[Noninvasive prediction of EGFR gene mutation in lung adenocarcinoma using multiparametric CT perfusion imaging]

Determination of epidermal growth factor receptor (EGFR) gene mutation status is essential for molecular targeted therapy in patients with lung adenocarcinoma (LUAD). Currently, tumor tissue for genetic testing is commonly obtained through invasive procedures such as surgery or needle biopsy. This study aims to explore the potential associations between multiple quantitative parameters derived from computed tomography perfusion imaging (CTPI) and EGFR gene mutations in patients with LUAD, and to provide preliminary evidence for the noninvasive prediction of EGFR mutation status in LUAD.

Patients with LUAD pathologically confirmed by surgery or needle biopsy at The Third Xiangya Hospital, Central South University, were enrolled. All patients underwent pulmonary CTPI within 1 week before surgery and subsequently underwent EGFR gene testing. According to the EGFR gene testing results, the patients were divided into an EGFR-mutant group and an EGFR-wild-type group. Differences in CTPI-derived quantitative parameters between the 2 groups were statistically analyzed. Logistic regression analysis was performed to identify independent predictors of EGFR mutation and to construct a combined prediction model. Receiver operating characteristic (ROC) curves and the area under the curve (AUC) were used to evaluate the predictive performance of each independent predictor and the combined model.

A total of 82 patients were ultimately included, comprising 45 women and 37 men. EGFR mutations were detected in 46 patients, while 36 had wild-type EGFR. Blood volume (BV), time to peak (TTP), and capillary permeability surface (PS) were significantly higher in the EGFR-mutant group than in the wild-type group (5.38±1.57 vs 3.08±1.17, P<0.001; 28.02±5.07 vs 24.47±7.14, P=0.014; 18.90±7.24 vs 12.11±6.11, P<0.001, respectively). No statistically significant differences were observed between the 2 groups in blood flow (BF), mean transit time (MTT), time to maximum of the tissue residue function (Tmax), positive enhancement integral (PEI), or mean slope of increase (MSI) (all P>0.05). Among the parameters showing significant between-group differences, BV (P<0.001) and PS (P=0.002) were independent predictors of EGFR mutation. ROC curve analysis revealed that the AUCs of BV and PS were 0.885 and 0.760, respectively, with corresponding cutoff values of 3.81 mL/100 g and 16.20 mL/(min·100 g). The combined CTPI prediction model incorporating BV and PS achieved an AUC of 0.923, which was higher than that of either parameter alone (both P<0.05).

The quantitative CTPI parameters BV and PS are independent predictors of EGFR mutation in patients with LUAD. The combined CTPI model incorporating BV and PS demonstrated good predictive potential in this cohort, although further validation in large, multicenter studies is required.