Definition of carcinoma at the EGJ and Barrett’s esophagus
We defined cancer at the EGJ according to the Japanese classification . In this classification, the area extending 2 cm above to 2 cm below the EGJ is designated as the EGJ area. Tumors having their epicenter in this area are designated as EGJ carcinomas irrespective of histological type. The location of an EGJ carcinoma is described using the symbols E (proximal 2 cm segment) and G (distal 2 cm segment), with the dominant area of invasion described ﬁrst; i.e., E, EG, E = G (both areas equally involved), GE, or G. Barrett’s epithelium was endoscopically diagnosed when columnar epithelium was
continuously observed from the stomach to the distal side of the esophagus. In the United States of America and most European countries, the diagnosis of Barrett’s epithelium requires histologically conﬁrmed intestinal metaplasia [21-22]. However, in England and Asian countries, including Japan, histological demonstration of goblet cell is not required [23-26]. In this study, we defined Barrett’s epithelium as the continuous columnar epithelium from the stomach with or without intestinal metaplasia. The presence of circular Barrett mucosa extending longitudinally for 3 cm or more in length was classified as long segment Barrett esophagus (LSBE). On the other hand, the presence of circular Barrett mucosa less than 3 cm in length or the presence of non-circular Barrett mucosa was designated as short segment Barrett esophagus (SSBE) .
Enzyme-activatable fluorescent targeting agent
The EP-HMRG probe was purchased from Goryo Chemical (Sapporo, Japan), resuspended in 10 mM dimethyl sulfoxide (Sigma-Aldrich, St. Louis, MO, USA) and then stored at -80°C. Before use, the EP-HMRG suspension was thawed to room temperature and diluted to 100 μM with phosphate-buffered saline (PBS, Life Technologies, Carlsbad, CA, USA).
This study prospectively reviewed early EGJ adenocarcinoma resected by ESD in 23 consecutive patients at five hospitals between May, 2016 and June, 2018. All ESD procedures were performed by experienced endoscopists. We included 21 cases that met the definition of EGJ carcinoma, and excluded the cases in which complete en bloc resection was not performed, as well as cases whose specimens were too damaged for histological investigation.
The resected specimen was immediately extended on a black rubber mat and fixed with pins, and then 100 μM EP-HMRG was sprayed onto the specimen. Fluorescence imaging was performed using a handheld fluorescence imaging system (Discovery; INDEC Medical Systems, Santa Clara, CA, USA) that captured white-light images and fluorescence images with 450–490 nm blue excitation light. The fluorescence images were recorded every minute for 10 minutes after the EP-HMRG administration. Subsequently, the specimens were washed with PBS and observed using an endoscope (H290Z, Olympus, Tokyo, Japan) under white light.
The fluorescence intensities were measured with ImageJ software (National Institutes of Health, Rockville, MD, USA). We set regions of interest (ROIs) at the area with the most fluorescence signal in the tumor lesion and in the non-tumor region adjacent to the tumor lesion. The mean fluorescence intensity of each ROI was measured as pixel intensity values ranging from 0 to 255, and the contrast-to-background ratio (CBR) was also measured.
The Ethical Review Committee of each hospital approved this ex vivo clinical study protocol. All patients provided informed consent to participate in this study.
Specimens were fixed in 40 g/L formaldehyde saline, embedded in paraffin, and cut into 5-μm sections. Tissue sections were stained with hematoxylin and eosin and then microscopically examined for the histological type, tumor size, depth of invasion, lymphovascular invasion, and resected margin by an experienced pathologist (KCH), according to the World Health Organization classification. Immunohistochemical analysis of DPP-IV expression was performed using an anti-DPP-IV antibody (Novus Biologicals, Littleton, CO, USA). The subtype of intestinal metaplasia in the non-tumor region was determined using the MUC5AC (Agilent, Santa Clara, CA, USA), MUC6 (Abcam, Cambridge, UK), MUC2 (Spring Bioscience, Pleasanton, CA, USA), and CD10 (Agilent) expression patterns. MUC5AC and MUC6 are markers of the gastric phenotype, whereas MUC2 and CD10 are markers of the intestinal phenotype. We defined the complete type as decreased expression of gastric mucin (MUC5AC or MUC6) and co-expression of MUC2 and CD10. The incomplete type was deﬁned as the expression of MUC5AC or MUC6 and MUC2 .
Receiver operating characteristic (ROC) curves were used to determine the sensitivity, specificity, and accuracy. All analyses were performed using GraphPad Prism version 6 (GraphPad Software, San Diego, CA, USA).