Does the histologic subtype of uterine carcinosarcoma correlate with outcomes?

Abstract

<h3>Objectives:</h3> To evaluate uterine carcinosarcoma (UCS) by epithelial component subtype and compare clinicopathologic characteristics, recurrence patterns, and survival. <h3>Methods:</h3> We conducted a retrospective medical record review of patients diagnosed with UCS from January 2008 through March 2019 at one institution. Cases of UCS were classified based on carcinoma subtype-endometrioid versus non-endometrioid. Non-endometrioid included serous, clear cell, mixed cell, squamous, poorly differentiated and un-differentiated epithelial components. If a predominant histologic carcinoma subtype was not identified with initial review, the slides were classified by a pathologist. Recurrence sites were pelvic, nodal, or distant. Data are presented as percent, median (interquartile range) or mean ± standard deviation. We compared characteristics between UCS subtypes and calculated hazard ratios (HR) and 95% confidence intervals (CI) for overall survival. <h3>Results:</h3> We identified 77 patients with UCS, 18 (23%) with endometrioid subtype, 58 (75%) with non-endometrioid subtype, and one who could not be classified, as slides were unavailable. Median length of follow-up was 2 years[MRH1]; mean age at diagnosis was 70 years. The cohort was primarily White (66%) and African American (18%). Measured demographic variables, risk factors, symptoms at presentation, surgery, and adjuvant treatments were similar between epithelial subtypes. Most patients underwent surgery (96%) with nodal evaluation in 67%, 77% received chemotherapy, and 51% underwent radiation treatment. Forty percent of patients had mismatch repair protein immunohistochemistry, and 33% of patients with endometrioid subtype had loss of expression of at least one protein compared with 0% in the non-endometrioid group (p=0.02). Of patients with endometrioid subtype, 72% presented in stage I/II compared to only 46% of those with non-endometrioid subtype (p=0.06). The endometrioid subtype was more likely to recur in the pelvis (33%) than non-endometrioid (12%), but of the 32 (46%) patients whose disease recurred, there was no differences in nodal or distant recurrence by subtype. Median overall survival for stage I/II UCS was 6[MRH2] years; patients with stage I/II endometrioid subtype demonstrated poorer overall survival than stage I/II non-endometrioid subtype (HR 3.0, 95% CI 0.98-9.2). Median overall survival for stage III/IV was 2 years; data survival for stage III/IV did not meet assumptions to calculate an HR. <h3>Conclusions:</h3> In this single-institution review of patients with UCS, those with stage I/II endometroid subtype had worse overall survival than those with stage I/II non-endometrioid subtype. Deficiency in mismatch repair protein expression and pelvic site of recurrent disease was more common in endometrioid than non-endometrioid subtypes. These findings suggest that the epithelial subcomponent in UCS has implications for biologic behavior of the disease and therapeutic choice.