PRO collection in CART therapy
We report the feasibility of collecting PROs in patients receiving CART therapy in a dynamic and complex outpatient setting. We were able to recruit 97.1% of patients approached to participate and 95% of patients who were alive and still on-study at 12 months completed their 12-month FACT-Lym assessment. Attrition was largely due to progressive disease rather than questionnaire burden or fatigue. FACT-Lym completion rates were lower at day 30 (63%) and 3 months (50%), likely a reflection of patients feeling unwell and less stable. Our results support the development of a larger study measuring long-term PROs in patients who received CART therapy.
In addition to feasibility, we report the longitudinal changes in HRQoL in patients who received tisagenlecleucel for relapsed or refractory large B-cell lymphoma. As expected, FACT-Lym scores initially worsened during the acute CART experience and took six months to improve beyond baseline scores. Clinically meaningful changes were seen in the composite scores, FACT-Lym total and FACT-TOI. PRO data from the JULIET trial also demonstrated clinically meaningful improvements in FACT-Lym scores in those who achieved and maintained at least a partial response from baseline scores at 3, 6, 12, and 18 months[9]. They were unable to assess non-responders as these patients either succumbed to their disease or withdrew from the JULIET trial to pursue alternative therapies. A recent scoping review of 14 studies of PROs (HRQoL and patient-reported symptoms) in CART therapy indicates that after a decline in the early phase after infusion, there is a general improvement in PROs, at least in some domains in responding patients[15].
Interestingly, our exploratory analysis does not demonstrate improvement in in the emotional or social well-being domains. Similar findings have been demonstrated in clinical trial studies [16, 17]. This is consistent with the fact that patients with cancer have higher rates of depression and anxiety compared to the general population [18] and individuals with hematologic cancers have some of the highest reported rates of depression and anxiety amongst cancer types [19]. Despite higher rates of depression and anxiety in patients with cancer, only 5% see a mental health professional [20]. Changes in cognition may impact PRO scores. Two studies evaluated long-term, defined as 6–12 months, cognitive changes following CART therapy [21, 22]. Ruark and others found that 50% of patients reported at least one cognitive difficulty or clinically meaningful depression or anxiety in long-term follow-up [21]. Barata and others found that cognition worsened from month 3 to month 12 following CART therapy [22]. Our study highlights a potential need for future studies to investigate the role of social and mental health interventions in CART therapy and potential barriers for accessing this care in patients with cancer.
To the best of our knowledge, outside of a feasibility study measuring the adoption of a wearable vital sign tracker to implement PRO completion[23], our study is the first to report the feasibility of measuring a PRO in patients who received commercial CART therapy. This is important as the CART journey is complex with patients moving from clinic to ODS to hospital and to sometimes geographically isolated homes. In the recent scoping review of PROs (HRQoL and patient-reported symptoms) in CART therapy (24), Efficace and others identified seven observational studies [21–27] measuring PROs in CART therapy. These observational studies were largely uncontrolled and heterogenous. In addition to the pilot study using wearables[23], two were a cross-sectional design and did not measure the collection of PROs at multiple time points [21, 26]. One study was retrospective, evaluating standard of care PRO measures[27]. The remaining three trials were prospective; however, they did not report feasibility, missing data, MID, or attrition[22, 24, 25], which are important outcomes in PRO research. There is currently insufficient data to understand the long-term effects of CART on PROs.
In addition to a PRO, we report our initial experience with administering tisagenlecleucel in the outpatient setting at a Canadian hospital and cancer centre. Compared to the JULIET trial, our real-world data demonstrated similar toxicity and response rates in a similar patient population [1]. However, we did observe a superior median one year (95% CI) overall survival of 65.6% (46.6, 79.3) compared to 40% (95% CI, 32 to 49)[1]. This may reflect the fact that there are more treatment options after CART therapy failure that were not available at the time of JULIET. Another explanation could be that we observed lower rates of bridging therapy (65.7%) compared to the JULIET trial (92%)[1]; knowing that bridging therapy is associated with inferior OS[28].
In our outpatient CART model of care, 80% of patients were admitted due to toxicity. To maintain a successful outpatient program and ultimately maintain a sustainable CART program in a strained socialized healthcare system, there needs to be a strategy to mitigate toxicity necessitating admission to hospital. To sustain an outpatient model of CART delivery, it is important to be able to identify those at risk of severe toxicity. It is important that future studies of outpatient CART delivery incorporating alternative models of care such as remote monitoring, and strategies to mitigate toxicity include PROs to evaluate the effectiveness from the patient perspective. Our study demonstrates that PRO collection in the dynamic administration of CART therapy is feasible.
Strengths and limitations
A major strength of our study is the prospective nature of data collection, allowing for robust and granular clinical data. A second major strength is that the attrition rate of HRQoL completion was largely due to progressive disease or death from the patient’s disease, rather than loss of follow-up. This implies that there was little concern about questionnaire fatigue or burden.
Despite these strengths, there are several methodological limitations to consider. First, we were unable to collect a number of HRQoL measures due to COVID. The beginning of the pandemic posed a unique barrier to PRO data collection. As a result of our experience, we are prepared for this type of barrier in future studies. To prevent missing data for future studies, we would recommend paper and electronic collection to allow for universal access to the PRO. Second, due to a small sample size, this study was underpowered to explore significant changes in HRQoL. We were therefore only able to perform exploratory analyses for the purposes of hypothesis generation to inform future studies.
Considerations for future studies include exploring the long-term effects on cognition that would not have been captured but may have impacted HRQoL scores, particularly the emotional and social well-being domains. In addition, PROs should be measured more frequently during the acute phase of treatment and height of toxicity. Lastly, there are nuances in the patient experience that cannot be captured using a questionnaire. Qualitative work to capture the patient perspective of CART therapy would be valuable to identify areas for improvement and possible intervention.