Construction and validation of frailty risk prediction model in elderly patients with colorectal cancer

Early identification of risk factors and timely interventions can significantly reduce the incidence of frailty among elderly colorectal cancer patients, thereby improving their quality of life. This study aimed to develop and validate a frailty risk prediction model for elderly patients with colorectal cancer.

Three hundred two elderly hospitalized colorectal cancer inpatients (158 males; age range: 60-79 years; mean age: 68.79 ± 5.27 years) from the Gastrointestinal Surgery Department at the Second Affiliated Hospital of Guangzhou Medical University were enrolled, and 31 frailty risk indicators were measured, encompassing sociodemographic, lifestyle, health status, cognitive, pain, psychological, and biochemical factors. A binary logistic regression model was constructed to predict frailty and subsequently validated in an independent cohort of 181 patients. The Hosmer-Lemeshow (H-L) goodness-of-fit test was used to evaluate the model's fit, while calibration curves assessed the prediction model's accuracy. Clinical decision curve analysis was employed to examine the model's clinical utility. The predictive performance of the model was analyzed using the receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC). The optimal cutoff score for frailty assessment was determined at the point where the Youden index (sensitivity + specificity - 1) reached its maximum.

The prevalence of frailty among colorectal cancer patients was 43.4%. Logistic regression analysis identified age (66-70 years: OR = 1.946, 95% CI 0.814-4.651; 71-75 years: OR = 2.858, 95% CI 1.038-7.870; 76-80 years: OR = 7.371, 95% CI 1.933-18.107), Clock Drawing Task (CDT) performance (OR = 3.025, 95% CI 1.165-7.853), Pittsburgh Sleep Quality Index (PSQI) score (OR = 2.674, 95% CI 1.204-5.938), Self-rating Anxiety Scale (SAS) score (OR = 5.157, 95% CI 1.628-16.337), and Charlson Comorbidity Index (CCI) score (OR = 1.936, 95% CI 1.229-3.049) as significant predictors of frailty. The model demonstrated excellent predictive performance, with an AUC of 0.895 (95% CI 0.858-0.930), sensitivity of 0.817, specificity of 0.848, and an optimal cutoff value of 0.665 for predicting preoperative frailty in colorectal cancer patients.

Age, comorbidity, sleep quality, anxiety, and cognitive function were identified as independent predictors of frailty in elderly colorectal cancer patients. The developed logistic regression model exhibited strong predictive performance for frailty occurrence, highlighting its potential clinical utility.