Summary
Although patient-reported outcomes (PROs) predict overall survival (OS) in solid cancer populations, there is little evidence around the prognostic value of PROs in patients with acute myeloid leukemia (AML) and other hematologic malignancies.
A new study presented at the 2020 ASCO Annual Meeting investigated whether scales from the Functional Assessment of Cancer Therapy – Leukemia (FACT-Leu) predicted OS beyond established prognostic factors and found that the Physical Wellbeing Scale (PWB) was a strong predictor of overall survival (Abstract 7532).
To explore additional resources offered by Oncology Times and to subscribe, visit https://journals.lww.com/oncology-times/pages/default.aspx
Original Article
Predicting Patient-Reported AML Outcomes With the FACT-Leu Scale in AML
Oncology Times
By Sarah LaCorte
Although patient-reported outcomes (PROs) predict overall survival (OS) in solid cancer populations, there is little evidence around the prognostic value of PROs in patients with acute myeloid leukemia (AML) and other hematologic malignancies.
A new study presented at the 2020 ASCO Annual Meeting investigated whether scales from the Functional Assessment of Cancer Therapy – Leukemia (FACT-Leu) predicted OS beyond established prognostic factors and found that the Physical Wellbeing Scale (PWB) was a strong predictor of overall survival (Abstract 7532).
There is a growing, international recognition that patient-reported outcome (PROs) are well-suited to reveal important aspects of cancer patient’s health status. According to John Devin Peipert, PhD, the quantitative lead of the Northwestern University Center on Outcomes Research and Education (NUCORE), at the Feinberg School of Medicine, this in turn has led to an interest in whether or not PROs could be useful tools for screening patients in cancer trials. Peipert and co-authors jointly commented on the study’s findings.
“Establishing that PRO scores recorded at the start of trials can predict trial endpoints like survival creates opportunities to use this information in trial design and can also help to better interpret trial results. There are many advantages around using PROs, including their relative ease of administration and comparative low cost,” said Peipert and authors. “Indeed, it should be noted that prognostic factors typically used in oncology are often based on laboratory data. Therefore, the fact that we can obtain independent prognostic information by simply asking patients (with well-validated tools) about how they feel and about functional limitations is a very pragmatic way of further improving prognostication.”
The researches collected data from 317 AML patients unfit for intensive therapy from a clinical trial comparing decitabine plus talacotuzumab versus decitabine alone (AML2002; NCT02472145). They used ridge-penalized Cox models to determine whether baseline (first cycle) FACT-Leu scales predicted OS. FACT-Leu scales significant in these models and factors from a validated prognostic model, the AML composite model (AML-CL; covariates), were entered into Cox proportional hazard models. Lastly, model selection procedures were run with 1,000 bootstrapped samples using all variables. The researchers then examined inclusion frequency of each FACT-Leu scale in the final models to evaluate prognostic value for OS (i.e., higher the % of inclusion, higher importance of the variable).
The researchers found that, of the ridge-penalized Cox models, the Physical Wellbeing Scale (PWB), Trial Outcome Index (TOI), and FACT-Leu Total scales were significant predictors of OS. After adjusting for the AML-CL factors, an important difference (2 pts) in PWB score was associated with a 9 percent decline in OS. Model validity was evidenced as the PWB scale appeared in a large majority of selected models (90%-97%), while the TOI (45%-73%) and FACT-Leu Total (41%-71%) appeared less often.
“Our central finding was that the Physical Wellbeing Scale (PWB) of the FACT-Leu scale was the strongest predictor of survival. We know that the PWB, and FACT measures in general, are commonly used in cancer trials and have been well-tuned to capture important health impacts. We also know from a recent systematic review that physical function is the most common type of PRO to demonstrate a significant prognostic relationship with outcomes in cancer trials [Lancet Oncol 2019;20(12):e685-e698]. Given this context, we were not surprised to find that the PWB was a strong predictor of overall survival.”
Peipert added that the population of AML patients not eligible for intensive therapy is often older, potentially frail, and typically has a poor prognosis. Therefore, patient-reported assessments of physical function and well-being may be particularly useful in tracking functioning.
“Of course, PROs have the added benefit of capturing impairment and well-being deficits from the patient’s perspective, which may add to their sensitivity as predictors of prospective health problems and poor outcomes,” Peipert said.
The authors concluded by stating that the results may indicate PROs' value as stratification factors in trials with AML patients and underscore the need to more systematically collect PRO data in routine care practice with AML patients.
“Our findings are most relevant for designing new trials of AML treatments for people who are not candidates for intensive therapy. The PWB is brief, with only 7 questions. It may be a practical assessment useful for determining inclusion or stratifying patients who are enrolled in trials.
“For example, these questions could play a similar role in stratification that is often filled by performance status ratings such as the ECOG or Karnofsky, but would have advantages over these ratings, providing more detailed, fine-grained information and sensitive assessment of functioning. One of the key next steps in this line of research would be to identify clinically relevant cut-off values of the PWB scale that could be used to better stratify patients in future RCTs,” Peipert concluded.
Sarah LaCorte is associate editor.