Abstract
1408
Objectives: The characterization of gastroesophageal junction carcinoma (GEJC) into subtypes based on genotype has evolved in the past decade. Many studies reported that the pathological classification, differentiation and the status of some pathological molecular indicators could determine the treatment and predict the prognosis of patients. Co-registering metabolic and morphologic data, 18F-FDG PET/CT has been widely used in the management of malignant tumors. In this study, we aim to investigate whether metabolic signature can predict pathological molecular indicators in patients with GEJC.
Methods: We retrospectively reviewed patients with histopathological diagnosis of GEJC between January 2010 and December 2015. A total of 66 patients (59 years, 63.97±7.5 years) with locally advanced gastroesophageal junction adenocarcinoma who were referred for preoperative 18F-FDG PET/CT scans were enrolled in this study. The maximum standardized uptake values (SUVmax), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) of the primary tumor were measured and calculated with the region of interest (ROI) technique. The predictive value of metabolic parameters in Lauren’s classification, histologic differentiation and the status of Ki-67, human epidermal growth factor receptor 2 (HER2), c-Met and epidermal growth factor receptor (EGFR) were evaluated in all of GEJC patients. Results: SUVmax was found significantly higher in the high Ki-67 index (≥75%) group (8.49±3.20 vs 6.97±2.15, P=0.045), the under curve area of receiver-operating characteristic (ROC) analysis was 0.664±0.081with an SUVmax cutoff of 7.27. Sensitivity and specificity for the prediction of Ki-67 expression were 68.2% and 70.0%. Higher SUVmax also refer to c-Met-negative status (P=0.036) with high sensitivity and specificity (85.7% and 61.7%, respectively). No significant correlation was found between metabolism parameters and the expression of Her2 and EGFR in GEJC, but we found close relationship between rates of Ki-67 and HER2 (P = 0.011), c-Met and EGFR expression (P = 0.047). Significant differences were observed between intestinal and no-intestinal (mixed and diffuse) adenocarcinoma in SUVmax (8.23±2.83 vs 6.29±2.41, P = 0.008), SUVmean (4.85±1.47 vs 3.93±1.22, P=0.017), MTV (24.96 cm3 vs 11.17cm3; P = 0.004) and TLG (97.38 cm3 vs 47.12 cm3, P = 0.005). SUVmax, MTV, and TLG of the moderately differentiated adenocarcinomas were significantly higher than those of the poorly differentiated ones, furthermore, correlation analysis showed that histologic types were closely related to SUVmax (r=0.33), MTV (r=0.47) and TLG (r=0.47). Conclusion: Metabolic signature can predict pathological molecular indicators, Lauren’s classification, and histologic differentiation in patients with GEJC. Metabolic imaging has the potential to become a useful complement for assessing the molecular profile of EGJC and for determining the therapeutic strategy.