Time from Diagnosis to Treatment Initiation Predicts Survival in Acute Myeloid Leukemia (AML).

Abstract

Abstract Background: Pts, particularly those age ≥ 60, with untreated AML and prognostically unfavorable cytogenetics have CR rates < 50% when given standard induction therapy, and a treatment-related mortality approaching 25%. Waiting until cytogenetic results become known and offering investigational therapy to high-risk pts may be beneficial. This approach is potentially problematic, though, as it entails delaying induction therapy, contrary to the notion that AML must be treated immediately. We tested this hypothesis by examining the effect of time from AML diagnosis to treatment (TDT) on complete remission (CR) rates and overall survival (OS), using patient characteristics readily available to every clinician at a patient’s diagnosis. Methods: 1660 pts were initially considered in the joint CC-MDA database. After excluding pts with APL (n=10), <17 yrs (n=1), with white blood cell count (WBC) >50,000/mm3 (n=48), diagnosed >3 months prior to therapy (n=24), not treated with Ara-C (n=177) and with incomplete data (n=19), the final data set consisted of 1381 pts, 224 (16%) from CC, and 1167 (84%) from MDA. The effect of TDT on CR and OS was evaluated using logistic regression models for CR, log-logistic for OS, in older (≥60 years (yrs)) and younger (< 60 yrs) adults, controlling for treatment center and pre-treatment variables that would factor in to a decision about intensive chemotherapy (age, baseline WBC, and AML etiology (de novo vs. secondary (sAML))). Results: Median values were: age 60 yrs (range, 17–87), pre-treatment WBC 5.2/mm3 (0.3–49.1), and TDT 4 days (0–78). 44% of pts had sAML. Cytogenetic risk distribution was similar to other large databases: favorable 8%, intermediate 66%, and unfavorable 26%. CR was achieved in 837 pts (61%), 67% of those <60 yrs and 54% of those ≥60 yrs. Median survival was 47 weeks, 67 weeks for those <60 yrs and 32 weeks for those ≥60 yrs. In pts <60 yrs, sAML and increasing age predicted for lower CR rates and OS (p<0.001 for CR and OS for both), though not increasing WBC. In pts ≥60 yrs, CR rates and OS were worse with increasing WBC (p=0.002 and 0.01, respectively) and increasing age (p=0.006 and 0.002, respectively), but not for sAML. Considered alone, longer TDT was associated with worse CR rates and OS in younger (p<0.001 and p=0.001), but not older (p=0.48, p=0.14) pts. In multivariable analyses, in pts <60 yrs, TDT had a negative impact on both CR and OS (p=0.006 and p=0.003), as did increasing age (p<0.001 and p<0.001) and sAML (p<0.001 and p=0.008). For pts ≥60 yrs, TDT had no impact on outcome, though CR rates and OS were negatively impacted by increasing age (p=0.007 and p=0.001). In both age groups, increasing WBC negatively impacted OS (p=0.04 for <60yrs, p=0.01 for ≥60 yrs). Conclusions: AML should be considered a medical emergency in younger AML patients, as delaying induction therapy significantly affects CR rates and OS. Delaying therapy does not seem harmful in older pts, particularly those closer to 60 yrs with a low WBC, and would allow therapy to be individualized, e.g. by accounting for cytogenetics and other biologic markers. Inclusion of other covariates may alter these results, and we are assessing whether the harm in delaying therapy in younger AML pts is offset by the potentially greater harm in giving standard therapy to those later found to have unfavorable cytogenetics.