Stage I uterine carcinosarcoma: Matched cohort analyses for lymphadenectomy, chemotherapy, and brachytherapy.

OBJECTIVE: To determine if lymphadenectomy, chemotherapy and radiotherapy are associated with survival benefit among women with stage I uterine carcinosarcoma.

METHODS: Women with stage I uterine carcinosarcoma (n=5614) were identified from the 1998-2013 National Cancer Data Base. Kaplan-Meier survival estimates and Cox proportional-hazards regression models were used to evaluate predictors of overall survival. Effects of these predictors were also estimated using propensity score matched analyses for lymphadenectomy, adjuvant chemotherapy, and radiotherapy.

RESULTS: 42.0% (2360/5614) of women in the cohort received no adjuvant radiation or chemotherapy. Black race and positive surgical margin status were associated with decreased survival by multivariable Cox regression. Among women with pathologically node-negative disease, the hazard of death increased 5% (4-7%) per each one centimeter increase in tumor size (P=1.9×10-10). From matched cohort analyses, omitting lymphadenectomy was associated with decreased median (interquartile range) survival: 45.2 (36.4-57.6) versus 73.9 (63.8-91.6) months, hazard ratio (HR) (95% CI) 1.38 (1.20-1.59), P=9.4×10-6. Hazard of death decreased by 3% (1-5%) for each five lymph nodes removed (P=0.01). Multiagent chemotherapy and vaginal brachytherapy were associated with decreased hazard of death (HR (95% CI) 0.62 (0.54-0.73), P=1.1×10-9 and HR (95% CI) 0.83 (0.70-0.97), P=0.02, respectively). Highest five-year survival was observed after brachytherapy and multiagent chemotherapy (74.1% (68.3-80.3%), P<2.0×10-16).

CONCLUSION: Lymphadenectomy to at least 15-20 removed nodes is associated with increased survival of women with node-negative uterine carcinosarcoma. Adjuvant "cuff and chemo" with vaginal brachytherapy and multiagent chemotherapy is associated with increased survival.