Impact of extranodal sites of lymphoma on clinical outcomes after CD19 CAR T cell therapy

Abstract

Abstract Background: Lymphomatous involvement of extranodal (EN) sites has been recognized as a negative prognostic factor for patients (pts) treated with chemoimmunotherapy and has been incorporated into risk stratification scores such as the IPI for large B-cell lymphoma (LBCL) and FLIPI2 for follicular lymphoma (FL). A recent study (Iacoboni et al. 2025) found decreased overall survival (OS) and progression-free survival (PFS) after chimeric antigen receptor (CAR) T cell therapy in LBCL pts with EN disease. The impact of different EN disease sites on response to CAR T in other lymphoma subtypes, such as FL and mantle cell lymphoma (MCL), and on CAR T complications, such as ICANS and CRS, is less understood, and this study aims to address these knowledge gaps. Methods: This is a retrospective analysis of all lymphoma pts treated at our center during January 2018–July 2023 with standard-of-care anti-CD19 CAR T cell therapy with axi-cel (LBCL or FL), tisa-cel (LBCL or FL), liso-cel (LBCL), or brexu-cel (MCL). Radiology reports from the PET/CT scans performed prior to CAR T infusion for each patient as part of standard clinical practice were reviewed by two independent data extractors to categorize patients as those with EN disease, nodal disease, or no evidence of disease (NED). The outcome metrics were OS, PFS, complete response rate (CRR) from Day+90 PET/CT, and the maximum grades of CRS and ICANS. Results: The study population included 522 pts, of whom 166 were female (32%), and the median age at time of CAR T was 63 years (range 18 to 90 years). Lymphoma histologies included LBCL (436 pts; 83%), MCL (56 pts; 11%), and FL (30 pts; 6%). Pre-CAR T baseline scans revealed EN lymphoma in 318 (61%) patients, while 147 (28%) had only nodal disease, and 57 (11%) had NED. Among the pts with EN disease, the most frequent sites of lymphoma involvement included soft tissue (155 pts; 49%), bone/bone marrow (109 pts; 34%), lung/pleura (99 pts; 19%), liver (43 pts; 8%), peritoneum (39 pts; 8%), gastrointestinal tract (37 pts; 7%), central nervous system (30 pts; 6%) and kidney/urinary tract (28 pts; 5%). The majority of patients with EN disease had a single anatomic type of EN site (152 pts; 48%), while 95 pts (30%) had two sites, and 70 (22%) had more than two sites. Univariable analyses revealed significantly lower CRR after CAR T in pts with EN disease (61%) versus those with only nodal (74%) disease or NED (90%) (p < 0.0001) and in pts with two or more anatomic types of EN sites compared to others (56% vs. 73%; p < 0.01); multivariable analyses detected lower CRR in EN disease vs. NED (p < 0.01). With a median follow-up of 23 months, univariable analyses found significantly lower OS and PFS in those with EN disease and in those with two or more EN sites. However, these differences were not significant in multivariable analyses. Multivariable analysis found significantly lower OS in those patients with EN disease involving the liver (p < 0.001; HR 2.13; 95% CI 1.45-3.13), peritoneum (p < 0.001; HR 2.09; 95% CI 1.42-3.08) or GI tract (p < 0.05; HR 1.67; 95% CI 1.09-2.57) after adjusting for factors such as sex, LDH, ECOG, type of CAR T product and lymphoma histology. Significantly lower PFS was also observed in pts with either liver (p < 0.001; HR 1.86; 95% CI 1.3-2.66) or peritoneal (p < 0.001; HR 1.92; 95% CI 1.32-2.79) sites in multivariable analysis. The frequencies of grade ≥3 CRS and grade ≥3 ICANS were both significantly higher (p < 0.001 and p < 0.05, respectively) in univariable analysis of those with EN disease versus those with only nodal disease or NED, but neither of these associations was significant in multivariable analyses. There were no significant associations noted for high-grade CRS/ICANS with specific EN disease sites. Conclusions: We performed a systematic characterization of EN disease sites on baseline imaging prior to standard of care anti-CD19 CAR T cell therapy for different lymphoma subtypes and analyzed the impact on clinical outcomes. Pts with EN disease had lower CRR compared to those without evidence of disease. Site-specific analyses revealed lymphomatous involvement of the liver, peritoneum, or GI tract as negatively impacting OS in multivariable analysis. These results have implications for pre-CAR T risk stratification and counselling, could inform targeted interventions such as bridging radiation, and identify high-risk pts for closer monitoring and prophylaxis for complications after therapy.