Endoscopy 2013; 45(09): 689-690
DOI: 10.1055/s-0033-1344616
© Georg Thieme Verlag KG Stuttgart · New York

Expanded criteria for endoscopic treatment of early gastric cancer: safe in the long term if feasible in the short term!

Pedro Barreiro
1  Department of Gastroenterology, Egas Moniz Hospital, Portugal
2  Department of Gastroenterology, Portuguese Oncology Institute of Porto, Portugal
Mário Dinis-Ribeiro
2  Department of Gastroenterology, Portuguese Oncology Institute of Porto, Portugal
3  Centre for Research in Health Technologies and Information Systems (CINTESIS), Medical Faculty, Porto, Portugal
› Author Affiliations
Further Information

Publication History

Publication Date:
29 August 2013 (online)

Endoscopic treatment of malignant superficial gastric lesions has undergone remarkable development in recent decades. Standard criteria for curative endoscopic treatment of these lesions are based on post-resection data and relate to complete resection (no lateral or deep margin involvement) of intramucosal well-differentiated tumors less than 2 cm in size, without ulceration and without lymphovascular invasion [1] [2]. However, the development of endoscopic submucosal dissection (ESD) and the feasibility of en bloc resection regardless of the size and shape of the lesions have led to a broadening of these criteria to the so-called “expanded criteria” – large lesions, ulcerated, undifferentiated or with minimal submucosal invasion (less than 500 μm). Nevertheless, according to the 2010 Japanese guidelines, the use of these criteria should be regarded as “investigational treatment” [1], while the European Society for Medical Oncology (ESMO) does not yet recognize them as safe [2].

The safety of endoscopic treatment of early lesions meeting the expanded criteria must rest on two pillars: data supporting the feasibility of successful resection of these lesions (short-term results), and data showing nil or negligible risk of lymph node invasion in these lesions and consequent favorable long-term outcomes.

Park et al. [3] and Ahn et al. [4] have shown excellent results in terms of the technical efficacy of resection by ESD of lesions meeting the expanded criteria, i. e., they were larger and/or ulcerated. Both groups of authors showed en bloc resection and complete resection (R0) rates higher than 85 %, with low rates of complications. These results, although slightly worse than those obtained after resection of lesions meeting the standard criteria, support the contention that experienced operators can obtain excellent short-term outcomes in the treatment of lesions that are technically more challenging.

On the other hand, Gotoda et al. [5] retrospectively analyzed more than 5000 surgical cases of gastrectomy for gastric cancer and showed no cases of nodal invasion in lesions that met the expanded criteria – but, importantly, they estimated a risk of lymph node involvement of up to 2.5 % and 2.6 % (upper limit of 95 % confidence interval) in lesions with minimal submucosal invasion and in undifferentiated carcinoma, respectively, so the risk should not be taken as zero. Moreover, Jee et al. [6] described the presence of lymph node metastasis in 2.8 % of patients with lesions that met the expanded criteria (2 carcinomas with minimal submucosal invasion, 2 small intramucosal undifferentiated carcinomas, and 1 intramucosal carcinoma with ulceration). Finally, two recent studies have shown 5-year survival rates between 83 % and 93 %, with similar values for endoscopically resected lesions meeting the standard or the expanded criteria. In neither study was there any death associated with gastric recurrence of the primary lesion [7] [8].

In this issue of Endoscopy, Sekiguchi et al. [9] and Abe et al. [10] present the short- and long-term outcomes of endoscopic treatment by ESD in two particular groups of patients with early gastric cancer (EGC) – patients with local recurrence after endoscopic resection of EGC and patients with undifferentiated carcinoma, respectively.

In both studies, the authors show rates of en bloc resection and R0 comparable to those generally obtained in the treatment of EGC by ESD according to standard criteria (> 84 %) [11]. We note, however, that Sekiguchi et al., despite excellent curative resection rates with ESD treatment for locally recurrent lesions (81 %), also have complication rates, particularly perforation (6 %), higher than those observed in most series of gastric ESD. This aspect becomes particularly concerning if we consider the average size of the lesions (only 12 mm). These data reveal the expected technical difficulty of ESD on these lesions, with the nonlifting sign, supporting the idea that this procedure, although technically possible, should be restricted to reference centers. In contrast, in the series of endoscopic treatment of undifferentiated carcinomas, Abe et al. report complications at expected rates (3 % perforation), but a lower rate of curative resection (64 %). As noted by the authors, this is probably due to the tight proposed criteria for curative endoscopic resection in undifferentiated tumors (< 2 cm, without ulceration, or submucosal or lymphovascular invasion) associated with more aggressive/diffuse biological behavior. An important finding from this study was the low rate of cases considered with noncurative resections according to positive horizontal margins recorded by the authors. Despite the difficulty in defining the margins on these lesions, ensuring extensive resection margins (> 5 mm from the macroscopic limit) will probably be of particular importance in this histological type of neoplasm [12]. Encouragingly, both studies present 5-year survival rates above 93 % in cases of curative resection according to the expanded criteria, with no deaths related to primary gastric cancer recurrence – similar rates to those obtained after surgical treatment of EGC.

One last point is the high rate of metachronous lesions recorded by these studies (> 10 % at 5 years) and others [13]. It is likely, however, that these values are an overestimation, since the so-called new lesions diagnosed during the first period after primary lesion resection may correspond to previously misdiagnosed lesions. Regardless, this number undoubtedly calls for a requirement for regular endoscopic surveillance following the endoscopic treatment of EGC – in fact the single case of gastric cancer-associated death after curative resection in this series was in a patient with an advanced metachronous lesion that was overlooked earlier, as has been observed in other series of long-term follow-up [7]. Patients should therefore be made aware that choosing an endoscopic treatment for their EGC will require regular endoscopy surveillance that, if not acceptable or not possible, should probably prompt them to choose the surgical option instead.

Taking it all together, we may consider that if R0 is achievable after recurrence or in lesions with expanded criteria, long-term results seem to show that it is a safe procedure, as it was previously described to be for standard criteria, as long as adequate surveillance is accepted by the patient and is carried out.

However, two particular problems are seen in Western countries. First, the available literature on which to base an estimate of lymph node invasion of EGC in these countries is very scarce. In 2009, Hölscher et al. [14] reported this risk in patients of a German center, on the basis of the depth of invasion of the lesions. Although this was pioneering work in the West, it did not provide a morphological description of the lesions, thus precluding an interpretation based on the expanded criteria. Second, the expanded criteria are based on precise histological features which are mostly described in the Eastern literature, but there appear to be some differences between Western and Eastern pathologists in the diagnostic criteria they employ [15].

It is therefore important to have longer follow-up data in Western countries to ensure overlapping results, making sure that Western and Eastern pathologists follow the same exact methodologies, and that these standards are summarized in European guidelines.