A novel computer based risk prediction model for vocal cord palsy before thyroidectomy.

Area of Science:

• Medical research
• Surgical outcomes
• Machine learning applications in healthcare

Background:

• Thyroidectomy is performed for various conditions, but vocal cord palsy (VCP) is a significant complication.
• Temporary VCP incidence ranges from 3.4% to 7.2%, and permanent VCP from 0.2% to 0.9%.

Purpose of the Study:

• To develop a predictive model for identifying patients at risk of VCP before thyroidectomy.
• To enable personalized surgical approaches for high-risk individuals, thereby minimizing VCP occurrence.

Main Methods:

• Utilized a dataset of 1039 thyroidectomy patients from 2015-2018.
• Employed random forest classification and proposed sampling methods to build a clinical risk prediction model.

Main Results:

• A novel prediction model for VCP with 100% accuracy was successfully developed.
• The model effectively identifies patients at high risk for post-operative vocal cord palsy.

Conclusions:

• The developed clinical risk prediction model aids physicians in pre-operative patient assessment.
• This tool can help reduce the incidence of VCP following thyroidectomy.