A laboratory-less nomogram predicting survival rates for hospice patients with advanced cancer.

INTRODUCTION

Cancer is the leading cause of death globally(1). According to the WHO's 2020 Global Cancer Report(2), China represented 23.7% of new cancer cases and 30.2% of cancer-related deaths worldwide in 2020. From 2015 to 2020, cancer cases made up 18.4% of the global total. Recent statistics show that in China, malignant tumors accounted for 23.91% of all deaths, with both incidence and mortality rates on the rise. Hospice patients in China often lack the measurement of laboratory indicators, which poses difficulties in their survival prediction. This is because almost all current survival prediction models include laboratory parameters. This study established a lab-free prediction model with an accuracy of approximately 73%-75% to predict the survival rates of patients at 30 days, 45 days, and 60 days. An online version has also been developed for wide applications.

MATERIALS AND METHODS

We conducted a retrospective analysis of data from patients who received hospice care between January 2008 and December 2018. A total of 4,229 patients were divided into a training set (70%) and a test set (30%). The training group was used to develop the nomogram and a web-based calculator using the least absolute shrinkage and selection operator (LASSO) technique. The test group was used to validate the nomogram, using metrics such as the area under the receiver operating characteristic curve, calibration curve, and decision curve analysis.

RESULTS

Our analysis included 4,299 patients, with 3,163 in the training group and 1,066 in the test group. Using the LASSO algorithm, we identified eight predictors, namely quality of life, Karnofsky performance score, gender, pain duration, anorexia, abdominal distention, tachypnea, and edema. A nomogram with an online version was constructed to predict survival rates at 30, 45, and 60 days for hospice patients with advanced cancer. In the test set, the area under the curve (AUC) values were 0.7538, 0.7342, and 0.7324 for 30-day, 45-day, and 60-day survival, respectively. The nomogram demonstrated excellent calibration, and the decision curve analysis (DCA) showed a significant clinical net benefit.

CONCLUSION

This study developed a laboratory-free nomogram and a web-based calculator for accurately predicting survival in hospice patients with terminal cancer.