Gastric epithelial dysplasia: characteristics and long-term follow-up results after endoscopic resection according to morphological categorization

The increased use of esophagogastroduodenoscopy recently has resulted in an increase in the diagnosis of GED and the subsequent endoscopic treatments of this lesion. Although several studies have previously reported on the clinical and endoscopic characteristics according to the morphologic types [15-19], numerous questions still remain concerning the clinical significance of each type of GEDs, including the long-term follow-up outcomes. In addition, these previous studies included only relatively small numbers of cases. In the present study, we compared the clinicopathologic characteristics of GEDs according to the morphologic types and evaluated the long-term follow-up outcomes of each type (synchronous and metachronous lesions).

In the present study, we found that GED was more prevalent in men (male:female ratio, 2.2:1), which is similar to the results of previous studies [6,7]. However, there was no difference in the male:female ratio according to the morphologic types.

The natural history of H. pylori infection in the stomach is to go through a cascade of events, including non-atrophic gastritis, atrophic gastritis, intestinal metaplasia, dysplasia, and finally cancer [28,29]; and the presence of H. pylori infection has showed to be associated with an increased risk of progression to dysplasia or gastric cancer, with an odds ratio of 1.8 [30]. In the present study, the frequency of H. pylori infection was 79.6% in patients with GED, which is substantially higher than the frequency reported in the general population (59.6%) [31,32]. However, there was no difference in the frequency of H. pylori infection according to the morphologic types. Although this was not an epidemiological study, our results support the close relationships between H. pylori infection and GED, irrespective of morphologic types.

It has been well established that GEDs occur throughout the stomach, with a slight antral predominance, and that they can range in size from a few millimeters to several centimeters [33-35], which is similar to our results (antrum:body ratio, 1.4:1). However, in this study, we moreover demonstrated that GEDs show distinct endoscopic and histopathologic features according to the morphologic types. We found that adenomatous GEDs were more likely to occur in the gastric body and lesser curvature side of the stomach, whereas foveolar GEDs were mainly located in the gastric antrum and non-lesser curvature side of the stomach. In addition, foveolar GEDs were smaller, and were normal/red colored and showed a flat/depressed shape more frequently than adenomatous GEDs. From a histologic viewpoint, foveolar GEDs were more likely to show high-grade dysplasia than were adenomatous GEDs, which is consistent with the results of our previous study [19]. An association between macroscopic shape and histologic grade has been suggested in a previous study [36], with GEDs with a depressed shape showing higher proportions of reddish color and severe atypia or carcinoma compare to GEDs with non-depressed shapes. These findings concur with our findings that foveolar GEDs predominantly showed normal/reddish color, flat/depressed shape, and high-grade morphology. Furthermore, in the present study, of the 8 metachronous gastric cancers detected during the follow-up period, only one case was a signet ring cell carcinoma, and this occurred in a patient with a foveolar GED, which is consistent with the results of previous studies suggesting that the foveolar type is associated with high-grade histologic features and poorly-differentiated adenocarcinomas [3,19,22,23]. We speculate that these differences observed between GED types might reflect differences in the tumorigenesis of each GED type. However, further large-scale studies are required to clarify this.

In the present study, the complete resection rate in adenomatous type lesions was 78.4%, which was significantly lower than in foveolar and hybrid type lesions (84.5% and 92.0%, respectively). We hypothesize that the cause for this phenomenon may be because most GEDs with adenomatous type were low-grade and considered more discrete lesions owing to their whitish color and elevated/protruded morphology compared to those of the other 2 types, and that, consequently the endoscopists may have tended to perform endoscopic resection less carefully, such as marking closely around the lesion.

Treatment strategies for incomplete resection after endoscopic resection of GEDs have not been well established. Considering the indolent behavior of GED, we performed annual follow-up for incompletely resected GEDs. During the follow-up period, local recurrence occurred in only 2 of 59 lesions with incomplete resection. In these 2 lesions, additional argon plasma coagulation was performed, after which further recurrence did not occur. Local recurrence is generally caused by small remnants being left on the margins after endoscopic resection. However, in the present study, the recurrence rate in cases with incomplete resection was very low (3.4%). The reason for this might be the uncertainty of judging a cut margin because of the burning effect of electrosurgical devices on the residual tumor cells in positive margins. Our results suggest that close endoscopic surveillance without additional treatment may be an acceptable option for GEDs with incomplete resection.

A major concern in the management of GEDs is the inability to predict in which patient cancer will occur. Because it is often difficult to distinguish between the histologic grades from endoscopic findings, and because there are commonly discrepancies between the pre-endoscopic resection and post-endoscopic resection diagnoses [37-39], treatment modalities for GEDs ensuring accurate diagnosis and potentially curative resection, such as ESD or EMR, are needed. However, ESD or EMR can only remove gastric neoplasms in the part of the gastric mucosa containing the visible lesion with minimal surrounding normal tissue. Therefore, there is always a risk of occurrence of synchronous or metachronous gastric neoplasms in other sites, as GEDs are known to be indicators of an increased risk of synchronous adenocarcinoma elsewhere in the stomach [12,40].

One or more primary carcinomas can coexist at the time of diagnosis (synchronous), or develop consequently (metachronous), sometimes years after resection of the first GEDs. In the present study, gastric cancers were found on 21 out of 149 patients with available long-term follow-up information (14.1%). This rate is very high compared to the incidence of gastric cancer in the general population (0.3%) and individuals participating in health-screening programs (0.44%) [41,42]. Conversely, there were no cases of tumor-related deaths due to synchronous or metachronous cancer in this study. The prognosis in patients with gastric cancer depends largely on the stage at diagnosis [43], and in the present study, the median time interval from endoscopic resection of GED to metachronous cancer was approximately 34 months, and all synchronous and metachronous cancers were found at an early stage. Although the purpose of this study is to evaluate differences in the follow-up according to the three histologic types, we added an analysis based on grading of the lesions (low- and high-grade). There were no statistically significant differences in the incidence of synchronous and metachronous lesions. Despite numerous reports on the characteristics of synchronous and metachronous lesions after endoscopic treatment, the appropriate surveillance strategies after gastric endoscopic treatment remain unclear. On the basis of our results, we believe that the meticulous endoscopic follow-up in patients with GEDs after endoscopic resection may lead to the diagnosis of gastric cancer at an early stage, thus improving the survival rate, and regular endoscopic examinations and histologic controls (for example annually) are advisable.

Furthermore, about two-thirds of synchronous and metachronous GEDs were found at similar locations as the primary lesion in the present study. Hence, careful observation around the site of the primary lesion may be helpful for detecting synchronous or metachronous lesions and to reduce the risk of overlooking these lesions during follow-up endoscopy.

To our knowledge, this is the largest study to show morphologic characterization of GEDs to date, and the first study to reveal the long-term outcomes of GEDs after endoscopic resection according to morphologic types. However, our study has some limitations. First, there may have been a potential selection bias resulting from the retrospective nature of this single-center study. In our study, many patients were not included in the analysis of long-term outcomes because they did not undergo endoscopy at our hospital and hence their results were not available. The reasons for loss to follow-up in this study could be explained by the fact that after endoscopic treatment of GEDs, some patients choose to undergo follow-up endoscopy at other hospitals for various reasons such as the cost of endoscopy, the distance to the hospital, personal reasons, or the National Cancer Screening Program. In Korea, the National Cancer Screening Program provides endoscopy free-of-charge every 2 years for individuals older than 40 years. If the data for the patients lost to follow-up within 12 months of ESD had been included in our analysis, the study may have shown different results. Second, the follow-up period (median 37.3 months) was relatively short for a study of long-term results, and further large-scale prospective studies involving a greater number of patients and longer follow-up periods will be needed to obtain more exact information on the clinicopathologic features of each subtype of GEDs and the long–term outcomes of GEDs after endoscopic resection. Third, we classified GEDs as adenomatous, foveolar, or hybrid types according to their morphologic features only; rthese types were determined by a single expert gastrointestinal pathologist. However, histologic evaluation alone, even if performed by an expert pathologist, may fail to evaluate the morphologic type correctly. Therefore, it could have been more accurate and more objective if at least two expert pathologists had analyzed the morphologic type. Furthermore, in such situations, mucin expression patterns as determined by immunohistochemistry are helpful to obtain detailed information on the differentiation of GEDs. Although we did not evaluate the mucin expression patterns in the present study, we previously found that foveolar and hybrid types are more often positive for MUC5AC, whereas the adenomatous type is more often positive for CD10 [19]. Since there results have already been reported previously, we did not perform immunohistochemical analysis to investigate the concordance of histologic and immunohistochemical classification of GEDs in the present study.