Risk, overall survival, and prognostic analysis of therapy-related acute myeloid leukemia in non-Hodgkin lymphoma survivors.

Abstract

e19037 Background: Therapy-related acute myeloid leukemia (tAML) is a serious complication in patients with Non-Hodgkin lymphoma (NHL) exposed to chemotherapy or radiation. tAML demonstrates high risk characteristics and poorer outcomes compared with de novo AML. We aimed to quantify the risk of tAML in NHL, determine factors associated with overall survival (OS) in tAML, and compare them with de novo AML. Methods: Patients with a histologic diagnosis of NHL and de novo AML from 2009 to 2018 were identified from the Surveillance, Epidemiology, and End Results (SEER) 18 database. AML that developed at least 1 year after the diagnosis of NHL was classified as tAML. Multiple primary standardized incidence ratio (SIR) sessions of the SEER*Stat software (version 8.3.9) were used to calculate SIR and absolute excess risk (AER) of tAML based on - age, sex, race, year of diagnosis, chemotherapy, radiotherapy, and interval from NHL diagnosis. The 95% confidence intervals (CI) and p-values were generated using multivariate Poisson regression model. OS of both tAML and de novo AML was assessed using Kaplan Meier curves and then compared using log rank test. The roles of various factors on OS in tAML and de novo AML were evaluated using multivariate cox proportional hazard regression. Results: A total of 373 patients with tAML (N for NHL = 301,903) and 23,360 patients with de novo AML were included in the analysis. More de novo AML cases were ≥70 years compared to tAML (41.1% vs 32.7%, p < 0.001). The risk of development of tAML was significantly higher in ages < 60 years compared to 60-69 years and ≥70 years (SIR 14.0, 95% confidence interval [CI] 11.79-16.51 vs SIR 4.87, 95% CI 4.00-5.86 vs SIR 2.80, 95% CI 2.32-3.34, p < 0.0001). Patients who received chemotherapy were more likely to develop tAML than the non- recipients (SIR 8.44, 95% CI 7.51-9.44 vs SIR 1.75, 95% CI 1.37-2.21, p < 0.0001). The risk of tAML was higher within 5 years of NHL diagnosis (SIR 5.05, 95% CI 4.49-5.67 vs SIR 4.39, 95% CI 3.49-5.45, p < 0.001). There was no statistically significant difference in SIR based on sex, race, receipt of radiotherapy, and year of diagnosis. The median OS and 5-year OS were- 8 months and 13.1% for tAML and 10 months and 27.6% for de novo AML. On multivariate analysis, tAML was not found to be an independent predictor of OS (HR 0.93, 95% CI 0.82-1.04, p = 0.21). Age ≥60 years (age 60-69 years: HR 1.53, 95% CI 1.08-2.15, p = 0.01, age ≥70 years: HR 1.93, 95% CI 1.40-2.66, p < 0.001) and no chemotherapy (HR 1.82, 95% CI 1.40-2.35, p < 0.001) were associated with poor OS in tAML subcategory. Conclusions: Our large population-based study shows increased risk of tAML within the first 5 years of NHL diagnosis, younger NHL survivors, and chemotherapy recipients. Older age and no chemotherapy predispose to dismal OS in tAML.