Effect of Adjuvant Radiation and Chemotherapy in Localized Uterine Carcinosarcoma: A National Cancer Database Analysis

Abstract

The optimal management of localized, resected uterine carcinosarcoma remains unclear. National guidelines list multiagent chemotherapy (MCT), external beam radiation (EBRT), vaginal cuff brachytherapy (VCB), a combination, or observation all as adjuvant treatment options. However, small randomized trials have not demonstrated an overall survival (OS) benefit for any adjuvant treatment. We hypothesized that adjuvant radiation therapy (RT) and MCT would be independently associated with improved OS in resected stage I-III uterine carcinosarcoma in a large, modern database.The National Cancer Database (NCDB) was queried for women diagnosed with uterine carcinosarcoma from 2004-2017 who had undergone total hysterectomy or greater revealing a pathologic stage of I-III. Women were grouped by stage and receipt of MCT, VCB, EBRT, or EBRT+VCB. OS was measured by Kaplan-Meier estimator. Cox proportional hazard modeling was done, adjusting for lymph node dissection (LND), tumor size, LVSI, margin status, age, comorbidity, treatment year, and several demographic factors. A propensity score weighted analysis was performed to see the effect of RT for women with Stage I, II, and III disease who received MCT and LND.13,899 women were analyzed (57% Stage I, 9% Stage II, 34% Stage III). As adjuvant treatment, 52% received MCT, 19% received EBRT, 14% received VCB, and 9% received EBRT+VCB. For the whole group, significant OS improvements were seen with adjuvant EBRT (HR 0.76), VCB (HR 0.75), EBRT+VCB (HR 0.63), and chemotherapy (HR 0.79) with Cox modeling (all P < 0.0001). However, when stages were analyzed individually, an OS benefit was seen for EBRT only in women with Stage III disease (HR for Stage I, II, and III 0.91 (P = 0.12), 0.95 (P = 0.61), 0.64 (P < 0.0001) respectively, compared to no RT). By contrast, an OS benefit for VCB was seen in all stages individually (HR for Stage I, II, and III 0.82 (P = 0.002), 0.66 (P = 0.002), 0.75 (P = 0.0005) respectively, compared to no RT). Similarly, an OS benefit for EBRT+VCB was seen in all stages individually (HR for Stage I, II, and III 0.72 (P = 0.0003), 0.69 (P = 0.002), 0.57 (P < 0.0001) respectively, compared to no RT). The OS benefit for chemotherapy was seen in Stages I and III (HR for Stage I, II, and III 0.82 (P = 0.0002), 0.84 (P = 0.10), 0.76 (P < 0.0001) respectively). Propensity score weighted analysis confirmed an additional OS benefit for women with stages I and III who received LND and MCT when they also received any RT.Adjuvant RT and MCT were both independently associated with improved OS in localized, resected carcinosarcoma in this NCDB database. However, RT was only associated with improved OS in stages I and II when VCB was part of the treatment regimen (VCB or EBRT+VCB), and not with just EBRT. For Stage III disease, each radiation modality (EBRT alone, EBRT+VCB, and VCB) was associated with improved OS. Randomized studies are warranted to further study the benefit of these treatments.K.C. Bylund: None. A.K. Chowdhry: Employee; University of Rochester, Massachusetts General Hospital. K. Nguyen: None. S. Geevarghese: None.