One hundred and four lesions in 103 consecutive patients were pathologically diagnosed as HGIEN of gastric mucosa based on biopsy samples obtained by EGD procedures in Shanghai Ruijin Hospital from January 2001 to March 2008. The charts, EGD reports and histologic sections were reviewed. Histopathologically, suspected stromal invasion was identified in 4 lesions. Endoscopically, lesions with an appearance strongly suggestive of advanced gastric cancer (i.e. Borrmann-type gastric cancer) were found in 21 patients, including fungating type in 1 patient, ulcerated in 7 and infiltrative ulcerated in 13 patients. These patients, excluded from the study, were later proved to have advanced gastric cancer. Seventy-nine superficial lesions were confirmed in 78 patients. Of these 78 patients, 6 refused interventional treatment while the remaining one was not a candidate for resection because of severe co-morbidities and poor overall condition. Seventy-one patients (59 males, 12 females, range 35-82 years, and a mean age of 60.5 ± 9.9 years) with 72 superficial neoplastic lesions were finally enrolled in the present study.

All patients gave their written consent to undergo endoscopic or surgical resection with the knowledge that the lesion may not be carcinomatous or infiltrative and the widely recognized guideline in treating GI neoplastic lesions with a biopsy diagnosis of HGIEN was not well established. Patients for endoscopic treatment were well informed about possibilities of incomplete resection, tumor recurrence, hemorrhage and other complications beforehand.

EGD procedures were carried out on each patient prior to resection using either EG 410HR or EG 590WR video endoscope (Fujinon, Saitama, Japan). Characteristics of the lesion, including its location, size, contour, and ulceration if present, as well as changes in mucosal rugae and endoscopic type, were documented. If an ulcer or ulcer scar was noted, descriptive terms and an estimation of the size of ulcer plague or scar was obtained. Lesions were classified according to the Paris endoscopic classification of superficial neoplastic lesions[11]. “Superficial” lesions are defined when endoscopic appearance suggests either a small cancer or a noninvasive neoplastic lesion[11]. Recognition of superficial gastric lesions was also based on our experience in the endoscopic diagnosis of early gastric cancer. Biopsy was obtained and all patients were diagnosed as HGIEN based on biopsy samples.

Abdominal US and enhanced CT scan were performed and distant metastasis was ruled out in all patients. All patients received en-bloc resection within two months after the initial diagnosis of HGIEN. Surgery was preferentially carried out given its advantage in curative resection irrespective of tumor size and invasion depth. Endoscopic resection was strictly limited to type 0-Ior 0-IIa lesions with fine deformability and no ulceration. Lesions less than 15 mm in size were subjected to EMR using a strip biopsy method[12] and otherwise to endoscopic submucosal dissection (ESD) using the insulated-tip knife[13].

The diagnosis of HGIEN was confirmed by two senior pathologists (Y-B. Z & Q. W). The IARC diagnostic criterion for HGIEN was applied[1], in which HGIEN is characterized by increasing architectural distortion with glandular crowding and prominent cellular atypia. An increased proliferative activity was present throughout the epithelium where no stromal invasion occurred. A lesion suspected of stromal invasion was ruled out of this series. Standardized sectioning of the resected specimen was carried out according to the guideline by JGCA[14]. Photograph of the specimen was taken before sectioning and a schematic map of sectioning was drawn. “True HGIEN” was diagnosed when each section of the en-bloc resected specimen met the criteria of HGIEN. Gastric carcinoma is diagnosed when the tumor invades to the lamina propria[1].

Endoscopic and clinicopathologic variables that might be predictive of gastric cancer were analyzed, including age, sex, history of melena or weight loss, time frame between the diagnosis of HGIEN and resection, endoscopic type (0-I/0-IIa, 0-IIc, 0-IIa + IIc/0-IIc + IIa, 0-IIc + III/0-III; mixed type), exact lesion size, location, mucosal rugae changes and presence of marked ulcer plaque or ulcer scar on the lesion. Continuous variables (age, time frame, exact lesion size) were compared using group t test when the variables were normally distributed and Wilcoxon test when a normal distribution was not observed. Categorical variables [sex, history of melena or weight loss, endoscopic type (0-I/0-IIa, 0-IIc, 0-IIa + IIc/0-IIc + IIa, 0-IIc + III/0-III; mixed type), lesion size with a cut-off value of 1.5 cm, location, mucosal ruga changes and presence of marked ulcer plaque or ulcer scar] were compared using χ² test or continuity-adjusted χ² test. Multiple logistic regression analysis was used to identify the factors independently associated with gastric cancer. P < 0.05 was considered statistically significant. All analyses were performed using the SAS software (SAS Institute, Cary, NC, USA).

Of the 71 patients, 69 were presented with symptoms while the other 2 were asymptomatic. The time of symptoms ranged 1 wk-11 years. The indications for EGD are presented in Table 1.

Of the 71 patients, 69 were proved to have isolated HGIEN lesions in the stomach, one was found to have two biopsy-proven HGIEN lesions both in antrum and in fundus, and the other one was found to have gastric polyposis involving antrum and corpus. Of the 72 lesions, 28 were located in antrum, 20 in angularis, 10 in corpus, 1 in fundus and 11 in cardia. Antrum, angularis and corpus were involved in another 2 lesions.

All patients received either surgical or endoscopic resection within two months of initial diagnosis. The time frame between initial diagnosis of HGIEN and resection was 19.2 ± 11.8 d. Open gastrectomy and laparoscopic-assisted gastrectomy were performed on 53 and 14 patients, respectively. Three patients received EMR and one received ESD. All surgical and endoscopic resections were proved curative with horizontal and vertical R0 margins.

Of the 72 lesions, 16 (22.2%) were pathologically confirmed as true HGIEN of gastric mucosa, 51 (70.8%) were early gastric cancer and 5 (6.9%) were advanced gastric cancer. Among the early gastric cancer, 34 were intramucosal and 17 were submucosal cancer.. Advanced gastric cancer in the 5 cases invaded the superficial layer of muscularis propria. Differentiated adenocarcinoma was found in 54/56 lesions (96.4%) while undifferentiated adenocarcinoma was found in 2/56 lesions (3.6%). A case of true HGIEN of gastric mucosa is presented in Figure 1.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 A case of true HGIEN of gastric mucosa. A: Endoscopic view of a sessile polypoid lesion (types 0-I) in the greater curvature of gastric corpus. The lesion is approximately 1.2 cm in size with a smooth contour. Biopsy pathology indicated high grade intraepithelial neoplasia; B: Low-power view of the en-bloc resected specimen, showing negative vertical and horizontal margins (HE staining, × 25); C: High-power view shows prominent cellular atypia and increased proliferative activity without stromal invasion, indicating high grade intraepithelial neoplasia (HE staining, × 200).

All lesions were divided into “true HGIEN” group (n = 16) and gastric cancer group (n = 56), according to their final diagnosis. A comparison of endoscopic and clinicopathological parameters of the two groups is shown in Table 2.

Twenty-three patients previously underwent EGD procedure within past 2 years. Of the 23 patients, one had normal EGD and 22 were identified to have 23 lesions. Targeted biopsy revealed high grade intraepithelial neoplasia in 17 cases, low-grade intraepithelial neoplasia in 2 and no neoplasia in 4 cases, respectively. Six out of the 23 lesions were finally proved to be true HGIEN of gastric mucosa.

Since endoscopy is an essential method in the diagnosis of gastrointestinal neoplasm and endoscopic appearance of a lesion plays an important role in predicting its characteristics and its invasion depth or metastasis[15], macroscopic findings under conventional endoscopy were analyzed in this study, including macroscopic type, lesion size, location, mucosal ruga changes (e.g. convergence, tapering, abruption, etc) and presence of ulcer plaque or ulcer scar, all of which are easily accessible in daily practices. Location was categorized as proximal (body, cardia and fundus of stomach) or distal (antrum and angularis). Presence of marked ulcer plaque/ulcer scar was defined as one or several ulcer plaques/ulcer scars at least 5 mm in size noted on the lesion.

The results of univariate analysis indicate that gastric cancer was associated with the size of lesion and the presence of marked ulcer plaque or ulcer scar (P < 0.05). Abruption of mucosal rugae was noted around 5 lesions, all of which were proved to be gastric cancer, so did the 8 lesions presenting with an ulcerated type (0-IIc + III or 0-III). However, no statistical significance was identified in these parameters.

Multivariate analysis revealed that a lesion larger than 1.5 in size (OR 18.400, 95% CI 3.364-100.653, P < 0.001) and the presence of marked ulcer plaque or ulcer scar (OR 10.000, 95% CI 1.062-94.111, P = 0.044) were associated with gastric cancer (Table 3). The value of ulcer plaque in the diagnosis of invasive cancer is demonstrated in Figure 2.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Presence of ulcer plaque is associated with gastric cancer. A: Retroflex view of a flat depressed lesion in the lesser curvature of cardia. The lesion is presented with a reddish area approximately 1.4 cm in size, scattered with irregular ulcer plaque; B: Forward view of the same lesion. The ulcer plaque can be clearly observed with a size larger than 5 mm (arrows); C: Low-power view of biopsy specimen shows irregular tubules with increased branching and architectural distortion. Prominent cellular atypia can be noted, indicating high grade intraepithelial neoplasia (HE staining, × 100); D: The entire lesion was removed by surgery which showed tumor invasion into lamina propria and partly the muscularis mucosae (HE staining, × 100). The final diagnosis was a type 0-IIc well-differentiated adenocarcinoma with muscularis mucosae invasion, T1 N0 M0.

A macroscopic criterion for true HGIEN of gastric mucosa was accordingly proposed as a lesion ≤ 1.5 cm in size, and absence of 5-odd mm ulcer plaque or ulcer scar on the premise that HGIEN was suggested by biopsy pathology. A total 20 lesions in this series met the above criteria, of which, 14 were finally proved to be “true HGIEN” of gastric mucosa, and 6 were proved to be intramucosal gastric cancer. The sensitivity, specificity and negative predictive value of multivariate analysis for predicting “true HGIEN” was 87.5%, 89.3% and 96.2%, respectively.

Intraepithelial neoplasia has been proposed by the WHO to replace dysplasia in gastrointestinal epithelial neoplastic changes. However, differential diagnosis between high grade intraepithelial neoplasia and invasive cancer is still confusing when the diagnosis is based on biopsy pathology.

Both intraepithelial neoplasia and early gastric cancer can be presented as superficial lesions. Conventional endoscopy provides direct observation of the lesion, yet the appearance of intraepithelial neoplasia and early gastric cancer resembles each other. It is unhelpful to macroscopically differentiate intraepithelial neoplasia from early cancer. Biopsy pathology may also fail to differentiate intraepithelial neoplasia from early cancer. This study focused on the macroscopic features of gastric lesions with a biopsy diagnosis of high grade intraepithelial neoplasia, suggesting that certain macroscopic features can help for better differentiation.

Conventional endoscopy observation has not been considered to play a role in differential diagnosis between intraepithelial neoplasia and early cancer. Likewise, gastric pit pattern observation either by magnifying endoscopy or by NBI lacks the accuracy to better characterize the lesion. Our study suggested that under typical conditions conventional endoscopy would be capable of differentiating between high grade intraepithelial neoplasia and gastric cancer.

The results of this study enhance the value of routine conventional endoscopy in the diagnosis of clinically confusing conditions. It could simplify the work-up of patients with a biopsy diagnosis of high grade intraepithelial neoplasia and save costs for patients and healthcare system.

Intraepithelial neoplasia means to replace dysplasia for precancerous status with cellular and structural atypia. High grade intraepithelial neoplasia denotes increasing architectural distortion with prominent cell atypia without invasion to lamina propria.

The authors investigated the macroscopic features of gastric superficial lesions with a biopsy diagnosis of high grade intraepithelial neoplasia and compared endoscopic findings with surgical results. It revealed that a lesion > 1.5 cm in size and the presence of ulcer/scar are associated with invasive cancer. The results are of interest and value for better patient work-up.