Fried's Frailty Phenotype Predicts Mortality and Survival for Newly Diagnosed Older Patients with Acute Myeloid Leukemia or High-Risk Myelodysplastic Syndrome

Introduction: Older patients with high-risk myeloid malignancies are now eligible for either induction chemotherapy (IC) or novel treatment strategies including liposomal cytarabine and daunorubicin (Vyxeos) or hypomethylating agent (HMA)/venetoclax. Vyxeos and HMA/venetoclax may be associated with less early mortality, but no trials have demonstrated significantly less mortality compared to IC. In this context, we need a better tool to determine a patient's risk of early mortality by strategy in order to inform therapy selection for older patients. The current methods to assess a patient's fitness for IC are the Eastern Cooperative Oncology Group (ECOG) or Karnofsky Performance Status (KPS) assessments, clinician subjective evaluations [i.e. gestalt (CG)], and often age itself (>75-80 years). None of these include objective measures of fitness. The Fried frailty phenotype (FP) uses both subjective (exhaustion and activity level) and objective measures (weight loss, gait speed, and grip strength) of frailty to categorize patients into fit, pre-frail, and frail. We hypothesized that FP would correlate with early mortality in this population and could be used to guide initial treatment selection. Methods: From September, 2018 to June, 2019 we prospectively enrolled 30 patients age 60 years or older with newly diagnosed, untreated acute myeloid leukemia (AML) or high-risk myelodysplastic syndrome (MDS) on an ongoing institutional review board approved clinical trial. We performed FP, CG, hematopoietic cell transplantation-comorbidity index (HCT-CI), and ECOG assessments on all patients prior to initiation of any therapy for their high grade myeloid disease. The primary endpoints were 60 and 100-day mortality. Results: Median patient age and follow up were 70.9 (range 61-82) years and 111 days, respectively. Patients had high-risk disease features as shown in Table 1, with the vast majority of AML patients being adverse risk by European Leukemia Network criteria and all MDS patients being high or very high-risk by the Revised International Prognostic Scoring System. The most common molecular mutations were TP53, ASXL1, RUNX1, IDH1/2, FLT3, and DNMT3A. IC consisting of either 7+3 or Vyxeos was used in 11 patients, HMA alone or in combination in 16, targeted therapy with enasidenib alone in 1, and best supportive care (BSC) in 2 (Table 1). Four patients who were frail by FP received IC (3 with Vyxeos). By FP assessments, 17% of patients were fit (score 0), 33% were pre-frail (score 1-2), and 50% were frail (score 3-5). In contrast, CG categorized 37% of patients as fit, 30% as pre-frail, and 33% as frail. Seven of 15 frail patients by FP were categorized as pre-frail or fit by CG. FP and ECOG scores also differed with 7 patients having low-risk ECOG scores of 0-1 and a high-risk FP of ≥3 (frail). For entire cohort, 60- and 100-day mortality rates were 15% and 25%, respectively. Frail patients by FP had 60-day and 100-day mortality rates of 32% and 51%, respectively, which was significantly worse than fit and pre-frail patients, all of whom were alive at 100 days (Figure 1/2, p=.016). Of the 6 out of 15 FP frail patients who died, 1 received a HMA alone, 3 HMA with venetoclax, and 2 BSC. In patients with ECOG scores of 0-1 or those categorized as fit by CG, 100-day mortalities were 8% and 10%, respectively. Mortality for patients with HCT-CI scores of 1-2 was unexpectedly higher (34% at both 60 days and 100 days) than for patients with higher risk HCT-CI scores ≥3 (13% and 29%, respectively). Conclusion: In newly diagnosed older patients with advanced myeloid neoplasms, being frail by FP was associated with increased 60 and 100-day mortality, owing to high mortality rates with HMA based therapies. While HCT-CI scores have been significantly associated with mortality after allogeneic transplantation, we observed that these scores were not linearly associated with survival or mortality in newly diagnosed AML and high/very-high risk MDS patients. By ECOG and CG, even the low risk groups contained patient deaths, suggesting that these standard metrics may not capture certain at-risk patients. CG tended to down grade frailty scores, categorizing patients as fitter than they were by FP. As such, FP may be a useful tool to predict early mortality. Further research is warranted to determine whether mortality differs by treatment within a given FP category, which may support the use of FP to select initial therapy.