Objective The aim of the present study was to estimate the risk of therapy-related acute myeloid leukemia (t-AML) in patients with gynecologic malignancies receiving chemotherapy using a population-based database.

Methods The National Cancer Institute’s Surveillance, Epidemiology, and End Results database was accessed, and a cohort of women diagnosed with a primary ovarian, uterine, or cervical malignancy between January 1, 1992, and December 31, 2014, who received chemotherapy was selected. Those who subsequently developed AML were identified. Standardized incidence ratio (SIR) with 95% confidence intervals (CIs) and excess risk (ER) per 10,000 persons were calculated. Median overall survival of women with t-AML was calculated following generation of Kaplan-Meier curves.

Results We identified 60,130 women who met the inclusion criteria; 56.4%, 19.4%, and 24.2% were diagnosed with ovarian, cervical, and uterine cancer, respectively. A total of 79 patients (0.13%) developed t-AML. The calculated SIR was 4.41 (95% CI, 3.49–5.50). For women with ovarian, cervical, and uterine cancer, the SIRs were 4.25 (95% CI, 3.13–5.66), 5.33 (95% CI, 2.92–8.95), and 4.26, (95% CI, 2.52–6.73), respectively. The highest risk was observed among women younger than 50 years (SIR, 11.69; 95% CI, 7.56–17.25). Median interval between gynecologic cancer and t-AML diagnosis was 40 months (range, 3–218 months), whereas median OS following the diagnosis of t-AML was 4 months (95% CI, 1.52–6.48 months).

Conclusions Therapy-related AML following chemotherapy treatment for a gynecologic malignancy is a very rare late treatment-related event associated with a poor prognosis.