RESUMO
Aims. Identification of factors that can predict the subtypes of lung adenocarcinoma preoperatively is important for selecting the appropriate surgical procedure and for predicting postoperative survival. Methods. We retrospectively evaluated 87 patients with lung adenocarcinomas ≤30 mm. Results. Preoperative radiological findings, serum CEA level, serum microRNA-183 (miR-183) level, and tumour size differed significantly between patients with adenocarcinoma in situ (AIS) or minimally invasive adenocarcinoma (MIA) and those with invasive adenocarcinoma (IAC). Receiver operating characteristic curves and univariate analysis revealed that patients who were older than 57 years or had a pure solid nodule or a tumour with mixed ground-glass opacity (mGGO), a tumour >11 mm, a serum CEA level >2.12 ng/mL, or a serum miR-183 level >1.233 (2-ΔΔCt) were more likely to be diagnosed with IAC than with AIS or MIA. The combination of all five factors had an area under the curve of 0.946, with a sensitivity of 89.13% and a specificity of 95.12%. Moreover, patients with a cut-off value >0.499 for the five-factor combination had poor overall survival. Conclusions. The five-factor combination enables clinicians to distinguish AIS or MIA from IAC, thereby aiding in selecting the appropriate treatment, and to predict the prognosis of lung adenocarcinoma patients.