Use of Geriatric Assessment and Genetic Profiling to Personalize Selection of Intensive Versus Low Intensity Chemotherapy in Older Adults with Acute Myeloid Leukemia (AML): Final Results of a Phase II Trial

Introduction: Geriatric assessment can predict the risk of toxicities of chemotherapy in older adults. Genetic risk categories are associated with survival following intensive chemotherapy in AML. Integrating geriatric assessment for patient profiling and genetic profiling of leukemic cells represents an innovative precision medicine approach to personalize therapy selection in older adults with AML. We report final results of a pragmatic phase II trial using such strategy, aiming to reduce early mortality (NCT03226418).

Methods: Patients ≥60 years with a new diagnosis of AML underwent geriatric assessment prior to initiation of treatment. Geriatric assessment of physical function, cognitive function and comorbidity burden were used to determine fitness for intensive chemotherapy (Table 1). Additional assessment included Karnofsky Performance Scale (KPS), Patient Health Questionnaire-9 (PHQ-9), and Mini Nutritional Assessment-Short Form (MNA). Genetic profiling for therapy selection relied on karyotyping and followed the 2017 European LeukemiaNet criteria. While available mutation test results were incorporated to risk stratify, the study did not require waiting for the results prior to therapy initiation, given an anticipated turnaround time of 1-2 weeks for mutation test results. Therapy selection followed the algorithm in Figure 1. Patients with good or intermediate-risk AML received intensive chemotherapy such as 7+3 +/- gemtuzumab or midostaurin if determined to be fit. Patients with high-risk AML received low-intensity chemotherapy such as a hypomethylating agent with or without venetoclax or novel drugs (based on timing of FDA approval of these drugs), or CPX 351 if they were fit and met the FDA-approved indications. Patients with organ dysfunction (e.g. creatinine ≥2 mg/dl) and those requiring chemotherapy for other malignancy were eligible for low-intensity chemotherapy. Chemotherapy could be administered in community oncology settings. Patients were followed for quality of life assessments, as well as functional and oncologic outcomes.

Results: Between July 2017-October 2022, 75 patients were consented; 2 patients were considered screen failure. Baseline characteristics of 73 eligible patients included a median age of 69 years (range 60-87 years), 49% female, 92% white, and a median KPS of 80 (range 60-100). As presented in Table 1, most had ≥3 comorbidities (55%), impaired physical function measured by Short Physical Performance Battery (70% had a score of 9 or less) and impaired cognition measured by Montreal Cognitive Assessment (64% had a score of 25 or less). Genetic risk categories included adverse (62%), intermediate (22%), and good-risk AML (16%). Patients had one or more of the following mutations: TET2 (26%), NPM1 (22%), RUNX1 (19%), ASXL1 (19%), TP53 (19%), DNMT3A (16%), IDH2 (15%), IDH1 (11%), FLT3 ITD (10%), FLT3 TKD (10%), biallelic CEBPA (4%), EZH2 (2%). Seven patients (10%) received intensive chemotherapy; CPX 351 (n=4) or 7+3 based regimen (n=3). Other patients received low-intensity chemotherapy: azacitidine or decitabine and venetoclax (n=43, 59%), decitabine or azacitidine alone (n=18, 24%) prior to approval of venetoclax, and other (n=5, 7%). The median time from diagnosis to therapy initiation was 6 days and from enrollment to therapy initiation was 1 day (range 0-13). Mortality at 30 days from diagnosis was 6.8% (95% confidence interval, CI 3.0-15.1%) and at 90 days was 21.9% (95% CI 14.0-32.7%). Mortality compared favorably to an unmatched historical cohort of patients ≥60 years treated at our center between 2011-2016, where 30-day mortality was 30% (95% CI 22-40%) and 90-day was 41% (95% CI 32-52%) (Future Oncol. 2019;15:1989-95).

Conclusions: Our model to personalize AML therapy selection represents an innovative approach to precision medicine that incorporates both geriatric assessment for patient profiling and genetic profiling of leukemia cells. Geriatric assessment demonstrated high frequency of impairment in objective physical and cognitive function. Patients were able to start therapy within a median of 1 day following enrollment. Pragmatic aspects of the trial included broad eligibility criteria and co-management of patients with community oncologists. The results from our final analysis appear promising with lower rates of early mortality compared to unmatched historical controls.