Results of an internal audit on the survival of patients with uterine sarcoma

Prognostic factors and survival of patients with uterine sarcoma: a German unicenter analysis

PURPOSE

Uterine sarcoma (US) as a histologically heterogeneous group of tumors is rare and associated with poor prognosis. Prognostic factors based on systematic data collection need to be identified to optimize patients' treatment.

METHODS

This unicenter, retrospective cohort study includes 57 patients treated at the University Hospital Freiburg, Germany between 1999 and 2017. Progression-free survival (PFS) and overall survival (OS) were calculated and visualized in Kaplan-Meier curves. Prognostic factors were identified using log-rank test and Cox regression.

RESULTS

44 Leiomyosarcoma (LMS), 7 low-grade endometrial stromal sarcoma (LG-ESS), 4 high-grade ESS and 2 undifferentiated US patients were identified. The median age at time of diagnosis was 51.0 years (range 18-83). The median follow-up time was 35 months. PFS for the total cohort was 14.0 (95%-Confidence-Interval (CI) 9.7-18.3) and OS 36.0 months (95%-CI 22.1-49.9). Tumor pathology was prognostically significant for OS with LG-ESS being the most favorable (mean OS 150.3 months). In the multivariate analysis, patients over 52 years showed a four times higher risk for tumor recurrence (hazard ratio (HR) 4.4; 95%-CI 1.5-12.9). Progesterone receptor negativity was associated with a two times higher risk for death (HR 2.8; 95%-CI 1.0-7.5). For LMS patients age ≥ 52 years (p = 0.04), clear surgical margins (p = 0.01), FIGO stage (p = 0.01) and no application of chemotherapy (p = 0.02) were statistically significant factors for OS.

CONCLUSION

Tumor histology, age at time of diagnosis and progesterone receptor status were prognostic factors for US. Unfavorable OS in LMS patients was associated with advanced FIGO stage, suboptimal cytoreduction and application of chemotherapy.

Development and Validation of a Personalized Survival Prediction Model for Uterine Adenosarcoma: A Population-Based Deep Learning Study

Background

The aim was to develop a personalized survival prediction deep learning model for adenosarcoma patients using the surveillance, epidemiology and end results (SEER) database.

Methods

A total of 797 uterine adenosarcoma patients were enrolled in this study. Duplicated and useless variables were excluded, and 15 variables were selected for further analyses, including age, grade, positive lymph nodes or not, marital status, race, tumor extension, stage, and surgery or not. We created our deep survival learning (DSL) model to manipulate the data, which was randomly split into a training set (n = 519, 65%), validation set (n = 143, 18%) and testing set (n = 143, 18%). The Cox proportional hazard (CPH) model was also included comparatively. Finally, personalized survival curves were plotted for randomly selected patients.

Results

The c-index for the CPH model was 0.726, and the Brier score was 0.17. For our deep survival learning model, we achieved a c-index of 0.774 and a Brier score of 0.14 in the external testing set. In addition, the limitations of the traditional staging system were revealed, and a personalized survival prediction system based on our risk scoring grouping was developed.

Conclusions

Our study developed a deep neural network model for adenosarcoma. The performance of this model was superior to that of the traditional Cox proportional hazard model. In addition, a personalized survival prediction system was developed based on our deep survival learning model, which provided more accurate prognostic information for adenosarcoma patients.