Purpose/Objective(s): In the era of rituximab, large B-cell lymphoma (LBCL) with secondary skeletal involvement has been associated with worse overall survival (OS) and higher relapse rates. Prospective data suggest an improvement in progression-free survival (PFS), and OS in patients with skeletal involvement treated with consolidative radiation (RT). Despite these positive results, real-world adoption of skeletal radiation and associated oncologic outcomes, need to be better characterized. We hypothesized that RT for patients with skeletal involvement from LBCL remains underutilized, leading to the detriment of OS.

Materials/Methods: We conducted a retrospective analysis using the National Cancer Database for patients with LBCL with bone involvement with chemotherapy +/- RT from 2016-2020. Due to limitations on how skeletal involvement was coded prior to 2016, we excluded data from previous years. Patients who did not receive rituximab+chemotherapy, had prior malignancy, unknown RT treatment, or unknown follow-up were excluded. Patients who received palliative doses of RT (<30 Gy) were also excluded. Receipt of radiation treatment was defined as skeletal (SRT), non-skeletal (NSRT), or none. Trends in the utilization of RT were assessed using linear and logistic regressions. Survival analysis was performed using Cox regression and Kaplan-Meier methods.

Results: A total of 641 patients met the inclusion criteria, with 8.7% receiving SRT, 7.2% receiving NSRT, and 84.1% without RT. In 2016, 80.6% and 13.9% of patients received no RT and SRT, respectively. In 2020, 84.9% and 7.9% of patients received no RT (slope: 1.1%/yr; p=0.141) and SRT (slope: -1.6%/yr; p=0.186), respectively. No patient or clinical factors were associated with receipt of either SRT or RT in general. Median follow-up was 2.6 years (IQR: 1.1-4.2 years). The 3-yr OS rates in the SRT, NSRT, and no RT cohorts were 87.0% (95% CI 74.7–93.6%), 83.4% (95% CI 68.1–91.8%), and 72.6% (95% CI 68.5–76.3%), respectively. Relative to no RT, receipt of SRT was associated with improved OS on univariable Cox regression (HR: 0.41; p=0.020) and multivariable (MVA) Cox regression (HR: 0.44; p=0.032). In comparison, relative to no RT, receipt of NSRT was not associated with improved OS on univariable Cox regression (HR: 0.54; p=0.087) or multivariable (MVA) Cox regression (HR: 0.55; p=0.098). Factors associated with OS on MVA included age and Charlson Deyo score. Other factors including race, ethnicity, insurance type, hospital type, and IPI score were not significant.

Conclusion: Consistent with prospective data, SRT conferred a statistically significant improvement in OS, with a numerical improvement similar to prior prospective analyses. Despite this benefit, rates of consolidative RT to skeletal sites are low and potentially decreasing. Future studies should further elucidate the benefit of consolidative RT for patients with LBCL with skeletal involvement and barriers to such treatment.