Lung cancer screening by low-dose computed tomography (LDCT) can improve mortality rates among high-risk individuals, especially adenocarcinoma cases with characteristically poor prognosis, although high false-positive rates have limited its clinical application. The objective of our study was to identify biomarkers for early-stage lung adenocarcinoma (ie, tumor diameter <2 cm) through extracellular vesicle long RNA (evlRNA) sequencing. High throughput evlRNA sequencing and support vector machine (SVM) identification of candidate diagnostic marker transcripts were performed using serum samples obtained before lung surgery. A total of 145 upregulated and 363 downregulated differential genes (P value <.05, fold change >1.5) were identified between lung adenocarcinoma (LUAD) patients and benign controls. An SVM model based on a 23-gene signature could distinguish EV samples of LUAD patients from those of control subjects with 86.49% sensitivity, 95.00% specificity and 92.31% accuracy in the training set and 93.75% sensitivity, 85.71% specificity and 88.24% accuracy in the validation set. A 17-gene signature was then identified that could distinguish AIS patient samples from those of MIA/IAD patients with 93.33% sensitivity, 98.00% specificity, and 96.25% accuracy in the trainingset and 83.33% sensitivity, 96.55% specificity, and 94.29% accuracy in the validation set. EvlRNAs in serum show considerable diagnostic value for screening LUAD patients with tumor sizes <2 cm in conjunction with LDCT, potentially reducing false positive rates while improving mortality rates.
Trial registration: ClinicalTrials.gov NCT03975504.
Keywords: diagnostic biomarkers; extracellular vesicle long RNA; lung cancer screening; lung nodules.
© 2022 UICC.