A novel model uses metabolic and volumetric parameters to predict less invasive lung adenocarcinomas

Abstract

OBJECTIVES: This study aimed to develop a prediction model for less invasive lesions of pathological Stage IA adenocarcinomas. METHODS: We retrospectively evaluated 121 lesions from 114 patients with pathological Stage IA adenocarcinoma who underwent surgery after fluorodeoxyglucose positron emission tomography and high-resolution computed tomography. Less invasive lesions were adenocarcinoma in situ and minimally invasive adenocarcinoma. The 3D parameter, solid tumour ratio, was the volume ratio of the solid part to the whole tumour. The 2D parameter was the consolidation-to-tumour ratio. The maximum standardized uptake value (SUVmax) in fluorodeoxyglucose positron emission tomography was the metabolic parameter. A volumetric analysis programme semiautomatically measured these 3 parameters. The cut-offvalues were 0.5, 0.125 and 1.0 for the consolidation-to-tumour ratio, solid tumour ratio and SUVmax, respectively. Multivariable logistic regression analysis was used to select the prediction model parameters. RESULTS: There were 34 (28.1%) less invasive lesions. A consolidation-to-tumour ratio < 0.5 was an insignificant predictive factor for less invasive lesions in the multivariable analysis. The prediction model had a total score of 3 points: 1 point for SUVmax < 1.0 and 2 points for the solid tumour ratio < 0.125. The area under the receiver operating characteristic curve in this model was 0.86 (95% confidence interval 0.78-0.94). The total score indicated 89.5% probability of possessing less invasive lesions. CONCLUSIONS: The solid tumour ratio and SUVmax effectively predicted less invasive lesions in early-stage lung adenocarcinomas. The prediction model generated by volumetric and metabolic parameters showed higher predictive power in this clinical setting.