Decision Criteria for the Futility of Intensive Chemotherapy in Older Patients with Acute Myeloid Leukemia (AML).

Aims and methods. We have recently reported results of intensive chemotherapy in 416 patients with AML aged 65 years or more (median, 72 years) treated in the ALFA-9803 trial (Gardin et al., Blood 2007). We show here the impact of pretreatment characteristics on short-term mortality in these patients (32% at 6 months in the whole population). A first objective was to evaluate the prognostic value of the Charlson Comorbidity Index (CCI) and Sorror Hematopoietic Cell Transplantation Comorbidity Index (HCTCI), but the main objective was to screen the most frequent characteristics individually or in combination, including comorbidities, for their sensitivity in short-term mortality prediction. The aim was to propose decision criteria to advice against the intensive approach in high-risk patients, defined here by the presence of at least one characteristic associated with a probability of death at 6 months of 50% or more. Value of comorbidity scores. Both comorbidity scores correlated pretty well (CCI 0/1/2 = 353/57/6; HCTCI 0/1/2/3+ = 268/85/42/21; P<0.001), but only HCTCI was predictive of mortality (P<0.001). Other independent factors were age, PS, and cytogenetics. As all these patients were previously selected as suitable for intensive chemotherapy, only four HCTCI comorbidities were, however, relatively frequent (prevalence ≥ 5%): coronary artery disease (10%), arrhythmia (8%), infection (8%), and diabetes (7%). Impact on mortality was only due to the demarcation of very few high-risk patients (N=21) limiting the clinical interest of HCTCI in treatment decision making. Of note, HCTCI did not correlate with advanced age or PS in this patient population. Definition of decision criteria. Further analysis of the predictive value of each characteristic or combination identified three decision criteria, each being predictive of 6-month mortality ≥ 50% (Table 1): high-risk cytogenetics, pre-treatment documented infection, and PS ≥ 2 if age ≥ 75 years. Taken together, these 3 criteria, which were validated in an independent set of 123 patients, allowed to demarcate 94 high-risk patients (23%) with a probability of death at 6 months of 57%, as compared to 26% in the remaining patients (P<0.001, by log-rank test). We propose thus to add these criteria to usual eligibilty criteria in order to better define the population of older AML patients who will draw a significant benefit from intensive chemotherapy. Table 1. Short-term mortality associated with most frequent Characteristic. Characteristic Prevalence Median OS (mo) 6-month mortality High-risk cytogenetics 12% 4.8 64% Documented infection 8% 4.7 63% PS≥2 and age≥75 years 7% 2.9 54% PS≥2 27% 7.0 47% Age≥75 years 20% 7.9 42% Coronary artery disease 10% 6.8 43% Diabetes 7% 14.2 41% Arrhythmia 8% 14.6 31% Post–MDS AML 15% 10.6 29%