The Combination of Computed Tomography Features and Circulating Tumor Cells Increases the Surgical Prediction of Visceral Pleural Invasion in Clinical T1N0M0 Lung Adenocarcinoma.

Jinghan Shi,Fei Li,Fujun Yang,Zhengwei Dong,Yan Jiang,Dania Nachira,Justyna Chalubinska-Fendler,Terence T. Sio,Yo Kawaguchi,Hiromitsu Takizawa,Xiao Song,Yang Hu,Liang Duan
DOI: https://doi.org/10.21037/tlcr-21-896
2021-01-01
Translational Lung Cancer Research
Abstract:Background: Visceral pleural invasion (VPI) is a clinical manifestation associated with a poor prognosis, and diagnosing it preoperatively is highly imperative for successful sublobar resection of these peripheral tumors. We evaluated the roles of computed tomography (CT) features and circulating tumor cells (CTCs) for improving VPI detection in patients with clinical T1N0M0 invasive lung adenocarcinoma. Methods: Three hundred and ninety-one patients were reviewed retrospectively in this study, of which 234 presented with a pleural tag or pleural contact on CT images. CTCs positive for the foliate receptors were enriched and analyzed prior to surgery. Logistic regression analyses were performed to assess the association of CT features and CTCs with VPI, and the receiver operating characteristic (ROC) curve was generated to compare the predictive power of these variables. Results: Patients mostly underwent either segmentectomies (18.9%) or lobectomies (79.0%). Only 49 of the 234 patients with pleural involvement on CT showed pathologically confirmed VPI. Multivariate logistic regression analysis revealed that CTC level >= 10.42 FU/3 mL was a significant VPI risk factor for invasive adenocarcinoma cases <= 30 mm [adjusted odds ratio (OR) =4.62, 95% confidence interval (CI): 2.05-10.44, P<0.001]. Based on CT features, subgroup analyses showed that the solid portion size was a statistically significant independent predictor of VPI for these peripheral nodules with pleural tag, while the solid portion length of the interface was an independent predictor of pleural contact. The receiver operating curve analyses showed that the combination of CTC and CT features were highly predictive of VPI [area under the curve (AUC) =0.921 for pleural contact and 0.862 for the pleural tag, respectively]. Conclusions: CTC, combined with CT features of pleural tag or pleural contact, could significantly improve VPI detection in invasive lung adenocarcinomas at clinical T1N0M0 stage prior to the patient's surgery.
What problem does this paper attempt to address?