Prediction of Early Mortality With Nonintensive AML Therapies Is Limited

For elderly patients with acute myeloid leukemia (AML) treated with less-intensive therapy, the ability to predict early death with routinely available clinical variables is limited, according to data presented at the American Society of Hematology’s 2021 Annual Meeting.

To better predict all-cause death by days 28, 42, 56, and 100 in patients with AML treated with less-intensive therapies, an international research team developed predictive models by using data from nearly 800 patients in the Less Intensive 1 (LI-1) trial. The models were then validated in a cohort of 540 patients taking treatment during 3 SWOG trials. Patients’ mean age was 75 years in the LI-1 cohort and 77 years in the SWOG cohort. 

To build the main multivariate logistic regression model, the team used routine clinical variables assessed at study enrollment that were available for both cohorts as covariates. These included age, performance status (0-1 vs 2-4), disease status (secondary AML vs de novo AML or high-risk myelodysplastic syndrome), leukocyte and platelet counts, and percentage of bone marrow blasts.

They then used the regression coefficients from the model fit in the LI-1 cohort to obtain a score that was applied to each patient in the SWOG cohort and assessed the predictive accuracy of the model by using the area under the receiver operating characteristic curve (AUC). Next, they added cytogenetic risk, mutation status, or patient-reported outcomes (PROs) as a covariate and determined how well a model with each of these enhancements could predict early death.  

Using these methods, the team found that the predictive ability of the models was limited. The model developed by using routinely available clinical covariates alone in the LI-1 cohort had optimism-corrected AUCs of 0.63 to 0.66. Including any of the additional covariates improved the model’s predictive accuracy only slightly.

 “Our ability to predict early death in older patients treated with lower intensity AML therapies is limited with routinely available clinical variables,” the team concluded. “Improving the clinical utility of these models may require more complete characterization of patient comorbidities.”

—Ellen Kurek

Reference:

Wang X, Thomas I, Ariti C, et al. Prediction of early mortality with non-intensive acute myeloid leukemia (AML) therapies: analysis of 1336 patients from MRC/NCRI and SWOG. Paper presented at: American Society of Hematology’s 2021 Annual Meeting; Atlanta, Georgia; December 11-14, 2021. https://ash.confex.com/ash/2021/webprogram/Paper151597.htm