Advances in endoscopic therapy have resulted in dramatic changes in the way early esophageal cancer is managed as well as in the palliation of dysphagia related to advanced esophageal cancer. Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are effective therapies for accurate histopathologic staging and provide a potential for complete cure. Mucosal ablative techniques (radiofrequency ablation and cryotherapy) are effective adjuncts to EMR and ESD and reduce the occurrence of synchronous and metachronous lesions within the Barrett esophagus. The successes of these techniques have made endoscopic therapy the primary means of management of early esophageal cancer.
Key Points
- •
Advances in endoscopic therapy have resulted in dramatic changes in the way early esophageal cancer is managed and in the palliation of dysphagia related to advanced esophageal cancer.
- •
Endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) are effective therapies for accurate histopathologic staging and provide a potential for complete cure.
- •
Mucosal ablative techniques are effective adjuncts to EMR and ESD and reduce the occurrence of synchronous and metachronous lesions within the Barrett esophagus.
- •
These techniques have placed the endoscopist at the multidisciplinary table in discussions of management of esophageal cancer, rather than solely being the diagnostician as in the past.
- •
Self-expanding metal stents (SEMS) has anchored itself as a durable and immediate modality for the relief of malignant dysphagia from esophageal cancer.
Introduction
Esophageal cancer is diagnosed in approximately 400,000 patients a year, but its true incidence has been challenging to estimate given the possibility of misclassifying gastroesophageal (GE) junction esophageal cancers as gastric cardia cancers. For all patients with esophageal cancer, the 5-year survival is approximately 20% or less. Given this strikingly poor prognosis, early detection and treatment offers the most promising prospect for improved survival or potentially even cure.
Esophageal cancers are characterized as squamous cell carcinoma or adenocarcinoma, the latter arising in association with Barrett esophagus. In the Western world, esophageal adenocarcinoma has surpassed squamous cell carcinoma as the predominant esophageal malignancy. These 2 cancer types differ in their rates of lymph node metastases. Lymph node metastases can be seen in up to 10% of patients with mucosal squamous cell tumors and in 50% of those with submucosal invasion. On the contrary, lymph node metastases are uncommon in patients with early esophageal adenocarcinoma. Independent of the type of esophageal cancer, endoluminal therapies are stratified into therapy for early-stage esophageal cancer and palliation of dysphagia symptoms in patients with metastatic disease or locally unresectable disease. Early esophageal cancers are those classified as Tis or high-grade dysplasia (HGD) (formerly called carcinoma in situ) or T1 tumors (involvement limited to the mucosa or submucosa).
Endoscopic therapies for early esophageal cancer continue to evolve. Advances in endoscopic technologies and techniques such as enhanced mucosal imaging, endoscopic ultrasound (EUS), and EMR have allowed for improved early cancer detection, staging, and eradication. It is imperative, however, to incorporate advanced endoscopic therapeutics for esophageal cancers in the context of a multidisciplinary setting to include surgical, medical, and radiation oncology expertise.
Endoscopic therapy for esophageal cancer can be stratified into management of early esophageal cancer (superficial cancers) and palliation of advanced cancer. This article begins with a brief review of the past management of early esophageal cancer, namely, esophagectomy and then reviews the available endoluminal therapies including EMR, ESD, thermal laser, argon plasma coagulation (APC), photodynamic therapy (PDT), and cryoptherapy. Finally, this article reviews endoluminal palliative options, namely, endoprosthetics (SEMS).
Introduction
Esophageal cancer is diagnosed in approximately 400,000 patients a year, but its true incidence has been challenging to estimate given the possibility of misclassifying gastroesophageal (GE) junction esophageal cancers as gastric cardia cancers. For all patients with esophageal cancer, the 5-year survival is approximately 20% or less. Given this strikingly poor prognosis, early detection and treatment offers the most promising prospect for improved survival or potentially even cure.
Esophageal cancers are characterized as squamous cell carcinoma or adenocarcinoma, the latter arising in association with Barrett esophagus. In the Western world, esophageal adenocarcinoma has surpassed squamous cell carcinoma as the predominant esophageal malignancy. These 2 cancer types differ in their rates of lymph node metastases. Lymph node metastases can be seen in up to 10% of patients with mucosal squamous cell tumors and in 50% of those with submucosal invasion. On the contrary, lymph node metastases are uncommon in patients with early esophageal adenocarcinoma. Independent of the type of esophageal cancer, endoluminal therapies are stratified into therapy for early-stage esophageal cancer and palliation of dysphagia symptoms in patients with metastatic disease or locally unresectable disease. Early esophageal cancers are those classified as Tis or high-grade dysplasia (HGD) (formerly called carcinoma in situ) or T1 tumors (involvement limited to the mucosa or submucosa).
Endoscopic therapies for early esophageal cancer continue to evolve. Advances in endoscopic technologies and techniques such as enhanced mucosal imaging, endoscopic ultrasound (EUS), and EMR have allowed for improved early cancer detection, staging, and eradication. It is imperative, however, to incorporate advanced endoscopic therapeutics for esophageal cancers in the context of a multidisciplinary setting to include surgical, medical, and radiation oncology expertise.
Endoscopic therapy for esophageal cancer can be stratified into management of early esophageal cancer (superficial cancers) and palliation of advanced cancer. This article begins with a brief review of the past management of early esophageal cancer, namely, esophagectomy and then reviews the available endoluminal therapies including EMR, ESD, thermal laser, argon plasma coagulation (APC), photodynamic therapy (PDT), and cryoptherapy. Finally, this article reviews endoluminal palliative options, namely, endoprosthetics (SEMS).
Management of early esophageal cancer
Operative Resection
The optimal treatment approach for patients with localized esophageal cancer remains controversial as new less-invasive endoscopic technology emerges to challenge traditional surgical esophagectomy. Before the emergence of advanced endoscopic techniques, esophagectomy was the treatment of choice for patients with localized squamous cell carcinoma, adenocarcinoma, intramucosal carcinoma (ImCa), and multifocal HGD in association with Barrett esophagus. Both transhiatal and transthoracic esophagectomy are frequently performed in the United States without a clear 5-year survival benefit of one over the other as reported from 2 large outcomes databases ; however, transhiatal esophagectomy is generally preferred in those with T1 cancers, based on its reduced morbidity (pulmonary complications, chylothorax, blood loss, shorter intensive care unit and hospital stay). In general, 5-year survival rates for esophagectomy for early esophageal cancer ranges from 63% to 83%. It is now clearly recognized that the outcomes of esophagectomy are a function of hospital and surgical volumes. At high-volume centers, the mortality associated with esophagectomy may be 2% (range of 0%–5%), but morbidity rates remain high, ranging from 30% to 50%. The introduction of a minimally invasive laparoscopic transhiatal approach (combined thorascopic–laparoscopic) seems to be associated with lower perioperative morbidity and mortality and may lead to improved overall survival and disease-free survival in expert hands.
The morbidity and mortality associated with surgical esophagectomy and the low rates of metastases associated with early esophageal cancer have led to the interest in local endoscopic therapy as an alternative to surgery. Few data are available directly comparing the surgical and endoscopic management of early esophageal cancer and what exists are retrospective and observational. The available literature suggests that long-term outcomes including median cancer-free survival with endoscopic therapy for early esophageal cancer is similar to that of surgical therapy with fewer complications, but a higher rate of recurrence. A recognized drawback of endoscopic monotherapy for early esophageal cancer in Barrett esophagus is the rate of recurrence (which can be as high as 22% in some series). However, the most recurrences can be endoscopically salvaged and do not lead to death from disease. These data suggest that endoscopic therapy can be used in intramucosal esophageal cancer with oncologic outcomes equivalent to surgery.
Staging
The overall outcome of esophageal cancer is strongly associated with its stage, therefore accurate staging is critical to providing estimations of prognosis and appropriate selection of treatment. Table 1 contains the T category definitions for esophageal cancer (both squamous cell and adenocarcinoma) and includes a more comprehensive subclassification scheme proposed for early esophageal cancers. Fig. 1 contains a graphical representation of the depth of invasion of esophageal tumors and their risk of lymph node metastases. Overall, mucosal-based lesions are associated with very low rates of lymph node metastases (<3%); however, tumors that invade the submucosa have a substantial risk of lymph node metastases. In general, esophageal cancer staging involves a multimodal approach including computed tomography (CT), fludeoxyglucose F18 positron emission tomography (FDG-PET), and EUS. After diagnosis, staging usually begins with a CT scan of the chest and abdomen to evaluate for distant metastatic disease. CT, however, has poor sensitivity for celiac lymph nodes and small metastases (particular peritoneal metastases) and cannot reliably assess for depth of tumor invasion. FDG-PET scans have improved sensitivity for small metastases and are routinely used in preoperative staging when traditional CT demonstrates no distant disease; however, it too has poor sensitivity and specificity for locoregional disease staging.
| Tx | Primary tumor cannot be assessed | |
| T0 | No evidence of primary tumor | |
| Tis, m1 | High-grade dysplasia, limited to mucosal layer | |
| T1 | Tumor invades lamina propria, muscularis mucosa or submucosa | T1a |
| T1, m2 | Tumor invades the lamina propria | |
| T1, m3 | Tumor invades into but not through the muscularis mucosa | T1b |
| T1, sm1 | Tumor invades into the shallowest one-third of the submucosa | |
| T1, sm2 | Tumor invades into the intermediate one-third of the submucosa | |
| T1, sm3 | Tumor penetrates the deepest one-third of the submucosa | |
| T2 | Tumor invades the muscularis propria | |
| T3 | Tumor invades adventitia | |
| T4 | Tumor invades adjacent structures | |
| T4a | Resectable tumor that invades pleura, pericardiam, or diaphragm | |
| T4b | Unresectable tumor invading aorta, vertebra, tracea, or other adjacent structures | |
EUS and EMR have allowed for more precise assessments of tumor depth and likelihood of regional lymph node metastases. In the past, early esophageal cancer was defined by the endoscopic criteria diameter less than or equal to 2 cm. However, this method clearly was unable to predict the depth of invasion or local lymph node involvement. At that time, only surgery was able to provide these additional assessments. These surgical resections have revealed that early esophageal cancers, those that do not penetrate the muscularis mucosa, are rarely associated with metastatic disease.
The development of EUS has obviated surgical T staging of esophageal cancer. EUS allows the endoscopist to visualize at 7.5 to 12.5 MHz transducer frequencies to delineate lesions and their depth of penetration into or through the 5-layer esophageal wall pattern. It also enables examination of periesophageal, celiac, and perigastric lymph nodes (all N1 categories) as well as most of the liver (M1 category), pleura, and vascular structures (T4 stage). Suspicious-appearing lymph nodes—those that are hypoechoic, round, in proximity to the tumor, and greater than 1 cm in diameter—may be sampled using the linear array echoendoscope for EUS-guided fine-needle aspiration. Higher-frequency probes that image at 20 to 30 MHz can image mucosal lesions at a higher resolution, if deemed necessary. These probes also eliminate artifact caused by compression of the endosonographic balloon that is traditionally used with EUS. However, higher-frequency EUS probes are limited by lack of deep tissue visualization. In most cases, traditional EUS (without the use of higher-frequency probes) is sufficient to stage esophageal cancer. Studies have demonstrated that preoperative staging by EUS performed at high-volume EUS centers (>50 EUS/endoscopist/year) is more sensitive and specific than when performed at low-volume centers. Therefore, EUS should be performed by experienced endosonographers to optimize patient selection for esophageal resection. EUS is considered the most accurate available locoregional staging modality, with an accuracy of 80% to 90% of T and N staging, whereas some data suggest that its accuracy is diminished to approximately 65% for superficial T1 cancers and that it cannot reliably differentiate between T1a and T1b lesions. On the contrary, a meta-analysis (19 studies) shows that EUS is accurate for staging of T1a and T1b tumors. The accuracy of EUS may be influenced by the experience of the endosonographer. Differentiating mucosal lesions from those that infiltrate the submucosa is much more difficult in the setting of Barrett esophagus than in squamous cell carcinoma given that crypts and villi in Barrett esophagus are more heterogeneous than the layered architecture of squamous epithelium. Barrett esophagous tends to be associated with more inflammatory changes that result in a double muscularis mucosae, and Barrett-esophagus-related neoplasms tend to be located close to the cardia, which is a difficult area for EUS interpretation. Some studies infer that there is limited added value of EUS for determining T category once tumor depth has been estimated by expert endoscopic inspection. Most authorities now, however, agree that histopathologic inspection of the suspected lesion obtained by endoscopic resection is now considered the most accurate means to assess the depth of tumor infiltration (pathologic T category) and the risk of lymph node involvement.
EMR has emerged as a diagnostic, staging, and therapeutic technique. This technique provides a greater volume of tissue for diagnostic purposes. The specimen yields the most accurate T staging for superficial esophageal cancers. Most importantly, for lesions confined to the mucosa and in certain cases the submucosa, it can be curative. At present, EMR is the standard therapy for treatment of early esophageal cancer in association with Barrett esophagus. Various techniques of EMR are discussed separately in the following sections.
Endoscopic Resection
Historically, surgery was considered the modality of choice for early esophageal cancer including ImCa and HGD in the setting of Barrett esophagus. Although cure rates were very high, they came at the cost of treatment-related morbidity and mortality. Given the very low risk of lymph node metastases associated with these conditions and the reported morbidity (30%–40%) and mortality (2%–5%) associated with esophagectomy (at high-volume centers performing >20 operations a year), endoluminal eradication and ablative therapies have surpassed surgery as the modality of choice. The goals of endoscopic therapy are to preserve the esophagus while achieving curative eradication. EMR, ESD, and mucosal ablative techniques including radiofrequency ablation (RFA), thermal laser therapy, APC, and PDT are discussed in the following sections.
Endoscopic mucosal resection (EMR)
EMR is an extension of standard snare polypectomy for the eradication of mucosal-based lesions that may not otherwise be amenable to endoscopic resection. Although the intent is often curative, it is used for enhanced diagnostic and staging purposes. EMR is also referred to as mucosectomy or endoluminal resection. The major advantage of EMR is that it delivers a large specimen, which provides a more accurate histologic diagnosis given the increased depth and volume of tissue sampling compared with standard biopsy forceps. Fig. 2 exemplifies a nodular lesion and the defect created after EMR. HGD and T1a tumors (m2 and m3) are the most appropriate candidate lesions for endoscopic resection given that lymph node metastasis is extremely rare.
When feasible, EMR is highly successful in achieving cure rates for HGD and ImCa in association with Barrett esophagus, with up to 96% complete response at a mean follow-up of 63 months in the largest observation series. The same series demonstrated that a significant drawback of EMR cure was the high rate of recurrent or metachronous lesions during follow-up, approaching 22%. Most of these patients can be adequately salvaged with endoscopic therapy alone, rather than surgery or death. Although no randomized controlled trial exists, the available literature demonstrates that there is no apparent difference in long-term outcomes (long-term complete response and overall survival) between EMR and surgery for early esophageal cancer.
There are several techniques available for EMR: cap-assisted, ligation-assisted, and injection-assisted EMR. The most commonly used techniques are the ligation-assisted and cap-assisted EMR. Irrespective of the EMR technique, before resection, a thorough white light and electronically enhanced (eg, narrow band imaging) examination is critical to identify and characterize suspect surface abnormalities. In order to sustain visibility of the targeted area for resection, the borders of the lesion may be marked using a cautery device such as multipolar coagulator (ie, tip of the snare).
Cap-assisted EMR
Cap-assisted EMR uses a transparent cap fitted over the tip of the endoscope. The lesion is lifted with a submucosal injection of saline or diluted epinephrine, and the cap is subsequently centered over the lesion and suction is applied to retract the lesion into the cap. A preseated snare is then immediately closed to capture the neopolypoid tissue, and suction is then released. The lesion is then snare-resected with electrocautery ( Fig. 3 B). This application is particularly useful for flat or nodular lesions. The cap is available in a completely flat, level lip, or an obliquely angulated cap end. The straight flat cap is favored when precision is required in the amount of tissue necessary for resection, whereas the oblique cap allows for larger resections. This technique was pioneered in Japan, where EMR is the standard of care for superficial esophageal cancer.
Ligation-assisted EMR
Ligation-assisted EMR or multiband mucosectomy uses a modified band ligation system similar to that of variceal band ligation. The endoscope is fitted with a friction cap that attaches to the tip of a standard or therapeutic endoscope. The cap is used to suction the target lesion into the banding chamber. A band is deployed, entrapping the tissue and thereby creating a neopolyp. The neopolyp is then resected using snare electrocautery below, above, or through the band (see Fig. 3 A). Some endoscopists use preligation submucosal injection to lift the mucosal-based lesion away from the musclaris propria and provide an elevated purchase around which a snare can be placed. This ligation-assisted EMR technique is attractive because the band routinely extrudes the muscularis propria layer entrapping only the mucosa and portions of the submucosa. Second, the system allows efficient repeated applications to enable wide-area confluent resections during a single session. The average diameter of the resected mucosa is approximately 15 mm when assessed by pathologists. A multiband mucosectomy device is commercially available (Duette; Cook Medical, Limerick, Ireland) for use with high-definition and therapeutic channel endoscopes. A prospective study of 243 procedures with a total of 1060 resections performed with ligation-assisted technique demonstrated complete resection in 91% with low rates of complication (bleeding in 3%, no perforations).
Ligation-assisted EMR has several advantages: (1) it does not require submucosal injection as with the cap-assisted technique because the muscularis propria will immediately retract when captured with a rubber band, whereas it may remain captured within the forcefully closed resection snare with cap-assisted EMR; (2) repeated withdrawal of the endoscope is not necessary with ligation-assisted EMR because the multiband device allows for 6 consecutive resections, thereby reducing cost and time; and (3) prelooping of the endoscopic resection snare in the ridge of the cap, which can be challenging, is not necessary. A randomized control trial of 84 patients compared cap-assisted EMR with ligation-assisted EMR for piecemeal resection of early cancer in association with Barrett esophagus. The ligation-assisted technique had significantly shorter procedure times (34 minutes vs 50 minutes) and was less costly than cap-assisted EMR. There was no difference in the depth of resection between the 2 techniques. A total of 3 perforations occurred in the cap-assisted group compared with 1 in the ligation-assisted EMR cohort. The time and cost savings of ligation-assisted EMR may be borne in the fact that preresection submucosal lift is not required and a single snare may be used for all resections.
Injection-assisted EMR
Injection-assisted EMR is akin to saline-assisted polypectomy of the colon. A solution is injected into the submucosal layer beneath the lesion to create a mound of tissue more suitable for ensnarement. This maneuver lifts the mucosal-based lesion away from the muscularis propria. Furthermore, the lift also reduces the electrosurgical resistance during snare resection with cautery. Injection-assisted EMR allows for resection of broad areas of tissue while minimizing the risk of perforation or transmural burn syndrome by protecting the muscularis propria with a saline cushion. The volume of fluid injected into the submucosa varies based on the lesion size and location. Although an ideal submucosal injectate remains to be determined, an admixture of methylene blue and normal saline solution is commonly used to aid in assessment of the lateral margins of the lesion and completeness and depth of resection. Some investigators also add diluted epinephrine to the injectate to reduce “back-bleeding” from the needle insertion and acute bleeding postresection.
Characteristics that are considered favorable for successful EMR are diameter less than 2 cm (for en bloc resection); lack of penetration of the muscularis mucosa; polypoid, flat, or elevated lesions; and well-differentiated histology ( Table 2 ). Ulcerated or depressed lesions are less likely to demonstrate appropriate lifting and are therefore not favorable for EMR. Lesions larger than 2 cm can be resected in a piecemeal manner to achieve wide-area confluent mucosal resection, but care must be taken to sufficiently overlap areas of resection so that small neoplastic residues are not left behind. Therefore, if possible, en bloc resection allows for the most accurate lateral and deep margin histologic evaluation.
| Characteristic | Favorable Outcome |
|---|---|
| Size | <2 cm |
| Depth of penetration | Muscularis mucosae is preserved |
| Grade of cancer | Well-differentiated cancer |
| Appearance | Elevated, polypoid, or flat |
In summary, if endoscopy with or without EUS identifies only the mucosal extent of neoplasia, then the patient is appropriate for initiation of endoscopic therapy. Curative EMR is determined by the depth of invasion precisely assessed at histopathologic inspection of the resected specimen. The final path of the EMR specimen, including tumor grade, depth of invasion, presence or absence of tumor at the resection margins, and lymphovascular invasion, ultimately determines whether endoscopic therapy alone is adequate or operative resection is indicated. If EUS demonstrates that the esophageal cancer invades beyond the muscularis mucosa or if there is evidence of pathologic lymph nodes, operative resection is recommended in fit candidates. An alternative school of thought to the above-mentioned algorithm supports the use of EMR for T staging without concomitant EUS, given the limitations of EUS in differentiating between T1a and T1b lesions. This practice, which the authors advocate, involves performing EMR for any endoscopically apparent superficial esophageal cancer with or without routine EUS examination.
Endoscopic submucosal dissection (ESD)
ESD was developed and popularized in Asia to obtain larger resected specimens en bloc with potential for complete excision of neoplastic lesions. ESD was initially described for the treatment of early gastric cancers in the stomach in Japan but is used in many parts of Asia for endoscopic therapy for early esophageal cancers.
The first step in ESD is to mark the area targeted for resection usually by using cautery to ensure optimal visualization after submucosal injection and to ensure that at least a 3-mm margin is resected. A submucosal bleb is formed in a manner similar to that described for EMR except that different injectates are commonly used. For ESD, a more durably persistent solution is used, given the approximate procedure duration of 1 to 2 hours. Examples of injectates include sodium hyaluronate, docium hydroxypropyl methylcellulose, sodium carboxymethylcellulose, hypertonic dextrose, hypertonic saline, or fibrinogen. Once the entire specimen has been lifted with the submucosal injectate, the lesion is usually circumscribed with one or more specialized electrosurgical “knives” to isolate the lesion from the normal surrounding mucosa. The submucosal layer beneath the isolated lesion is then meticulously dissected with electrosurgical devices to be removed en bloc; this part of the procedure is the most challenging and requires expert control and skill.
Multiple ESD cutting accessories have been developed. The devices have considerable variations and are championed by various enthusiasts, and some are best suited for certain portions of the procedure. Several ESD knives are depicted in Fig. 4 .
