A Survival Prognosis Prediction Model for Locally Advanced Laryngeal Cancer Based on Feature Selection Through
Jiangmiao Li1, Feng Zhao1, Junkun He1
1Department of Otolaryngology Head and Neck Surgery, The First Affiliated Hospital of Guangxi Medical University, Nanning, China.
Objective:
This study aimed to explore the high-risk factors associated with survival outcomes in patients with locally advanced laryngeal cancer (LALC) and to develop and validate a prognostic prediction model. This model aims to identify high-risk patients, assisting in the selection of appropriate treatment options for each individual.
Methods:
We included 283 patients who were diagnosed with LALC. The LASSO method, XGBoost algorithm, and random forests (RF) were used to screen essential features associated with the prognosis of LALC. A nomogram was then developed based on the COX regression model. Model validation was conducted internally using the bootstrap method. Receiver operating characteristic (ROC), the area under the ROC curve (AUC), the concordance index (C-index), and decision curve analysis (DCA) were used to evaluate model performance. Kaplan-Meier curves compared survival outcomes between different groups and the effectiveness of different treatment methods. All statistical analyses were performed using R statistical software (version 4.3.1).
Results:
A total of 484 patients with LALC were followed up. The mean follow-up time was (39.07 ± 30.85) months. The 1-, 3-, and 5-year survival rates of LALC were 79.13%, 62.82%, and 54.34%, respectively. After applying inclusion and exclusion criteria, 283 patients with LALC were finally included. Seven significant variables were identified, and the nomogram incorporating these predictors demonstrated favourable discrimination and calibration. Additionally, the nomogram successfully distinguished patients into low- and high-risk groups. The AUC values for predicting 1-, 3-, and 5-year OS rates were 0.852, 0.850, and 0.829. DCA indicated that the nomogram was clinically useful. The COX model, based on seven features, demonstrated superior performance in predicting 5-year survival outcomes compared to models based on AJCC 8th TNM stage, with NRI as 0.914 and IDI as 0.24.
Conclusions:
The Cox regression model developed based on seven independent factors, including 'Age', 'Treatment', 'Surgery', 'DAA', 'K+', 'LNR', and 'TCIS', can effectively predict OS in LALC patients. For LALC patients, especially those in the high-risk group, surgery or surgery combined with adjuvant radiotherapy may offer improved survival benefits.
More Related Videos
06:19Constructing and Visualizing Models using Mime-based Machine-learning Framework
Published on: July 22, 2025
04:09Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
Published on: October 10, 2018
