Acute Myeloid Leukemia in Elderly Patients: Experience of the Antoine Lacassagne Center and a New Score to Define Fit Patients

Management of Acute Myeloid Leukemia (AML) in the elderly is particularly difficult as life expectancy is highly variable and the benefit-risk ratio of treatment depends on comorbidities and age-related pharmacological specificities.

                                                                                                       

Objective:

To define the prognostic factors for overall survival (OS) and complete remission (CR) and establish a feasible and efficient new prognostic scoring system to assist clinicians in an age-adapted treatment strategy.

Methods:

From January 2000 to December 2014, 163 patients (pts) presenting an AML in a French regional cancer center in Nice were retrospectively analyzed. According to functional status, patients were treated with induction chemotherapy, azacitidine or palliative care. Complete remission rate (CR) and early-death rate were calculated. Six-month, 1-year and 2-year overall survival (OS) were analyzed with the Kaplan-Meier method and log-rank test. Univariate and multivariate logistic regression analyses were done and a p-value <0.1 and 0.05, respectively, were considered statistically significant.

Population:

From January 2000 to December 2014, 163 pts ³ 65 years were included; median age was 73.3 years (65-94.6), with 94 men (57.7%). PS was 0, 1, 2, 3 or 4 for 33 pts (20.3%), 77 pts (47.2%), 22 pts (13.5%), 20 pts (12.3%) and 6 pts (3.7%), respectively. Modified Charlson Score (calculated without considering age and leukemia criteria) was 0-1 and ³2 for 83 pts (50.9%) and 77 pts (47.2%), respectively. Secondary AML represented 44.2% of this elderly population. Induction chemotherapy was either full-dose cytarabin-daunorubicin 3+7, or cytarabin-idarubicin or cytarabin-clofarabine. Palliative care consisted of hydroxyurea and etoposid chemotherapy and best supportive care. Then, 112 pts (68.7%), 21 pts (12.9%) and 30 pts (18.4%) were treated with induction chemotherapy, azacitidine and palliative care, respectively. Among the 112 pts (68.7%) induced, 56 pts (35.9%) reached CR, 42 pts (25.8%) with a first induction and 14 pts (10.1%) with salvage chemotherapy. Otherwise, 39 pts (23.9%) ended up relapsing. Taking into account the entire treatment, 42 pts (25.8%) received azacitidine.

Results:

The 6-month, 1-year and 2-year OS for the entire cohort of 163 patients were 64.3%, 46.7% and 24.4%, respectively. Administration of induction chemotherapy was significantly predictive of CR (50% vs 2.0%, p<0.001) and almost predictive of 2-year OS (29% vs 12.5%, p=0.07). With or without induction chemotherapy, early-death rate (<8 weeks) was 12.5% and 37.3%, respectively. Thus, induction remained the best treatment for physically-fit elderly patients. Using azacitidine at any time of the treatment was a predictive factor of longer survival (1-year OS: 67.8% vs 36.9%, p=0.004). Otherwise, no impact was found on 2-year OS (25.6% vs 25.9%, NS). We tried to define a prognostic score to select elderly patients likely to benefit from induction chemotherapy. In univariate analysis, the significant prognostic factors of OS in the elderly population treated with induction chemotherapy was : age³ 75, unfavorable karyotype, creatinine clearance using the MDRD equation< 60ml/min/1.73m2, Modified Charlson Score ³2 and PS³2. In a multivariate analysis, only Charlson score and PS remained significant in terms of OS. Patients with Charlson Score ²1 and PS ²1 (0 unfavorable factor, n=60) or Charlson Score ³2 or PS ³2 (1 unfavorable factor, n=69) were considered to be at low risk (fit) for induction chemotherapy whereas patients with Charlson Score ³2 and PS³2 (2 unfavorable factors, 29 pts) were considered to be at high risk (frail). In the entire cohort (n=158 pts, data missing for 5 pts), Fit patients and Frail patients had an early-death rate of 10.8% vs 30.8% (p=0.048), a CR rate of 40.3% vs 3.8% (p<0.001), six-month OS 69.7% vs 39%, 1-year OS 53.7% vs 0% and 2-year OS 28.8% vs 0% (p<0.001), respectively. No survivors were observed in the frail group after one year of follow-up. Sub-group analysis found the same results for patients initially treated with induction chemotherapy or palliative treatment.

Conclusion:

We present a prognostic model composed of two easy-to-operate parameters that stratify patients into Fit or Frail groups. Patients with a Modified Charlson Score ³2 and a PS ³2 did not benefit from induction chemotherapy and had a high risk of death. In such cases, treatment should be azacitidine or palliative treatment. External validation is needed.