Top Line: With 3D imaging and auto-segmentation, a common question has become whether we can predict cardiac risk from adjuvant breast radiation with more granularity than just mean heart dose.
The Study: Not to knock mean heart dose, which has served as a reliable risk prediction model for the past decade. But this was established when mean heart dose was 3-fold higher than what we see with today’s use of deep inspiratory breath hold (DIBH) and sometimes even intensity-modulated radiation (IMRT). The Dutch, who have consistently made big headways in these investigations, are back at it with an auto-segmentation algorithm trained on contrast-enhanced cardiac scans and validated on 3D simulation scans. Cardiac structures including the heart as a whole (as we would normally do), individual cardiac chambers, the three main coronary arteries, and the aorta and pulmonary vessels were retroactively auto-contoured for 5800 patients receiving adjuvant breast radiation between 2005-2016. This was then correlated with data on subsequent hospitalizations and death due to heart disease, which occured in 135 patients (2.3%). In the end, risk of hospitalization/death from heart disease for the 5-10% of patients receiving the highest dose to any of the cardiac substructures (except for right atrium). Ok, but did any of these metrics reliably identify at-risk patients who didn’t also have a high tried-and-true mean heart dose? Nope. What was a stand-out predictor of poor heart disease outcomes? Presence of coronary artery calcification, with no interaction with radiation dose. In other words, we can auto-contour structures to our heart’s desire, but getting back to the basics with appropriate patient selection for inclusion of internal mammary node coverage that takes into account baseline cardiac risk as much as risk of cancer recurrence may prove the best way to maximize our therapeutic window.
TBL: Consistent with the original Darby data we all know and love, mean heart dose continues to be a good surrogate for conferred relative increased risk of future cardiac events following adjuvant breast radiation, meaning understanding a patient’s baseline risk is the most important predictor of all. | van Velzen, Int J Radiat Oncol Biol Phys 2022