Comparison of Endoscopic and Surgical Resection of Intramucosal Carcinoma in Barrett's Esophagus

Wesley D Leung; Jennifer Chennat
Posted: 10/27/2011; Expert Rev Gastroenterol Hepatol. 2011;5(5):575-578. © 2011 Expert Reviews Ltd.
Abstract and Introduction

Abstract

Evaluation of: Pech O, Bollschweiler E, Manner H et al. Comparison between endoscopic and surgical resection of mucosal esophageal adenocarcinoma in Barrett's esophagus at two high-volume centers. Ann. Surg. 254(1), 67–72 (2011).
The optimal management of intramucosal carcinoma in Barrett's esophagus continues to be controversial. To date, there has not been a lot of research directly comparing endoscopic versus surgical management of intramucosal carcinoma. Previous studies have shown that both modalities to have excellent outcomes. The reviewed article is a matched retrospective cohort study from two high-volume centers, which shows both modalities to be effective, but an associated higher morbidity rate and risk for procedure-related mortality in the surgical group, and higher recurrence rate in the endoscopic therapy group. The study is discussed in the context of the current state of knowledge regarding Barrett's esophagus and intramucosal carcinoma, in particular outcomes and limitations of the present study.

Barrett's Esophagus & Intramucosal Carcinoma

The annual incidence of adenocarcinoma from Barrett's esophagus (BE) is approximately 0.5% and appears to be rising.[1–3] Identification of different stages of BE (intestinal metaplasia to low-grade dysplasia, high-grade dysplasia (HGD), intramucosal carcinoma (IMC) and invasive adenocarcinoma) can be clinically challenging, and has profound treatment implications owing to the different prognostic profiles between early neoplasia and more advanced stages.

Increasing awareness of the risk of cancer from BE and enhanced and novel endoscopic imaging techniques have led to earlier detection of early curable esophageal adenocarcinoma. However, optimal management of patients with IMC from BE continues to be controversial and is still evolving.

Traditionally, esophagectomy has been the gold-standard treatment for BE with HGD, owing to the suspected risk of harboring occult invasive carcinoma, which has been estimated to be as high as 40%.[4,5] Our previous analysis of the published literature demonstrated that the true prevalence of submucosal invasive carcinoma in the setting of HGD was actually 12%, which was much lower than the pooled reported historical rate of 40%.[6] Esophagectomy has also been routinely performed for BE with IMC, despite the low incidence of lymph node metastasis of less than 1% that is associated with noninvasive disease limited to the mucosa (T1a).[7] In addition, esophagectomy is associated with significant morbidity and mortality rates even in high-volume centers.[8,9]

In contrast, endoscopic therapies have entered the clinical forefront as acceptable nonsurgical alternatives for HGD and IMC. The goal of endoscopic therapy for HGD or IMC is to ablate all BE epithelium (both dysplastic and nondysplastic) owing to the risk of synchronous/metachronous lesion development in the remaining BE segment.[7] Endoscopic therapies can be further subdivided into tissue-acquiring and nontissue-acquiring modalities. Tissue acquisition can be achieved through endoscopic mucosal resection (EMR), while photodynamic therapy, radiofrequency ablation and cryotherapy all ablate tissue without the benefit of histological specimen retrieval. Modalities such as argon plasma coagulation, multipolar electrocoagulation and laser therapies are thought to have more limited utility owing to high BE relapse rates, infrequent usage or significant risk of buried gland development.[10]

Summary of Methods & Results

The article by Pech and colleagues is a retrospective cohort study examining the outcomes of surgical versus endoscopic resection (ER) among patients with IMC.[11] The primary outcome of the study was to determine the complete remission rate in surgical and endoscopic groups.

A total of 114 patients with Barrett's IMC between 1996 and 2009 were included in the study. Patients having neoadjuvant chemoradiation or chemotherapy were excluded. A total of 38 patients were treated surgically at an expert center in Cologne University (Cologne, Germany; 26 direct referrals and 12 with unsuccessful endoscopic treatment). Of the 967 patients, 76 were treated with ER at Wiesbaden Hospital (Wiesbaden, Germany). They were randomly matched in a 2:1 fashion to the surgically treated patients with blinding to therapeutic outcome. Matching criteria included age, gender, tumor infiltration depth (pT1m1–3), tumor differentiation grade (G1/2 vs G3) and duration of follow-up. Circumferential EMR was not performed according to the authors, so the ER specimens were acquired from focal regions.

The endoscopic group had a pretreatment workup consisting of endoscopy with biopsies, CT scans of the chest and abdomen, and endoscopic ultrasound for lymph node staging. Complete remission was confirmed by two negative follow-up endoscopies and was followed by ablation of remaining nondysplastic BE using argon plasma coagulation and a strict follow-up surveillance program.

The surgical group had en-bloc esophagectomy after open or laparoscopic gastric mobilization and two-field lymphadenectomy of mediastinal and abdominal lymph nodes. Histopathologic examination of all resected specimens for the 38 surgical patients was re-evaluated by two experienced pathologists. Surgical patients also underwent a standardized follow-up.

The endoscopic and surgical resection cohorts were matched in a 2:1 fashion by the prespecified matching criteria. This resulted in no significant differences between the two cohorts regarding age, gender, Charlson index of comorbidity grade, infiltration depth, tumor differentiation grade or follow-up period. The only significant difference was that the ER group had a shorter Barrett's length (3 cm) than those treated surgically (6 cm; p < 0.02). There were no lymphovascular invasion or lymph node metastases in either group. Complete remission of neoplasia was achieved in all but one endoscopically treated patient (98.7%) versus all patients in the surgical group (100%).

Median follow-up in the ER and surgical groups were 4.1 years and 3.7 years, respectively. There was a trend towards increased overall rate of recurrence/metachronous neoplasia in the ER group versus the surgical group (6.6 vs 0%), but the difference was not significant (p = 0.17). Patients with metachronous neoplasia in the ER cohort all had long-segment BE (mean: 8.0 cm) versus those without (mean: 3.7 cm; p = 0.08).

The rate of minor complications in the ER group was 17%. Significantly more major complications were seen in the surgical versus ER cohort (32 vs 0%; p < 0.001), for example, anastomatic leakage, pneumonia, cardiac problems and sepsis. There was no tumor-related mortality in either group but surgery was associated with a procedure-related mortality rate of 2.6% and zero in the ER group, however, there was no significant difference between these groups (p = 0.333).

Multivariate analysis of survival revealed that only age was an independent prognostic factor for survival (hazard ratio: 1.13; p = 0.005). Multivariate analysis of disease-free survival could not identify risk factors for recurrence/metachronous neoplasia.

Discussion & Significance

There have been few published studies that have compared outcomes of endoscopic and surgical resection of Barrett's-related IMC. This paper is the first study comparing single treatment modalities, ER and transthoracic esophageal resection in this population. Because the study design matched the two treatment cohorts by various criteria, this made them more amenable to comparison than previous studies.

Traditionally, radical esophageal resection has been regarded as the treatment of choice in patients with BE-related IMC. Recent publications by various groups, including the author's and our own group, have reported excellent long-term results with endoscopic treatment and very low complication rates.[7,12–15]

Endoscopic resection is an important therapeutic option for removal of BE-related IMC because, unlike ablative methods, ER allows complete pathologic assessment of the resected specimen, including depth of cancer invasion, cellular differentiation and lymphovascular involvement. This pathological staging allows prediction of risk of lymph node metastasis and guides whether the patient can be cured endoscopically or should proceed towards surgical treatment (e.g., when there is deep submucosal or lymphovascular infiltration).

The authors report excellent complete remission rates in both treatment modalities for BE–IMC with no differences with regards to overall survival and disease-free follow-up. These rates are comparable to rates seen in other studies. However, there was a trend towards more recurrences in the ER group; however, the ablation modality used – argon plasma coagulation – has fallen out of favor in most centers. Thus, if more current therapies, such as radiofrequency ablation (RFA), were used in this study, the recurrence rates in the ER group might even be lower, as suggested by study by Pouw, where 41 patients with BE with early neoplasia underwent RFA with or without ER had 98% eradication of all BE tissue/dysplasia and no dysplasia after a mean 21 months follow-up.[16] Similar results have been seen for early neoplasia in BE longer than 10 cm.[17] In a more recent study, Pouw and colleagues also showed pre-RFA ER for visible lesions and serial RFA with focal escape ER for post-RFA residual tissue achieved 100% eradication of BE tissue after a median 22 months follow-up.[18]

In addition, the ER used in this study was not circumferential and may have left remaining neoplastic tissue behind to contribute to the higher recurrence rate in this subgroup. Lack of circumferential ER (or piecemeal resection) has been demonstrated to be a risk factor frequently associated with BE tissue recurrence.[7] Our group and other select centers have utilized complete Barrett's eradication EMR with the curative intent of removing all Barrett's epithelium with HGD or IMC to potentially reduce the risk of developing synchronous and metachronous lesions. Our center reported a 96.9% remission rate from Barrett's epithelium and no reported synchronous or metachronous lesions over a mean of 22.9 months.[14] As mentioned by the authors, the BE length in the ER group was lower than the surgical group and a higher recurrence rate of carcinoma may have been observed in the ER group if the patients had a comparable length to the surgical group.[11]

The complication rates of ER were similar to those that have been reported in other studies. Complication rates include perforation (<0.5%), bleeding (14%; usually easily controlled with epinephrine and/or metal clips) and esophageal stenosis (0–30%).[19] However, the complication rates from esophagectemy in this study appear higher than other centers. There is a major difference between resections performed for advanced esophageal cancer and those performed for T1 lesions. In fact, many series have shown that the mortality rate for early cancer is close to 0% and that after the immediate postoperative period, the quality of life is excellent.[20,21] For instance, the Luketich group has shown that esophagectomy for T1 lesions can be performed with a 0% mortality rate.[20] In addition, the use of minimally invasive techniques has brought about an improvement in the immediate postoperative course because it is associated with less postoperative pain, a shorter hospital stay and a faster recovery time. Chang and colleagues showed that postoperative symptoms caused by esophagectomy are common but mild, and do not interfere with quality of life, which is excellent and similar to that of the general population.[19]

Another argument favoring ER over esophagectemy for BE-related IMC was put forth in a decision analysis by Pohl et al..[22] Concerning BE-related early cancer (T1 with mucosal or minimal submucosal infiltration), a decision tree model compared EMR of the cancerous lesion with RFA of the remaining Barrett's segment to esophagectomy. During the 5-year interval of the study, endoscopic therapy cost US$17,000 and yielded 4.88 quality-adjusted life years, compared with US$28,000 and 4.59, respectively, for esophagectomy. The overall outcome was not changed by varying the recurrence rates of cancer or BE metaplasia after endoscopic therapy.

The study remains observational and contrasts with prior reports comparing the two modalities. In a study by the Mayo group, the ER and surgical groups were difficult to compare, but showed a higher recurrence rate in the ER compared with the surgical group (12 vs 2%) and lower cancer-free survival.[23] Overall survival was similar in the two groups. In a study by Schembre et al., two heterogeneous populations of HGD and IMC were treated by ER and surgery.[24] The ER group had a higher tumor progression rate and the surgical group had higher complication rates and was more expensive than the ER group.

There were a number of limitations to this study. For example, the follow-up time in this study may not have been long enough to detect differences between the two groups. Furthermore, despite using two separate centers and matching criteria to minimize selection bias between the two groups, selection bias could still have influenced the results. Finally, the study remains retrospective and observational, thus making it difficult to draw definitive conclusions from the study, and a randomized trial might still be required.

Expert Commentary & Five-year View

While observational studies can be useful, more definitive data will require randomized controlled trials. However, implementing this type of research methodology would be challenging to accomplish in the BE-related IMC population, as it would be difficult to persuade patients to be randomized to esophagectemy given the already excellent published results in patients that have undergone ER. In addition, blinding treatment groups would be impossible in a randomized trial and placebo treatment would be unethical. As a result of these and other factors, randomized placebo-controlled studies are difficult to achieve in BE-related IMC.

Future studies should include outcome and cost-analysis data comparing endoscopic treatment combining EMR and ablation therapies such as RFA or complete Barrett's eradication EMR, compared with esophagectemy. Despite the excellent efficacy of ER seen in this study, esophagectemy should still be considered in patients with IMC with lymph node involvement seen by endoscopic ultrasound, patients that are young and healthy (making 20–40 year surveillance less feasible or desired), patients not willing to be compliant with rigorous endoscopic follow-up, patients with multifocal disease and long-segment BE, and esophageal adenocarcinoma extending into and beyond the submucosa.

Sidebar

Key Issues

The incidence of esophageal adenocarcinoma has increased eightfold in the last 30 years and is thought to be secondary to gastroesophageal reflux disease through the development of Barrett's esophagus (BE).
The study by Pech and colleagues is a retrospective observational cohort study examining the outcomes of endoscopic and surgical resection in patients with BE-related intramucosal carcinoma. The primary outcome of the study was complete remission of cancer after treatment.
Both treatment groups had excellent complete remission rates.
The authors observed a higher morbidity rate and a risk for procedure-related mortality in the surgical group and a trend towards higher recurrence rate in the endoscopic group.
This study supports the role of endoscopic treatment of BE-related intramucosal carcinoma as an excellent alternative to esophagectemy and suggests a better safety profile, but confirmation from other studies may be required.