An online tool for predicting the prognosis of cancer patients with SARS-CoV-2 infection: a multi-center study.

ABSTRACT

PURPOSE: During the 2019 coronavirus disease (COVID-19) pandemic, oncologists face new challenges, and they need to adjust their cancer management strategies as soon as possible to reduce the risk of SARS-CoV-2 infection and tumor recurrence. However, data on cancer patients with SARS-CoV-2 infection remains scarce. METHODS: We conducted a retrospective study on 223 cancer patients with SARS-CoV-2 from 26 hospitals in Hubei, China. An individualized nomogram was constructed based on multivariate Cox analysis. Considering the convenience of the nomogram application, an online tool was also created. The predictive performance and clinical application of nomogram were verified by C-index, calibration curve and decision curve analysis (DCA). RESULTS: Among cancer patients with SARS-CoV-2, there were significant differences in clinical characteristics between survivors and non-survivors, and compared with patients with solid tumors including lung cancer, patients with hematological malignancies had a worse prognosis. Male, dyspnea, elevated PCT, increased heart rate, elevated D-dimers, and decreased platelets were risk factors for these patients. Furthermore, a good prediction performance of the online tool (dynamic nomogram https//covid-19-prediction-tool.shinyapps.io/DynNomapp/ ) was also fully demonstrated with the C-indexes of 0.841 (95% CI 0.782-0.900) in the development cohort and 0.780 (95% CI 0.678-0.882) in the validation cohort. CONCLUSION: Overall, cancer patients with SARS-CoV-2 had unique clinical features, and the established online tool could guide clinicians to predict the prognosis of patients during the COVID-19 epidemic and to develop more rational treatment strategies for cancer patients.