Introduction

Around 30%-40% of patients with diffuse large B-cell lymphoma (DLBCL) will experience disease progression or relapse after first line (1L) chemoimmunotherapy with most patients relapsing within the first 12 months. Relapse may be systemic, isolated to the central nervous system (CNS), or both with CNS relapse occurring in 2-5% of patients. The purpose of this study is to evaluate the characteristics and outcome differences of Veterans diagnosed with DLBCL within the VHA who experienced systemic relapse, isolated CNS relapse or both.

Methods

A retrospective chart review of 6,266 lymphoma patients seen in the VHA nationwide between 01/01/2011 and 12/31/2021 was performed by trained abstractors. Patients diagnosed with DLBCL were included. Patients were excluded if they had a diagnosis other than DLBCL, had incomplete records or were diagnosed and treated outside the VHA, had baseline CNS involvement, did not receive treatment, or died before receiving a scan to assess response to 1L therapy. Outcomes including response to 1L treatment, time to relapse and overall survival (OS) defined as time from diagnosis to death were evaluated. Early relapse was defined as having primary refractory disease (stable or progressive disease (SD or PD) as best response to 1L therapy) or disease progression in ≤12 months in patients achieving a complete response (CR) or partial response (PR). Late relapse was defined as disease progression >12 months.

Results

A total of 2,658 patients met inclusion criteria for analysis with 97% of the population being male, 65% with stage III-IV disease, and 90% with an ECOG performance status of 0-2. Of the total cohort, 74% of patients received a CHOP-based regimen, 74% had an intermediate- or high-risk CNS International Prognostic Index (CNS-IPI) score, 81% did not receive a baseline diagnostic lumbar puncture (LP) and 86% did not receive any CNS prophylaxis. The median follow-up time for subjects in this study was 54 months. Relapse occurred in 29% of the total cohort with systemic only relapse occurring in 27% of patients, isolated CNS in 1%, and CNS + systemic in 1%.

Of the 761 patients who experienced relapse, 77% had an early relapse with a median time to relapse of 4.4 months (IQR 2.8-6.7). For the 23% of patients with late relapse, the median time to relapse was 25 months (IQR 18-41). The characteristics of patients who experienced relapse were similar regardless of type of relapse or time to relapse (table 1). A higher percentage of patients with early relapse had a high-risk CNS-IPI score, bone marrow involvement, and renal or adrenal involvement. Autologous stem cell transplant was performed in 11% of patients, predominately in those with systemic only relapse.

The median OS (mOS) was similar between relapse types with systemic, isolated CNS, and CNS + systemic having a mOS of 23.5 months (95% CI: 21-27), 30.6 months (95% CI: 15-71), and 29.6 months (95% CI: 20-45), respectively (p=0.75). Figure 1 describes mOS by relapse type and time to relapse. Overall, patients with early relapse had a poorer prognosis regardless of relapse type with a mOS of 17 months (95% CI: 15-18) compared to a mOS of 68 months (95% CI: 59-88) in patients with late relapse (p<0.001). The 24-month survival rate of those with early versus late relapse was 37% and 78%, respectively. Patients with a late systemic relapse had the longest mOS of 74 months (95% CI: 60-90).

Conclusion

This is the largest study reporting outcomes of Veterans who experienced relapsed DLBCL within the VHA. Although earlier data suggests specific features may correlate with early versus late relapse, all characteristics were similar in our population across relapse types regardless of time to relapse. The rate of relapse in the total cohort and the rates of CNS relapse were lower than previously reported, adding to recent data suggesting the use of CNS prophylaxis may have limited benefit. Patients with early relapse performed poorly and may benefit from novel therapies including chimeric antigen receptor T-cell therapy, which could not be evaluated with this study. The retrospective nature of this analysis, and a predominately male veteran population may limit external application.

Horowitz:Bristol Myers Squibb: Speakers Bureau.

This content is only available as a PDF.
Sign in via your Institution