Regimen-intensity per count-recovery and hospitalization index: A new tool to assign regimen intensity for AML

Mohamed L. Sorror, Shirali Agarwal, Megan Othus, Ylinne Lynch, Jennifer E. Nyland, Mary Elizabeth Percival, Pamela S. Becker, Frederick R. Appelbaum, Elihu H. Estey

Research output: Contribution to journalArticlepeer-review

Abstract

Background: Low-intensity regimens have been increasingly used to treat older patients with acute myeloid leukemia (AML). Recent studies, however, suggest older patients can tolerate and potentially benefit from intensive chemotherapeutic regimens. The ability to compare the utility of varying regimen intensities in AML is hindered by the lack of a standardized definition of “regimen intensity.”. Methods: We conducted a survey asking AML physicians which of 38 regimens they would consider intensive vs less-intensive. Electronic medical records of 592 patients receiving many of these regimens were used to design a model characterizing regimens as intensive vs less-intensive as identified by ≥75% physician consensus. Variables included frequency and length of hospitalizations, intensive care unit admissions, severe gastrointestinal toxicities, time to nadir, and recovery of neutrophil/platelet count. Results: Physicians agreed at a rate of 75%-100% on the assignment of degree of intensity to the majority (n = 28) of these regimens, while the level of agreement was <75% for the remaining 10 regimens (26%). Logistic regression analyses identified number and length of hospitalizations to be significantly associated with intensive regimens and count recovery with less-intensive regimens. We created the “regimen-intensity per count-recovery and hospitalization” (RICH) index with an AUC of 0.87. Independent model validation yielded an AUC of 0.75. Conclusions: We were able to generate a novel model that defines regimen intensity for many therapies used to treat AML. Results facilitate a future randomized study comparing intensive vs less-intensive regimens.

Original languageEnglish (US)
Pages (from-to)6515-6523
Number of pages9
JournalCancer medicine
Volume9
Issue number18
DOIs
StatePublished - Sep 1 2020

All Science Journal Classification (ASJC) codes

  • Oncology
  • Radiology Nuclear Medicine and imaging
  • Cancer Research

Fingerprint Dive into the research topics of 'Regimen-intensity per count-recovery and hospitalization index: A new tool to assign regimen intensity for AML'. Together they form a unique fingerprint.

Cite this