significant via signed rank test. Conclusion: Linear correlations between cardiac and pulmonary dosimetric parameters and several metrics were identified and determined to be robust for left‐sided breast cancer patients planned at our institution, which facilitates reliable selection of patients who will benefit from a mDIBH technique.