Predictive and Prognostic Markers in Adults With Acute Myeloid Leukemia: A Single-Institution Experience

Acute myeloid leukemia (AML) is a heterogeneous malignancy with diverse genetic abnormalities, clinical presentations, and outcomes. Known predictive and prognostic factors in AML include age, performance status, comorbidities, cytogenetics, and molecular mutations. Identifying prognostic and predictive factors can inform the choice of induction therapy and outcomes prediction.

A retrospective review was performed of 137 adult AML patients from 2010 to 2015. Predictors of complete remission (CR) and overall survival (OS) were determined for patients treated with 3+7 (3 days of anthracycline and 7 days of cytarabine) or hypomethylating agent. Variables associated with CR or OS were assessed using univariate Cox regression and a multivariate Cox model.

The average age was 65 years and 91 patients (66%), sample size is 137 patients had primary AML. Patients in the 3+7 induction group were younger, had a higher bone marrow blast percentage, and more de novo AML compared with those in the hypomethylating agent group (P < .001, P < .001, P = .005, respectively). Univariate logistic regression for CR showed a significant association between age (P < .001), choice of induction (P < .001), and monosomy (P = .015), although only induction with 3+7 (P < .001) and absence of monosomy (P = .042) remained significant in multivariate analysis. Univariate Cox regression indicated that age (P = .003), AML status (de novo or secondary; P = .0277), choice of induction (P = .030), and monosomy (P = .010) had a significant association with OS. Only younger age (P = .018) and absence of monosomy (P = .022) were predictive of OS in multivariate Cox analysis.

Positive predictors of CR in adult AML include absence of monosomy and induction treatment with 3+7; whereas positive predictors of OS are younger age and absence of monosomy.