The Patient Journey and Treatment Outcomes Comparing Inpatient Versus Outpatient Axicabtagene Ciloleucel in Non-Hodgkin's Lymphoma (NHL) - a Large, Multicenter Study

Patient journeys and outcomes were evaluated for those receiving axicabtagene ciloleucel (axi-cel), an FDA-approved anti-CD19+ chimeric antigen receptor T-cell (CAR T) therapy, for non-Hodgkin lymphoma (NHL) across a community-based transplant and cell therapy network comprising 5 centers. In total, 167 axi-cel recipients were identified in the study period (2019-2023) with 63% of infusions occurring in 2022 and 2023. Axi-cel was infused either inpatient (IP) or in the outpatient setting (OP) supported by remote patient monitoring. Median age at referral was 61 years (IQR 53, 69), with 55% having a comorbidity index of 0 or 1. Follicular accounted for 24 (14%) and diffuse large B-cell lymphoma (DLBCL) accounted for 143 (86%) of the NHL axi-cel-treated cohort. Axi-cel was predominantly (>60%) delivered as 4L-6L for follicular and 3L-4L in DLBCL, which is inclusive of holding/bridging therapy. The overall DLBCL cohort showed an incidence of 73% and 41% for any grade CRS and ICANS, respectively, corresponding to use of tocilizumab, for CRS, and anakinra, for steroid refractory ICANS, at a frequency of 66% and 4.9%, respectively. Of those developing CRS or ICANS, 86% (CRS) and 74% (ICANS) were grade 1 or 2. In the DLBCL population, 47 (34%) were OP overall in the evaluated time period; however, OP comprised 43% and 62% of axi-cel treatment in 2022 and 2023, respectively. An important distinction is that administrative policy limited the use of OP for all CAR T at one center within the network. Socioeconomic factors and clinicodemographics such as sex, race, ethnicity, age, BMI, smoking status and pack years, primary language, marital status, comorbidity index, performance status, and prior autologous transplant were similar between IP and OP. OP and IP showed similar time from referral to infusion (111 days vs. 95 days, p=0.10). Furthermore, OP was not inferior to IP in regards to providing access to axi-cel with respect to distance to treatment facility (42 km [IQR 19, 276] vs. 33 [IQR 18, 70]), national and state area deprivation indices (on scales of 100 and 10, respectively: 42 and 4 vs. 37 and 3), and insurance coverage (Medicare 38% vs. 11%, Medicaid 13% vs. 3%, VA/Tricare 4% vs. 0%, Commercial 36% vs. 67%). Bendamustine and flu/cy were used 57% and 43% for lymphodepletion in OP, respectively; whereas flu/cy (93%) was predominant in IP. Of note, this pattern of lymphodepletion regimen was driven by the fludarabine shortage. Prophylactic dexamethasone on days 0, 1, and 2, was utilized more frequently in OP compared to IP (51% vs. 17%) but did not delay onset of CRS (4 days [IQR 4, 6] vs. 4 days [IQR 2, 6], p=0.2) or ICANS (6 days [IQR 5, 12] vs. 5 days [IQR 4, 11], p=0.2). Importantly, 6 (13%) OP patients avoided subsequent hospitalization and OP had similar ICU rates (28% vs. 23%) as IP. Median follow-up for the overall DLBCL, OP, and IP cohorts was 419 days (IQR 258, 679), 383 (IQR 258, 482), and 461 days (IQR 262, 932), respectively. Median length of stay (LOS) for OP was 7 days (IQR 3, 12) compared to 15 days (IQR 11, 21) for IP (p<0.001), with no difference in median ICU LOS for those admitted to ICU (8 days vs. 7 days, p=0.5). Complete or partial response at 30 days was achieved in 47% and 48% of OP and IP cases, respectively. More importantly, median progression-free survival of OP was significantly better than IP (not reached vs. 365 days, p=0.033). Finally, there was no difference in median overall survival for OP compared to IP (not reached vs. 723 days, p=0.49). In conclusion, despite having similar baseline characteristics, OP supported by remote patient monitoring maintained access to axi-cel therapy, yet reduced healthcare utilization by reducing overall hospitalization rates and duration and showed outcomes that were no worse or even better than those in IP. Conflict-of-Interest Disclosure: The views expressed in this publication represent those of the author(s). None of the authors declare any conflict of interest related to the current study beyond employment by HCA Healthcare and/or Kite. This study was funded, in whole, by Kite. Analysis was performed by teams within or affiliated with HCA Healthcare including the Sarah Cannon Cancer Network, the HCA Healthcare Research Institute and Genospace.