andrew k. roorda, m.d., series editor radiofrequency

PRACTICAL GASTROENTEROLOGY • JUNE 201020

Radiofrequency Ablation: An Overview

INTRODUCTION

Barrett’s esophagus (BE) is defined as a metaplas-tic change from the squamous epithelium tocolumnar epithelium containing goblet cells in

the lining of a portion of the distal tubular esophagus,and occurs as a result of chronic injury and inflamma-tion from acid and bile exposure characteristic of gas-troesophageal reflux disease (GERD). If untreated,approximately 10% of BE patients may developesophageal adenocarcinoma (EAC), a condition that isincreasing in the developed world.1 The prevalence ofBE has been difficult to estimate due to varying defin-

itions, but estimated rates of 3–5.6% have beendescribed in the general population.2, 3 More recentdata have shown that the prevalence of BE is on therise.4, 5 Increased detection may be due to more fre-quent use of endoscopy or a higher prevalence of riskfactors associated with the development of BE. Theserisk factors include long standing symptomatic reflux,male gender, older age, Caucasian race, alcohol use,smoking, and obesity.6

Prior to the development of EAC, BE is thought totransition from intestinal metaplasia to low- and/orhigh-grade dysplasia (LGD, HGD). The current man-agement strategy of endoscopic biopsy surveillanceaims to detect disease progression at an early enoughstage to allow for intervention. Previously, esophagec-tomy was the primary management option for HGDand intramucosal cancer. However, esophagectomy is

ENDOSCOPY: OPENING NEW EYES, SERIES #4

Omar S. Khokhar, M.D., Senior Gastroenterology Fel-low and Firas H. Al-Kawas, M.D., Professor of Medi-cine, Georgetown University Hospital, Washington,DC.

Andrew K. Roorda, M.D., Series Editor

Barrett’s esophagus (BE) is a premalignant condition with a small but real risk of pro-gression to esophageal adenocarcinoma. Development of cancer is thought to arise as thefinal step in the metaplasia-dysplasia-carcinoma sequence. Therefore, recent effortshave focused on endoscopic therapies that intervene early, with the goal of arresting pro-gression of intestinal metaplasia to dysplasia, and eventually, to carcinoma. To this end,several ablative techniques have been applied. This article summarizes recent dataregarding the efficacy and safety of radiofrequency ablation in the treatment of BE. Theindications, technical aspects, and cost-effectiveness are also reviewed.

Omar S. Khokhar Firas H. Al-Kawas

PRACTICAL GASTROENTEROLOGY • JUNE 2010 21


Radiofrequency Ablation

associated with significant mortality and morbidityeven in experienced centers.7 Therefore, over the lastdecade, various endoscopic methods have beenemployed for the treatment of HGD in an effort toavoid surgery. These include ablative techniques suchas argon plasma coagulation (APC), laser therapy, pho-todynamic therapy (PDT), and more recently, radiofre-quency ablation (RFA) and cryotherapy. Endoscopicmucosal resection (EMR) is another increasingly usedtechnique for focal lesions.8 Until recently, PDT wasthe most studied modality, however, photosensitivity,stricture formation and incomplete eradication of BEhave prohibited PDT from becoming a definitive ther-apy. Focal EMR is beneficial through its provision of atissue sample for diagnosis and staging, however, theresidual Barrett’s tissue remains at risk for progression,and cancer recurrences have been reported in up to 30%of patients.9,10 With stepwise radical EMR, the removalof all pathologic tissue is appealing. However, the pro-cedure is technically demanding and associated withsignificant stricture rates.9,11,12

More recently, endoscopic RFA has emerged as animportant modality in the management of patients withBE. This article will focus on the appropriate indica-tions, procedural aspects, potential complications, andefficacy of RFA.

RADIOFREQUENCY ABLATION OF BARRETT’S ESOPHAGUSAs it pertains to BE, ablation is synonymous withdestruction and removal of intestinal metaplasia andall grades of dysplasia. The metaplastic tissue is sub-sequently replaced with neo-squamous epithelium.The intent of RFA in this setting is to permanentlyremove all cells that may include or progress to cancer.

Prior to ablation, a meticulous white light examina-tion of the lower esophagus must be performed. Thegastroesophageal (GE) junction and extent of Barrett’stissue should be carefully assessed and recorded. Biop-sies should be taken in a four quadrant manner fromevery centimeter of suspected Barrett’s tissue. Any focalabnormalities should be sampled. Nodules should beevaluated further with endoscopic ultrasound andremoved with EMR. Our policy is to perform EMR inall patients with HGD on biopsy even if no focal lesions

are present. Limited data suggest that this approach willfurther improve histologic staging.13 If available,advanced imaging techniques such as high definitionendoscopy, dye chromoendoscopy or digital chromoen-doscopy (i.e. narrow band imaging [Olympus], i-SCAN[Pentax], or FICE [Fujinon]) may be employed to accu-rately delineate the extent and margin of BE or anyabnormal focal areas. In cases of HGD and early EAC,an expert gastrointestinal pathologist should be con-sulted for review and confirmation of diagnosis. Aspirinand other platelet-inhibiting medications should be heldfor one week prior to the procedure.

RADIOFREQUENCY ABLATION SYSTEM

Circumferential Ablation SystemThe ablation system (BÂRRX Medical, Inc., Sunny-vale, CA) consists of a high-power energy generator, aballoon-based sizing catheter (HALO360) (Figure 1A)with a maximum diameter of 33.7 mm, and a balloon-based ablation catheter (HALO360+) (Figure 1B). Theablation catheter diameters range in size from 18 to 31mm. The energy generator provides automated, pres-sure-regulated inflation of the sizing balloon and abla-tion catheters. Prior to ablation, the esophageal innerdiameter is measured with a sizing balloon (4 psi) toensure appropriate ablation catheter selection. Once anaccurate measure of esophageal diameter is made andrecorded, an appropriately sized ablation catheter isselected and exchanged over the guidewire. The abla-tion catheter consists of a 4 cm long noncompliant balloon, over which a 3 cm long flexible bipolarmicroarray circuit is attached. The array consists of 60 independent electrodes that encircle the balloon. Theelectrodes are tightly spaced (250 μm) and alternate inpolarity (plus/minus). The ablation catheter is inflated toa maximum pressure of 4.5 psi, which causes transientflattening of esophageal folds and tension along theesophageal wall. This close approximation of theinflated ablation catheter with the esophageal wall, cou-pled with the bipolar and spacing characteristics of theelectrode, allow optimal energy current delivery over aprecise area and depth. Energy delivery is initiated witha foot-activated pedal. The energy generator deliversnearly 300 W of RF energy over a very short period of




time (300 ms). The energy delivered is recorded injoules per unit surface area of the electrode (J/cm2). Thisamount of standardized energy generally does not pene-trate the muscularis mucosae layer.14

Focal Ablation SystemAblation may also be delivered focally through asmaller electrode (HALO90) (Figure 2). As with cir-cumferential ablation, energy is produced by a genera-tor and delivered to the focal ablation device. Incontrast to circumferential ablation, a sizing step is notnecessary. The focal ablation catheter consists of anarticulated platform (20 × 13 mm) which is coveredwith an electrode array similar to that of the circumfer-ential catheter. After mapping endoscopy is performed,the focal ablation catheter is placed on the tip of theendoscope, which is then re-inserted into the esopha-gus. Once the area of interest is identified, the tip of theendoscope is deflected upwards to approximate the

electrode surface with the targeted tissue. Energy isthen transmitted to the tissue through a foot-pedal sys-tem similar to that employed in circumferential abla-tion. Focal ablation may be used as primary ablation forshort (≤2 cm) segment BE or disease localized to theGE junction, and as secondary ablation for any residualtongues or islands of BE after primary treatment withcircumferential RFA, EMR, or photodynamic therapy.

INDICATIONS FOR RADIOFREQUENCY ABLATIONThe circumferential and focal radiofrequency ablationsystems received 510(k) clearance from the FDA in2001 and 2006, respectively. These systems are indi-cated for use in the coagulation of bleeding and non-bleeding sites in the gastrointestinal tract including,but not limited to, the esophagus (Table 1). RFA isaccepted as an appropriate management option in BEcontaining dysplasia. Evaluation of RFA in the settingof non-dysplastic BE (NDBE) is ongoing, with somephysicians offering RFA therapy to select patients (i.e.younger, long-segment disease, family history of EAC,significant patient anxiety, etc.).15 The emergence oflong-term data in this patient group will further clarifyoptimal clinical management.

TECHIQUES OF RADIOFREQUENCY ABLATION

Circumferential RFACircumferential ablation is most commonly performedin the outpatient setting. During endoscopy, the top ofthe intestinal metaplasia (TIM, i.e. Z-line) and top of the

(continued on page 24)

Figure 1A. Sizing balloon catheter (Courtesy of BÂRRX Medical, Sunnyvale, CA).

Figure 1B. Circumferential balloon-based ablation catheter(Courtesy of BÂRRX Medical, Sunnyvale, CA).

Figure 2. Focal ablation device with endoscope-mounted,articulated electrode (Courtesy of BÂRRX Medical, Sunnyvale, CA).




gastric folds (TGF) should be recorded. The esophagusis then irrigated with 1% N-acetylcysteine to removemucus that may interfere with energy transfer. This irri-gant is evacuated from the stomach prior to initiation ofablation. No lubricant is to be used during the proce-dure. A guidewire is inserted through the biopsy chan-nel into the stomach and the endoscope is removed. Thesizing catheter is passed over the guidewire and posi-tioned 12 cm above the TGF (TGF-12), which is thestarting point for sizing measurements. Sizing is gener-ally performed blindly, however, the endoscope may beintroduced in a side-by-side manner to size under directvisualization, if desired. Advancing proximal to distal,sizing is performed every 1 cm. At each level, theenergy generator inflates the sizing catheter up to a pres-sure of 4 psi, and measures the esophageal inner diame-ter through the application of a pressure:volumealgorithm. The generator displays the inner diameter ofthe esophagus and the associated recommended ablationcatheter size, which are recorded on a sizing worksheet.After inflation and sizing, the balloon automaticallydeflates. The endoscopist then advances the sizingcatheter 1 cm distally, and repeats the sizing steps (Figure 3A). An abrupt increase in diameter indicatesthat the sizing balloon has passed into a hiatal hernia orthe gastric cardia. After the last sizing step, the deflatedsizing catheter is then exchanged over the guidewire.

The physician reviews the sizing worksheet andselects the smallest sized ablation catheter that wasrecommended over the course of sizing. Even if multi-ple sizes were recommended during sizing, it is imper-ative to select the smallest sized ablation catheter tooptimize safety. This ablation catheter is passed overthe guidewire and positioned such that the top of theelectrode array is 1 cm above the TIM (Figure 3B).The endoscope is reintroduced next to the catheter andpositioned approximately 1–2 cm proximal to the bal-loon so as to enable visualization of ablation. Thephysician selects the appropriate energy density to usebased on grade of BE (12 J/cm2 for dysplastic BE, 10J/cm2 for non-dysplastic BE). When the ablationcatheter and endoscope are satisfactorily positioned,the operator initiates ablation. Energy is then deliveredat the pre-determined dose (Figure 3C). After the ini-tial ablation, the electrode will automatically deflateand is then advanced distally under direct visualiza-tion. It is critical to align the proximal margin of theelectrode with the distal margin of ablated tissue toensure continuity of ablation (Figure 3D). The processof ablation is repeated until the GE junction is reached(Figure 3E), after which the ablation catheter, endo-scope and guidewire are slowly removed under directvisualization. After removal, the electrode is cleanedto remove mucus and coagulum. A small transparent

(continued from page 22)

Table 1. Indications for Radiofrequency Ablation

Condition Type of Ablation

Intestinal metaplasia (Barrett’s esophagus) Circumferential/focal• High grade dysplasia (HGD)• Low grade dysplasia (LGD)*• Non-dysplastic intestinal metaplasia*

Gastrointestinal bleeding Focal• Gastric antral vascular ectasia (GAVE)• Radiation proctitis (RP)

Squamous cell carcinoma (early)* Circumferential/focal

*Data is accruing regarding long-term efficacy and cost-effectiveness in these conditions. The decision to use RFA in these disease states should be individualized.




3A.

3B.

3C.

Figure 3. Sizing commences 12 cm above the top of gastricfolds, and measurements of the esophageal inner diameterare taken every centimeter as the balloon is moved distally(Courtesy of BÂRRX Medical, Sunnyvale, CA).3A. Sizing balloon in position. Esophageal inner diameter

and recommended ablation catheter size are reported by the generator and recorded on a worksheet. A suddenincrease in diameter indicates passage of balloon intogastric lumen.

3B. Ablation catheter exchanged over guidewire and posi-tioned 1 cm above top of intestinal metaplasia.

3C. Inflation of electrode under endoscopic visualizationand delivery of energy.

3D. Deflation and repositioning of electrode distally toablate additional tissue.

3E. Final results after one pass.

3E.

3D.


plastic cap (HALO CAP) is affixed to the tip of theendoscope, which is then reintroduced into the esoph-agus. Using the plastic cap, the previous area of abla-tion is cleaned by gently pushing the coagulum into thestomach, applying small volumes of plain water irriga-tion as needed. Removal of coagulum from both theelectrode and esophageal surface ensures an optimalsecond ablation. After cleaning, the scope is removedand the ablation steps are repeated so that the abnormal

tissue receives a total of two ablation applications.Endoscopic images from a representative circumferen-tial RFA case are shown in Figure 4.

Focal RFAThe focal RFA catheter consists of a flexible siliconesleeve that fits on the tip of a standard endoscope. Thecatheter is attached in such a manner that the electrodesurface is positioned at 12 o’clock as depicted on thevideo monitor. Similar to circumferential ablation, theesophagus is irrigated with 1% N-acetylcysteine, whichis evacuated from the stomach prior to initiation of abla-tion. The electrode is positioned under the targeted tis-sue, and contact is achieved by deflecting the endoscopetip upwards. Once tissue contact has been established,energy is delivered twice in a standardized manner (12J/cm2). After all targeted tissue has undergone ablation,the tip of the electrode is used to remove the overlyingcoagulum. The endoscope is then removed to clean theelectrode surface. The endoscope and electrode are rein-troduced, and abnormal tissue is ablated again twice, fora total of four energy applications.



4A. Abnormal Barrett’s tissue in the distal esophagus priorto ablation.

Figure 4. Endoscopic appearance before and after treatmentwith circumferential ablation catheter.

4C. Lower esophagus at three month follow-up. Endoscopyreveals an irregular Z-line and two focal islands of per-sistent IM, which were treated with focal ablation. Onfollow-up 3 months later there was no endoscopic evi-dence of metaplastic columnar tissue and all biopsieswere negative for IM (CR-IM).

4B. Typical appearance of white coagulum after first ablationpass.

POST-ABLATION MANAGEMENTThe standard medication regimen on dischargeincludes maintenance high dose proton pump inhibitortherapy, liquid acetaminophen with codeine, liquidantacid/lidocaine, sucralfate, and antiemetics. Dietshould be fully liquid for 24 hours, which may beadvanced to soft diet for seven days. Platelet inhibitingagents (i.e., aspirin, clopidogrel, non-steroidal anti-inflammatory agents) and anti-thrombotic agents (i.e.,heparin, warfarin) may be resumed one week after pro-cedure. Proton pump inhibitor therapy should be con-tinued indefinitely for symptomatic relief andprevention of disease recurrence.

After initial ablation (generally circumferential),follow-up focal ablation may be performed. Ablationshould be performed at least two months after previousablation to allow adequate healing of esophagealmucosa. After complete endoscopic resolution is seen,biopsies should be taken per surveillance protocol toconfirm histological eradication.

OUTCOMES OF RADIOFREQUENCY ABLATION:CLINICAL DATA SUMMARY

Optimal Ablation Settings and SafetyAn early multi-phase study by Ganz and colleaguesreported the effect of ablation on esophageal epithe-lium.14 Using a porcine model, the study evaluated theenergy density necessary to completely ablateesophageal epithelium, as well as the ablation depthand stricture rate at various energy density and powerdensity settings. The final phase assessed the extentand degree of ablation in patients undergoingesophagectomy for EAC. Phase I of this trial demon-strated that energy density settings of 9.7 to 29.5 J/cm2

completely eliminated esophageal epithelium. Phase IIshowed that stricture frequency and severity weredirectly proportional to energy density, with >20 J/cm2

resulting in universal stricture formation, and 10.6J/cm2 or lower not causing any stricture formation.Phase III showed that the amount of energy densityapplied was directly proportional to the depth of abla-tion confirmed on histology. Phase IV suggested thatenergy densities of 10 and 12 J/cm2 were sufficient to

completely ablate epithelium without direct injury tosubmucosa or stricture formation.

In an effort to determine optimal RFA parameters(energy density and number of applications) necessaryto complete ablation of squamous esophageal epithe-lium without damage to deeper layers, 13 EACpatients underwent RFA prior to esophagectomy.16

Subjects were randomized to receive energy densitysettings of 8, 10, or 12 J/cm2 prior to surgery. Aftersurgery, histologic examination of the ablation zoneswas performed to assess depth and completeness ofablation, and thickness of residual ablation effect aftertissue necrosis. The results demonstrated that completeablation was achieved consistently at 10 J/cm2 deliv-ered twice (2×) and 12 J/cm2 1×–2×, whereas 8 J/cm2

1×– 2× and 10 J/cm2 1× only partially ablated theepithelium.

In order to evaluate optimal settings for ablation ofIM with HGD, eight patients with IM-HGD underwentRFA prior to esophagectomy.17 Energy was applied at10, 12, or 14 J/cm2 for 2×, 3×, or 4× applicationsrespectively. After surgery, esophageal specimensfrom ablation zones were inspected for maximumablation depth and complete ablation of all IM-HGD.The deepest ablation was achieved at an energy dis-bursement of 14J/cm2 4×; this setting led to edema ofthe superficial submucosa. Complete ablation wasachieved in 9 of 10 ablation zones. At 12 J/cm2 2×there was a single focus of HGD that remained at themargin of 1 out of 10 ablation zones. This was attrib-uted to incomplete overlap of the second ablationapplication. No specimens showed submucosal injury.The authors concluded that the maximum ablationdepth increased with the amount of energy densityapplied and number of applications.

The Ablation of Intestinal Metaplasia (AIM) trialsought to determine the dose-response and safety ofdelivering 6 to 12 J/cm2 with a single application. Out-comes at 1 and 3 months demonstrated similar histo-logic response rates for the 10 and 12 J/cm2 groupsthat were superior to those of the 6 and 8 J/cm2 groups.In addition, the adverse events were transient and mild(chest pain, focal scarring, fever, etc.), with rates com-parable for all groups. As a result, the initial dosimetryphase allowed for the selection of 10 J/cm2 2× for theeffectiveness phase of the trial.







Clinical Outcomes in BE PatientsThe effectiveness phase of the AIM Trial evaluatedcircumferential RFA in 70 patients with NDBE. At 12months 70% of patients demonstrated a completeresponse (CR), defined as all biopsy specimens nega-tive for BE, after a mean of 1.5 ablation procedures.There were no adverse events, such as stricture forma-tion, or buried glandular mucosa.18 The AIM-II studyfollowed these patients prospectively to assess long-term safety and efficacy of stepwise circumferentialablation followed by focal ablation.19 After undergo-ing a mean of 1.9 focal ablation sessions, CR-IM wasattained in 98% of patients (60 of 61 patients; per pro-tocol analysis) with 2.5 years of follow-up. There wereno strictures or buried glands at either 12 or 30 months.

A large multicenter trial recently assessed thesafety and efficacy of circumferential ablation fortreating BE with HGD.20 142 patients with HGD(median length of 6 cm) received circumferential abla-tion (median 1 treatment session) with median follow-up of 12 months. Follow-up biopsies were performedand evaluated for histological response, defined asnegative for HGD (CR-HGD), negative for dysplasia(CR-D), and negative for IM (CR-IM). Outcomes fromthis study were CR-HGD 90.2%, CR-D 80.4%, andCR-IM 54.3%. Limitations of this particular studyincluded non-randomized study design, lack of controlarm, lack of centralized pathology review, and lack offollow-up beyond 12 months.

A single-center prospective study evaluated step-wise circumferential and focal ablation for the eradica-tion of IM and LGD over 2 years.21 Ten patients withLGD underwent circumferential ablation at baselinewith repeat circumferential ablation at 16 weeks forany residual IM. Endoscopy with biopsies was per-formed at 1, 3, 6, 12 and 24 months. After 1 year, focalablation was performed for any residual IM. Patientswere also prescribed lansoprazole 30 mg twice daily.At 2 years, CR-D was 100% and CR-IM was 90%.There were no strictures or evidence of buried glands.One patient with persistent IM at 2 years received oneadditional focal ablation (off-trial) and was eventuallycleared of IM.

Gondrie and colleagues examined a prospectivecohort of 11 patients with BE.22 After visible abnor-malities were resected with EMR (6 patients), 2

patients remained with LGD and 9 with HGD. Afterundergoing a median of 2 circumferentialand 2 focalablations, and an ‘‘escape’’ EMR in one patient, all 11patients had complete eradication of dysplasia and IM.Another study published by the same group evaluatedthe safety and efficacy of using stepwise circumferen-tial and focal RFA for BE containing flat HGD or BEwith residual dysplasia after EMR for HGD and/orintra-mucosal carcinoma (IMC).23 A total of 12patients were enrolled, 11 with HGD and 1 with IMC.After a median of 1 circumferential ablation, 2 focalablations, and an ‘‘escape’’ EMR in one patient, CR-Dand CR-IM was achieved in all 12 patients.

A recent multicenter, prospective cohort studyassessed the efficacy of combination therapy withfocal EMR and RFA.24 Pouw and colleagues examined24 patients with BE containing HGD or early cancer(median length 8 cm). Visible lesions were removedvia EMR followed by RFA of the residual flat Barrett’s >6 weeks later. Focal escape EMR was usedfor persistent Barrett’s mucosa after RFA. Neoplasiaand intestinal metaplasia were eradicated in 95% and88% of patients, respectively; after escape EMR in 2patients, rates improved to 100% and 96%, respec-tively. Follow-up (median 22 months) revealed norecurrence of neoplasia.

A recent single center experience was reported inpatients with all grades of BE.25 Sixty-six patients witha median Barrett’s length of 3 cm (range, 1–14) under-went stepwise circumferential and focal RFA. Twelvepatients (18%) had HGD and 7 patients were post Nis-sen fundoplication. No immediate procedure-relatedcomplications were reported. Four strictures werereported in patients with >6 cm segments of Barrett’s.Six patients had a Nissen fundoplication performed atthe same time of ablation, leading to the conclusionthat RFA did not hinder fundoplication.

A randomized, multicenter, sham-controlled trialassessed the utility of RFA in 127 patients with dys-plasia (AIM Dysplasia Trial).26 Patients were random-ized to RFA or sham in a 2:1 ratio, and stratifiedaccording to grade of dysplasia and BE length (<4 vs.4–8 cm). Using intention-to-treat analysis, completeeradication of dysplasia following ablation was seen in



90.5% of patients with LGD compared with 22.7% inthe control group. Among patients with HGD, com-plete eradication was seen in 81% in the ablation groupcompared with 19% in the control group. Overall,77.4% of patients in the ablation group had completeeradication of intestinal metaplasia in contrast to 2.3%of those in the control group. In addition, patients whounderwent ablation had less disease progression (3.6%vs. 16.3%) and fewer cancers (1.2% vs. 9.3%) thanthose that received the sham procedure at 12-monthfollow-up.

For long-segment BE containing neoplasia, Her-rero and colleagues performed RFA in 26 patients.27 Onbaseline EMR, 11 patients had EAC, 6 HGD, and 1LGD; after EMR and prior to RFA 16 had HGD and 10LGD. All patients had BE segment length ≥10 cm.Patients underwent focal endoscopic resection for visi-ble lesions, followed by circumferential and focal RFAevery 2–3 months until complete remission wasachieved. Follow-up endoscopy with biopsies was per-formed at 2, 6, and 12 months. Twelve percent ofpatients showed poor healing after RFA, and stenosisand scarring after EMR resulted in superficial lacera-tion after circumferential RFA in 27% of patients. Aftera median follow up of 9 months, no recurrence of neo-plasia was found. No buried BE was found in 752 neo-squamous biopsies. A complete response was seen in14/17 (83%) patients, suggesting that long segment BEmay be safely and effectively treated with RFA.

Lyday and colleagues recently demonstrated afavorably safety and efficacy profile of RFA in a com-munity setting.28 Four hundred twenty-nine patientsunderwent a total of 788 ablations for NDBE, LGD,and HGD. Efficacy was reported in 2 cohorts; Effi-cacy-A (subjects with any post-RFA biopsy despitesome not having completed therapy) and Efficacy B(subjects with post-RFA biopsy >1 year post-RFA).Complete response to dysplasia and IM was seen in89% and 72% of subjects in Cohort-A, respectively,and in 100% and 77% of subjects, respectively, inCohort-B. The investigators concluded that data fromcommunity endoscopy centers compared favorably todata from specialized centers.

A recently published study examined 27 patientswith either NDBE or LGD. The average length of Barrett segment treated was 4.6 cm. Ablation was per-

formed with follow up EGD at 8 weeks. In all patients,the BE was replaced with normal squamous epithe-lium. GERD symptoms regressed in 16 patients (60%)with RFA and proton pump inhibitor therapy.29

Durability of CureThe longest term follow-up to date is from the AIM IIpatient cohort. Fifty patients with baseline NDBE andCR-IM at 2.5 years recently underwent extensive 5-year endoscopic biopsy follow-up, with a mean of 31biopsy specimens obtained per session. CR-IM wasmaintained in 92% of patients. There was no diseaseprogression; in the four patients with recurrence, eachhad NDBE. After one focal ablation session andbiopsy follow-up 2 months later, each of these 4patients resumed CR-IM status.30 Two- and three-yeardurability data from the AIM Dysplasia Trial hasrecently been reported by Shaheen et al. Of the 62patients who were CR-IM at 1 year, 59 remained CR-IM (95.2%, per protocol) at 2 year follow-up, and ofthe 13 patients with CR-IM at 1 year who reached 3-year follow-up, 13 (100%) remained CR-IM.31

Effect of Ablation on Esophageal Cancer ProgressionThe extent of reduction of esophageal adenocarcinoma(EAC) incidence in BE patients after endoscopic abla-tion was the subject of a recent meta-analysis.32

Patients with NDBE, LGD, or HGD and follow-up ofat least 6 months were included. Using weighted-aver-age incidence rates, the rate of cancer in patientsundergoing ablative therapy was compared to cancerrates calculated from natural history data. The naturalhistory vs. post-ablation progression rates to cancerper patient year of follow-up were 6.6% vs. 1.7% forHGD, 1.7% vs. 0.16% for LGD, and 0.6% vs. 0.16%for NDBE. The investigators concluded that ablationmay be associated with a reduction in cancer inci-dence, with greatest benefit observed in BE patientswith HGD.

Role of Acid SuppressionTo examine the role of acid suppression in ablation




efficacy, Roorda and colleagues combined circumfer-ential RFA and twice daily PPI therapy.33 Thirteenpatients were included, three with HGD, four withLGD, and six with NDBE. Circumferential RFA wasfollowed by twice daily PPI and surveillance every 3months for a total of 12 months. Patients received asecond ablation if any metaplasia or dysplasia wasseen at follow-up. After a mean of 1.4 ablation ses-sions, complete eradication was achieved in 6/13NDBE patients and 5/7 dysplastic BE patients. A 24-hrambulatory esophageal pH of less than 4 was seen lessthan 4% of time in only 5 of 13 patients. The data sug-gest that optimal pH control leads to more favorableresponse to RFA.

Effect of Ablation on Genetic MarkersTwo clinical studies have evaluated the effect of abla-tion on genetic abnormalities present in abnormalepithelium. In the first study, 16 patients with BE-LGDreceived a single ablation, while 5 received two abla-tions.34 Following a 2.5-year follow-up period, 51micro-dissection specimens were analyzed for a panelof 16 allelic imbalance mutational markers affecting1p, 3p, 5q, 9p, 10q, 17p, 17q, and 21q. RFA achievedCR-IM in 15/16 patients (94%), and previously pre-sent mutations were no longer detectable in anypatients. In the one patient with persistent metaplasia,only highly expanded mutations were present, sug-gesting that highly clonally expanded mutations weremore resistant to regression and required follow-upablation. In addition, the investigators noted that muta-tional regression was time-dependent and could occur6–12 months following treatment. In the second study,Pouw and colleagues assessed the elimination of pre-existing genetic abnormalities in dysplastic BE afterstep-wise circumferential and focal ablation.35

Twenty-two patients with dysplastic BE characterizedby abnormal Ki67 and p53 staining along with fluo-rescent in-situ hybridization (FISH) abnormalities atbaseline underwent stepwise circumferential and focalablation. Two months after the focal ablation, CR-Dand CR-IM were achieved in all patients (100%), andsubsequent biopsies of the neo-squamous epitheliumrevealed complete eradication of all oncogeneticmarkers seen at baseline.

Initial Ablation Data for Gastrointestinal BleedingData from two tertiary care centers have suggested thatfocal segmental ablation using HALO90 is safe andefficacious for treatment of gastric antral vascularectasia (GAVE) and chronic hemorrhagic radiationproctitis (RP). Ablation in both of these settings wasassociated with increased baseline hemoglobin anddecreased transfusion requirement.36,37 Six patientswith GAVE (4 men, mean age 58 years, range 47–65years) underwent ablation of antral lesions (mean focalablation treatments 1.7). The mean Hgb levelimproved from 8.6 to 10.2 g/dl (mean 2 months afterlast ablation). Five of six no longer required bloodtransfusions to maintain stable hemoglobin levels.Endoscopic images from a representative GAVE focalRFA case are shown in Figure 5.

Three patients with chronic hemorrhagic RPunderwent 1–2 focal ablation sessions. The authorsreported endoscopic findings of re-epithelialization ofsquamous mucosa over areas of prior hemorrhage. Nopatients required transfusions after ablation therapy.No stricturing or ulceration was seen on follow-up upto 19 months after RFA treatment. Endoscopic imagesfrom a representative RP focal RFA case are shown inFigure 6.

31



Figure 5A. Endoscopic view of GAVE prior to treatment(Courtesy of David L. Diehl, M.D. of Geisinger Health System,Danville, PA).



Initial Ablation Data for Squamous Cell CarcinomaPouw and colleagues reported the case of a 66-year-old male with a 35-mm flat-type early squamous cellcarcinoma surrounded by high grade dysplasia. Abla-tion resulted in complete endoscopic and histologicaleradication of cancerous tissue. No adverse eventswere reported.38 More recently, Zhang et al. havedescribed their early clinical experience with RFA forsuperficial esophageal squamous neoplastic lesions,with focus on patient selection and technique. Theauthors report that appropriate RFA candidates haveflat-type esophageal squamous cell neoplasia (type 0-IIb) on endoscopy, and a pre-treatment biopsy diagnosis of medium- or high-grade intraepithelialneoplasia (MGIN, HGIN), or well-to-moderately dif-ferentiated invasive cancer limited to the lamina pro-pria (m2).39

Clinical trials are currently underway in China, the U.S., and the Netherlands, with results eagerlyawaited.

COMPLICATIONS OF RADIOFREQUENCY ABLATIONComplications inherent to RFA include chest pain,chest tightness, dysphagia, and odynophagia. Gener-ally, these are self-limited (<3 days) and can be man-aged symptomatically.18,19 In rare instances, stricturemay develop at site of ablation requiring dilation. Thestrictures are generally described as easily amenable todilation, with resolution after 1–3 dilation ses-sions.20,26,28 Other complications that are common toall endoscopic procedures include esophageal perfora-tion, aspiration pneumonia, bleeding, and sedation-related adverse events. In larger trials in both academicand community practices (total ablations performed =1,170), complication rates were as follows: stricture(1.4%), mucosal injury (0.9%), chest pain (0.7%),bleeding (0.7%), dysphagia (0.4%), and fever(0.2%).28 Other reported complications included seda-tion-related transient airway obstruction in one patient,




Figure 5B. HALO90 ablation catheter being applied to thestomach with white coagulum evident (Courtesy of David L.Diehl, M.D. of Geisinger Health System, Danville, PA).

Figure 5C. Near complete resolution of disease following twoRFA treatments (Courtesy of David L. Diehl, M.D. ofGeisinger Health System, Danville, PA).


and sedation-related hypotension in one patient.18

There have been no RFA procedure-related deaths.

COST EFFECTIVENESS OF RADIOFREQUENCYABLATIONTwo recent economic decision analysis models haveexplored the cost utility of RFA vis-à-vis endoscopicsurveillance. Inadomi and colleagues examined patientpopulations with NDBE, LGD, or HGD.40 Strategiescompared were no endoscopic surveillance; endo-scopic surveillance with ablation for incident dyspla-sia; immediate ablation followed by endoscopicsurveillance in all patients or limited to patients inwhom metaplasia persisted; and esophagectomy. Abla-tion modalities that were modeled included RFA,APC, PDT, and multipolar electrocoagulation. Theauthors concluded that if RFA permanently eradicatedat least 28% of LGD or 40% of NDBE, ablation wouldbe preferred to surveillance. Overall, cost effectivenesswas dependent on the long-term effectiveness of abla-tion and whether surveillance endoscopy could safelybe discontinued after successful ablation. The secondanalysis by Das and colleagues employed similar com-

peting strategies but focused on patients with NDBEonly. Evaluation included natural history of NDBEwithout surveillance; surveillance performed per ACGpractice guidelines; and endoscopic ablation. This par-ticular model was biased against ablation with a con-servative estimate of complete response and continuedstandard surveillance even after complete ablation. Allpotential complications were accounted for, and anincomplete histological response after ablation waspresumed to have the same risk of progression asuntreated Barrett’s. Threshold analysis revealed thecritical determinants of ablation cost-effectiveness tobe rate of complete response to ablation, total cost ofablation, and risk of progression to dysplasia. Theauthors concluded that ablation for NDBE was morecost-effective than endoscopic surveillance.41

SUMMARYAvailable data suggest that RFA is a promising optionfor the management of dysplastic and nondysplasticBarrett’s esophagus. To date, RFA has been shown tobe highly effective in eradicating intestinal metaplasiaand associated dysplasia. Complication rates have




Figure 6A. Endoscopic appearance of radiation proctitis seenon retroflexion (Courtesy of F. Scott Corbett, M.D., of Gas-troenterology Associates of Sarasota, FL).

Figure 6B. After ablation of target site using HALO90 ablationcatheter (Courtesy of F. Scott Corbett, M.D., of Gastroen-terology Associates of Sarasota, FL.).

been very low in specialized centers, and data is nowaccruing that rates may be similar in trained commu-nity practitioners. Technically, in trained hands, RFA isstraightforward to perform with minimal side effects.

Questions, however, remain regarding optimalpatient selection and durability of cure. Long-term fol-low-up data are required to assess if RFA is adequateenough to either obviate the need for endoscopic sur-veillance or lengthen surveillance intervals in thesepatients. For patients with IMC, HGD, and LGD, RFA± EMR appears to be a clinically effective, cost-effec-tive, and safe management strategy. In patients withNDBE, the role of RFA is evolving. The decision toperform RFA in such patients should be individualizedin the context of the overall clinical scenario. Furtherstudies should focus on long-term follow-up data, anddevelopment of biomarkers to predict malignant pro-gression of IM. n

AcknowledgmentsThe authors wish to thank BÂRRX Medical, Inc. forimages and information discussed in the manuscript.

References1. Pohl H, Welch HG. The role of overdiagnosis and reclassification

in the marked increase of esophageal adenocarcinoma incidence.J Natl Cancer Inst 2005;97:142-6.

2. Hirota WK, Loughney TM, Lazas DJ, et al. Specialized intestinalmetaplasia, dysplasia, and cancer of the esophagus and esopha-gogastric junction: prevalence and clinical data. Gastroenterol-ogy 1999;116:277-85.

3. Hayeck TJ, Kong CY, Spechler SJ, et al. The prevalence of Bar-rett’s esophagus in the US: estimates from a simulation modelconfirmed by SEER data. Dis Esophagus 2010.

4. Kendall BJ, Whiteman DC. Temporal changes in the endoscopicfrequency of new cases of Barrett’s esophagus in an Australianhealth region. Am J Gastroenterol 2006;101:1178-82.

5. Gerson LB, Shetler K, Triadafilopoulos G. Prevalence of Barrett’sesophagus in asymptomatic individuals. Gastroenterology2002;123:461-7.

6. Wong A, Fitzgerald RC. Epidemiologic risk factors for Barrett’sesophagus and associated adenocarcinoma. Clin GastroenterolHepatol 2005;3:1-10.

7. Sihvo EI, Luostarinen ME, Salo JA. Fate of patients with adeno-carcinoma of the esophagus and the esophagogastric junction: a population-based analysis. Am J Gastroenterol 2004;99:419-24.

8. Wani S, Sayana H, Sharma P. Endoscopic eradication of Barrett’sesophagus. Gastrointest Endosc 2009.

9. Peters FP, Kara MA, Rosmolen WD, et al. Endoscopic treatmentof high-grade dysplasia and early stage cancer in Barrett’s esoph-agus. Gastrointest Endosc 2005;61:506-14.

10. May A, Gossner L, Pech O, et al. Local endoscopic therapy forintraepithelial high-grade neoplasia and early adenocarcinoma inBarrett’s oesophagus: acute-phase and intermediate results of anew treatment approach. Eur J Gastroenterol Hepatol2002;14:1085-91.

11. Peters FP, Kara MA, Rosmolen WD, et al. Stepwise radical endo-scopic resection is effective for complete removal of Barrett’sesophagus with early neoplasia: a prospective study. Am J Gas-troenterol 2006;101:1449-57.

12. Chennat J, Konda VJ, Ross AS, et al. Complete Barrett’s eradi-cation endoscopic mucosal resection: an effective treatmentmodality for high-grade dysplasia and intramucosal carcinoma—an American single-center experience. Am J Gastroenterol2009;104:2684-92.

13. Alan M, Michael JB, Luke FH, et al. Endoscopic Mucosal Resec-tion (EMR) for Barrett’s High Grade Dysplasia (HGD) and EarlyEsophageal Adenocarcinoma (EAC): An Essential Staging Pro-cedure with Long-Term Therapeutic Benefit. GastrointestinalEndoscopy 2009;69:AB348.

14. Ganz RA, Utley DS, Stern RA, et al. Complete ablation ofesophageal epithelium with a balloon-based bipolar electrode: aphased evaluation in the porcine and in the human esophagus.Gastrointest Endosc 2004;60:1002-10.

15. Fleischer DE, Odze R, Overholt BF, et al. The Case for Endo-scopic Treatment of Non-dysplastic and Low-Grade DysplasticBarrett’s Esophagus. Dig Dis Sci 2010.

16. Dunkin BJ, Martinez J, Bejarano PA, et al. Thin-layer ablation ofhuman esophageal epithelium using a bipolar radiofrequency bal-loon device. Surg Endosc 2006;20:125-30.

17. Smith CD, Bejarano PA, Melvin WS, et al. Endoscopic ablationof intestinal metaplasia containing high-grade dysplasia inesophagectomy patients using a balloon-based ablation system.Surg Endosc 2007;21:560-9.

18. Sharma VK, Wang KK, Overholt BF, et al. Balloon-based, cir-cumferential, endoscopic radiofrequency ablation of Barrett’sesophagus: 1-year follow-up of 100 patients. Gastrointest Endosc2007;65:185-95.




Practical Points• RFA ± EMR is a potential alternative to esophagectomy

or alternative endoscopic management strategies forintramucosal esophageal adenocarcinoma, high-gradedysplasia and low-grade dysplasia.

• The role of RFA in non-dysplastic metaplasia is evolving. The decision to perform RFA in these patientsshould be individualized.

• Short- and mid-term data has not demonstrated residual or recurrent intestinal metaplasia after RFA.Long-term data regarding durability of cure will guidefuture surveillance guidelines. Presently, patientsshould remain under endoscopic surveillance.

• Preliminary data have demonstrated superior cost-effectiveness and safety of RFA when compared toother eradication/ablation techniques.

• Future studies should focus on optimal patient selection, durability of cure, and development of biomarkers to predict malignant progression of Barrett’s esophagus.


19. Fleischer DE, Overholt BF, Sharma VK, et al. Endoscopic abla-tion of Barrett’s esophagus: a multicenter study with 2.5-year fol-low-up. Gastrointest Endosc 2008;68:867-76.

20. Ganz RA, Overholt BF, Sharma VK, et al. Circumferential abla-tion of Barrett’s esophagus that contains high-grade dysplasia: aU.S. Multicenter Registry. Gastrointest Endosc 2008;68:35-40.

21. Sharma VK, Kim HJ, Das A, et al. A prospective pilot trial ofablation of Barrett’s esophagus with low-grade dysplasia usingstepwise circumferential and focal ablation (HALO system).Endoscopy 2008;40:380-7.

22. Gondrie JJ, Pouw RE, Sondermeijer CM, et al. Stepwise circum-ferential and focal ablation of Barrett’s esophagus with high-grade dysplasia: results of the first prospective series of 11patients. Endoscopy 2008;40:359-69.

23. Gondrie JJ, Pouw RE, Sondermeijer CM, et al. Effective treat-ment of early Barrett’s neoplasia with stepwise circumferentialand focal ablation using the HALO system. Endoscopy2008;40:370-9.

24. Pouw RE, Wirths K, Eisendrath P, et al. Efficacy of Radiofre-quency Ablation Combined With Endoscopic Resection for Barrett’s Esophagus With Early Neoplasia. Clin GastroenterolHepatol 2009;8:23-9.

25. Velanovich V. Endoscopic endoluminal radiofrequency ablationof Barrett’s esophagus: initial results and lessons learned. SurgEndosc 2009;23:2175-80.

26. Shaheen NJ, Sharma P, Overholt BF, et al. Radiofrequency abla-tion in Barrett’s esophagus with dysplasia. N Engl J Med2009;360:2277-88.

27. Herrero LA, Pouw R, Van Vilsteren F. What Are the Outcomesof Endoscopic Radiofrequency Ablation for Very Long Segmentsof Barrett Esophagus Containing Neoplasia? GastrointestinalEndoscopy 2009;69:AB116.

28. Lyday W, Corbett FS, Kuperman DA, et al. EndoscopicRadiofrequency Ablation of Barrett’s Esophagus: Safety andEfficacy Outcomes in 429 Patients Treated in a Multi-CenterCommunity Practice Registry. Gastrointestinal Endoscopy2009;69:AB114.

29. Eldaif SM, Lin E, Singh KA, et al. Radiofrequency ablation ofBarrett’s esophagus: short-term results. Ann Thorac Surg2009;87:405-10; discussion 10-1.

30. Fleischer D, Overholt BF, Sharma V, et al. Endoscopic Radiofre-quency Ablation for Barrett’s Esophagus: Five-Year DurabilityOutcomes from a Prospective Multi-Center Trial. GastrointestEndosc 2010;71:AB117-8.

31. Shaheen NJ, Fleischer DE, Eisen GM, et al. Durability of Epithe-lial Reversion After Radiofrequency Ablation: Follow-up of theAIM Dysplasia Trial. Gastroenterology 2010 138:S16-7.

32. Wani S, Puli SR, Shaheen NJ, et al. Esophageal adenocarcinomain Barrett’s esophagus after endoscopic ablative therapy: a meta-analysis and systematic review. Am J Gastroenterol 2009;104:502-13.

33. Roorda AK, Marcus SN, Triadafilopoulos G. Early experiencewith radiofrequency energy ablation therapy for Barrett’s esoph-agus with and without dysplasia. Dis Esophagus 2007;20:516-22.

34. Finkelstein SD, Lyday WD. The molecular pathology of radiofre-quency mucosal ablation of Barrett’s esophagus. Gastroenterol-ogy 2008;134:A-436.

35. Pouw RE, Gondrie JJ, Rygiel AM, et al. Properties of the neo-squamous epithelium after radiofrequency ablation of Barrett’sesophagus containing neoplasia. Am J Gastroenterol 2009;104:1366-73.

36. Gross SA, Al-Haddad M, Gill KR, et al. Endoscopic mucosalablation for the treatment of gastric antral vascular ectasia withthe HALO90 system: a pilot study. Gastrointest Endosc2008;67:324-7.

37. Zhou C, Adler DC, Becker L, et al. Effective treatment of chronicradiation proctitis using radiofrequency ablation. TherapeuticAdvances in Gastroenterology 2009;2:149-56.

38. Pouw RE, Gondrie JJ, Curvers WL,et al. Successful balloon-based radiofrequency ablation of a widespread early squamouscell carcinoma and high-grade dysplasia of the esophagus: a casereport. Gastrointest Endosc 2008;68:537-41.

39. Zhang YM, Bergman JJ, Weusten B, et al. Radiofrequency abla-tion for early esophageal squamous cell neoplasia. Endoscopy2010;42:327-33.

40. Inadomi JM, Somsouk M, Madanick RD, et al. A cost-utilityanalysis of ablative therapy for Barrett’s esophagus. Gastroen-terology 2009;136:2101-14 e1-6.

41. Das A, Wells C, Kim HJ, et al. An economic analysis of endo-scopic ablative therapy for management of nondysplastic Barrett’sesophagus. Endoscopy 2009;41:400-8.



Fellows’ Corneris open to

Trainees and Residents ONLY.

Section Editor: C. S. Pichumoni, M.D.

Send in a brief case report. No more than one double-spaced page. One or two illustrations, up to

four questions and answers and a three-quarter to one-page discussion of the case.

Case to include no more than two authors. A $100.00 honorarium will be paid per publication.

Case should be sent to: C. S. Pitchumoni, M.D.

Chief, Gastroenterology, Hepatology and Clinical Nutrition

St. Peter’s University Hopsital254 Easton Avenue, Box 591

New Brunswick, NJ 08903E-mail: [email protected]

PPRRAACCTTIICCAALL GGAASSTTRROOEENNTTEERROOLLOOGGYY

R E P R I N T SR E P R I N T S

For details on rates or to place an order, visit our Web site at: www.practicalgastro.com

andrew k. roorda, m.d., series editor radiofrequency

Documents