Development and Validation of a Nomogram for Predicting Specific Mortality Risk After Surgery for Locally Advanced...

Area of Science:

• Oncology
• Biostatistics
• Epidemiology

Background:

• Rectal cancer is a significant global health concern with high incidence rates.
• Locally advanced rectal cancer (LARC) requires precise prognostic tools for effective management.
• Existing predictive models may not fully capture cancer-specific mortality risks in LARC.

Purpose of the Study:

• To develop and validate a competing risk model for estimating cancer-specific mortality in LARC patients.
• To assess the predictive accuracy and calibration of the developed model.
• To identify factors influencing non-cancer mortality in LARC patients.

Main Methods:

• Retrospective cohort study using the SEER database (2010-2015) for training and validation.
• Inclusion of 7608 LARC patients (stages III-IV), with an external validation cohort from Tianjin People's Hospital.
• Development of a competing risk model, evaluated using ROC curve (AUC) and C-statistic for predictive accuracy and calibration curves.

Main Results:

• The developed competing risk model demonstrated strong predictive performance with AUC values ranging from 0.779 to 0.792 across different time points in both training and validation cohorts.
• The C-statistic confirmed the model's robust discriminative ability for cause-specific mortality.
• Older age, black race, and single marital status were associated with an increased risk of non-cancer death.

Conclusions:

• The developed competing risk model offers high accuracy and reliability for predicting cancer-specific mortality in LARC patients.
• The model serves as a valuable tool for physicians and patients in clinical decision-making and follow-up planning.
• Log odds of positive lymph nodes (LODDS) is not a sufficient standalone predictor for LARC-specific mortality.