Feasibility of Patient-Reported Outcome Assessment in the Acute Phase after CAR T-Cell Therapy in Lymphoid Malignancies: A Pilot Study

Introduction: Although acute toxicities of chimeric antigen receptor T-cell (CAR T-cell) therapies are well described, there is a paucity of data on patient-reported outcomes (PROs) in the acute phase after CAR T-cell infusion. We performed a prospective pilot study to assess the feasibility of measuring PROs and trajectory of health-related quality of life (HRQoL) in the first 3 months after CAR T-cell therapy in lymphoid malignancies.

Methods: We enrolled patients with aggressive B-cell non-Hodgkin lymphoma receiving CD-19 directed CAR T-cell therapy. PROs were assessed using 2 instruments: PROMIS-29 v2.0 (7 domains and 29 items) and FACT-Lym (6 domains and 48 items). PROs were assessed at the following time-points: baseline/prior to lymphodepleting chemotherapy, day#0 (day of CAR T-cell infusion [before infusion]), and post-CAR days 7 (±3), 14 (±3), 21 (±3), 28 (±3), 60 (±10), and 90 (±10). The primary endpoint was compliance rate of PRO instruments at each time-point. The key secondary endpoint was trajectory of QoL domains, as measured by PROMIS-29 and FACT-Lym. The raw score of each PROMIS-29 domain was converted to standardized T-score, which has a mean of 50 (±10) for general US population and a minimal clinically important difference (MCID) of 5 .

Results: A total of 23 patients were enrolled, among which 3 died prior to baseline PRO assessment. In 20 evaluable patients, median age was 66 years (range, 50-78) and all had ECOG performance status 0 or 1. The CAR T product was tisa-cel in 11, axi-cel in 9, and experimental in 1 patient. Two patients died during the study period, one at day#35 and one at day#89 after CAR T-cell infusion. Proportion of patients with any missing PROMIS data increased from 0% at baseline to the highest of 50% at day#14 post CAR (p-value for trend=0.0001), and subsequently decreased to 22% at day 90 (p-value for trend [day#14 vs 90]=0.01) (Figure 1a). Proportion of patients with any missing FACT-Lym increased from 15% at baseline to the highest of 50% at day#14/21, and then decreased to 22.2% at day#90.

The three domains of PROMIS-29 that showed significant change over time were physical function, fatigue, and satisfaction with social role (Figure 1b). The physical function T-score (46.5 [baseline]-->38.8 [day#21]-->43.6 [day#90]) and satisfaction with social role T-score (50.2 [baseline]-->43.0 [day#21]-->49.3 [day#90]) significantly decreased from baseline to day#21, and then significantly increased from day#21 to day#90, with fatigue T-score showing a significant increase-decrease pattern (50.6-->55.9-->49.1, at respective time-points). There were no significant difference between baseline and post-CAR day#90 for all three scores. There was also no significant change in pain over time, both by PROMIS pain interference domain and pain intensity item. The only FACT-Lym domain that showed significant change over time was Functional Well-Being (FWB). It decreased from baseline to day#21 (LS-mean 0.37 to 0.30, p=0.04 , and significantly increased from day#21 to day#90 (0.30 to 0.38, p=0.03).

Conclusion: Our study adds novel data on feasibility of using PROMIS-29 for PRO assessment in CAR T-cell therapy recipients. The rate of missing data with both PROMIS-29 and FACT-Lym was greatest at day#14 and day#21 post-CAR, which correlated with a drop in HRQoL. Notably, PROMIS-29 demonstrated significant variability in PROs in the domains of physical function, fatigue, and satisfaction with social roles, whereas only one FACT-Lym domain (functional wellbeing) showed significant variability. Future clinical trials of CAR T-cell therapy should incorporate CMS-recommended PROMIS-29 for PRO assessment to reduce heterogeneity and enable comparison across products and trials.