SIE/Sies/GITMO Criteria in Elderly Patients with Acute Myeloid Leukemia (AML): Useful Also in the New Drug Era? a Multicentric Real-Life Study By the Hematological Network Rete Ematologica Lombarda (REL)

Introduction: In elderly acute myeloid leukemia (AML) patients (pts) the debate about the use of intensive chemotherapy (iC) as opposed to non-intensive therapy (niT) or best supportive care (BSC) is a hot topic, particularly after the introduction of venetoclax (VEN) use. Age, comorbidities, functional impairment, therapy benefits and risks, pts preferences, and AML features influence this choice. The SIE/SIES/GITMO (“fitness”) criteria (Ferrara, 2013) are a valuable and comprehensive tool, but a revalidation is needed in the light of the new treatment options.

Methods: Within the Rete Ematologica Lombarda (REL), we evaluated 1) the concordance between “fitness” of pts, and the type of treatment they actually received, 2) the overall survival (OS) according to the “fitness”, to the prognostic European Leukemianet (ELN17) stratification and to the treatment received, and 3) the presence of other parameters, as Charlsons Comorbidity Index (CCI), potentially useful in therapy decision. Pts were classified as fit to iC (FIT), unfit to iC (UNFIT), or unfit even to niT (FRAIL), as defined by the SIE/SIES/GITMO criteria.

Results: From Jan 20 to Dec 22, 503 AML pts (53% de-novo), with a median age of 76 years (y) (range, 65-93) were diagnosed consecutively at 11 Hematology Units.

According to “fitness” criteria, 25% of pts were FIT, 61% UNFIT, and 14% FRAIL (Table1). Median age was significantly lower in FIT pts (70 y), as compared to UNFIT (78 y) and FRAIL (79 y) (p <0.0001). Overall, the concordance between “fitness criteria” and the treatment actually received by pts was 75.9% (71% in FIT, 76% in UNFIT and 84% in FRAIL). After a median follow-up of 12 months (m), median OS of the whole population was 7.4 m. Median OS of FIT, UNFIT and FRAIL pts was 13.7, 7.2 and 1.4m, respectively (p <0.0001, Fig 1a).

According to physicians' decision, 18% of pts received iC, 55% niT, mainly hypomethylating agents (HMA: azacytidine or decitabine) + VEN, and 27% BSC. Median OS was 17m, 10.2m, 1.4m in iC, niT and BSC pts, respectively (p<0.0001, Fig 1b). According to ELN17, evaluable in 73% of pts, 19% were favorable (fav), 20% intermediate (int) and 61% adverse (adv), without differences into “fitness” groups. Median OS was significantly worse in ELN adv than in ELN fav and int (p: 0.001 and p: 0.002, respectively, Fig 1c).

In FIT pts, median OS was 17m with iC and 11.4m with niT (VEN-HMA) (p: 0.3). The use of iC was associated with younger age (median age: 69 in iC vs 73 y in niT, p<0.0001), lower CCI (CCI< 2: 88.5% in iC vs 64% in niT, p: 0.0032) and not significantly with better ELN risk (ELN adv 49% in iC pts vs 70% in niT, p: 0.07). The use of iC or niT did not affect remission rate (64% vs 71%) nor OS in different ELN risk groups.

In UNFIT pts, median OS was 14m with niT and 3.8m with BSC pts (p<0.0001). The use of niT compared to BSC was associated with younger age (median 77 vs 81 y; p<0.0001), not adv ELN risk (41% vs 7%, p: 0.0001), lower ECOG performance status (PS) and CCI (PS<2: 81% vs 65%, p: 0.0001; CCI<2: 66% vs 50%, p: 0.038). In UNFIT pts treated with VEN-HMA median OS was 11m vs 7.3m with HMA only (p: 0.019). The use of VEN-HMA rather than HMA only was associated with younger age (median: 76 vs 80 y; p: 0.0045), better PS and CCI (PS<2: 94% vs 84%, p: 0.026; CCI<2: 68% vs 50%, p: 0.008). ELN risk distribution was similar. In non-adv risk pts, OS was 17m with VEN-HMA, significantly better compared to HMA only (7m) (p: 0.03) as well as to adv risk pts treated with VEN-HMA (p: 0.0017). Age >82y, number of comorbidities, lack of caregiver were the most frequent reasons for the physician's preference for HMA therapy over VEN-HMA.

InFRAIL pts, median OS was 2.9 m with niT and 1.1m with BSC pts (p: 0.04). The use of niT was associated with a relatively better PS and younger age (<73 y).

At multivariable analysis, ELN fav and iC favorably affected survival, whereas a FRAIL fitness status, PS>2, and CCI>2 proved to be independent adverse prognostic factors.

Conclusion: Fitness stratification of SIE/SIES/GITMO correlated to pts's survival and were useful in the decision-making approach of treatment also in the new drugs era.

Within the “fitness” categories, age, CCI and particularly ELN risk, had an impact on the choice of the type of induction therapy, in FIT pts (iC vs niT) without changing the outcome, while in UNFIT pts an under-treatment worsened OS.

These preliminary data should be implemented with a longer follow-up and a larger sample size.