Prediction Models of Cardiac Sparing during Deep Inspiration Breath-Hold in Left-sided Breast Cancer with Internal Mammary Node Irradiation

Abstract

Purpose To evaluate the cardiac dosimetric advantages of deep inspiration breath-hold (DIBH) in postoperative left-sided breast cancer patients undergoing internal mammary node irradiation (IMNI) and construct prediction models using factors influenced cardiac sparing through DIBH. Methods Forty-five postoperative left-sided breast cancer patients receiving IMNI were collected in this study continuously, and 2 simulated computed tomography (CT) scans were performed in the free breathing (FB) and DIBH. Anatomical parameters were measured based on the FB planning CT scans. Dosimetric parameters were then compared between DIBH and FB, and predictors of cardiac sparing during DIBH were analyzed. Results Compared to FB, DIBH reduced the dose parameters of Dmean, Dmax, and V5-V30 for the heart and left anterior descending coronary artery (LAD) significantly (P < 0.05). Breast volume (BV), age and parasagittal cardiac contact distances (FB-CCDps) were related to mean heart dose reduction independently (P < 0.05), while initial LAD-to-manubrium distance (FB-LMD), BV, and age were identified as independent predictors of mean LAD dose reduction (P < 0.05). Mathematical prediction models were developed to predict cardiac sparing. Conclusions DIBH could effectively reduce the radiation dose for both of heart and LAD in postoperative left-sided breast cancer patients undergoing IMNI. The mathematical prediction models using BV, age, FB-LMD and FB-CCDps may help identify patients most likely to benefit from DIBH.