Objective To investigate the CT and ¹⁸FDG-fluorodeoxyglucose (FDG) PET/CT imaging features of pulmonary neuroendocrine tumors (PNETs) and to compare the accuracies of PNETs diagnosis based on PET/CT and the multimodal diagnostics of PET/CT combined with enhanced CT and high-resolution CT.

Methods The clinical, CT, and PET/CT data of 44 patients with PNETs diagnosed via histopathological examination in the Affiliated Hospital of Southwest Medical University from January 2010 to May 2019 were analyzed retrospectively. The patients comprised 34 males and 10 females aged 14–78 (57.3±10.0) years. All patients were divided into the carcinoid group (8 cases), the large-cell neuroendocrine carcinoma group (LCNEC, 15 cases), and the small-cell lung cancer group (SCLC, 21 cases). The CT and PET/CT features of the PNETs were investigated and analyzed. The diagnostic efficacy and critical value of maximum standardized uptake value (SUV_max) were analyzed and calculated by receiver operating characteristic (ROC) curve. By taking the results of histopathological examination as the gold standard, the accuracies of PNET diagnosis based on PET/CT and the multimodal diagnostics of PET/CT combined with enhanced CT and high-resolution CT were compared. Measurement data were compared by using one-way analysis of variance, and the least significant difference method was used to compare the two groups. The qualitative data were compared by applying χ² test or Fisher's exact probability method.

Results The age of the carcinoid group was lower than that of the LCNEC and SCLC groups (46.62±8.09 vs. 61.47±8.03 vs. 58.52±9.39), and the difference was statistically significant (F=6.186, P=0.004). However, no significant difference in sex and smoking history (χ²=1.220, 4.539; both P>0.05) was found. Significant differences were discovered in the location, lobulation, obstructive pneumonia or atelectasis, mediastinal lymph node metastasis, hilar lymph node metastasis, simultaneous mediastinal and hilar lymph node metastasis, distant metastasis, and vascular invasion in patients in the carcinoid, LCNEC, and SCLC groups (χ²=6.662–9.877, all P<0.05). However, no significant difference was found in maximum diameter, shape, density, enhancement degree, calcification, spiculation, necrosis and cystic degeneration, pleural effusion, bronchial invasion, pleural thickening (F=0.370, χ²=0.298–8.472, all P>0.05). The SUV_max of the LCNEC and SCLC groups was significantly higher than that of the carcinoid group (13.79±3.06 vs. 9.51±2.49 vs. 4.52±1.77), and the difference was statistically significant (F=32.43, P<0.01). For differentiating LCNEC from SCLC, the cutoff value of SUV_max was 12.25, the area under curve was 0.860 (95%CI: 0.729−0.991, P<0.01), the sensitivity was 80.00%, and the specificity was 81.00%. The accuracy of PNETs diagnosis based on PET/CT was 65.91%(29/44) and that of the multimodal diagnostics of PET/CT combined with enhanced CT and high-resolution CT was 87.80%(36/41). The difference was statistically significant (χ²=5.655, P=0.017).

Conclusions The CT and PET/CT manifestations of PNETs have certain characteristics, and the multimodal diagnostics of PET/CT combined with enhanced CT and high-resolution CT can improve the accuracy of diagnosing PNETs.