OBJECTIVE: We aimed to compare the prognoses of patients with pathologically true negative (P-TN) N2 and PET/CT false negative (FN) results in stage T1 non-small cell lung cancer (NSCLC). MATERIALS AND METHODS: Our institutional review board approved this retrospective study with a waiver of informed consent. The study included 184 patients (124 men and 60 women; mean age, 59 years) with stage T1 NSCLC who underwent an integrated PET/CT and surgery. After estimating the efficacy of PET/CT for detecting N2 disease, we determined and compared disease-free survival (DFS) rates in three groups (P-TN [n = 161], PET/CT FN [n = 12], and PET/CT true positive [TP, n = 11]) using the Kaplan-Meier analysis and log-rank test. RESULTS: Pathologic N2 disease was observed in 23 (12%) patients. PET/CT had an N2 disease detection sensitivity of 48% (11 of 23 patients), a specificity of 95% (153 of 161), and an accuracy of 89% (164 of 184). The 3-year DFS rate in the PET/CT FN group (31%, 95% confidence interval [CI]; 13.6-48.0%) was similar to that of the TP group (16%, 95% CI; 1.7-29.5%) (p = 0.649), but both groups had significantly shorter DFS rates than the P-TN group (77%, 95% CI; 72.0-81.2%) (p < 0.001). CONCLUSION: The PET/CT shows a high specificity, but low sensitivity for detecting N2 disease in stage T1 NSCLC. Patients with PET/CT FN N2 disease have survival rates similar to PET/CT TP N2 disease patients, which are both substantially shorter than the survival rate of P-TN patients.
Adult ; Aged ; Aged, 80 and over ; Carcinoma, Non-Small-Cell Lung/mortality/*radiography/*radionuclide imaging ; Disease-Free Survival ; Female ; Fluorodeoxyglucose F18/diagnostic use ; Humans ; Lung Neoplasms/mortality/*radiography/*radionuclide imaging ; Lymphatic Metastasis ; Male ; Mediastinum ; Middle Aged ; *Positron-Emission Tomography ; Prognosis ; Radiopharmaceuticals/diagnostic use ; Sensitivity and Specificity ; Survival Rate ; *Tomography, X-Ray Computed

OBJECTIVE: To identify CT and FDG-PET features associated with epidermal growth factor receptor (EGFR) protein overexpression, and to evaluate whether imaging features and EGFR-overexpression can help predict clinical outcome. MATERIALS AND METHODS: In 214 patients (M : F = 129 : 85; mean age, 63.2) who underwent curative resection of stage I non-small cell lung cancer, EGFR protein expression status was determined through immunohistochemical analysis. Imaging characteristics on CT and FDG-PET was assessed in relation to EGFR-overexpression. Imaging features and EGFR-overexpression were also evaluated for clinical outcome by using the Cox proportional hazards model. RESULTS: EGFR-overexpression was found in 51 patients (23.8%). It was significantly more frequent in tumors with an SUVmax > 5.0 (p < 0.0001), diameter > 2.43 cm (p < 0.0001), and with ground glass opacity < or = 50% (p = 0.0073). SUVmax > 5.0 (OR, 3.113; 95% CI, 1.375-7.049; p = 0.006) and diameter > 2.43 cm (OR, 2.799; 95% CI, 1.285-6.095; p = 0.010) were independent predictors of EGFR overexpression. Multivariate analysis showed that SUVmax > 4.0 (hazard ratio, 10.660; 95% CI, 1.370-82.966; p = 0.024), and the presence of cavitation within a tumor (hazard ratio, 3.122; 95% CI, 1.143-8.532; p = 0.026) were factors associated with poor prognosis. CONCLUSION: EGFR-overexpression is associated with high SUVmax, large tumor diameter, and small GGO proportion. CT and FDG-PET findings, which are closely related to EGFR overexpression, can be valuable in the prediction of clinical outcome.
Adult ; Aged ; Aged, 80 and over ; Carcinoma, Non-Small-Cell Lung/metabolism/mortality/*radiography/*radionuclide imaging ; Chi-Square Distribution ; Female ; Fluorodeoxyglucose F18/therapeutic use ; Humans ; Lung Neoplasms/metabolism/mortality/*radiography/*radionuclide imaging ; Male ; Middle Aged ; Neoplasm Staging ; Positron-Emission Tomography/*methods ; Proportional Hazards Models ; ROC Curve ; Radiopharmaceuticals/diagnostic use ; Receptor, Epidermal Growth Factor/*metabolism ; Statistics, Nonparametric ; Survival Rate ; Tomography, X-Ray Computed/*methods