Karp, J. E., Blackford, A., Visvanathan, K., Rugo, H. S., Moy, B., Goldstein, L. J., Goldstein, K. Stockerl, Neumayer, L., Langbaum, T. S., Hughes, M. E., Weeks, J. C., and Wolff, A. C.
BACKGROUND: Cooperative group trials and institutional series reported the risk of MDS and/or AML after adjuvant chemotherapy for early stage breast cancer (eg, NSABP 0.21%, Smith JCO 2003; ECOG 2197 0.5%, Goldstein ASCO 2012). We examined the incidence of MDS and/or AML in breast cancer survivors that are part of the NCCN Breast Cancer Outcomes Database. METHODS: The NCCN Database prospectively identifies all incident breast cancers diagnosed at participating cancer centers, and patients that continue their care at an NCCN site are followed annually with information collected on subsequent treatment, recurrence, and new cancers. Mortality is also ascertained for all diagnosed patients every 2 years through the National Death Index. Women with stages 1-3 breast cancer (BC) diagnosed between 1997- 2008 were included in this analysis. Univariate chi square analyses were used to compare characteristics between patients with BC and BC&MDS and/or AML. Time from BC to MDS and/or AML and 5-/10-year cumulative incidence rates were analyzed using proportional subdistribution hazards regression models that account for death as a competing risk to MDS and/or AML. RESULTS: Among 21563 stage 1-3 BC patients from 8 NCCN sites, 39 (0.18%) subsequently developed MDS and/or AML. Median overall follow-up was 4.7 years and median time from BC to MDS and/or AML was 2.8 years (range, 0.6 to 8.3 years). Cumulative MDS and/or AML incidence was 0.19% at 5y and 0.27% at 10y. Overall, cancer characteristics at BC diagnosis within BC&MDS and/or AML and within BC groups were similar (Caucasian 82%v82%, AA 8% v8%; ER+ 79%v75%, HER2+ 14%v21%, TNBC 15%v14%; contralateral BC 3%v2%; pancreas or ovarian cancer 0%v1%). 13191 had adjuvant chemo (75%v67%; anthracycline 89%v91%, AC-taxane 46%v35%, dose-dense 6%v5%). Trends observed included more patients in BC&MDS and/or AML group with prior radiation (85%v70%, p = 0.06) and older age at BC diagnosis (median age 59v54 years, p = 0.09). Risk for developing MDS and/or AML after BC (adjusted for age at BC and race) was increased in those receiving any chemotherapy versus no chemotherapy (HR 2.16, 95% CI 0.98-4.77, p = 0.06), any RT versus no RT (HR 2.44, 95% CI 1.01-5.88, p = 0.04), and if chemo/RT given versus just chemo or just RT (HR 2.2, 95% CI 1.05-4.59, p = 0.04). No increased risk was observed if a taxane was added to an anthracycline regimen (HR 1.27, 95% CI 0.57-2.81, p = 0.56). CONCLUSIONS: In a large national prospective registry across NCCN sites, risk of MDS and/or AML is increased after adjuvant chemotherapy and/or adjuvant radiation therapy for stages 1-3 breast cancer, especially if both are used. Specific population subsets will be further examined. These rates are similar to those observed in clinical trials, but may underestimate actual numbers due to patients lost to follow-up. In addition, an MDS diagnosis was routinely under-reported in population databases until recently. Strategies to limit risk of subsequent cancers in patients likely to survive early stage breast cancer should be considered at the time of initial diagnosis. Predictors of MDS and/or AML in breast cancer survivors are needed. [ABSTRACT FROM AUTHOR]