Development of a miRNA-Based Model for Lung Cancer Detection

Background:
Lung cancer is the leading cause of cancer-related mortality globally, with late-stage diagnoses contributing to poor survival rates. While lung cancer screening with low-dose computed tomography (LDCT) has proven effective in reducing mortality among heavy smokers, its limitations, including high false-positive rates and resource intensiveness, restrict widespread use. Liquid biopsy, particularly using microRNA (miRNA) biomarkers, offers a promising adjunct to current screening strategies. This study aimed to evaluate the predictive power of a panel of serum miRNA biomarkers for lung cancer detection.

Patients and Methods:
A case-control study was conducted at two tertiary hospitals, enrolling 82 lung cancer cases and 123 controls. We performed an extensive literature review to shortlist 25 candidate miRNAs, of which 16 showed a significant two-fold increase in expression compared to the controls. Machine learning techniques, including Random Forest, K-Nearest Neighbors, Neural Networks, and Support Vector Machines, were employed to identify the top six miRNAs. We then evaluated predictive models, incorporating these biomarkers with lung nodule characteristics on LDCT.

Results:
A prediction model utilising six miRNA biomarkers (mir-196a, mir-1268, mir-130b, mir-1290, mir-106b and mir-1246) alone achieved area under the curve (AUC) values ranging from 0.78 to 0.86, with sensitivities of 70–78% and specificities of 73–85%. Incorporating lung nodule size significantly improved model performance, yielding AUC values between 0.96 and 0.99, with sensitivities of 92–98% and specificities of 93–98%.

Conclusions:
A prediction model combining serum miRNA biomarkers and nodule size showed high predictive power for lung cancer. Integration of the prediction model into current lung cancer screening protocols may improve patient outcomes.