Despite significant improvements in the field of lung oncology, the overall survival rate of patients with lung cancer remains to be poor. Ineffectual detection of lung cancer in early stages is considered to be a key reason for these reduced survival rates. Multivariable molecular signatures, like miRNA, could have high rates of specificity and sensitivity in detecting lung cancer, though more studies on the miRNA signatures are needed. This study aims to evaluate the efficacy of blood-borne miRNAs in detecting lung cancer in symptomatic patients.
This multicenter, cohort study included a total of 3,046 patients with different types of lung cancers. Genome-wide miRNA profiles of blood samples of the participants were assessed. The primary outcomes of the study were the sensitivity and specificity of liquid biopsy using miRNA signatures for the detection of lung cancer.
Three classification scenarios were evaluated by splitting the samples into the following sets: one with a 15-miRNA signature and two 14-miRNA signatures to distinguish patients with lung cancer from patients with different lung diseases. The sensitivity and specificity of the three sets were 82.8% & 93.5%, 96.4% & 88.6%, and 76.3% & 97.5%, respectively. The accuracy of the three sets was above 90%.
The research concluded that the identified miRNA patterns could be used for minimally invasive lung cancer and biopsy tests.