Clinical calculator predictive of chemotherapy benefit in stage 1A uterine papillary serous cancers

Abstract

Objective: Determine the utility of a clinical calculator to predict the benefit of chemotherapy in stage IA uterine papillary serous cancer (UPSC).

Patients and methods: Data were collected from NCDB from years 2010-2014. Based on demographic and surgical characteristics, a clinical score was developed using the random survival forest machine learning algorithm.

Results: Of 1,751 patients with stage IA UPSC, 1,012 (58%) received chemotherapy and 739 (42%) did not. Older age (HR 1.06), comorbidities (HR 1.31), larger tumor size (HR 1.27), lymphovascular invasion (HR 1.86), positive peritoneal cytology (HR 2.62), no pelvic lymph node dissection (HR 1.51), and no chemotherapy (HR 2.16) were associated with poorer prognosis. Compared to no chemotherapy, patients who underwent chemotherapy had a 5-year overall survival of 80% vs. 67%. To better delineate those who may derive more benefit from chemotherapy, we designed a clinical calculator capable of dividing patients into low, moderate, and high-risk groups with associated 5-year OS of 86%, 73%, and 53%, respectively. Using the calculator to assess the relative benefit of chemotherapy in each risk group, chemotherapy improved the 5-year OS in the high (42% to 64%; p < 0.001) and moderate risk group (66% to 79%; p < 0.001) but did not benefit the low risk group (84% to 87%; p = 0.29).

Conclusion: Our results suggest a clinical calculator is useful for counseling and personalizing chemotherapy for stage IA UPSC.

Keywords: Chemotherapy; Machine learning; Personalized medicine; Uterine papillary serous.