Real-World Treatment Patterns and Clinical Outcomes Among Patients with Diffuse Large B-Cell Lymphoma in a US Healthcare Claims Database

Background: Novel agents, including chimeric antigen receptor T-cell (CAR-T) therapy and other targeted therapies, have dramatically expanded treatment options available to patients with diffuse large B-cell lymphoma (DLBCL). We conducted a real-world analysis of DLBCL patients to describe the evolution of treatment patterns and clinical outcomes by line of therapy (LOT). Methods: The study utilized Optum's de-identified Clinformatics® Data Mart database, which includes administrative claims data for members of large private and Medicare Advantage health plans. A DLBCL LOT algorithm was applied from January 1, 2008 to December 31, 2023. The index date was defined as the date of LOT start, stem cell transplant (SCT), or CAR-T infusion. Study inclusion criteria were DLBCL diagnosis before or on index, ≥365 days continuous enrollment before and ≥29 days after index, age ≥18 years on index, and index between October 1, 2015-December 31, 2023. Subgroups specifying LOT also required ≥183 days continuous enrollment before DLBCL diagnosis. Exclusion criteria were clinical trial participation ≤6 months before index or select hematologic malignancies ≤12 months before index. Demographic and clinical characteristics and treatment patterns were reported descriptively. Overall survival (OS), time to next treatment or death (TTNT), and treatment failure rate (next treatment or death within 24 months) were reported using the Kaplan-Meier estimator. Results: LOTs from 7852 patients met study criteria (first line [1L] n=6718; second line [2L] n=2309; third line [3L] n=749); additional subgroups by treatment category of interest were included (1L rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone [R-CHOP], n=3880; 2L SCT, n=179; CAR-T n=330). Demographic composition remained largely consistent from 1L to 3L (in 1L, patients had a median age of 73; 54% were male; 73% were White; 78% were covered by Medicare Advantage, and 22% were commercially insured; median Quan-Charlson comorbidity index was 4.0). Median time from DLBCL diagnosis to 1L index was 1.3 months; median follow-up was 15.7 months after 1L index. In 1L, 66% of patients received an R-CHOP-based regimen and 13% received non-R-CHOP chemoimmunotherapy (CIT) regimens. Most remaining 1L patients were treated with either rituximab monotherapy or other therapies. Treatment patterns in 2L evolved substantially; use of conventional CIT and chemotherapy changed from 79% in 2016 to 45% in 2023, while use of polatuzumab vedotin (pola-), tafasitamab-, and CAR-T-based regimens increased steadily, with 35% of 2L patients treated with one of these therapies in 2023. In 3L, conventional CIT decreased from 19% in 2016 to 10% in 2023, with 51% treated in 2023 with pola-, tafasitamab-, CAR-T-based, and other novel immunotherapy regimens. Over the study period, median OS was 58 months (95% CI 54-61) in 1L, 29 months (26-32) in 2L, and 18 months (14-21) in 3L. Treatment failure rate by 24-month follow-up was 45% (95% CI 43-46%) in 1L, 64% (62-67%) in 2L, and 71% (67-75%) in 3L. Median TTNT was 36 months (95% CI 33-40) in 1L, 11 months (10-13) in 2L, and 8 months (7-10) in 3L. For treatments of interest, median OS was 68 months (95% CI 63-73) in 1L R-CHOP, not reached (NR) in 2L SCT, and 25 months (20-28) in CAR-T. Treatment failure rate was 37% (95% CI 35-39%) in 1L R-CHOP, 34% (25-43%) in 2L SCT, and 59% (51-66%) in CAR-T. Median TTNT was 58 months (95% CI 53-66) in 1L R-CHOP, NR in 2L SCT, and 18 months (13-24) in CAR-T. The CAR-T subgroup included 36% axi-cel, 16% liso-cel, 7% tisa-cel, and 41% unspecified brand. 63% of patients utilized bridging therapy (BT). The top 3 drug classes for BT included corticosteroids (46%), monoclonal antibodies (38%), and chemotherapy (28%). The top 3 BT regimens were pola + rituximab (11%), rituximab (7%), and pola + bendamustine + rituximab (7%). The top regimen used in the next LOT after CAR-T was loncastuximab tesirine (13%). Conclusions: Despite the rapid uptake of novel agents in 2L and 3L, these results suggest that an unmet medical need persists for patients with DLBCL. 1L failure rate within 24 months was 45%, and both OS and duration of treatment response (proxied by TTNT) declined rapidly for patients receiving 2L and 3L treatment. Persistently poor outcomes in later LOTs highlight the need for more effective and longer-lasting frontline and relapsed/refractory treatment options that can prevent further disease progression.