Advances in Management of Early Esophageal Carcinoma

1 Adjunct Professor in the Department of Endoscopic Surgery at the Instituto Alfa de Gastroenterologia in the Faculty of Medicine of the Universidad Federal de Minas Gerais in Belo Horizonte, Brazil 2 Director of the Endoscopy and Gastroenterology Service of the Hospital de la Policía in Bogotá, Colombia 3 Associate Professor in the Department of Gastroenterology at the Hospital of the University of Kobe in Kobe, Japan


INTRODUCTION
Esophageal cancer is the third leading cause of death from gastrointestinal cancer (1), and has a survival rate at fi ve years of only 15% (2).Squamous cell carcinoma (SCC) of the esophagus is predominant in men (3.6:1) between the fi ft h and seventh decades of life (3).When there is an early diagnosis of esophageal SCC, the prognosis is bett er, with survival rates at 5 years up to 95% (4).
Th is review aims to describe the major recent advances which have occurred in early esophageal cancer treatment, with emphasis on the current role of high-resolution endoscopy and chromoendoscopy for the diagnosis of early forms of the disease and the importance of new Japanese techniques of endoluminal endoscopic resection called endoscopic submucosal dissection (ESD).

METHODS
A review was done of the most important original articles found in the English language literature since 1990 on Medline (PubMed).It took into account the keywords "esophageal squamous cell carcinoma," "endoscopic submucosal dissection," "management, "and" Lugol chromoscopy." For the technical description of the ESD procedure we gave priority to articles published by the co-author of this manuscript (TT ), who has one of the largest personal series of ESD (4,000 cases performed up to 2012), and a line of research dedicated to developing new accessories and perfecting the ESD technique.

DIAGNOSIS
Th e great challenge in Latin America and in Western countries is to establish the diagnosis of esophageal cancer at an early stage when patients are asymptomatic and macroscopic changes are diffi cult to recognize.Th e best method for the detection of esophageal cancer is endoscopy, especially if it is combined with chromoscopy techniques (5,6).Endoscopy for screening for esophageal cancer SCC in the general population is not justifi ed because of the costs of the procedure, but population an endoscopic screening program can be cost eff ective in a high-risk.Th ere is an association of esophageal SCC with certain risk factors, especially, abuse of alcohol abuse and tobacco, drinking mate, a history of SCC in the upper aerodigestive tract or caustic injury of the esophagus, megaesophagus and human papilloma virus infections (5,6).Among all these factors, a history of SCC in the upper aerodigestive tract is most consistently associated with synchronous and metachronous neoplasia and therefore upper endoscopy for screening is routinely recommended in this population (5,6).
Chromoscopy using Lugol's solution is the method of choice for diagnosis of early esophageal SCC (3).Lugol's solution is a reaction dye in which iodine strongly stains esophageal squamous cells which are rich in glycogen.It does not stain dysplastic and neoplastic cells which are poor in glycogen.An important recently described aspect of chromoscopy using Lugol's solution, which is not oft en valued by some endoscopists, is the "pink color sign" (7,8).Th is signal is the transformation of the yellow color of neoplastic lesions which are Lugol'snegative into pink aft er 3 to 5 minutes.Th is transformation is due to the low levels of glycogen existing in the neoplastic cells of the esophagus.Its appearance indicates existence of highgrade dysplasia or squamous cell carcinoma with high specificity (7,8).In 2009, one of the authors (VA) began a screening program for cancer of the esophagus in patients with SCC in the upper aerodigestive tract.Unsedated transnasal endoscopy combined with digital chromoendoscopy using Lugol's solution was used.Among the 106 patients examined up to 2011, esophageal SCC was identifi ed in 13 cases (12.3%) of which 77% were at early stages which allowed treatment by endoscopy in 8 cases (9).Th is high incidence illustrates the importance of monitoring these high-risk patients in Latin America.

CLASSIFICATION AND STAGING OF SUPERFICIAL ESOPHAGEAL CANCER
According to the endoscopic classifi cation of superfi cial lesions published in the Paris Consensus (10), a surface neoplasia is defi ned as a lesion whose morphological appearance suggests involvement of the mucosal and submucosal layers without infi ltration of the muscularis propria.In Japan the superfi cial lesions are conventionally referred to as type 0 of the Borrmann classifi cation of advanced gastric tumors (12).Th ere are three subtypes of superfi cial lesions: protruding (type 0-I), superfi cial, fl at or elevated (Type 0-II) and excavated (type 0-III).Protruding lesions are further classifi ed into pedunculated (0-Ip), subpedunculated (0-ISP) and sessile (0-Is).In the esophagus, fl at superfi cial neoplasms predominate.Th ey are subdivided into lesions that are elevated compared to adjacent mucosa (IIA), lesions that are fl at (IIb) and lesions that are depressed (IIC).Protruding and excavated forms are rare (10).Superfi cial neoplasms are further subdivided according to transmural penetration.M1 corresponds to the epithelium and the basal layer, M2 to the lamina propria, and M3 to the muscularis mucosa.Compromises of the submucosa are also subdivided into SM1 (upper third), SM2 (average) and SM3 (lower third).
For histopathological description of superfi cial tumors the current recommendation is to use the Vienna classifi cation (11), by which the cancer is classifi ed according to the TNM-p ("p" of pathology) classifi cation.In the absence of invasion of the lamina propria, the lesion is called low or high-grade intraepithelial neoplasia, but the term carcinoma in situ (pTis) can also be used.When there is invasion of the lamina propria, esophageal carcinoma neoplasia is called micro-invasive or intramucosal (pT1m).When there is infi ltration of the submucosa, the tumor is considered to be invasive (pT1sm).
Th e importance of these subdivisions is due to the fact that the risk of lymph node metastasis in superfi cial cancer is closely related to the depth of vertical invasion of the organ wall.Th is is a key criterion for the selection of candidates for endoscopic therapy with curative intent.When the compromise is limited to M1 and M2 which is the neoplastic epithelial surface, the risk of lymph node involvement is close to zero, and realization of endoscopic excision is suffi cient for cure (10,12).When cancer invades the muscularis mucosa (M3) and the proximal portion of the submucosa to a depth of 200 μ (SM1), risk can conceptually vary from 9% to 19%, respectively.In cases situated at the edge of healing endoscopic treatment, further evaluation is essential.Th e following parameters must be carefully monitored: tumor size, presence of invasion and lymphatic vascular horizontal extension (width) of the invasion in the lamina muscularis of the mucosa.In the multivariate analysis study of Choi et al. (13), in which 190 pieces from esophagectomies were evaluated, the authors found that when the neoplastic lesion size was less than 3 cm for tumors aff ecting M3/SM1 focally, there was no lymph node or vascular invasion.If the compromise of the lamina muscularis of the mucosa had a width of less than 3 mm, the risk of lymph node metastases was very low (1 in 63 cases, or 1.5%).In the presence of massive infi ltration of the submucosa (SM2 and SM3), there was metastasis to the lymph nodes in approximately 40% of cases.
For initial staging of esophageal cancer endoscopic ultrasonography can be used in addition to detailed assessment of morphological characteristics found through endoscopy.Th e conventional echoendoscope operates at low frequencies (5-12 MHz) which allow division of the wall of the gastrointestinal tract into 5 layers.Due to its greater penetration the mediastinum and the celiac trunk can be fully appreciated in the search of lymph nodes.High resolution miniprobes have higher frequencies (15 to 30 MHz) which allow a more detailed diff erentiation of the gastrointestinal tract wall by dividing it into up to 9 layers.A fairly accurate assessment of the vertical extent of tumor invasion can be obtained with these miniprobes.Buskens et al. (14) analyzed the accuracy of endoscopy in the staging of 77 patients with dysplastic esophageal lesions and early carcinomas who had undergone esophagectomies.Th ey compared the accuracy of the method against the gold standard of an examination of the surgical specimen.Th e authors found that endoscopic ultrasound correctly demonstrated the absence of lymph node metastases in 93% of cases.Th e negative predictive value of endoscopic ultrasound for detecting the absence of submucosal invasion was 95%.To maximize the accuracy of endoscopic ultrasonography for esophageal SCC, we recommend combining the use of a high frequency miniprobe for T stage and a conventional echoendoscope for evaluation of lymph nodes.Th e presence of lymph node metastases suggests that neoplastic lesions cannot be cured with only local resection.Despite the good results mentioned above (14), the main limiting factor in endoscopic ultrasound evaluation of early esophageal neoplasias is the desmoplastic reaction caused by the infl ammatory response of the underlying neoplasia.Th is can lead to over-staging of the compromise which in turn can lead to incorrect routing of patients to treatment other than endoscopic.Absence of invasion of the submucosal layer combined with absence of malignant lymph nodes supports endoscopic resection.Th e ultimate outcome will be dictated by systematic histological evaluation of the resected specimen.If this assessment identifi es M3 or SM invasion, lymphatic or vascular invasion, or a tumor that is deep, the patient's treatment should be redirected to chemotherapy and radiation therapy, or to surgery.Considering that histological evaluation is the decisive factor in the fi nal defi nition of treatment, every eff ort should be made to maximize the quality of the sample and to prevent that the esophageal neoplasia is fragmented (as in "piecemeal" mucosectomy).Unlike complete removal of a tumor, a fragmented sample does not allow a proper analysis of margins and prevents characterization of endoscopic resection.

INDICATIONS FOR ENDOSCOPIC TREATMENT OF SUPERFICIAL ESOPHAGEAL NEOPLASIA
Th e classically accepted criteria for endoscopic resection with curative intention in superfi cial esophageal cancer include: depth of compromise restricted to M1 and M2 layers (epithelium and lamina propria); maximum length of 3 cm, maximum lateral extent of less than ¾ of the circumference, and a maximum of 4 lesions (15).With improved techniques for endoscopic resection, especially aft er the advent of ESD, these criteria have tended to widen to accept the endoscopic treatment of lesions larger than 3 cm which occupy the entire circumference of the esophagus and to eliminate the limit on the number of lesions as long as they are only intramucosal.
Endoscopic Treatment modalities include esophageal cancer resection techniques (mucosal resection or ESD) and ablation techniques.Ablation methods include photodynamic therapy, argon plasma, YAG laser, multipolar electrocoagulation, and most recently radiofrequency ablation.All these methods have the drawback that they make histopathological analysis impossible because they eradicate the neoplastic lesion.However, the information obtained from such an analysis is crucial for estimating curability with the endoscopic procedure.Th erefore, current ablative methods should not be used for the endoscopic treatment of esophageal SCC.In this paper, we discuss in detail the role of ESD and mucosectomy in the treatment of superficial esophageal cancer.

TECHNICAL PRINCIPLES OF ENDOSCOPIC MUCOSAL RESECTION
Th e gut wall is has by two main components: the mucosa and muscularis propria.Th ese elements are joined by the submucosa which is a layer of loose connective tissue.Since the thickness of the esophageal wall is between 3.5 mm and 4 mm, removing a superfi cial neoplastic lesion entails a risk of unintentional damage to the muscle layer and visceral perforation.To avoid these complications, a submucosal injection is necessary to separate the lesion from the muscularis propria.A 0.9% saline solution is used for these injections.Th e disadvantage of saline solution is that it dissipates rapidly into the gastrointestinal wall making removal of lesions larger than 1 cm impossible.For these cases viscous solutions have been developed to promote a longer lasting bubble eff ect (16).Th e best known are sodium hyaluronate, glycerin and hydroxypropyl methylcellulose (HPMC) (16).Elevation of the neoplastic lesion aft er infi ltration of the submucosa indicates that there is no deep invasion by cancer.Aft er injection of a suffi cient volume of solution into the submucosa, elevated lesions can be captured by loop diathermy which allows resection with an adequate margin of safety: this procedure is called mucosectomy.
Th ere are variations of mucosectomy techniques: injection and loop; injection, cap-assisted mucosectomy with suction, and mucosectomy aft er application of elastic bands.Cap-assisted mucosectomy is the recommended for application in esophagus (15,17).

Results of the esophagus mucosectomy
Th e most consistent results from endoscopic mucosal resection for esophageal cancer are found in the Japanese experience.Inoue (17) published a series of 142 patients with esophageal cancer who had undergone mucosectomy with curative intent who had been followed up on for up to nine years.When all the selection criteria were met, there were no cases of local recurrence or metastasis.Th e survival rate at 5 years was 95% and there were no deaths related to cancer.Endo et al. (18) reported the results of esophageal mucosal resection for intramucosal cancer smaller than 2 cm and occupying less than one third of the circumference.Th e survival rate at 5 years was 100%, local recurrences occurred in 7%.In all of these cases endoscopic treatment was repeated.Yoshida et al. (19) found no diff erences in 5 year survival rates between patients with esophageal cancer with involvement of M1 and M2 treated by mucosal resection (86% survival at 5 years) and similar patients treated with esophagectomy (83.2% survival at 5 years).
Th e effi cacy of endoscopic mucosal resection for curative treatment when all inclusion criteria are met has been well demonstrated.Moreover, there is no benefi t from esophagectomy for this subgroup of patients (19).However, even if certain criteria are violated, some authors have proposed alternative treatments to surgery.Shimizu et al. (20) reported a series of 16 patients who underwent endoscopic treatment of mucosal lesions that were invading the muscularis mucosa and submucosa and refused to undergo esophagectomy.Th ese patients received adjuvant radiotherapy and chemotherapy aft er mucosectomy.Th e author compared the outcomes of these patients with another 39 patients with esophageal cancer who were treated with esophagectomy.No patients in the mucosectomy and chemotherapy/radiation therapy group presented local recurrence or metastasis.Th e 5-year survival rate in the group that was treated endoscopically was 100% and that of the patients who underwent surgery was 87.5%.Although this was not a randomized controlled trial, the results of this research suggest that when the criteria for curative endoscopic resection is incomplete, additional therapy with chemotherapy and radiotherapy appears to be an equally eff ective alternative to esophagectomy.Currently there is an ongoing Japanese multicenter study to evaluate the therapeutic effi cacy of ESD associated with adjuvant chemoradiotherapy in patients with neoplasia M3 and SM1.

ENDOSCOPIC SUBMUCOSAL DISSECTION
Th e ESD technique was developed in Japan in the last 10 years to allow for resection of neoplastic lesions that are larger than 2 cm (21-23).ESD's secondary benefi ts include production of a sample that is adequate for histological study.From the clinical point of view, ESD achieves local resection with the greatest potential of healing and the lowest recurrence rate (24).Th is technique was originally designed for application in the stomach.Due to the smaller luminal space of the esophagus, compared to that of the stomach, the process is more complex more diffi cult to implement technically in the esophagus, so the use of ESD in developed more slowly in this organ.Refi nement and standardization of ESD and the development of new accessories has made it possible to extend the application of this modality to the management of esophageal SCC.Besides these factors, increased detection of tumors at early stages through increasing use of high-resolution endoscopy and chromoendoscopy coupled with high morbidity and mortality associated with esophagectomy have stimulated the development and improvement of endoluminal therapeutic interventions which allow preservation of the organ and maintain the life quality of patients.Currently, ESD is the preferred method for endoscopic resection of early esophageal cancer in Japan, and this procedure is being gradually introduced and accepted in Latin America and other Western countries.

Equipment and Accessories
Th e following equipment is recommended for performance of ESD: • A high resolution and magnifi cation endoscope with a waterjet (a specifi c channel for irrigation water) for demarcation of the resection margins and therapeutic work.Double-channel endoscopes are not recommended.• A pump for water infusion at controlled pressure.• Plastic caps (fi xing devices at the tip of the endoscope).• Calipers for foreign bodies and loops with nets for retrieval of the specimen.• Tubes with air leakage control valves.

ESD Technique
Th e patient should undergo preoperative and surgical risk evaluation since the majority of patients with esophageal cancer are chronic alcoholics and smokers.Patient should be hospitalized during treatment.In Japan, ESD procedu-res are performed under sedation.However, for beginners in the art, or when the estimated execution time exceeds two hours, we recommend the assistance of an anesthesiologist and the use of general anesthesia with orotracheal intubation (25).Th is allows the endoscopist to focus on the endoscopic procedure without risk of pulmonary aspiration.Cardiopulmonary monitoring and pulse oximetry in all cases.Th e use of a urinary catheter is also mandatory for prolonged procedures.Although there is no scientifi c evidence to justify the use of antibiotic prophylaxis for esophageal ESD, this practice is widespread in the Japanese centers where the use of second-generation intravenous cephalosporins is maintained for 3 days.Several ESD techniques have been described.Th is review briefl y describes two diff erent techniques.Th e fi rst technique was proposed and developed by one of the co-authors (TT ) at the University of Kobe and described previously (26)(27)(28)(29).Initially, the extent of the lesion must be thoroughly inspected using an endoscope with optical magnifi cation and digital chromoscopy that also has a short 4mm long plastic cap att ached to its distal end.Th en, proceed to chromoscopy with 0.8% Lugol solution for accurate tracing of the lesion's boundaries.In the esophagus, the procedure is done with short needle stilett o that is 1.5 mm long with a rounded tip (and which was designed by one of the authors (TT )).Th is can be used for all the steps of the ESD: marking, incision, submucosal dissection, simultaneous injection of saline solution and hemostasis of blood vessels (Flush-Ball-tipped Knife 1.5, Fujifi lm Fujinon Co., Japan) (27).Use of the 200D or 300D VIO electrosurgical unit (ERBE Elektromedizin, Turbingen, Germany) is recommended to.Table 1 shows the parameters used for surgical esophageal electrocautery through ESD at Kobe University.Chromoendoscopy is followed by marking the limits of resection with a Flush Knife (FK) respecting the minimum margins of 2-5 mm (parameters: Soft coagulation, Outcome 5, 100 watt s).In the esophagus, the proximal and distal margins must be large (5 mm), whereas the side margins can be more conservative (2 mm) to minimize the risk of circumferential esophageal stenosis secondary to resection.Next, proceed to submucosal injection with a 25 gauge needle (TOP Corporation, Japan).Initially Th e procedure begins by injecting a fi rst bubble of 0.9% saline solution into the submucosa.Th is is followed by injection of a viscous solution of hyaluronic acid (Muco-Up®, Seikagaku, Japan) that maintains the submucosa for a longer period.
Submucosal injection should be initiated at the proximal (oral) margin of the lesion, and should continue with successive injections of 1 to 3 ml in the lateral margins (left and right).Finally, injections should be made on the distal (anal) edge.A reasonable elevation of the lesion (rising sign) should be seen.Transfi xing injections in the center of the neoplastic lesion should be avoided in order to minimize the risk of the tumor seeding into the muscular layer.Next the protocol proceeds to an incision in the mucosa with an FK (Eff ect 4 Endocut I, cutt ing extent 2, cutt ing interval 3).Th e incision should create the outline of a "C" by starting at the anal margin, proceeding with a transverse incision to the lateral margin, then continuing with a longitudinal incision to the oral margin, and fi nally making another transverse incision to complete the "C".Once this is done, the submucosal layer should be dissected with the FK in the oral-anal direction (forced coagulation eff ect 2, 40 watt ).Th is will, create an incision fl ap in the lateral direction towards the center of the lesion.Every electric cut should be preceded by application of more saline solution injections into the submucosa.Th e dissection should be performed in the deep submucosal layer between the muscular and vascular submucosal layers of the esophagus to make it more effi cient and to optimize vascular control.As previously described, perfect hemostasis of submucosal vessels is essential for safe procedure and a bloodless fi eld (30,31).Perforated esophageal vessels should be identifi ed and isolated, and hemostasis should be performed by applying FK Soft Coagulation, eff ect 5, at 100 watt s for 3 to 5 seconds on each side of the vessel.Th is should be followed by section of the vessel with forced coagulation.If the hemostatic maneuver with the FK is not successful aft er 3 att empts, the practitioner should proceed to hemostasis with forceps (COAG grasper, Olympus Co. Japan).Aft er dissection of 70% to 80% of the lesion from its outside oral edge to the anal margin, sodium hyaluronate should be injected into the submucosa on the side margin that has not yet been isolated.Th is should be followed by a longitudinal incision in the in the lateral margin of the mucosa in an oral-anal direction.At this point the submucosal dissection fl ap previously created will be completed by using the cap to expose the submucosal space.FK movements should always be parallel to the axis of the esophageal wall and never perpendicular to the muscular layer.Aft er completion of the excision, the sample must be recovered with foreign body forceps.Take care to grasp the sample from its submucosal side in order not to damage the mucosa of the sample portion.Th e resection site should be reexamined.At this point protruding vessels must be coagulated and lacerations of the muscle layer must be approximately closed with clips.If the resection site's extension exceeds 75% of the circumference, a 4 ml injection of triamcinolone acetonide 10 mg/ml can be added via a catheter injector using 20 punctures with aliquots of 0.2 ml per puncture directed at both the edge and the center of the resection site.Be careful to inject only in the superfi cial portion of the muscle layer.Th is measure aims at minimizing the risk of esophageal stenosis.In a recently published study comparing 41 patients with circumferential esophageal resections that were divided into two groups (with and without injection of triamcinolone), esophageal stenosis occurred in 19% of the treated group as opposed to 75% of the control group (30).In Kobe, triamcinolone injections are repeated aft er 7 and 14 days.Recently, oral doses of 30 mg of prednisolone have been administered to prevent esophageal stricture aft er circumferential ESD.Th is begins on the third day following surgery, and treatment continues for 8 weeks.Preliminary results are promising (31).
Another variant of esophageal ESD is known as the tunneling method.Th is technique was initially developed by Yamamoto.Aft er training obtained by one of the authors (VA) with the inventor of this method, this technique was used in 22 consecutive patients with early esophageal SCC in Brazil.Aft er defi ning the margins, the fi rst step of the tunneling technique consists of a submucosal injection followed by FK incision in the anal margin of the lesion.Th is is followed by another submucosal injection and a transverse incision in the oral margin of the lesion.Th e incision should be deep enough to target the submucosal edge and the muscularis propria.Th en the endoscope is inserted through the submucosal layer using an oblique cap over the tip of the endoscope (ST Hood, Fujinon Co., Japan).A tunnel is created in the submucosa in the oral-anal direction through visually directed dissection of submucosal fi bers and clott ing of perforated esophageal vessels back into the muscularis propria.Aft er the tunnel has been created and the anal margin has been reached by the transverse incision, dissect the submucosa in the transverse direction from one end to the other of the lesion.Th e fi nal steps more submucosal injections in the right and left side margins of the lesion followed by lateral incisions in an oral-anal clockwise direction to complete the procedure.Figures 1-9 are representative of the tunneling technique.classifi cation Vienna (11) which indicates the degree of tumor diff erentiation, invasion depth and whether or not resection was complete.In any case, the proximal, distal, lateral and vertical margins must be carefully evaluated.For surgical specimens containing the mucosa, submucosa, muscularis and adventitial, a semi-quantitative analysis of the submucosal invasion depth is reliable since the submucosa can be divided into three segments of equal thickness (SM1, SM2 and SM3).In pieces of endoscopic resection, the submucosa is not always complete in which case this distinction is less reliable.In these cases, a quantitative measurement is taken by measuring the invasion of the submucosa in microns (μ) from the last layer of the muscularis mucosa.A cut-off point above which it is considered there is an increased risk of metastasis in lymph nodes (SM2) must be established.In the esophagus this point is located at below 200μ (10).

ESD results in early esophageal cancer
Th ere are few studies in the literature on the use of ESD to resect esophageal tumors.A summary of the results of the main publications are described in Table 2. Th e fi rst experiments with the use of ESD in the esophagus and colon were carried out only fi ve years aft er the fi rst use of ESD in the stomach.In 2005, Toyonaga et al. were among the fi rst to report the use of ESD for esophageal tumors (26).Th e authors reported 20 patients with esophageal SCC who underwent ESD with FK.Successful resections with free margins were performed in 100% of these cases.Th e average diameter of the samples was 47 mm, and the average time of the procedures was 65 minutes.In this series the sole complication, mediastinal emphysema, was managed clinically.Fujishiro et al. (24) showed similar results in 43 patients.En bloc resections were performed in 100% of these cases, but in 22% of the cases endoscopic resection was not curative due to compromises at the vertical or lateral margins.Th ere were four cases of perforations which were treated conservatively.Ishihara et. ( 32) Published an analysis comparing the results of ESD and mucosectomies performed on patients with esophageal cancers of less than 20 mm.Th e en bloc resection and curative resection rates for ESD cases were superior to those for endoscopic mucosal resection with cap and with strip biopsies (100% and 97%, 87% and 71%, 71% and 46 % respectively).Th ere were no diff erences in complication rates among the three techniques.It should be noted that when the endoscopist is suffi ciently trained, the ESD complication rate is similar to that for mucosectomies.
Only recently have the fi rst two experiments with ESD for esophageal cancer been done in Europe and Brazil (Table 2) (33,34).

Recovery and preparation of the resected specimen
Th is is a fundamental step in the endoscopic treatment of superfi cial tumors that is oft en ignored by Western endoscopists, but that is carried out systematically by the Japanese.Th e recovered sample is fi xed onto Styrofoam or rubber with pins and placed in formalin.Th e pathologist must cut the sample into parallel fragments that are 2 mm wide.Biopsy samples should be evaluated according to the At the Alfa Institute of Gastroenterology, ESD procedures in the esophagus began to be performed routinely in 2009.Th is began with training of the primary author (VA) of this article at the National University Hospital of Singapore and the Kobe University Hospital in Japan.So far, 22 ESD procedures have been conducted for esophageal cancer at this center.Th e tunneling method has been used for these procedures.It has achieved an en bloc resection rate of 91% and a curative resection rate of 77.3%.Th ese are in line with the results found in the world's literature (24,33,35,36).Th ere were two complications in this series.One patient developed mediastinal and subcutaneous emphysema without apparent perforation.Th is patient was treated conservatively.Th e other case had esophageal perforation which was treated by applying clips and clinical management.No cases of local recurrence occurred during the follow-up periods which ranged from 3 months to 2 years (data submitt ed for publication).

FINAL CONSIDERATIONS
ESD is a reality in Asia where it is considered the treatment of choice for early SCC.Th e experience is still insuffi cient to conclude whether it is feasible to apply ESD for large scale treatment of superfi cial esophageal cancer in western centers as usually happens today in Japan and South Korea.In Western countries, especially in Latin America, the greatest challenge is to promote early diagnosis of esophageal neoplasia through proper training of endoscopists and the establishment of screening programs for high-risk patients.It is also essential to build referral centers with well-trained specialized human resources supported by a complete infrastructure so that the use of ESD can expand in these countries with the obvious benefi ts it can bring for quality of life and reduced morbidity and mortality.
• A CO 2 insuffl ator.• An electrosurgical unit especially for use in endoscopy.Currently the only validated models for use in ESD are ERBE electrocautery units (ICC 200, 200 and VIO VIO 300) which have the pulsating endocut mode in addition to soft ware for dry-cut, soft coagulation, coagulation and swift forced coagulation.Each operator willing to perform ESD must have deep knowledge of electrosurgical properties and of the specifi c parameters for each step of the procedure.A number of stilett os and products developed by Japanese experts are now among the accessories used for ESD.Th ey include: • Stilett os: Flush-Knives (straight and ball-tipped), IT-Knives, Hook Knives, Flex Knives, Dual-Knives, Hybrid Knives, Safe Knives, and Swam Blade Knives (Figure 1).• 25 gauge catheters for submucosal injection.• Hemostasis: hot biopsy forceps (hot biopsy) or coagulation forceps.• Disposable endoclips for perforation management.

Figure 1 .
Figure 1.Chromoscopy using Lugol's solution revealing three fl at synchronous neoplastic lesions from the distal third of the esophagus which were Lugol's negative.Biopsies were consistent with squamous cell carcinoma.

Table 1 .*
Parameters recommended by the University of Kobe for endoscopic submucosal dissection of the esophagus using a balltipped fl ush Knife and VIO 200 or 300 electrosurgical unit (ERBE Elektromedizin, Turbingen, Germany).In case of fi brosis use Swift Coagulation.

Figure 3 .
Figure 3. Transverse incision in the anal margin of the lesion.

Figure 4 .
Figure 4. Transverse incision in oral margin of the lesion.

Figure 5 .
Figure 5. Positioning oblique and long cap (ST Hood) at the tip of the endoscope in preparation for creation of the submucosal tunnel in the esophagus.

Figure 6 .
Figure 6.Creation of the submucosal tunnel reaching the anal margin incision of the lesion.

Figure 7 .
Figure 7. Final appearance of the ESD completed in 90 minutes with completed en bloc resection.Note that the resection site corresponds to 80% of the circumference of the esophagus.

Figure 8 .
Figure 8. Specimen stained with Lugol and fi xed on Styrofoam plate with pins for histological analysis.Histology showed squamous cell carcinoma invading into the muscularis mucosa (M3) with deep free lateral margins.

Figure 9 .
Figure 9. Endoscopic follow-up 60 days later.Chromoscopy using Lugol's solution showed regular scar with no residual lesions or stenosis.