OBJECTIVE: Our aim was to evaluate the Japan Gastroenterological Endoscopy Society criteria for endoscopic submucosal resection of early gastric cancer (EGC) based on the experience in a Brazilian cancer center. METHODS: We included all patients who underwent endoscopic submucosal resection for gastric lesions between February 2009 and October 2016. Demographic data and information regarding the endoscopic resection, pathological report and follow-up were obtained. Statistical calculations were performed with Fisher’s exact test and chi-square tests, with 95% confidence intervals. RESULTS: In total, 76% of the 51 lesions were adenocarcinomas, 16% were adenomas, and 8% had other diagnoses. The average size was 19.9 mm (±11.7). The average procedure length was 113.9 minutes (±71.4). The complication rate was 21.3%, with only one patient who needed surgical treatment (transmural perforation). Among the adenocarcinomas, 39.5% met the classic criteria for curability, 31.6% met the expanded criteria and 28.9% met the criteria for noncurative resection. Analysis of the indication criteria and curability revealed differences among cases with “only-by-size” expanded criteria (64.28%), other expanded criteria (40%) and classic criteria (89.47%), with a p-value of 0.049. During follow-up (15.8 months; ±14.3), 86.1% of the EGC patients had no recurrence. When well-differentiated and poorly differentiated lesions or lesions included in the classic and expanded criteria were compared, there were no differences in recurrence. The noncurative group presented a higher recurrence rate than the classic group (p=0.014). CONCLUSION: These results suggest that the Japanese endoscopic submucosal resection criteria might be useful for endoscopic treatment of EGC in Western countries.

RESUMO CONTEXTO: A ressecção endoscópica da mucosa é tratamento aceito para o tratamento do câncer gástrico precoce em casos selecionados. Os critérios histopatológicos favoráveis à ressecção endoscópica curativa são adenocarcinomas intramucosos, bem diferenciados, com margens lateral e profunda livres, ausência de ulceração ou de embolização angiolinfática. Taxas de recorrência local próximas a 5% têm sido descritas mesmo quando se cumprem tais critérios. Por outro lado, a expressão antigênica por células tumorais tem sido relacionada com o comportamento biológico de diversos tumores. OBJETIVO: Avaliar se a imunoexpressão de mucinas, p53 e Ki-67 podem predizer a recorrência tumoral após mucosectomia endoscópica no câncer gástrico precoce, mesmo se critérios de cura histopatológicos forem atingidos. MÉTODOS: Vinte e dois pacientes com critérios de cura para ressecção endoscópica e sumetidos a mucosectomia foram selecionados. A recorrência local ocorreu em 5/22 (22,7%). O estudo imunohistoquímico foi realizado em 18 (81,8%) espécimens. Os pacientes foram divididos em grupos com e sem recorrência local. Foram comparados quanto a dados demográficos, endoscópicos, histológicos e fatores imunohistoquímicos para MUC2, MUC5A, CD10, p53, e Ki-67. RESULTADOS: A imunoexpressão de mucinas permitiu a reclassificação dos adenocarcinomas gástricos em intestinal (10), gástrico (2), e de fenótipo misto (4) e nulo (2). Os fenótipos mistos (positivos tanto para MUC2 quanto para MUC5A) foram encontrados em 80% dos casos no grupo de recorrência local, enquanto tipos intestinais (MUC2 positivo e MUC5A negativo) foram identificados em 76,9% dos casos sem recorrência (P=0,004). Os outros fatores observados não se relacionaram com a recorrência tumoral. CONCLUSÃO: O fenótipo misto do câncer gástrico precoce está associado a maior probabilidade de recorrência local após a mucosectomia.

ABSTRACT BACKGROUND: Endoscopic mucosal resection is still considered an accepted treatment for early gastric cancer for selected cases. Histopathologic criteria for curative endoscopic resection are intramucosal well-differentiated adenocarcinoma, lateral and deep margins free of tumor, no histological ulceration, and no venous or lymphatic embolism. A 5% local recurrence rate has been described even when all the above-mentioned criteria are met. On the other hand, antigen expression by tumoral cells has been related to the biological behavior of several tumors. OBJECTIVE: To evaluate whether early gastric cancer mucin immunoexpression, p53 and Ki-67, can predict recurrence after endoscopic mucosal resection, even when standard histopathologic criteria for curative measures have been attempted. METHODS: Twenty-two patients with early gastric cancer were considered to have been completely resected by endoscopic mucosal resection. Local recurrence occurred in 5/22 (22.7%). Immunohistochemical study was possible in 18 (81.8%) resected specimens. Patients were divided in two groups: those with and those without local recurrence. They were compared across demographic, endoscopic, histologic data, and immunohistochemical factors for MUC2, MUC5a, CD10, p53, and Ki-67. RESULTS: Mucin immunoexpression allowed a reclassification of gastric adenocarcinoma in intestinal (10), gastric (2), mixed (4), and null phenotypes (2). Mixed phenotype (positive for both MUC2 and MUC5a) was found in 80% of cases in the local recurrence group, while the intestinal type (positive MUC2 and negative MUC5a) was found in 76.9% of cases without local recurrence (P=0.004). Other observed features did not correlate with neoplastic recurrence. CONCLUSION: The mixed phenotype of early gastric adenocarcinoma is associated with a higher probability of local recurrence after endoscopic mucosal resection.

Racional: Sabe-se que os maiores riscos na ressecção endoscópica da submucosa são o sangramento e a perfuração. A criação de um coxim submucoso duradouro é essencial para a remoção da lesão completa e com segurança. Atualmente não se conhece um modelo experimental adequado para avaliação da durabilidade do coxim submucoso com diferentes soluções. Objetivo : Descrever um modelo experimental melhorado para avaliar soluções de injeção de submucosa. Métodos: Foram utilizados quatro porcos domésticos para avaliar dois tipos diferentes de soluções para injeção na submucosa gástrica. Após laparotomia mediana, a parede gástrica anterior foi aberta no sentido corpo-antro e sua superfície mucosa foi exposta por eversão da abertura gástrica. Dois tipos diferentes de solução (manitol a 10% e solução salina normal) foram injetados na submucosa da parede gástrica anterior de corpo distal. Todos os coxins submucosos foram injetados até que alcançassem o mesmo tamanho, padronizado como 1,0 cm de altura por 2,0 cm de diâmetro. Foram aplicados régua e compasso para garantir a acurácia das medidas. Resultados: O experimento foi completo nos quatro animais. Todos os coxins submucosos tinham o mesmo tamanho, medido com régua e compasso. Com o uso da solução de manitol, a duração média do coxim submucoso foi maior que a da solução salina: 20 e 22 min (média 21 min) vs 5 e 6 min (média 5,5 min). Conclusões: Este modelo experimental é simples e permite analisar duração, tamanho e efeito do coxim submucoso, tornando-se mais confiável que outros modelos que empregam estômagos ressecados de porcos ou imagens endoscópicas de modelos porcinos vivos.

Background: Endoscopic submucosal dissection carries an increased risk of bleeding and perforation. The creation of a long lasting submucosal cushion is essential for the safe and complete removal of the lesion. There is not a suitable experimental model for evaluation of the durability of the cushioning effect of different solutions. Aim: To describe an improved experimental model to evaluate submucosal injection solutions. Methods: A total of four domestic pigs were employed to evaluate two different submucosal fluid solutions in the gastric submucosa. After midline laparotomy, the anterior gastric wall was incised from the gastric body to the antrum and its mucosal surface was exposed by flipping inside out the incised gastric wall. Two different solutions (10% mannitol and normal saline) were injected in the submucosa of the anterior wall of the distal gastric body. All submucosal cushions were injected until they reach the same size, standardized as 1.0 cm in height and 2.0 cm in diameter. A caliper and a ruler were employed to guarantee accuracy of the measurements. Results: All four animal experiments were completed. All submucosal cushions had the exact same size measured with caliper and a ruler. By using the mannitol solution, the mean duration of the submucosal cushion was longer than the saline solution: 20 and 22 min (mean, 21 min) vs 5 and 6 min (mean, 5.5 min) Conclusions: This experimental model is simple and evaluate the duration, size, and effect of the submucosal cushion, making it more reliable than other models that employ resected porcine stomachs or endoscopic images in live porcine models.