Background: Nearly 40% of patients with diffuse large B-cell lymphoma (DLBCL) are either refractory to or relapse (R/R) after initial first-line (L1) treatment. These patients frequently receive subsequent lines of treatment, although to achieve long-term remission requires aggressive chemoimmunotherapy followed by autologous bone marrow transplant (BMT). Few real-world studies have examined the treatments used in R/R DLBCL patients over the entire disease trajectory. Methods: Using the VA Cancer Registry System and electronic healthcare records (EHR), we identified Veterans diagnosed with DLBCL between January 1, 2000 to December 31, 2016 who relapsed or progressed after L1 RCHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone ± rituximab) ± etoposide and then proceeded to receive at least one additional line of treatment. Human annotation of EHR notes confirmed DLBCL diagnosis, treatment regimen(s) received (including BMT), and treating physician response assessment after each line. Treatment regimens were classified as aggressive (with intent to proceed to BMT) or non-aggressive per National Comprehensive Cancer Network (NCCN) guidelines. Eligible patients were followed until loss to follow-up, death or the end of the study period (March 31, 2019). Patients with a cancer diagnosis other than DLBCL were excluded from the study. Results: We identified 270 R/R DLBCL patients who received a second-line (L2) treatment after having previously received L1 with RCHOP ± etoposide. The mean age of patients was 64.6 years; 97.4% of patients were male. Of the 270 patients, 166 (61.5%) patients received an aggressive L2 treatment regimen. RICE (ifosfamide, carboplatin, and etoposide ± rituximab) and BR (bendamustine ± rituximab) were the most commonly used aggressive and non-aggressive regimens, accounting for 39.3% and 8.9% of L2 treatment, respectively. Compared with patients who received non-aggressive L2 treatment, aggressive L2 treatment patients were younger (61.4 versus 69.7 years). Following aggressive L2 treatment, 87 patients (52.4%) achieved complete or partial response (CR/PR), while CR/PR was achieved in 43 (41.3%) patients who received non-aggressive L2. Of the 29 (10.7%) patients who received BMT, 28 received an aggressive L2 regimen. Approximately half of all patients who received L2 therapy (121/270) proceeded to third-line (L3) treatment, of which 47 (38.8%) received an aggressive L3 regimen. Nearly half of those patients who received an aggressive L2 treatment (36, 47.4%) proceeded to an aggressive L3 regimen, while only 11 (18.8%) non-aggressive L2 patients proceeded to aggressive L3 treatment. Following L3 treatment, 7 (5.8%) patients proceeded to a BMT. The median overall survival (OS) in patients receiving L2 was 9.7 months (CI: 8.2-11.7 months). The median OS for patients treated with aggressive L2 was 9.7 months (CI: 8.1-12.2 months) compared with 9.7 months (CI: 6.8-12.9 months) in patients treated with a non-aggressive L2. The median progression-free survival (PFS) for patients treated with aggressive L2 was 5.6 months (CI: 4.6-7.9 months) and 4.9 months (CI: 3.5-8.4 months) for patients treated with non-aggressive L2. Median OS in patients receiving L3 was 6.3 months (CI: 5.1-8.7 months). The median OS for patients treated with aggressive L3 was 6.8 months (CI: 5.1-12.5 months) compared with 6.6 months (CI: 4.8-9.7 months) in patients treated with a non-aggressive L3. The median PFS for patients treated with aggressive L3 was 4.5 months (CI: 2.6-6.9 months) and 2.9 months (CI: 1.8-5.7 months) for patients treated with non-aggressive L3. Conclusions: Patients who receive aggressive L2 treatment tend to be younger, more likely to receive another aggressive therapy (if L3 treatment is needed), and were much more likely to proceed to BMT than patients who received non-aggressive L2 therapy. Patients who received aggressive L2 treatments resulted in higher response rates compared to less aggressive treatment. We did not find a difference in OS between patients who received aggressive and non-aggressive treatment, the cause of which may be multifactorial. Future studies will examine factors that may impact OS in these patient groups. Acknowledgments: The study was sponsored by Genentech, Inc. Masaquel, Halloran, DeLong-Sieg, Schulz, and Li are employees of Genentech and may receive stock options from Roche. Disclosures Halwani: Pharmacyclics: Research Funding; Immune Design: Research Funding; Seattle Genetics: Research Funding; Amgen: Research Funding; Takeda: Research Funding; Kyowa Hakko Kirin: Research Funding; Miragen: Research Funding; Genentech, Inc.: Research Funding; Bristol-Myers Squibb: Research Funding; AbbVie: Research Funding. Masaquel:Roche: Equity Ownership; Genentech: Employment. Halloran:Genentech, Inc.: Employment; Roche: Equity Ownership. Delong-Sieg:Genentech/Roche: Employment, Equity Ownership. Schulz:Roche: Equity Ownership; Genentech, Inc.: Employment. Li:Genentech: Employment; Roche: Equity Ownership. Sauer:University of Utah and SLC VA Medical Center: Employment.