Factors Affecting the Survival of Therapy Related AML/MDS, Secondary AML and De Novo AML

Introduction:

Patients exposed to cytotoxic agents are at a higher risk of developing therapy related AML and MDS (tAML/tMDS), and have poor survival as compared to de novo AML due to high risk of adverse features. Secondary AML (sAML)includes patients with progression from myeloproliferative neoplasms (MPN) and myelodysplastic syndrome (MDS) to AML, considering this progression could be natural history of disease process. Patients with tAML are considered to have an inferior outcome compared with de novo AML. In this retrospective chart review study, we aimed to look at factors affecting the survival of tAML/tMDS, sAML and de novo AML.

Method:

This retrospective analysis included 219 AML patients treated at Wake Forest Baptist Medical Center between January 2010 and December 2016. Kaplan-Meier estimation was used to evaluate survival at one and two-year period in these three types of AML. Multiple Cox proportional hazards models were used to examine the interaction between baseline characteristics (Table 1) and AML type (tAML/tMDS, sAML, de novo AML) on survival. Backward selection method was used to identify important predictors for a final model. Hazard ratios and 95% CI of all-cause mortality were based on the final Cox model.

Results:

We analyzed 219 patients with AML diagnosis. Of those 151 (69%) were de novo AML, 25 (11%) sAML and 43 (20%) tAML/tMDS, with mean age of 60.7, 70.7, and 69.7 years respectively. 88% of sAML and 72% of tAML/tMDS were ≥ 65 years compared to 50% of de novo AML patients (p=0.0009). More patients were in underweight/normal BMI (< 24.9) category of sAML (50%) compared to 36% of de novo AML and 21% of tAML/tMDS although this was not statistically significant (p=0.10). There were more females with tAML/tMDS (51%) compared to de novo AML (48%) and sAML (40%) (p=0.52). Most patients in all three groups of AML were white with 79% of de novo AML, 88% of sAML and 88% of tAML/tMDS.

Almost one-third of sAML (33%) and tAML/tMDS (38%) were in adverse risk category group with 24% of de novo AML in this category. Most of de novo (62%), sAML (67%) and tAML/tMDS (45%) were in intermediate risk category. There were 5 patients with tAML/tMDS in favorable risk category with zero sAML and 18 de novo AML in this category. 54% of our patients had ECOG performance score of 0-1. A majority of sAML (63%) had a positive smoking history compared to 47% of de novo AML and 44% of tAML/tMDS. Majority of patients in the three categories denied any alcohol use. Incidence of FLT-3 mutation was 23% in de novo AML, 0% in sAML and 9% in tAML/tMDS (p=0.0001). NPM1 mutation was present in 19% of de novo AML, none of sAML and 5% in tAML/tMDS (p = 0.0016). CEBPa mutation was present in 6% of de novo AML, 4% of sAML and 2% of tAML/tMDS.

Median survival was 18.5 months for de novo AML (95% CI 14.9- 23.7), 7.2 months for tAML/tMDS (95% CI 3.3- 11.5) and 7.0 months for sAML (95% CI 3.4-15.6). The median survival was longer among males, compared to females with de novo AML (23.2 months in males; 95% CI 18.3-37.1 vs. 14.6 months in females; 95% CI 10.3-19.0) compared to sAML (13.5 months in males; 95% CI 3.8-55 vs, 3.3 months in females; 95% CI 0.2-8.1) and tAML/tMDS (6.3 months in males; 95% CI 4.5-17.8 vs. 7.2 months in females; 95% CI 2.3-11.5) p=0.06. Patients with adverse risk category had a shorter median survival compared to those with favorable risk category, especially in tAML/tMDS but this was not significant.

After adjusting for age, risk category and FLT-3 status, the type of AML was not a significant predictor of survival. However, when compared with de novo AML, patients with sAML and tAML/tMDS appear to have a somewhat increased risk of death (HR 1.3; 95% CI 0.7-2.4 and 1.6; 95% CI 0.9, 2.7 respectively). Mortality was 4.5 (95% CI 1.8, 11.3) and 9.3 (95% CI 3.6, 24.0) times higher in patients with intermediate and adverse risk category respectively, when compared to the favorable risk group. Patients with positive FLT-3 had 1.6 times mortality compared to negative FLT-3 (95% CI 1.0, 2.7).

Conclusion:

We found that median survival was better in de novo AML compared to sAML and tAML/tMDS. There was no difference in survival between sAML and tAML/tMDS. Advancing age increases the odds of death across three types of AML. It is important to note that the effect of patient characteristics on survival is mostly consistent across the AML types and that once the major survival predictors are accounted for, the type of AML is no longer significant.

View largeDownload slide

View largeDownload slide

Close modal