Rethinking the role of mean heart dose in breast cancer radiotherapy

Breast cancer radiotherapy has significantly improved outcomes by reducing locoregional recurrence and mortality.[1] However, as survivorship in breast cancer increases, long-term treatment-related toxicities, especially cardiac complications, are coming into sharper focus.[2] While mean heart dose (MHD) has traditionally served as the benchmark for estimating cardiac exposure, its adequacy in capturing the true extent of risk, particularly to vulnerable cardiac substructures, is increasingly questioned.[3,4] The study by Dash et al.,[5] examining whether MHD meaningfully predicts substructure doses in patients treated with adjuvant radiotherapy postmastectomy, adds a valuable prospective Indian dataset to the growing body of literature in this area. This prospective observational study enrolled 50 postmastectomy patients with breast cancer (25 left-sided and 25 right-sided) treated between July 2021 and July 2023 using 3D conformal radiotherapy (3D-CRT) with tangential fields.[5] The authors meticulously contoured the heart, cardiac chambers, and coronary vessels according to established guidelines. The central finding was that MHD poorly correlated with the mean doses (Dmean) received by individual cardiac substructures. In left-sided cases, the correlation between MHD and Dmean to the left anterior descending artery (LAD) was modest (R2, 0.196), and to the left ventricle (LV), even weaker (R2, 0.061). In right-sided cases, these associations were similarly limited. No R2 value for any cardiac structure exceeded 0.70, and importantly, some patients with relatively low MHDs still received alarmingly high focal doses, up to 40 Gy, to parts of the LAD and LV.[5] These results cast doubt on the adequacy of MHD alone to assess and mitigate cardiac risk. Nonetheless, the study has important limitations. It was restricted to patients treated with 3D-CRT tangents without any cardiac shielding or use of more advanced modalities such as deep inspiration breath-hold (DIBH), volumetric modulated arc therapy (VMAT), or proton therapy. As such, the findings cannot be extrapolated to more contemporary techniques that exhibit markedly different dose distributions. The study did not report on clinical cardiac outcomes, which limits the ability to directly translate dosimetric findings into patient risk stratification. Additionally, while all contours followed standard guidelines, interobserver variability in substructure delineation was not addressed, and the relatively small sample size further limits generalizability. Still, the uniform technique and consistent planning process add credibility to the observed patterns. The relevance of these findings becomes even more apparent when considered in light of evolving evidence. Darby et al.’s[6] seminal analysis demonstrated a linear 7.4% relative increase in major cardiac events per 1 Gy of MHD among patients with breast cancer treated decades ago. However, that work reflected an era of 2D planning and unshielded tangents, very different from current practice. Today, we understand that cardiac damage is more often related to localized high-dose regions than to uniformly distributed exposure. The LAD, for example, is frequently located in the high-dose volume of tangential beams and is particularly susceptible to injury.[7] Wennstig et al.,[8] showed that LAD dose correlated more strongly than MHD with coronary interventions in breast cancer survivors. Jacob et al.,[9] reported that 56% (50/89) of patients with left-sided breast cancer and an MHD <3 Gy still received LAD doses >40 Gy, highlighting the disparity between MHD and LAD dose. Furthermore, recent evidence indicates that equivalent MHDs across different techniques (3D-CRT versus VMAT versus proton therapy) can result in profoundly different internal dose distributions. VMAT, for instance, may lower LAD dose but increase low-dose exposure to the rest of the heart, sometimes leading to a higher MHD.[10] Proton therapy appears especially promising, often reducing both LAD and MHD simultaneously.[11] These insights have important clinical implications. First, there is an urgent need to adopt standardized, reproducible contouring guidelines for cardiac substructures to allow meaningful comparisons across studies and treatment centers.[12] Second, structure-specific dose constraints must be developed and validated, not just for the LAD and LV but also for the left atrium and right ventricle, which are increasingly recognized as relevant across various dose-fractionation schemes in breast cancer radiotherapy. Third, future research must integrate dosimetric analysis with clinical follow-up to establish dose-response relationships for substructures, enabling more accurate toxicity prediction models. Finally, in settings like India and other low- and middle- income countries, where baseline cardiovascular risk is high and long-term cardiology follow-up may be limited, upfront cardiac sparing is not merely an academic concern but a public health necessity. Dash et al.,[5] offer a compelling case for moving beyond MHD as the sole marker of cardiac risk in breast cancer radiotherapy. Their study underscores the limitations of population-level averages and highlights the importance of individualized, anatomy-aware dosimetry. Mean heart dose has served us well, but the time has come to outgrow it.