Patient-reported outcomes may predict overall survival in AML
Click Here to Manage Email Alerts
Researchers found that scales from a leukemia-targeted patient-reported outcomes, the Functional Assessment of Cancer Therapy – Leukemia, and especially its physical well-being scale, predicted OS among patients with acute myeloid leukemia not suitable for intensive therapy, according to a presentation from ASCO20 Virtual Scientific Program.
Patient-reported outcomes (PROs) often play a role in screening patients in cancer trials, according to John Devin Peipert, PhD, from the department of medical social sciences at Northwestern University Feinberg School of Medicine.
“We know that PRO scores from the start of trials can predict trial endpoints like survival, and this information can be used in trial design, as well as to help interpret trial results,” he told Healio. “A lot of research has focused on this topic in other cancer types; however, there is relatively little prognostic research around PROs for AML patients who are not candidates for intensive therapy, a population for whom the development of new therapies is needed.”
Researchers examined data from 317 patients with AML from a clinical trial comparing decitabine plus talacotuzumab vs. decitabine alone. They assessed whether baseline (1st cycle) FACT-Leu scales predicted OS using ridge-penalized Cox models, then entered the FACT-Leu scales significant in these models and factors from the AML composite model into Cox proportional hazard models. In addition, they ran model selection procedures with 1,000 bootstrapped samples using all variables, according to the abstract.
Peipert and colleagues examined inclusion frequency of each FACT-Leu scale in the final models to assess the prognostic value for OS — meaning the higher the percentage of inclusion, the higher the importance of the variable.
Analysis revealed that the Physical Wellbeing Scale (PWB), Trial Outcome Index (TOI), and FACT-Leu Total scales significantly predicted OS in the ridge-penalized Cox models.
When the researchers adjusted for the AML composite model covariates, they observed a 2-point difference in PWB score was linked to a 9% decline in OS (HR = 0.91; 95% CI, 0.86-0.97), according to the data. The PWB scale emerged in up to 97% of selected models, but the TOI and FACT-Leu Total scales emerged less often (45%-73% and 41%-71%).
“Since the PWB is brief, with only seven questions, it is a practical assessment useful for determining inclusion or stratifying patients who are enrolled in trials,” Peipert told Healio.
Collapse
Peipert JD, et al. Abstract 7532. Presented at: ASCO20 Virtual Scientific Program; May 29-31, 2020.
Click Here to Manage Email Alerts