Facility and Individual-Level Factors Contribute to Racial Disparities in Heart Dose Among Breast Radiotherapy Patients
Recommended Citation
Chapman CH, Jagsi R, Griffith K, Laucis AM, Moran JM, Hayman JA, Walker EM, Abu-Isa EI, Dominello MM, and Pierce LJ. Facility and Individual-Level Factors Contribute to Racial Disparities in Heart Dose Among Breast Radiotherapy Patients. Int J Radiat Oncol Biol Phys 2019; 105(1):E53.
Document Type
Conference Proceeding
Publication Date
8-2019
Publication Title
Int J Radiat Oncol Biol Phys
Abstract
Purpose/Objective(s): Racial disparities in breast cancer presentation (stage/subtype), baseline cardiac risk, and access to care highlight the need to quantify and address predictors of racial variation in heart dose among women who receive radiotherapy (RT). We sought to explore factors that contribute to heart dose, including intrinsic patient factors and facility level factors. Materials/Methods: We queried a statewide consortium database to examine racial differences in mean heart dose among women treated with whole breast RT at 25 institutions. We generated separate models of heart dose based on disease laterality and receipt of conventional (CWBI) versus accelerated whole breast irradiation (AWBI). We included demographic, disease, and treatment characteristics expected to affect heart dose, as well as facility type. We created a model with only patient-level characteristics followed by a multi-level model to account for clustering within facilities. Results: Among 9042 women in the analytic sample, estimated mean heart doses (Gy) were: 1.74 for left-sided AWBI, 1.60 for left-sided CWBI, 0.61 for right-sided AWBI and 0.66 for right-sided CWBI. On patient-level multivariable analysis, race was an independent predictor of higher heart dose for women with left-sided disease: Black patients receiving ABWI or CWBI and Asian patients receiving AWBI had higher heart doses than White women. Higher heart dose was also associated with separation, breast volume, inclusion of internal mammary nodes, use of intensity modulated RT, supine positioning, dose to 50% of the breast volume, treatment at an academic center, decreasing obesity, decreasing comorbidities, absence of deep inhalation breath hold, and earlier treatment year. Multilevel modeling revealed that 22-30% of the variability in heart dose was attributed to patient clustering within facilities. Multilevel models suggest that heart dose is elevated for Black and most Asian (v. White) patients, with average increase between 3-13% and 6-22%, respectively, and statistically significant (p<0.02) depending upon laterality and fractionation. Multilevel modeling uncovered disparities for Black and Asian right-sided AWBI patients that were not observed in patient-level models. Conclusion: Mean heart doses were higher for Black and Asian women in this sample, even when accounting for relevant patient-level factors. Accounting for treatment facility decreased, but did not eliminate this disparity for left-sided disease and uncovered disparities for right-sided disease. These findings suggest that disparities in heart dose may be influenced by patient factors and the facilities at which women obtain care. Further research is needed to clarify whether disparities for Black and Asian women are due to unmeasured, unmodifiable anatomic or clinical factors versus modifiable individual or system-level factors.
Volume
105
Issue
1
First Page
E53