Endoscopy 2018; 50(07): 655-656
DOI: 10.1055/a-0596-7447
© Georg Thieme Verlag KG Stuttgart · New York

Risk-stratified clinical management of superficially invasive esophageal squamous cell carcinoma after endoscopic resection: finding the sweet spot

Referring to Takahashi K et al. p. 662–670
Marnix Jansen
Department of Pathology, University College London Hospital NHS Trust, London, United Kingdom
Department of Pathology, UCL Cancer Institute, London, United Kingdom
Jacques Bergman
Department of Gastroenterology and Hepatology, Academic Medical Center, Amsterdam, The Netherlands
› Author Affiliations
Further Information

Publication History

Publication Date:
27 June 2018 (online)

In this issue of Endoscopy, Takahashi et al. report on the long-term outcomes of a risk-stratified management approach for patients with superficially invasive esophageal squamous cell carcinoma (ESCC) [1]. The key factor driving management decisions for superficial ESCC after endoscopic resection is depth of tumor invasion, as this is linked to the risk of lymph node metastasis (LNM) in these patients [2] [3]. For in situ ESCC (stage T1a-M1) and ESCC limited to the lamina propria (T1a-M2), the risk of co-incident LNM is estimated to be less than 5 %. This marginal risk is deemed acceptable for endoscopic therapy, given the morbidity and mortality associated with esophageal resection. For lesions invading the muscularis mucosae (T1a-MM), the risk of co-incident LNM has been variously reported as 0 % – 9 %, whereas the LNM risk for lesions infiltrating the superficial submucosa (< 200 μm from the muscularis mucosae, stage T1b-SM1) has been reported to lie between 8 % and 16 % [4] [5] [6]. These superficially invasive ESCC lesions (stage T1a-MM and T1b-SM1) have therefore been considered a “borderline” indication for endoscopic treatment, where management should be individualized and discussed in a multidisciplinary setting in centers with a tertiary referral service for esophageal cancer.

However, these data on LNM risk for stage T1a-MM and T1b-SM1 ESCC have been derived from retrospective histopathologic analyses of resection specimens, which do not reveal the natural history of conservatively treated superficially invasive ESCC. Indeed, studies have suggested that long-term survival of patients with ESCC lesions infiltrating the muscularis mucosae without further poor prognostic indicators, such as lymphovascular invasion (LVI) or poor tumor differentiation, is excellent (> 95 % tumor-specific survival) [7] [8]. As a result, there is now a move toward expanding the eligibility criteria for endoscopic therapy to include ESCC patients with infiltration of the muscularis mucosae and superficial submucosa who lack poor prognostic indicators, so-called low-risk disease [3] [9].

“This report provides strong support for the safety of an expanded indication for conservative management of superficially invasive ESCC after endoluminal resection.”

Takahashi et al. carried out a retrospective analysis of outcomes of patients treated at their institution for superficially invasive ESCC between August 2004 and November 2013 [1]. From an initial endoscopy database of 694 patients who underwent endoscopic submucosal dissection (ESD) for ESCC during the study period, the authors retained 108 patients who met the study inclusion criteria (i. e. histologically proven ESCC, either stage T1a-MM or T1b-SM1, without clinically documented LNMs at the time of ESD). After excluding only six patients because of loss to follow-up, the authors included 102 patients in the final study analysis. All patients had been uniformly clinically evaluated using computed tomography (CT) following endoluminal resection. Patients were offered conservative management with strict endoscopic follow-up and CT every 6 months if the lesion lacked poor prognostic indicators and initial CT did not reveal locoregional LNMs. If metastases were detected during follow-up, patients were offered salvage surgery and/or (chemo)radiotherapy depending on their clinical condition. The median length of follow-up was close to 6 years.

Retrospective analysis of this cohort showed that the overall 5-year disease-specific survival was 97.5 %, and the 5-year tumor-free survival was 82.1 %. A total of 23 patients in this cohort (22.5 %) showed poor prognostic indicators (either LVI or poor tumor differentiation, or both), and of this group, 12 patients were offered surgery and/or (chemo)radiotherapy, whereas 11 patients were not offered additional treatment because of comorbidities or treatment refusal. Only two patients across the cohort died of ESCC, and metastases were detected in 12 patients on follow-up. LVI was shown to be a strong independent predictor of the development of metastases on follow-up, although the confidence intervals (CI) were wide (hazard ratio 5.42, 95 %CI 1.39 – 21.18), suggesting significant heterogeneity. Indeed, 6 of 12 patients who developed metastases were initially assessed as low risk.

The authors then carried out a subgroup analysis to investigate the combination of patient risk stratification and further treatment on outcome. In this analysis three groups of patients were compared: patients with either T1a-MM or T1b-SM1 disease without any poor prognostic indicators (low-risk group); patients with poor prognostic indicators who underwent surgery and/or (chemo)radiotherapy (high-risk surgery group); and those with poor prognostic indicators who did not undergo additional treatment (high-risk conservative management group). Analysis of survival and tumor recurrence rates showed that tumor-free survival was significantly poorer in high-risk patients who did not undergo additional therapy compared with low-risk patients. Tumor recurrence was comparable between low-risk patients and high-risk patients who underwent additional therapy following ESD.

So, what are the key points from this retrospective data analysis? The data clearly signal that patients with invasion into, but not beyond, the muscularis mucosae and without further poor prognostic indicators, such as LVI or poor differentiation, are likely to have a low risk of loco-regional LNMs. This suggests that conservative endoscopic follow-up is a safe and efficacious management option for these patients. Future prospective studies will need to support or refute this contention. There is also a suggestion that patients with poor prognostic indicators (regardless of invasion depth) who undergo surgery decrease their risk of disease recurrence; however, the numbers are small and, again, this needs stringent verification in larger studies. The natural history of patients with lesions that invade the superficial submucosa (T1b-SM1) but lack poor histopathologic indicators is less clear. In the Takahashi study, 13 patients occupied this middle ground. The authors suggest that, in this cohort, there is no principal difference between groups stratified for invasion depth. However, there is a trend toward more frequent LVI in patients with T1b-SM1 tumors, suggesting that a diligent histopathologic search for signs of LVI, possibly aided by D2 – 40 immunohistochemistry to highlight lymphatic vessels, is indicated in these patients. Management decisions for these patients will have to be made on an individual basis, also taking into account comorbidities and patients’ expectations. In this regard, the survival data are reassuring and suggest that, again, many of these patients will fare well.

This report by Takahashi et al. provides strong support for the safety of an expanded indication for conservative management of superficially invasive ESCC after endoluminal resection. Future prospective studies are essential to define which high-risk features in these cancers predict the need for additional treatment. This will further narrow down the category of patients who require additional aggressive management after local endoluminal excision.