Assesment of Frailty Scores Validating Fitness of Elderly AML Patients for Intensive Chemotherapy: A Single Centre Study

Background: Although acute myeloid leukemia (AML) occurs most commonly in adults ≥60 years, the treatment of AML in older patients remains a significant challenge due to both, more aggressive disease biology as well as patient-related risk factors that limit tolerance of intensive chemotherapy. Several studies demonstrated improved survival for older patients receiving intensive induction chemotherapy. Therefore intensive chemotherapy, as long as patient fitness allows, is aimed. Defining the subset of patients that is eligible or "fit" for intensive chemotherapy involves a great deal of subjectivity. Criteria yet have to be standardized across or within institutions.

Aim: The aim of the present study was to investigate the power of three validation scores in assessing patient fitness at diagnosis in parallel to physician evaluation. Further patient outcome according the respective assessment was compared.

Methods: A total of 130 clinically and molecularly well characterized consecutive elderly (≥60 years) patients with newly diagnosed AML were treated from 2012 to 2018 according to age, performance status and co-morbidities in a single hematology center. Initial haematologist evaluation followed by discussion of the patient case in an interdisciplinary board led to decision of therapy intensity of each patient. In parallel, independently from the medical board decision, three scores were performed: i) a local geriatric G8 screening tool, consisting of seven items from the Mini Nutritional Assessment (MNA) questionnaire and age, ii) the Sorror Index used for hematopoietic stem cell transplantation (HSCT) evaluation as well as iii) the AML score proposed by the German Acute Myeloid Leukemia Cooperative Group, predicting the probability of complete remission (CR) and early death (ED). Therapy response was defined according to ELN criteria. Overall survival from diagnosis was compared between groups using the Cox model.

Results: Median age was 71,2 (range: 60-86) years. A total of 75 (57,7%) patients were evaluated "fit" by the medical board and treated by intensive chemotherapy ("7+3" regimen), whereas 41 patients (31,5%) underwent semi-intensive therapy (based on hypomethylating agents or low-dose Cytarabine s.c.) and 14 patients (10,7%) received best supportive care. Fifty patients (38,5%) achieved a complete remission after induction chemotherapy, could follow consolidation chemotherapy and ten of them underwent allogeneic HSCT. Sixty-four (49,2%) were non responders and 16 patients (12,3%) died during the first cycle.

Overall, the median survival time was 11,2 months (95% CI 7,9-17,8). Primary physician care evaluation was able to define a "fit" from an "unfit" patient in a statistically significantly way. Median survival time from the "fit" patients was 17,6 months (95%CI 9-28,9) compared to the "unfit" evaluated patients with 3,6 months (95%CI 1,8-8,8), p<0.001 with a HR (unfit vs. fit) of 3,04 (95% CI of 1,80-5,15). Even the distinction by the local G8 screening tool resulted significant distinguishing "fit" with median OS of 18,7 months from "unfit" patients with a median OS of 7,9 months, HR (unfit vs. fit) 2,1 (95% CI 1,23-3,58), p=0,007. The Sorror Index (HR 2,14 with a 95% CI of 1,27-3,59) as well as the AML score ED (HR 1,94 with 95%CI of 1,16-3,24) resulted also significant in their power of separating "fit" from "unfit" patients, p=0,004 and p=0,011, respectively. Although a certain degree of correlation is expected between scores, this was limited to levels below 0.35 according to the pairwise spearman correlation coefficient.

Conclusion: In conclusion, the frailty scores G8, Sorror Index and the AML Score, that have been applied to elderly patients at AML diagnosis seem to discriminate patients quite well in terms of overall survival in this single centre patient cohort and may represent objective instruments for the haematologist in the validation of the fitness of the elderly AML patients for intensive chemotherpy. In order to validate the discrimination ability arising from the performed analysis, a multi-centre study is planned.