Objective To comparatively analyze the ¹⁸F-fluorodeoxyglucose (FDG) PET metabolic characteristics and multislice spiral CT imaging features of pulmonary invasive adenocarcinoma appearing as ground-glass nodules (GGN) with different risk levels and to evaluate the value of ¹⁸F-FDG PET/CT in the diagnosis of risk levels of GGN.

Methods Retrospective analysis was performed on 143 patients (54 males, 89 females, 30−79(60.2±8.9) years old) with pulmonary invasive adenocarcinoma confirmed by histopathological examination or follow-up. All patients underwent ¹⁸F-FDG PET/CT whole body imaging (including 50 cases of ¹⁸F-FDG PET/CT dual-phase imaging) and surgical resection of solitary GGN of the lung. In accordance with the adenocarcinoma growth pattern, the patients were further divided into two groups. Patients with lesions with lepidic predominant adenocarcinoma and/or acinar predominant adenocarcinoma and/or papillary predominant adenocarcinoma were assigned to the low-risk group, and those with lesions with solid predominant adenocarcinoma and/or micropapillary predominant adenocarcinoma were classified into the high-risk group. The recorded data included gender, age, lesion location, size, density, maximum standardized uptake value (SUV_max), mean standardized uptake value (SUV_mean), retention index (RI) in dual phase imaging, the SUV_max ratio of tumor to contralateral normal lung background (T/N), the rate of change in the ratio of tumor to contralateral normal lung background based on the SUV_max (ΔT/N_max), lobulation sign, spiculation sign, vocule sign, air bronchgram, pleural indentation, and vascular convergence sign. Qualitative factors were analyzed by using independent-sample t test, whereas quantitative variables were analyzed by using χ² test. Multivariate unconditional Logistic regression analysis was utilized to test the correlation factors with statistical differences before treatment. Receiver operating characteristic (ROC) curve analysis was performed in accordance with the Logistic regression analysis results.

Results In 143 patients, lesion size ((14.33±4.18) mm vs. (17.61±4.48) mm), SUV_max (1.32±1.07 vs. 2.00±1.25), SUV_mean (1.07±0.85 vs. 1.66±1.11), RI (0.01±0.36 vs. 0.20±0.07), lobulation (76.1%(89/117) vs. 92.3%(24/27)), and pleural indentation (39.3%(46/117) vs. 69.2%(18/26)) showed statistically significant differences between low-risk group (117 cases) and high-risk group (26 cases) (t=−3.242 to −2.392; χ²=4.773, 6.766; all P<0.05). In 50 patients underwent ¹⁸F-FDG PET/CT dual-phase imaging, delayed imaging SUV_max (1.18±0.63 vs. 2.85±1.82), delayed imaging SUV_mean (0.92±0.43 vs. 2.72±1.69), delayed imaging T/N (2.55±1.33 vs. 5.84±3.83) showed statistically significant differences between low-risk group (40 cases) and high-risk group (10 cases) (t=−2.867, −3.359, −2.678; all P<0.05). Among these factors, SUV_mean, lesion size, and pleural indentation were the independent influencing factors for differentiating the two groups. When the value of SUV_max was 1.625, the area under the ROC curve was 0.699. The sensitivity, specificity, and accuracy of differentiating the two groups were 57.7%(15/26), 78.6%(92/117), and 74.8%(107/143), respectively. When the value of SUV_mean was 0.845, the area under the ROC curve was 0.698. The sensitivity, specificity, and accuracy of differentiating the two groups were 80.8%(21/26), 43.6%(51/117), and 50.3%(72/143), respectively. When the lesion size was 13.765 mm, the area under the ROC curve was 0.716, and the sensitivity, specificity, and accuracy of differentiating the two groups were 80.8%(21/26), 54.7%(64/117), and 59.4%(85/143), respectively. The combined diagnosis with SUV_max+SUV_mean+lesion size+pleural indentation+lobulation sign has the highest efficiency in differentiating the two groups compared with single diagnosis.

Conclusion In the diagnosis of pulmonary invasive adenocarcinoma appearing as GGN, ¹⁸F-FDG PET/CT contributes to risk levels.