Protein-Based Risk Model Versus Existing Lung Cancer Prediction Tools

By Kaitlyn Kosko - Last Updated: December 26, 2023

A protein-based risk model is superior in predicting lung cancer compared with two existing prediction tools, according to Xiaoshuang Feng, PhD, and colleagues.

Advertisement

The case-control study, published in the Journal of the National Cancer Institute, evaluated six prospective cohorts that included 624 patients with lung cancer who were then matched with a cancer-free participant. They were matched carefully by the following criteria: study center, gender, date of blood collection, date of birth, smoking status, quit years in two categories for former smokers (<10 and ³10 years since quitting), and intensity in two categories for current smokers (<15 and ³15 cigarettes smoked per day).

From these pairs, four cohorts were designed to select proteins and train the risk model. Then, two cohorts that included 154 pairs were used to compare the risk-discriminatory performance of the protein-based model with the Early Cancer Detection Test (EarlyCDT)-Lung model and the PLCOm2012, a smoking-based risk model.

The findings showed good discrimination between the lung cancer participants and smoking-matched cancer-free participants, with an overall area under the curve of 0.75 compared with 0.64 for the PLCOm2012 model.

In addition, the protein-based model had a greater sensitivity at 49% compared with the PLCOm2012 (30%) and EarlyCDT-Lung (14%) models. All three models had 86% specificity.

Since the prediction model demonstrated positive results, the authors hope for further evaluation in a larger sample size.

Advertisement