TRANSCEND CLL 004 Analysis Proposes Meaningful Change Thresholds for PRO Measures in R/R CLL

Among 62 patient-reported outcome (PRO)–evaluable patients with relapsed/refractory chronic lymphocytic leukemia (CLL) from the phase 1/2 TRANSCEND CLL 004 trial (NCT03331198), meaningful within-patient change (MWPC) thresholds were identified and proposed for the symptom burden, physical condition/fatigue, and worries/fears on health and functioning domains of the EORTC QLQ-CLL17, according to findings from an analysis published in the Journal of the British Society for Haematology.¹

Results from the MWPC analysis showed that a substantial improvement or worsening in a symptom burden at the patient level may be best represented by a responder definition (RD) threshold of –11/+11 based on standard error of the mean (SEM) and minimum state changes, exceeding the medium effect size of 0.5 × 9.01. This threshold reflects substantial changes in symptom burden.

Similarly, for the physical condition/fatigue domain, a −16/+16 RD threshold was proposed, surpassing the SEM and exceeding the medium effect size of 0.5 × 11.09. In the worries/fears domain, thresholds of 16-point decrease was proposed for improvement and at least a 13-point increase was proposed as the deterioration threshold; the latter was the next possible actual change that may occur at the patient level and exceed the SEM of 11.42 for this domain.

At the group level, thresholds primarily ranged between 0.3 and 0.5 of the standard deviation (SD) of baseline scores.

“To our knowledge, this is the first analysis to propose meaningful change thresholds for interpreting improvement and deterioration in EORTC QLQ-CLL17 domain scores at the patient and group levels in patients with relapsed/refractory CLL,” study authors wrote in the paper. “These thresholds may be useful in clinical studies to define treatment responders [using MWPC/RD] with respect to health-related quality of life [HRQOL] or aid in the interpretation of the meaningfulness of within-group mean changes from baseline [CIC] and between-group difference in mean changes [CID] based on this instrument.”

TRANSCEND CLL 004 PRO Analysis: Design and Methodology

TRANSCEND CLL 004 was an open-label, multicenter study that evaluated the efficacy and safety of lisocabtagene maraleucel (liso-cel; Breyanzi) in adults with relapsed/refractory CLL/SLL. Following enrollment, patients underwent leukapheresis followed by lymphodepletion with 30 mg/m² of fludarabine and 300 mg/m² of cyclophosphamide for 3 consecutive days. Liso-cel was administered 2 to 7 days post-lymphodepletion at a dose level of 100 x 106 CAR T cells. Following a protocol amendment on February 16, 2021, the duration of follow-up was extended to 48 months. Notably, patients who maintained an ongoing response per International Workshop on Chronic Lymphocytic Leukemia 2018 criteria after 2 years were monitored for an additional 2 years or until disease progression.

Investigators explained that the current analysis of TRANSCEND CLL 004 was conducted to address the lack of published guidance informing the interpretation of score changes in each of the EORTC QLQ-CLL17 domains.

PROs were assessed using the EORTC QLQ-CLL17, EORTC QLQ-C30, and EQ-5D-5L instruments. PRO data were collected at baseline (≤7 days before lymphodepletion), prior to dosing on the day of liso-cel infusion, and at multiple time points post-infusion up to 24 months. Subsequent thresholds for each of the three EORTC QLQ-CLL17 domains were triangulated according to estimates obtained from anchor- and distribution-based analyses. All analyses focused on the PRO-evaluable analysis set, which comprised patients with an evaluable assessment of EORTC QLQ-CLL17 at baseline and at least one post-baseline evaluable PRO assessment.

“Thresholds for interpreting meaningful changes at the group level have been recommended to be based on the magnitude of effect size being considered for a clinical study, and at least a small effect size should be considered, “investigators noted.

The primary objective of the analysis was to establish MWPC thresholds for the EORTC QLQ-CLL17, enabling the identification of patients who experienced clinically significant changes in QOL across its domains as a clinical end point. The secondary objective was to determine the clinically important change (CIC) and clinically important difference (CID) thresholds for each domain of the EORTC QLQ-CLL17 to interpret the meaningfulness of within-group changes and between-group differences for patients with relapsed/refractory CLL.

Of the patients who underwent leukapheresis (n = 112), 55% were included in the PRO-evaluable population. Investigators explained that this lower rate of inclusion could be attributed to the addition of PRO assessments during an amendment to study protocol following trial initiation, as well as COVID-19–related restrictions.The anchor-based analysis included 62 patients with changes in baseline score pooled from 240 observations between 1 and 18 months.

The mean age of patients was 64.3 years of age and 27% were female. Most patients were not Hispanic or Latino (90%), were White (90%), had CLL (94%), had high-risk cytogenetics (87%), and had a baseline ECOG performance status of 1 (71%). Regarding disease status, 81% of patients were refractory to their last treatment; 19% had unknown status and no patients had relapsed disease. Among patients who previously received a BTK inhibitor, 87% were refractory, 29% were intolerant, and 13% were intolerant only. Prior venetoclax (Venclexta) was administered to 85% of patients, 76%, 16% and 8% of whom were refractory, intolerant, and intolerant only to the agent, respectively.

According to EORTC QLQ-CLL17 data at baseline, the mean and median symptom burden was 25.0 and 22.2, respectively (min-max, 0.0-61.1). For physical condition/fatigue, respective mean and median scores were 31.0 and 25.0 (min-max, 0.0-100.0). The mean score for worries or fears on health and functioning was 31.1 and the median score was 27.6 (min-max, 0.0-100.0)

Within- and Between- Group Changes in EORTC QLQ-CLL17 Domains

Further evaluation of CIC and CID thresholds indicated that the CIC threshold for improvement in symptom burden domain ranged from −9.06 to −18.72 points, corresponding to effect sizes from medium (−0.51) to large (−1.05). For deterioration, the CIC threshold was between 6.94 and 10.10 points, with effect sizes ranging from small (0.39) to medium (0.57). Distribution-based methods suggested that CIC thresholds between 5.40 and 9.01 points may be appropriate. The CID threshold for improvement was between −4.39 and −13.33 points (effect sizes: −0.26 to −0.77); for deterioration, the CID threshold ranged from 9.52 to 10.88 points (medium effect size, 0.55–0.63).

For the physical condition/fatigue domain, the CIC threshold for improvement ranged from −15.58 to −17.28 points, with effect sizes from medium (−0.71) to large (−0.79). For deterioration, the CIC threshold was between 12.50 and 12.88 points, with medium effect sizes (0.57–0.59). Distribution-based methods suggested CIC thresholds of 6.66 to 11.09 points. The CID threshold for improvement ranged from −9.19 to −9.69 points (small effect sizes of −0.42 to −0.44). For deterioration, estimates were between 11.47 and 14.95 points (medium effect sizes of 0.53 to 0.68).

For the worries/fears on health and functioning domain, the CIC threshold for improvement ranged from −12.93 to −15.80 points, with effect sizes from medium (−0.68) to large (−0.83). For deterioration, the CIC threshold was between 1.26 and 4.16 points, corresponding to effect sizes from trivial (0.07) to small (0.22). Distribution-based methods suggested CIC thresholds of 5.54 to 9.24 points. The CID threshold for improvement was between −8.29 and −10.10 points (small to medium effect sizes of −0.44 to −0.54). For deterioration, estimates ranged from 7.37 to 7.87 points (small effect sizes of 0.39 to 0.42).

Investigators added that, “All CIC and CID threshold estimates for improvement (deterioration) in any domain from the anchor-based analysis represent at least a small effect size, except for the CIC threshold for deterioration in the worries/fears on the health and functioning domain.

Limitations and Future Directions

The analysis was limited by its small sample size, investigators explained. To mitigate this, data from multiple post-baseline visits for individual patients were pooled. Additionally, commonly recommended anchors, such as the Patient Global Impression of Severity and Patient Global Impression of Change, were not included. However, the anchors used met key FDA criteria and generated thresholds consistent with those reported using other EORTC quality-of-life instruments.

"Although the thresholds derived from this analysis are reasonable and consistent with thresholds for other PRO measures developed by the EORTC Quality of Life Groups, they should be further validated using a larger sample size, in the frontline CLL setting and with different treatment modalities," study authors concluded.

Reference

Eliason L, Fofana F, Wang L, et al. Establishing meaningful change thresholds for European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire CLL-specific module domain scores: An analysis based on the TRANSCEND CLL 004 study in patients with relapsed or refractory chronic lymphocytic leukaemia or small lymphocytic lymphoma. eJHaem. 2024; 1–8. doi: 10.1002/jha2.1007