Uterine carcinosarcoma: a retrospective clinical cohort analysis

Background

Carcinosarcomas are rare, heterogeneous tumours with a poor prognosis where carcinomatous and sarcomatous elements co-exist. There is no well-defined treatment pathway. Through analysis of University College London Hospital (UCLH) patients, we correlated survival with treatment and patient's characteristics to assess potential prognostic factors.

Methods

Women with uterine carcinosarcoma treated at UCLH from 2003 to 2014 were retrospectively identified and analyzed. Clinico-pathological data included poor prognostic factors and treatment. Kaplan-Meier Survival curves were generated using Stata version 14.1; survival differences were estimated using the long-rank test, p

Results

73 patients were included, median age was 69 and all underwent surgery. 69.8% were FIGO stage I/II, 62.2% had heterologous elements and 60.0% had lymph vascular space invasion (LVSI). Adjuvant chemotherapy and radiotherapy (RT) was received by 48% of patients, 27.4% had RT alone and 8.2%, chemotherapy alone. 16.4% patients declined adjuvant therapy. Median OS was 38.9 months and PFS 25.5 months. Two year OS and PFS were higher in lower stage (OS 68.3%, PFS 58.7%), homologous component (OS 72.7%, PFS 60.8%) and absence of LVSI (OS 75.3%, PFS 66.9%), compared to high stage (OS 48.9%, PFS 36.1%), heterologous component (OS 41.7%, PFS 35.7%) and presence of LVSI (OS 53.7%, PFS 42.5%). The OS and PFS for patients who received adjuvant chemotherapy and RT was 64.9% and 58.2% respectively, which was similar to survival rates of those who received RT alone (OS 68.4% and PFS 57.4%). Patients who didn't receive adjuvant treatment had a definitely poor survival rates (32.7% OS and 11.4% PFS).

Conclusions

Patients receiving RT had similar survival to those receiving RT and chemotherapy. Patients who didn't receive adjuvant treatment had the poorest survival. We recommend adjuvant therapy for carcinosarcomas if fitness allows and RT may be sufficient. LVSI had a statistically significant survival impact and is therefore a prognostic factor. The heterologous component had a trend towards significance.

Clinical trial identification

Legal entity responsible for the study

Carla Perna

Funding

University College of London Hospital

Disclosure

All authors have declared no conflicts of interest.