A predictive model for early death in elderly colorectal cancer patients: a population-based study.

PURPOSE

The purpose of this study is to determine what variables contribute to the early death of elderly colorectal cancer patients (ECRC) and to generate predictive nomograms for this population.

METHODS

This retrospective cohort analysis included elderly individuals (≥75 years old) diagnosed with colorectal cancer (CRC) from 2010-2015 in the Surveillance, Epidemiology, and End Result databases (SEER) databases. The external validation was conducted using a sample of the Chinese population obtained from the China-Japan Union Hospital of Jilin University. Logistic regression analyses were used to ascertain variables associated with early death and to develop nomograms. The nomograms were internally and externally validated with the help of the receiver operating characteristic curve (ROC), calibration curve, and decision curve analysis (DCA).

RESULTS

The SEER cohort consisted of 28,111 individuals, while the Chinese cohort contained 315 cases. Logistic regression analyses shown that race, marital status, tumor size, Grade, T stage, N stage, M stage, brain metastasis, liver metastasis, bone metastasis, surgery, chemotherapy, and radiotherapy were independent prognostic factors for all-cause and cancer-specific early death in ECRC patients; The variable of sex was only related to an increased risk of all-cause early death, whereas the factor of insurance status was solely associated with an increased risk of cancer-specific early death. Subsequently, two nomograms were devised to estimate the likelihood of all-cause and cancer-specific early death among individuals with ECRC. The nomograms exhibited robust predictive accuracy for predicting early death of ECRC patients, as evidenced by both internal and external validation.

CONCLUSION

We developed two easy-to-use nomograms to predicting the likelihood of early death in ECRC patients, which would contribute significantly to the improvement of clinical decision-making and the formulation of personalized treatment approaches for this particular population.