Page 169 - 2020_02-Haematologica-web

Comparing front-line therapies in older AML patients
Figure 2. Treatment responses based on various treatment modalities. CR/Cri: complete response or complete response with incomplete count recovery; HI: high intensity; HMA: hypomethylating agent; LI: low intensity; TRM: treatment-related mortality.
also shown to be superior to supportive care with respect
to overall survival. Interestingly, among the high-intensity
and low-intensity treatment cohorts, overall survival sig-
nificantly favored high-intensity treatment. Superior out-
comes with high-intensity chemotherapy versus lower
intensity chemotherapy and supportive care have been
previously reported in older AML patients.1,3,24,25 Together
with our results, it is apparent that providing any treat-
ment is superior to no treatment (supportive care) and
these data may provide support to select intensive
chemotherapy over lower intensity treatment in eligible
patients. However, given the heterogeneity of the disease,
risk stratification based on biological features of disease,
functional status, comorbidity assessment, and cytogenet-
ics rather than age alone should help guide treatment deci- sions.2,6,10,26,27
Although the superiority of high-intensity treatment over supportive care or low-intensity treatment was evi- dent, high-intensity treatment failed to show survival superiority versus HMA in both our univariate and multi- variate analyses. We found that high-intensity therapy conferred at least 35% higher risk of mortality than treat- ment with HMA. To eliminate the selection bias in our retrospective non-randomized study and to be able to accurately estimate the effects of treatment by reducing the bias due to confounding variables (such as baseline CCI among other co-variates), we implemented the propensity score matching method. Even with this method, the overall survival benefit was upheld with HMA treatment compared with high-intensity treatment (Online Supplementary Table S2). Our findings contrast somewhat from data previously reported by Quintas- Cardama et al. that indicated therapeutic equivalence between HMA and high-intensity therapy, including within the intermediate-risk cytogenetic group.28 However, our data focused on a somewhat older popula-
tion and used propensity score matching to minimize selection bias for front-line treatment options.
The higher rate of TRM that we observed with high- intensity treatment compared with HMA treatment (7.2% vs. 1.5%) may be implicated as a potential cause for the overall inferior survival, although it cannot be the sole rea- son for overall inferiority. Distinct disease biology of AML in older patients (compared with younger patients with AML) is certainly a contributing factor for suboptimal treatment responses. Secondary AML originating from a prior myelodysplastic syndrome is common in the elderly and portends a poor prognosis. In our patient cohort, a sig- nificant proportion (56.9%) had secondary AML, primari- ly stemming from myelodysplastic syndromes. For this subgroup of patients, induction with intensive chemother- apy is frequently utilized, but the duration of response and long-term outcomes continue to remain poor.29
Treatment with a prior HMA has been previously shown to be an independent negative predictive factor for responses and overall survival in patients with secondary AML.30 In our analysis of this high-risk subgroup with prior HMA exposure, treatment with a high-intensity reg- imen did not produce significantly superior overall out- comes (HR=1.25, 95%CI: 0.68-2.27; P=0.47) compared with HMA. Moreover, low-intensity treatment also failed to produce improved outcomes compared with HMA treatment (HR=1.42, 95%CI: 0.76-2.66; P=0.28). Interestingly, the supportive care cohort had far inferior outcomes than patients in the HMA group (HR=2.29, 95%CI: 1.30-4.02; P=0.004). These results further rein- force the notion that some therapy may be superior to supportive care only, including that a clinical trial should be strongly considered whenever possible for this group.
A small minority (12.1%) of the patients in the HMA cohort had been previously treated with HMA. Typical approaches for such patients at our institution include
haematologica | 2020; 105(2)
403