Prognostic Score for Adults with Acute Myeloid Leukemia in First Complete Remission

Background

Current prognostic scoring systems in acute myeloid leukemia (AML) help predict outcomes in newly diagnosed patients at the time of diagnosis. In AML, achievement of complete remission (CR) is essential for the success of treatment. Therefore, development of a prognostic score for patients who enter first CR would be an important tool for physicians to guide their decisions for further treatment at the time of CR.

Patients and Methods

We developed a prognostic score based on data from 938 newly diagnosed adult AML (non-APL) patients (median age 59 years, range [18 - 88]) diagnosed and treated at the University of Texas MD Anderson Cancer Center, USA between [1999-2012], who achieved 1^st CR/CRp and were not transplanted in 1^st CR. Median follow-up was 22.3 months [range 1.4-165.1 months]. Information on MRC 2010 classification was available for 896 patients (156 (17.4%) patients in favorable, 550 (61.4%) in intermediate, and 190 (21.2%) in adverse group); ELN classification could be determined for 478 patients (160 (33.5%) in favorable, 88 (18.4%) in intermediate-1, 103 (21.5%) in intermediate-2, and 127 (26.6%) patients in adverse risk group.

For survival analysis Kaplan-Meier method was used. Survival distributions were compared using log-rank tests. Univariable and multivariable Cox regression analyses were performed to evaluate an impact of clinically significant prognostic variables (P<0.05). Each significant variable in the multivariate analysis was then assigned points based on the weight of the coefficient. A sum of the points led to a score that segregated patients into different prognostic subgroups (Table 1 and 2).

Results

Based on multivariable analysis, age and MRC/ELN subgroup were statistically significant prognostic factors (p<0.001) and were subsequently included in the prognostic score model. Since age (<60 and ≥60 years old) is an important prognostic factor not directly accounted for in the original MRC2010 and ELN classification systems, this was added to each of the new prognostic scores: "MRC+age" and "ELN+age".

In "MRC +age" prognostic score four groups were identified: favorable, intermediate-1, intermediate-2 and poor (Table 1B). In "ELN+age" score five groups were identified: favorable, intermediate-1, intermediate-2, poor and very poor (Table 2B).

According to "MRC+age" prognostic score, the 3- and 5-year overall survivals (OS), respectively were: 80% and 77% in the favorable group; 53% and 45% in the intermediate-1 group; 31% and 19% in the intermediate-2 group; and 8% and 6 % in the poor group. According to "ELN+age" prognostic score, the 3- and 5-year OS, respectively, were 78% and 77% in the favorable group; 68% and 50% in the intermediate-1 group; 46% and 37% in the intermediate-2 group; 27% and 22% in poor group; and 3% and 2% in the very poor group (Figure 1; A and B).

Conclusion

Despite the new discoveries in AML (DNA sequencing, genomic mutations, etc.) many clinics worldwide still don't have access to those tools and they stratify patients according to MRC or ELN classifications. The prognostic score presented above could be a powerful tool for physicians to risk-stratify patients with AML in first CR for post-remission therapy based on age and basic cytogenetic and/or molecular testing.

Figure 1.

View largeDownload slide

Cumulative rates of overall survival in patients with newly diagnosed AML in 1^st complete remission according to A) MRC + Age prognostic score, and B) ELN + Age prognostic score

Figure 1.

View largeDownload slide

Cumulative rates of overall survival in patients with newly diagnosed AML in 1^st complete remission according to A) MRC + Age prognostic score, and B) ELN + Age prognostic score

Close modal