The simple c-reactive protein (CRP) to albumin ratio (CAR) stratifies treatment response and survival following chimeric antigen receptor T-cell therapy for non-Hodgkin's lymphoma

Abstract Introduction: CAR T therapy has revolutionized treatment landscape for patients with relapsed refractory (R/R) non-Hodgkin's lymphoma (NHL), leading to long-term survival for a significant portion of patients. However, an equally significant portion of patients relapse after CAR T therapy and often require additional therapies. Pre-CAR T disease burden has been shown to be a reliable indicator of CAR T therapy efficacy and more recently, a pro-inflammatory state prior to CAR T infusion, as measured by a composite of 14 routine laboratory values, is also emerging as a novel predictor. In this retrospective study, we examine whether the simple c-reactive protein (CRP) to albumin ratio (CAR), which is known to be associated with treatment response and survival across multiple cancer and treatment types including chemotherapy and immunotherapy, could stratify CAR T treatment efficacy and survival in a contemporary cohort of patients with NHL. Methods: The single institution cohort includes patients with a diagnosis of R/R NHL after ≥1 line of prior therapy treated with axicabtagene ciloleucel, brexucabtagene autoleucel, lisocabtagene maraleucel, or tisagenlecleucel. CAR T therapy eligibility criteria, clinical care, and monitoring of CRS and ICANS follows standard institutional guidelines. Patient characteristics are summarized using descriptive statistics. Associations between CAR and baseline characteristics are estimated using Wilcoxon rank sum test for binary variables and Pearson correlation for continuous variables; and the development of ICANS or CRS using Wilcoxon rank sum tests. Overall survival (OS) is defined from infusion until death or end of follow-up, whereas progression-free survival (PFS) is defined from infusion until progression, death, or end of follow-up in the absence of an event; patients who received subsequent treatment for NHL before or in absence of post-CAR T relapse are censored at next treatment start for PFS. Survival endpoints are estimated using Kaplan Meier methodology and compared using log-rank tests. Cox proportional hazards models are used to evaluate univariable and multivariable associations between outcomes and baseline characteristics. Results: From January 1st, 2020 to December 31st, 2024, 375 patients with NHL have undergone commercial CAR T therapy. Of these, 287 patients have pre-infusion CAR values and are included in this analysis. The median age at infusion is 66 years (IQR 56, 74) and 48% of patients received a 4-1BB-based CAR T product (lisocabtagene maraleucel or tisagenlecleucel). In addition, 48% of patients are in stable or progressive disease and 33% have elevated LDH pre-lymphodepletion. The median follow-up among survivors is 24.9 months, IQR: (9-36.3). Pre-lymphodepletion CAR is associated in variable degree with several baseline characteristics including patient age, KPS, and LDH. In multivariable analysis of PFS, however, higher pre-lymphodepletion CAR (HR 1.14, 95% CI 1.04 – 1.26, p=0.007) and LDH (HR for normal LDH: 0.39, 95% CI 0.27 – 0.56, p < 0.001) are independently associated with outcome after adjusting for KPS and type of CAR T product. Similarly, in multivariable analysis of OS, higher pre-lymphodepletion CAR (HR 1.15, 95% CI 1.03 – 1.29, p = 0.011), LDH (HR for normal LDH: 0.27, 95% CI 0.18 – 0.42, p < 0.001), and greater number of prior lines of therapy (HR 1.17, 95% CI 1.03 – 1.33, p = 0.018) are significant predictors of outcome after adjusting for age and KPS. Higher pre-lymphodepletion CAR is also associated with increased cumulative incidence of relapse in multivariable analysis, but not with the development of all-grade or high-grade CRS or ICANS. Conclusions: In this contemporary, large, single-institution cohort study, we show that pre-lymphodepletion CAR is associated with relapse and overall survival after adjustment for other well-established risk factors. Since CAR is a dynamic, readily available estimate of the inflammatory and nutritional state as demonstrated previously in other cancer types, these results suggest that it can also be applied to the setting of CAR T therapy for R/R NHL, in combination with pre-treatment disease burden as measured by LDH. Our results support our previous findings on the crucial role of inflammation in CAR T and cancer immunotherapy outcomes, and lend support to a rational target for therapeutic and behavior intervention.