Longitudinal assessments of simple frailty tools can help predict outcomes of patients undergoing chimeric antigen receptor T-cell (CAR-T) therapy: A prospective pilot study at Princess Margaret Cancer Centre

Abstract Introduction Chimeric Antigen Receptor T-Cell (CAR-T) therapy has become the preferred treatment for patients with relapsed/refractory B-cell lymphomas, however outcomes remain suboptimal, and patient selection is key. There is no upper age limit for treatment eligibility, and frailty may be an important factor in assessing fitness for treatment. This study aims to determine if frailty assessments pre-CART can identify those at higher risk for acute toxicities, and predict PFS, and OS, as well as evaluate changes in frailty over time. Methods We performed a cohort study of consecutive patients with lymphoma undergoing CAR-T therapy at our institution, from April 2021 to date. Frailty was evaluated using ECOG performance status, Clinical Frailty Scale (CFS), Grip Strength, Gait Speed, Mini-Cog, and Patient Health Questionnaire (PHQ-2/PHQ-9) at 5 time points: baseline (clearance visit), and 1, 3, 6 and 12 months post-CART. At baseline, additional frailty assessments were completed: the Vulnerable Elders Survey (VES-13), Hematopoietic Cell Transplantation-specific Comorbidity Index (HCT-CI) and Cumulative Illness Rating Scale (CIRS). Clinical variables were also collected. Both standard and time-dependent Cox models (incorporating time-varying covariates) were used to identify predictors of toxicity, and overall and progression-free survival (OS; PFS). Results Seventy-nine patients have been included. Median age is 60 yrs (range 22-83) and 59% were male. 49% had de novo diffuse large B-cell lymphoma, 30% transformation from follicular lymphoma (FL), 9% HGBCL, 8% PMBCL and 4% FL. Most patients (85%) received axicabtagene ciloleucel. Median follow-up was 12.9 months (range 1.5-49 mo). All 79 patients completed assessments at baseline, 63 at 1-mo, 40 at 3-mo, 28 at 6-mo and 19 at 12-mo. Median scores (and range) for the HCT-CI, CIRS, VES-13, Mini-Cog, and CFS at baseline were 1.0 (0-7), 4.0 (0-13), 1.0 (0-7), 4.0 (1-5) and 3.0 (1-7) respectively. The median 4m walk test speed was 1.1 m/s (range 0.5-1.7). There were clinically significant changes observed in CFS over time (p=<0.001), with mean scores of 3.2 at baseline, 3.6 at 1 month, 2.8 at 3 months and 2.4 at 12 months. There were also significant changes (p=0.001) between timepoints for gait speed. Twenty-two patients (28%) experienced immune effector cell neurotoxicity syndrome (ICANS) (4% Grade 4, 4% Grade 3, 6% Grade 2, 14% Grade 1) and forty-nine patients (87%) experienced cytokine release syndrome (CRS) (1% Grade 3, 55% Grade 2, 35% Grade 1) during the 30 days following cell re-infusion. Eight patients (10%) were admitted to the ICU. Of the frailty and clinical variables tested (sex, bridging therapy, LDH, CRP), most were not significantly associated with the development or grade of CRS or ICANS, except for lower risk of CRS with increasing age: OR 0.91 (95%CI 0.84-0.98, p=0.018). The 12-mo OS of the whole cohort is 66% and PFS is 58.1%. On univariable analysis (UV) of baseline data, ECOG, LDH level, and the CFS score were predictive of PFS, while ECOG, LDH, CRP, VES-13 score, CFS, HCT-CI, CIRS score and the 4m walk speed were all predictive of OS. On multivariable (MV) analysis of PFS, only baseline LDH (p=0.009) remained significant. Models were tested on MV analysis of OS, and most frailty measures lost their significance, except for CFS (HR 1.39, 95%CI 1.08-1.79, p=0.011). Analyses incorporating repeated measures were performed, and on UV analysis, the CFS, 4m walk test, LDH, VES-13 (≥3 vs. <3) and Mini-Cog were significantly associated with PFS; the same variables, HCT-CI and CRP were associated with OS. On MV repeated measurements analyses, the CFS (1.508, 95%CI: 1.11-2.05, p=0.008) and Mini-Cog (0.68, 95%CI: 0.50- 0.92, p=0.012) were still associated with PFS, as well as with OS (CFS: HR 1.68, 95%CI 1.26-2.22, p<0.001; Mini-Cog: HR 0.67, 95%CI 0.51-0.89, p=0.005). Conclusions Conducting serial frailty assessments in patients undergoing CAR-T therapy is feasible, but their use and interpretation is complex. Significant longitudinal changes were seen in CFS and gait speed, suggesting an element of reversible functional impairment related to patients' lymphoma. Within the limits of our sample size, baseline measures of frailty were not predictive of CRS or ICANS; however the combination of the simple measures CFS and Mini-Cog was predictive of overall survival, which could help clinical decisions; these observations need to be validated in larger number of patients.