Patterns and Disparities of Ablative Radiation Therapy Use in Patients with Metastatic Cancer: a Study of the National Cancer Database

Abstract

In this study, we assess demographic, tumor, and treatment factors associated with the use of ablative radiation therapy (ABRT) in patients with metastatic cancer. We also explore the outcomes associated with ABRT use. Using the National Cancer Database (NCDB), we identified adult patients with newly diagnosed metastatic prostate, breast, colorectal, and lung cancer from 2010 to 2014. We defined ABRT as RT received within 120 days of diagnosis with BED10 ≥ 50 Gy. The number of organ systems with metastatic disease was based on the NCDB variables for metastatic involvement of bone, liver, lung, and brain. Demographic, tumor, and treatment characteristics were compared between patients who did and did not receive ABRT. Cox proportional hazards (CPH) models were used to identify factors associated with ABRT use. We also performed exploratory analyses of overall survival (OS) using CPH models with a landmark time of 120 days to account for immortal time bias. Propensity score matching (PSM) was used to account for group differences between known confounders to reduce selection bias. Sensitivity analyses were performed to assess the stability of results. We identified 93,334 patients who met our inclusion criteria. 11,193 (12.0%) patients received ABRT and 82,141 (88.0%) did not. Most patients (62%) receiving ABRT had metastases to one organ. In multivariable CPH models, age ≤ 60 years, male sex, non-Hispanic white race, private or non-Medicare/Medicaid government insurance, and non-academic treatment facility were associated with ABRT use. Compared to patients of non-Hispanic white race, patients of Hispanic (HR 0.82; p<0.001), black (HR 0.88, p<0.001), and Asian/other race (HR 0.80; p<0.001) were less likely to receive ABRT. Compared to patients with private insurance, uninsured patients (HR 0.87, p=0.001) and those with Medicaid (0.93; p=0.02) and Medicare (HR 0.87; p<0.001) were less likely to receive ABRT. The median follow-up time for survivors from the landmark time of 120 days post-diagnosis was 25.2 months (range, 0.04-67.5 months). In an exploratory analysis, ABRT was significantly associated with improved OS (HR 0.91; 95% CI, 0.88-0.93; p<0.001) in a multivariable CPH model, as were receipt of surgery and chemotherapy. ABRT also was associated with improved OS among 6,577 PSM pairs (HR 0.90; 95% CI, 0.85-0.95; p<0.001). The OS improvement was robust in sensitivity analyses altering the landmark time and additionally altering/controlling for other covariates. In disease-specific CPH models, ABRT was associated with improved OS for all histologies except breast cancer. In disease-specific PSM models, ABRT was associated with improved OS in colorectal cancer only. Many patients with metastatic prostate, breast, lung, and colorectal cancer receive ABRT, predominantly younger, non-Hispanic white, and privately insured patients with fewer organs of metastatic disease. Patients selected for ABRT appear to have improved OS.