Page 170 - 2020_02-Haematologica-web

C. Talati et al.
Table 3. Multivariate analysis and comparisons of entire cohort versus patients without prior exposure to hypomethylating agent prior to diagnosis of acute myeloid leukemia and versus patients who received hypomethylating agent prior to the diagnosis of acute myeloid leukemia.
Clinical Parameter
Type of AML
De novo
Secondary
Karyotype
Favorable or intermediate
Adverse ECOG PS
0-1
2-4
Front-line therapy
HMA
HI therapy
LI therapy Supportive care
Age at diagnosis (per 5-year increase) WBC, per 1 log increase
Platelets, per 1 log increase
Hemoglobin, per 1 log increase
All Patients (n=980)
P Hazard 95%CI 95%CI Ratio Lower Upper
Without Prior Exposure to HMA
With Prior Exposure to HMA
P
(n=795)
Hazard 95%CI 95%CI
Ratio Lower Upper
P
(n=185) Hazard 95%CI Ratio Lower
95%CI Upper
Reference Reference
Excluded 1.34 1.13 1.59
Reference
0.001 1.87 Excluded
Reference
0.47 1.25
0.28 1.42
0.004 2.29
0.001 1.38 <0.0001 1.34 0.002 0.80 0.035 0.87
<0.0001 1.44 1.23
Reference
<0.0001 1.92 1.64
Reference
<0.0001 1.80 1.48
Reference
0.004 1.35 1.10
<0.0001 2.01 1.53
<0.0001 2.94 2.39 0.002 1.14 1.05 <0.0001 1.19 1.13 <0.0001 0.81 0.75
<0.0001 0.91 0.86
1.69 0.001
Reference
2.25 <0.0001
Reference 2.18 <0.0001
Reference 1.65 0.024
2.62 <0.0001
3.61 <0.0001
1.23 0.036
1.25 <0.0001
0.87 <0.0001
0.95 0.0003
2.02 1.69
1.82 1.47 2.26
1.29 2.72
0.68 2.27
0.76 2.66
1.30 4.02
1.14 1.67
1.16 1.54
0.69 0.92
0.77 0.99
2.41
1.29 1.03 2.12 1.54 3.02 2.40 1.10 1.01 1.18 1.11 0.80 0.73
0.91 0.86
1.61
2.91
3.81
1.21
1.25
0.88
0.96
AML: acute myeloid leukemia; CI: confidence interval; ECOG PS: Eastern Cooperative Oncology Group Performance Status; HI: high intensity; HMA: hypomethylating agent; LI: low intensity; WBC: white blood cell.
clinical trials (if available), intensive chemotherapy, lower- intensity approaches, or best supportive care. If HMA therapy is continued, we may utilize a different dosing schedule (10-day decitabine or 7-day azacitidine if treated with a 5-day schedule) or switch to the alternative HMA agent. Although limited, our data suggest that patients treated previously with HMA do not benefit from any specific standard-of-care approach, indicating the impor- tance of clinical trials for this subpopulation.
Since the Food and Drug Administration approval of CPX-351 for secondary AML (AML with myelodysplasia- related changes and therapy-related AML) that established a new standard of care for this distinct high-risk AML sub- group, the treatment landscape for AML has become increasingly complex.31 CPX-351 is considered an inten- sive chemotherapy and is demonstrated to have similar early TRM as “7+3”. However, CPX-351 has not been compared head-to-head with HMA-based therapies and were not included in our study. But such a comparison is warranted to determine the optimal treatment choice for older AML patients aged ≥70 years.
While the results of our study are potentially practice- changing, there are several limitations. Although this is the largest single-institution series of AML patients ≥70 years of age, a referral bias affecting baseline disease char- acteristics is expected. In our cohort, 50% of the patients had prior hematologic malignancy and >90% of these patients had diagnosis of MDS. In addition, treatment out- comes of patients seen at a tertiary care center may not reflect outcomes of the general community, thereby limit- ing its general applicability. For instance, per the SEER reg- istry studies, only 10-20% of elderly patients are treated with HMA or intensive chemotherapy, compared with
58% of the patients in our cohort.11,13 The non-random- ized retrospective nature of this study also does not allow for definitive conclusions to be made as there might be some inadvertent, inherent biases introduced that we did not consider, although we attempted to account for such bias via utilization of propensity score matching.
Although most patients in our cohort had cytogenetic results, the lack of molecular data in our analysis is anoth- er study limitation. Prior studies have shown that with advanced age there is an increase in the incidence of unfa- vorable cytogenetics, aberrant karyotypes, and molecular abnormalities.32-36 Testing newly diagnosed AML patients irrespective of their age for key molecular markers (includ- ing FLT3, NPM1, and KIT) should be universally done given their prognostic and therapeutic implications. Unfortunately, molecular testing and routine testing of all elderly patients with AML have only become a standard practice during the past ten years. The lack of available testing likely explains the lack of available molecular data in our database, which incorporates patients dating back to 1995.
Accurate assessment of baseline performance status and comorbidity measurements in elderly patients with AML can provide useful prognostic information and help guide treatment decisions. Functionality and comorbidity are independent prognostic variables and should be measured independently in elderly patients.37 A retrospective study by Etienne et al. reported that patients with a CCI score >1 had significantly lower rates of obtaining a CR than those having CCI scores <1.38 Analyses of the SEER data have also shown that survival of those with CCI of 0-1 improved with therapy, whereas those with CCI >2 expe- rienced early death and had minimal improvements in
404
haematologica | 2020; 105(2)