Prediction of distant metastasis in patients with early ~ locally advanced uterine sarcoma

Abstract

Introduction Uterine sarcoma is difficult to diagnose before the surgery and sometimes it presumed as a benign mass. And uterine sarcoma shows high recurrence rate and low overall survival rate even in the early stage. It means that the current system cannot discriminate the patients who has worse prognosis and recurrent possibility. There is a need of a novel uterine sarcoma nomograms to predict overall survival and distant metastasis at the time of diagnosis for distinguishing the patients group who require further adjuvant treatment, and which is expected to attributed to improve prognosis.

Methods From January 1, 2007 to December 31, 2017, patients who had at least one operation for uterine sarcoma(leiomyosarcoma, low grade endometrial stromal sarcoma, high grade endometrial stromal sarcoma, malignant mixed Mullerian tumor) at out institution were identified. The patients’ medical record were analysis to predict 5-year overall survival and distant metastasis. Variables for nomogram were selected based on the statistical significance and clinical evidence. Final model included age, histology, mass size, metastatic site. Calibration curve was gained and bootstrap technic was used for internal validation

Results 169 patients were included in the study among the 189 patients. The median follow-up was 28.9 months (range, 0.6-166.3 months). The median progression free duration was 19 months (range, 0-129 months). Of the 169 patients, 80(51.5%) patients experienced recurrence or disease progression and 60(35.5%) patients were died for the disease. The estimated median overall survival was 99 months, with 95% confidence interval 66.7-131.3 months and estimated 5-year overall survival rate was 61%. For nomogram to predict 5-year survival, all of 169 patients were included and for the model to predict distant metastasis, 161 patients were analyzed except 8 patients who had residual lesion after the first-line treatment. After statistical analysis, nomograms were developed with 6 variables. Concordance probability of nomogram for 5-year overall survival was 0.838 and 0.839 by the bootstrap-corrected estimates and CP of nomogram for distant metastasis was 0.801 and 0.808 respectively.

Conclusions There nomograms can predict 5-year overall survival and distant metastasis more accurate than the current staging system. With these nomograms, a new patients group that who need adjuvant treatment could be discriminated. With modification though the further study and external validation, these nomograms could be an option to improve prognosis of uterine sarcoma