Oesophagectomy within 30 days after noncurative endoscopic resection for oesophageal cancer: are we able to follow this ‘golden’ interval? Free

Endoscopic mucosal resection (EMR) is considered the first-line treatment for early-stage intramucosal lesion pT1a (m1-3) and carefully selected pT1b (sm1) non-circumferential oesophageal carcinomas [1]. This endoscopic treatment strategy has great advantages compared to surgical oesophageal resection, with a reduced invasiveness, decreasing morbidity and mortality rates and allowing the organ preservation with subsequent improvement in patients’ quality of life. Moreover, recent data demonstrated that EMR has equivalent long-term outcomes when compared to both oesophagectomy for intramucosal tumour (m1-3) or limited to the invasion of few microns of the muscularis mucosa with favourable histopathological factors (sm1), with an acceptable low risk of positive lymph node (LN) occurring in <5% of cases [2].

However, there are many circumstances in which EMR has to be deemed as noncurative, leading to consider radical oesophagectomy because of the high risk of recurrence or LN metastases. These circumstances are: tumour size >2 cm, invasion >sm1, excavated or polypoid lesion, circumferential lesion, positive lateral and/or vertical margin, poorly differentiated tumour, evidence of lymphovascular or perineural invasion and suspicion of positive LN [3]. If oesophagectomy is the recommended treatment option for noncurative EMR, the question that remains unclear is the optimal interval between the 2 procedures regarding the surgical risk if the delay is too short due to post-resection inflammation, and the risk of a pathological upstaging if the delay is too long due to local disease progression.

In this issue of the European Journal of Thoracic and Cardiovascular Surgery, Huang et al. [4] addressed the question of the optimal time interval from noncurative EMR to oesophagectomy for oesophageal cancer. Based on a multicentre retrospective series over 10 years, they assessed 155 patients (squamous cell carcinoma in 83%) who underwent noncurative EMR for cT1N0M0 oesophageal cancer and then received subsequent oesophagectomy. Overall survival and disease-free survival (DFS) were analysed to find an optimal interval of time cut-off from EMR to oesophagectomy. Among all, 106 (68%) received oesophagectomy within 30 days, while 49 (31%) had an interval over 30 days. The greatest difference of DFS was found in the group who underwent oesophagectomy before 30 days with better long-term outcomes (P = 0.016). Comparing the pathological stage between EMR and oesophagectomy, upstaging occurred in 26 patients and thus leading to worse DFS (HR = 3.780, P = 0.042). Multivariate analysis highlighted 4 independent risk factors for pathological upstaging: age, T1b invasion, lymphovascular invasion (LVI) and interval >30 days. The authors concluded that oesophagectomy should be performed within 30 days following noncurative EMR, otherwise it might cause pathological upstaging worsening prognosis, especially in patients with submucosal invasion and LVI.

In brief, this study brings arguments to support that oesophagectomy is the recommended option after noncurative EMR. The majority of patients have been treated with minimally invasive surgery and previous EMR should not be seen as a barrier. Surgery resulted in a very short length of hospital stay (median 9 days) with apparently very low morbidity and mortality rates, but not reported in the article. Most importantly, 17 patients were found to have N+ disease (2 T1a, 12 T1b and 3 >T1b tumour) and in all 26/155 (16%) patients were pathologically upstaged towards stage II or III cancer.

The authors have to be commended for addressing this important issue that specifically interrogates on the ability of all eastern or western centres to follow this ‘golden period’ of 30 days in performing oesophagectomy after noncurative EMR.

The first limit preventing us to respect this ‘golden period’ after EMR is the pathologist expertise. It is recommended that pathological analysis of mucosectomy should be performed by at least 2 independent pathologists because of the poor intra- and interobserver agreement [5]. Moreover, the third pathologist is sometimes required in case of controversy. It objectively seems difficult to obtain clear and definitive conclusions for final discussion and decision at the tumour board in this 30-day delay.

The second reason for the extended time between EMR and surgery may be the need for endoscopic ultrasonography, Positron emission tomography (PET)/computed tomography (CT) scan, pulmonary function test and cardiovascular assessment before surgery. Accessibility to hospital facilities, especially considering the ongoing pandemic situation, emphasized the difficulty to perform this preoperative work up in due time. In daily practice, the majority of patients are referred to EMR without the full recommended preoperative staging that is obviously conducted thereafter. Unsurprisingly, the authors had probably have the same issue because one-third of their patients have been operated after 30 days. However, they did not provide the reasons why these patients were delayed.

The third reason that may extend the interval between EMR and surgery is potential complications related to EMR. Stenosis is reported as the most frequent complication [pooled incidence: 5%, 95% confidence interval (CI) 3–8%], followed by perforation (1%, 95% CI 0–1%) and mediastinal emphysema (0% 95% CI 0–1%). Bleeding is another possible complication but with an incidence of lower than 2% [6, 7]. Even if these complications are relatively infrequent (in all <10%), they have the potential to significantly impact the short-term outcomes after EMR and thus potentially delay surgery. Unfortunately, these complications have not been reported in the present study.

At last, the authors determined the cut-off value of 30 days on their cox regression model. However, one of the main weaknesses of this study is the very small number of patients followed after 3 years regarding DFS. In fact, only a third of patients at risk are present in the Kaplan–Meier model after 3 years, when the difference between the two curves becomes observable (42 patients in the 0–30d group and only 9 patients in the 30–90d group). This suggests an underpowered study on the primary outcome and a high risk of alpha error to conclude a significant difference between groups that does not really exist.

Rather than focusing on 30 days as the sole determining factor, the most important message and conclusions that can be achieved from this study are to recommend a prompt preoperative workup followed by an earlier oesophagectomy if one pathological criterion is unfavourable such as patients with age >70 years, T1b invasion or LVI.

REFERENCES