Additive Roles of F-18 FDG PET/CT for the Prediction of Silent Brain Metastasis in Patients with T1 and T2 Adenocarcinoma of Lung

Abstract

Objective: F-18 fluorodeoxyglucose (FDG) PET/CT is widely used for the diagnosis and staging in patients with non-small cell lung cancer (NSCLC). However, its relatively low sensitivity to detect brain metastasis is one of limitations of F-18 FDG PET/CT, as brain metastases are not uncommon even in patients with early stage NSCLC. The purpose of this study was to evaluate an additive role of F-18 FDG PET/CT to predict brain metastasis in patients with T1 and T2 adenocarcinoma of lung. Methods: A total of 395 neurologically asymptomatic lung adenocarcinoma patients with T1 and T2 on chest CT from 2011 to 2014 were enrolled, consecutively. All patients underwent chest CT, F-18 FDG PET/CT and brain magnetic resonance imaging (MRI) as a part of initial staging. TNM-staging (TCT, NCT and MCT) were determined on diagnostic CT scans according to the AJCC staging system (7th edition). Standardized uptake value (SUV) and metabolic-volumetric parameters of primary tumors (TPET) were obtained and N/M re-staging (NCT+PET, MCT+PET ) were determined on F-18 FDG PET/CT. Brain metastasis was determined on initial brain MRI and/or follow-up imaging studies up to 6 months. EGFR mutation and other clinical status were determined by medical record reviews. Receiver operating curve (ROC) analyses were performed to evaluate the optimal cutoff value for the continuous parameters. Logistic regression analyses were done to evaluate both clinical and PET metabolic parameters to predict brain metastases in lung adenocarcinoma. T, N and M factors by CT and FDG PET were compared with each other by NcNemar test. Result: Of 395 patients enrolled, 51 patients (13%) had brain metastasis on brain MRI (n=43) and/or follow-up imaging studies (n=8). Optimal cut-offs and area under the curves (AUCs) to predict brain metastasis for metabolic parameters of primary tumors (TPET) on F-18 FDG PET: SUVpeak, TLG, and SUVmean were 7.5

0.684, 40.5

0.718, 4.5

0.646, respectively. The rates of silent brain metastasis were 6.9%, 5.2%, 8.3% in SUVpeak, TLG, SUVmean of the cut-offs or less, whereas the rates were 20.5%, 20.4% and, 21.2% in above the cut-offs, respectively (P<0.001). The sensitivity and specificity of F-18 FDG PET for detecting mediastinal lymph node metastases were 52.1% and 81.1%, whereas those of chest CT were 23.3% and 92.1%, respectively (P<0.01 by McNemar test). F-18 FDG PET/CT detected distant metastases in 123 with extrathoracic metastasis except brain in 26 patients in addition to the thoracic metastasis (n=97) on chest CT. In univariate analysis, metabolic parameters of primary tumors (TPET) on FDG PET, that is, SUVpeak, TLG, and 2.5SUVmean had hazard ratios (HRs) of 3.69 (P<0.001), 4.71 (P<0.001), and 2.95 (P<0.001), respectively, while T-staging by CT (TCT) having hazard ratio (HR) of 1.98 (P=0.027). HRs for N-staging by CT (NCT) and FDG PET (NCT+PET) were 12.6 (P<0.001) and 13.5 (P<0.001), and HRs of intrathoracic M-staging by CT (MCT) and additional extrathoracic M-staging by FDG PET (MCT+PET) were 11.1 (P<0.001) and 57.4 (P<0.001), respectively. In multivariate analysis, NCT and MCT staging on CT (HR 5.41

P<0.001, HR1.98

P<0.001) and NCT+PET and MCT+PET staging (HR 3.1

P=0.056, HR37.3

P<0.001) had significant associations with the occurrence of brain metastasis, respectively. The rate of silent brain metastasis was 0.6% in M0 and N0 and 0.5% in Stage IIA or less by F-18 FDG PET/CT, and those were 0.9% and 1.3% by CT in patients with T1 and T2 lung adenocarcinoma on chest CT. Conclusion: Silent brain metastasis was not uncommon and its incidence rate was more than 1%, even in the early stage of lung adenocarcinoma in our data. Additive roles of F-18 FDG-PET/CT to predict silent brain metastasis in patients with T1 and T2 lung adenocarcinoma on chest CT were definitely achieved by detecting extrathoracic metastasis (MCT+PET). N-staging (NCT+PET) and higher metabolic activity (TPET) on F-18 FDG PET also had significant associations with silent brain metastasis, but their additive roles to predict brain metastasis had marginal statistical significances, in this retrospective study. NCCN guideline (ver 2.2018) optionally recommends brain MRI with contrast in stage IB of non-small cell lung cancer (NSCLS). Further study with a large population of stage I NSCLC and a prospective design will be needed to define the additive roles to guide optional use of brain MRI study in early stage of NSCLC.