Comparison of Bayesian analysis and PET-CT findings with respect to likelihood of malignancy in solid and subsolid solitary pulmonary nodules.
Objectives: To determine whether Bayesian Analysis (BA) predicts malignancy with similar specificity and sensitivity values in both subgroups of solitary pulmonary nodules (SPNs) and to compare PET-CT findings in solid and subsolid subgroups of PET-CT scanned SPNs. Study Design: An observational study. Setting: Department of Chest Diseases, Ankara Chest Diseases and Chest Surgery Training and Research Hospital. Period: February 2013 to February 2016. Materials and Methods: 156 patients with SPNs and whose histopathological diagnosis confirmed by fiberoptic bronchoscopy biopsy, transthoracic tru-cut biopsy or surgical methods were evaluated retrospectively. BA and PET-CT findings of nodules were evaluated. Results: 73.3% of male patients and 80% of females with subsolid SPN were diagnosed malignant. BA was statistically significantly found to be consistent with definitive diagnosis in Kappa compliance analysis in solid and nonsolid nodules (p <0.005 kappa = 0.604; p = 0.023 kappa = 0.358). The sensitivity of BA in solid and subsolid nodules was 63.6% and 80%, respectively, while their specificity was 93.4% and 73%, respectively. Positive predictive values (PPVs) were found to be 84% in solid nodules and 36% in subsolid nodules. Negative predictive values (NPVs) were calculated as 83% in solid nodules and 95% in subsolid nodules. In the patients with SPN included in our study, Kappa compliance analysis was performed separately in the solid and subsolid subgroups of patients who underwent PET-CT. When the cutoff value of Kappa compliance analysis in solid nodules was taken 2.5, definitive diagnosis and suvmax uptake were found to be statistically consistent (p <0.005 kappa = 0.638). When the cutoff value of Kappa compliance analysis in subsolid nodules was taken to be 2.5 as malignancy value, definitive diagnosis and suvmax uptake were found to be statistically consistent as in subgroup (p=0,011 kappa=0,399). When we took PET-CT suvmax cutoff value as 2.5, sensitivity uptake and specificity of PET were found in solid nodules, to be 68.4% and 93.1%, respectively. PPV was 86.7% and NPV was 82%. The sensitivity and specificity values of subsolid subgroup were 70% and 75% respectively, whereas the PPVs and NPVs were 50% and 87.5%, respectively. Conclusion: In subsolid SPNs, as in BA, PET-CT seems to be more reliable when used exclusively in malignancy exclusion. Although a significant suvmax cut-off value was determined for malignancy, the PPV of PET-CT is lower than that of solid SPNs.