Advanced characterization and grading of invasive lung adenocarcinoma: integrative analysis with spectral CT and...

Lung cancer is the leading cause of cancer-related mortality globally. The novel classification system for invasive lung adenocarcinoma (LUAD), which integrates predominant histologic and high-grade patterns, has been shown to correlate with prognosis. This study aimed to evaluate and compare the discriminative ability of spectral computed tomography (CT), ¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT), and their combined model for identifying high-grade invasive non-mucinous adenocarcinoma (INMA) based on the 2021 World Health Organization (WHO)/International Association for the Study of Lung Cancer (IASLC) classification.

A total of 135 patients with 144 lung nodules who underwent preoperative spectral CT were evaluated retrospectively. Of these, 55 patients with 60 lung nodules underwent additional PET/CT imaging. CT morphological features, spectral CT and PET/CT parameters of the tumors were compared between the high-grade group (grade 3) and the low-grade group (grade 1 and 2). Univariate and multivariate analyses were performed using binary logistic regression with spectral CT and PET/CT parameters. The diagnostic efficiencies of spectral CT parameters, PET metabolism parameters, and a combination of CT features were computed by receiver operating characteristic (ROC) curve analysis. The discriminative power and calibration of the nomogram were evaluated.

There were significant differences in morphological CT (nodule attenuation), spectral CT parameters [iodine concentration of the lesion in arterial phase (ICL_a), normalized iodine concentration in arterial phase (NIC_a), slope of spectral Hounsfield unit curve in arterial phase (λHU_a), iodine concentration of the lesion in venous phase (ICL_v), normalized iodine concentration in venous phase (NIC_v), slope of spectral Hounsfield unit curve in venous phase (λHU_v)] and PET/CT parameters [maximum standardized uptake value (SUVmax), mean standardized uptake value (SUVmean), total lesion glycolysis (TLG)] between the grade 3 group and non-grade 3 group (P<0.05). According to the ROC analysis, the area under the curve (AUC) of the spectral CT, PET/CT, and comprehensive model was 0.930, 0.864, and 0.949, respectively. The performance of the comprehensive model was not superior to that of the spectral model [DeLong test, Z=0.819, P=0.413; integrated discrimination improvement (IDI) =0.042, 95% confidence interval (CI): -0.015 to 0.099, P=0.149]. Furthermore, our spectral CT and comprehensive model significantly outperformed the PET/CT model in predictive ability (IDI =0.148, 95% CI: 0.016-0.280, P<0.05; IDI =0.190, 95% CI: 0.085-0.295, P<0.001, respectively). The multivariate regression spectral CT model based on 144 lesions in 135 patients showed that ICL_a [odds ratio (OR) =0.005, 95% CI: 0.001-0.027, P<0.001] and solid nodule (OR =26.757, 95% CI: 6.843-104.623, P<0.001] were independent predictors for grade 3 LUAD. The nomogram had good calibration power.

The spectral CT model, including nodule attenuation type and ICL_a, showed diagnostic performance comparable to that of the comprehensive model and superior to PET/CT, showing strong potential for histopathological high-grade INMA diagnosis.