AML: Fitness, Not Age, Predicts IC Eligibility at EHA 2026

Acute Myeloid Leukemia (AML) predominantly affects older individuals, with a median age at diagnosis of approximately 68 years.¹ Intensive chemotherapy (IC) regimens, typically involving an anthracycline and cytarabine, offer the best chance for complete remission and long-term survival.² However, concerns regarding treatment-related mortality and morbidity in older patients have historically led to age-based exclusion from IC, often without a nuanced assessment of physiological reserve. This practice has resulted in a significant proportion of older, fit patients being denied potentially curative treatment, while some frail younger patients may receive therapy they cannot tolerate.³ The EHA 2026 presentation directly addresses this disparity by evaluating the utility of fitness criteria beyond chronological age. The rationale for this study stems from a growing understanding that physiological age, rather than chronological age, is a more accurate predictor of tolerance to aggressive therapies and subsequent outcomes in various oncological settings. This is particularly relevant in AML, where treatment decisions carry significant risks and potential benefits.

What the study did

The presented analysis pooled data from several prospective trials and real-world registries, encompassing N=3,870 patients with newly diagnosed AML.⁴ The cohort included patients aged 60-80 years, stratified by age (60-69, 70-75, >75 years) and by a composite fitness score. This score integrated Eastern Cooperative Oncology Group (ECOG) performance status, the Charlson Comorbidity Index (CCI), and a modified Geriatric Assessment (GA) score, which included assessments of instrumental activities of daily living (IADL) and cognitive function.⁴ Patients were categorized as 'fit,' 'intermediate-fit,' or 'unfit' based on these composite scores, irrespective of their chronological age. The primary endpoints were overall survival (OS) and complete remission (CR) rates following IC. Secondary endpoints included early mortality (30-day and 60-day) and incidence of grade 3 or higher non-hematologic toxicities.⁴ The specific IC regimens generally followed standard 7+3 induction protocols, though minor variations existed across contributing studies, reflecting real-world clinical practice. Patient characteristics such as cytogenetics and molecular mutations were also collected and considered in subgroup analyses, although the primary focus remained on fitness and age.

The study demonstrated that among patients aged 70-75 years, those classified as 'fit' (ECOG 0-1, CCI ≤2, intact IADL and cognitive function) achieved a complete remission rate of 68% (95% CI, 64-72%) following IC, which was comparable to the 72% (95% CI, 69-75%) observed in fit patients aged 60-69 years.⁵ The median overall survival for fit patients aged 70-75 was 14.5 months (95% CI, 12.8-16.2 months), not significantly different from the 16.1 months (95% CI, 14.7-17.5 months) in fit patients aged 60-69 years (Hazard Ratio [HR] for mortality: 1.08; 95% CI, 0.95-1.23; P=.24).⁵ These results suggest that chronological age alone does not dictate treatment outcomes for physiologically robust older individuals.

Conversely, patients aged 60-69 years classified as 'unfit' (ECOG ≥2, CCI >2, or significant IADL/cognitive impairment) experienced a significantly lower CR rate of 42% (95% CI, 38-46%) and a median OS of 6.3 months (95% CI, 5.1-7.5 months).⁶ This 'unfit' younger cohort had a 30-day mortality rate of 18% (95% CI, 15-21%), which was higher than the 12% (95% CI, 10-14%) observed in fit patients aged >75 years (HR for 30-day mortality: 1.56; 95% CI, 1.25-1.95; P<.001).⁶ The incidence of grade 3 or higher non-hematologic toxicities was primarily driven by fitness status rather than age. Fit patients, regardless of age, had lower rates of severe infections, mucositis, and cardiac events compared to unfit patients. For instance, severe infection rates were 35% in fit patients aged 70-75 versus 58% in unfit patients aged 60-69.⁶ These findings underscore the importance of comprehensive fitness assessments in identifying patients at higher risk of treatment-related complications, irrespective of their birth year.

The study's limitations include its retrospective component from registry data, which may introduce selection bias, and the heterogeneity in specific GA tools used across different contributing studies. While the composite fitness score aimed to standardize assessment, variations in data collection could influence outcomes. For example, the precise instruments used to assess IADL or cognitive function might have differed, potentially leading to subtle inconsistencies in fitness categorization. Furthermore, the study did not specifically evaluate the impact of novel AML therapies, which have emerged in recent years, on fitness-based treatment decisions. Future prospective trials are warranted to validate these findings with a uniformly applied, standardized fitness assessment tool. Nevertheless, the consistent trend across multiple datasets strengthens the argument for a fitness-based approach.