A novel risk-scoring model for predicting cancer-specific mortality and instituting chemotherapy in elderly patients with esophageal cancer

Chemotherapy is indispensable to the treatment of esophageal cancer (EC), but its benefits in elderly patients remain unclear due to severe toxicity and side effects. The study aims to develop a risk-scoring model to predict which older patients are likely to benefit from chemotherapy. A total of 10,655 patients aged ≥70 years with pathologically diagnosed with EC were extracted from the Surveillance Epidemiology and End Results database, among whom 5,684 received chemotherapy. Sub-distribution hazard function were used to identify independent risk factors related to prognosis from the perspectives of the whole population, chemotherapy population, and non-chemotherapy population. Multivariate analyses based on the optimal model identified age, marital status, histology, tumor grade, T stage, N stage, M stage, lymph node positive rate, tumor size, surgery therapy, and radiotherapy as prognostic factors for elderly patients with EC. The concordance index of the best model was 0.772. X-tile software was utilized to classify the whole cohort into high-risk, medium-risk, and low-risk according to the total risk score. Sub-distribution hazard ratios was utilized to explore the effect of chemotherapy according to different risk stratifications. Chemotherapy did not impart a survival benefit to low-risk elderly patients with EC (Score ≤202), but did improve survival in patients in the medium-high risk groups (Score>202). Based on the results of the survival nomogram, a novel risk scoring system was constructed to screen for medium-high risk patients, for whom chemotherapy was more likely to bring additional benefit.