Utility of a Clinical Decision Support System in Weight Loss Prediction After Head and Neck Cancer Radiotherapy

Area of Science:

• Oncology
• Medical Informatics
• Radiotherapy

Background:

• Significant weight loss is a common complication in head and neck cancer patients undergoing radiotherapy.
• Accurate prediction of weight loss is crucial for timely intervention and improved patient outcomes.
• Existing methods for predicting weight loss may not fully leverage clinical, dosimetric, and radiomics data.

Purpose of the Study:

• To evaluate the effectiveness of a clinical decision support system (CDSS) incorporating a weight loss prediction model.
• To assess whether the CDSS improves physicians' ability to predict significant weight loss in head and neck cancer patients.
• To determine if the CDSS's utility varies based on physician experience level.

Main Methods:

• A prediction model for significant weight loss was developed using logistic regression with clinical, dosimetric, and radiomics predictors.
• Four physicians evaluated 100 head and neck cancer patients twice: once without and once with the CDSS.
• Physicians predicted weight loss development and probability, with performance compared against actual outcomes.

Main Results:

• Physician accuracy in predicting significant weight loss increased from 0.58 to 0.63 with the CDSS (P = .06).
• The area under the curve (AUC) for predicting weight loss probability significantly improved from 0.56 to 0.69 with the CDSS (P < .05).
• Less-experienced physicians showed significantly greater improvement in performance when using the CDSS (P < .01).

Conclusions:

• A weight loss prediction-based CDSS shows potential to enhance physician decision-making in oncology.
• The CDSS appears particularly beneficial for improving the predictive accuracy of less-experienced clinicians.
• Further research with larger cohorts is recommended to confirm the utility of CDSS in managing patient weight loss during radiotherapy.