Barrett's esophagus: Treatment with radiofrequency ablation

INTRODUCTION — Barrett's esophagus (BE) occurs when an abnormal, intestinal-type epithelium called "specialized intestinal metaplasia [IM]" replaces the stratified squamous epithelium that normally lines the distal esophagus. The condition develops as a consequence of chronic gastroesophageal reflux disease and predisposes to the development of adenocarcinoma of the esophagus.

Traditionally, high-grade dysplasia (HGD) and intramucosal cancer arising from BE were treated with esophagectomy, while nondysplastic BE and BE with low-grade dysplasia (LGD) were managed with endoscopic surveillance. Problems associated with these approaches included significant morbidity and mortality from esophagectomy and the risk of missed or interval development of cancer in patients undergoing surveillance. To address these issues, less invasive endoscopic treatments have been developed.

Radiofrequency ablation (RFA) is an endoscopic treatment modality for eradication of BE. Primary circumferential ablation is performed using a balloon-based bipolar electrode, while secondary treatment of residual BE is performed using an endoscope-mounted bipolar electrode on an articulated platform. Studies suggest that this ablation technique is highly effective in removing Barrett's mucosa and associated dysplasia and in preventing progression of disease, while minimizing the known drawbacks of photodynamic therapy and argon plasma coagulation, such as esophageal stenosis and subsquamous foci of BE ("buried Barrett's").

This topic will review the use of RFA for the treatment of BE. Other issues related to BE, including alternative treatments, are discussed separately.

●(See "Barrett's esophagus: Epidemiology, clinical manifestations, and diagnosis".)

●(See "Barrett's esophagus: Pathogenesis and malignant transformation".)

●(See "Barrett's esophagus: Evaluation with optical chromoscopy".)

●(See "Barrett's esophagus: Surveillance and management".)

●(See "Barrett's esophagus: Treatment of high-grade dysplasia or early cancer with endoscopic resection".)

INDICATIONS FOR RADIOFREQUENCY ABLATION

Following removal of visible lesions containing HGD or IMC — Patients with visible abnormalities in a segment of BE that contains intramucosal carcinoma (IMC) or high-grade dysplasia (HGD) may be treated with RFA, but only after endoscopic resection (ER) of the IMC or visible lesion. (See "Barrett's esophagus: Treatment of high-grade dysplasia or early cancer with endoscopic resection" and 'Combining endoscopic mucosal resection and RFA' below.)

ER, which can be performed with either endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD), provides a relatively large tissue specimen that allows for histopathologic staging of a lesion, enabling selection of patients for endoscopic treatment who have HGD or IMC and a low risk of lymph node involvement [1-3]. Patients found to have submucosal invading lesions on histology (>T1sm1) have a 15 to 30 percent risk of positive local lymph nodes and should be referred for surgery. On the other hand, the risk of lymph node involvement is minimal in patients with IMC, making these patients candidates for endoscopic management [4,5].

A reason for combining ablative therapy with ER is that ER only removes a focal area from the BE, leaving the patient at risk for metachronous lesions arising from the residual Barrett's mucosa [6]. The addition of ablative therapy to ER helps overcome this limitation.

In addition to staging a lesion, ER is also done prior to RFA to provide a flat mucosa for RFA, which helps ensure that the ablation reaches the muscularis mucosa (picture 1).

Flat high-grade dysplasia — Patients with BE and HGD seem to be ideal candidates for RFA since successful eradication of their dysplastic BE prevents the development of cancer [7]. However, proper patient selection is critical. Patients should have no visible lesions as these require ER for optimal staging and treatment. (See "Barrett's esophagus: Treatment of high-grade dysplasia or early cancer with endoscopic resection".)

To ensure that only patients with flat HGD are being treated with RFA monotherapy, several studies have required that patients undergo at least two high-resolution endoscopies with four-quadrant biopsies every 1 to 2 cm within two months prior to RFA to exclude cancer [1,2,8].

Low-grade dysplasia — Our approach is to offer RFA to most patients who have low-grade dysplasia (LGD) that has persisted over time and has been confirmed by two expert pathologists. Treatment of LGD is associated with reduced rates of progression to HGD or esophageal adenocarcinoma (EAC) [9,10]. Other factors that influence the decision of whether or not to proceed with RFA for LGD include the patient's age and comorbidities, risk of adverse events associated with RFA, and patient preferences.

The natural history of LGD in BE is unclear. While some studies have reported that neoplastic progression rates for LGD are low (comparable with those reported for nondysplastic BE) [11-13], other studies have found that if the LGD diagnosis is confirmed by an expert pathologist, the risk of progression to HGD or EAC may be as high as 27 percent within two years of follow-up [14-18]. In a study of 255 patients with LGD who underwent a median of four endoscopies during a median follow-up of 42 months, risk factors associated with progression to HGD or esophageal cancer in 45 patients (18 percent) included agreement on the diagnosis of LGD by three expert pathologists (odds ratio [OR] 47, 95% CI 13-170) and persistence of LGD at subsequent endoscopy (OR 9.3, 95% CI 4.4-19.6) [19]. (See "Barrett's esophagus: Surveillance and management", section on 'Dysplasia as a marker of risk' and "Barrett's esophagus: Surveillance and management", section on 'Low-grade dysplasia'.)

Patients who are recently diagnosed with flat LGD are evaluated prior to RFA therapy at a referral center to exclude higher grades of dysplasia. In a study including 248 patients with an initial diagnosis of flat BE with LGD, fifty-seven patients (23 percent) had HGD or cancer during re-staging endoscopy that was performed at a referral center [20].

The use of RFA decreases the risk of progression of LGD to HGD or EAC. In a meta-analysis of three studies (two of which were randomized trials) that compared RFA versus surveillance in patients with BE with LGD, the risk of progression to HGD or EAC was lower in the RFA group (relative risk 0.14 percent, 95% CI 0.04-0.45) [9]. In a systemic review of mostly observational studies that included 2700 patients with BE and LGD who had either RFA or surveillance, the rate of progression to HGD or EAC was also lower in the RFA group (1.7 versus 12.6 percent) [9]. Endoscopic surveillance for patients after a confirmed diagnosis of LGD is presented separately. (See "Barrett's esophagus: Surveillance and management", section on 'Low-grade dysplasia'.)

Nondysplastic BE — The risk of progression to cancer in patients with nondysplastic BE is small, and no objective markers are yet available to identify patients with an increased risk of developing cancer, though research looking at the risk-stratification of nondysplastic BE has shown promising results. (See "Barrett's esophagus: Pathogenesis and malignant transformation" and "Barrett's esophagus: Surveillance and management", section on 'Cancer risk'.)

Whether to offer RFA to patients with nondysplastic BE is highly controversial and is influenced by many factors. An argument against RFA in these patients is that the annual risk of malignant progression is low and many patients with BE are older adults with significant comorbid conditions that limit their life expectancy. Factors that favor treatment include the efficacy and safety profile of RFA. (See "Barrett's esophagus: Pathogenesis and malignant transformation" and 'Efficacy' below and 'Adverse events' below.)

For most patients with nondysplastic BE, the net health benefit of RFA may be too low to justify its use. However, RFA could be considered for selected patients (eg, <50 years of age, a long BE segment, and a positive family history for EAC). (See "Barrett's esophagus: Surveillance and management", section on 'Summary and recommendations'.)

ABLATION PROCEDURES — RFA of BE generally starts with a stepwise circumferential ablation procedure, followed by focal ablation for any residual BE (figure 1). RFA is performed using the Barrx FLEX system (previously HALOFLEX system), which is comprised of two distinct ablation catheters: the Barrx360 Express ablation balloon for primary circumferential RFA and the Barrx90 device for secondary focal RFA of BE. Three alternative focal ablation devices have become available: the Barrx90 Ultra, the Barrx60, and the Channel RFA device.

Circumferential ablation — Historically, circumferential ablation of BE was a two-step procedure that consisted of determining esophageal size with a sizing catheter, followed by selecting an ablation catheter with an appropriate diameter for ablation therapy. The Barrx360 system has been replaced by the Barrx360 Express system, which contains a 4 cm-long bipolar electrode wrapped around an inflatable self-sizing balloon. The device features the ability to self-adjust to the size of the esophageal lumen (picture 2). The Barrx360 Express catheter uses the Barrx FLEX energy generator (previously, the HALO360 generator was used).

The standard procedure for RFA includes [21]:

●Landmark determination – Historically, the first step in circumferential ablation was cleaning the esophageal wall with 1 percent acetylcysteine and flushing with water to remove excessive mucus. However, randomized trials have suggested that standard water rinses through the water jet channel of the endoscope are just as effective [21]. We have therefore abandoned the cleaning with acetylcysteine. Next, the locations of the top of the gastric folds and the proximal extent of the BE (including islands) are recorded for reference during the ablation procedure. A stiff guidewire or metal wire is then introduced, and the endoscope is removed, leaving the guidewire in place.

●Under endoscopic visualization, the balloon is positioned at the proximal end of the Barrett's segment, allowing approximately 1 to 2 cm overlap with the normal squamous esophageal tissue. The catheter is inflated by pressing the foot pedal, and once adequate mucosal contact is made, the foot pedal can be used to activate the electrode and ablate the mucosa. After the ablation, the balloon deflates and the electrode curls up again so it can be advanced distally. The catheter is repositioned distal to the prior ablation zone, allowing an overlap of ±5 mm with the previous ablation zone, and a subsequent ablation can be performed. With this method of working from the proximal to the distal esophagus, the entire segment of BE can be ablated.

●The entire length of BE is treated from proximal to distal areas once, and then this method is repeated. The cleaning step performed in between the two ablations passes is necessary to remove the coagulated tissue. Without this cleaning step, it is impossible to distinguish the edges of the ablation zone made during the second pass. These edges need to be identified for correct re-placement of the ablation catheter when moving from proximal to distal to avoid unnecessary overlap. For this cleaning step, the endoscope and balloon catheter are removed from the patient after the first ablation pass. The balloon can then be cleaned outside of the patient with a wet gauze. The endoscope, with a distal attachment cap at its tip, is reintroduced to clean the esophageal wall by rinsing water and scraping off the coagulum using the rim of the cap. Then, the guidewire is reintroduced and the endoscope is removed. Next, the ablation catheter is reintroduced, followed by reintroduction of the endoscope, after which a second ablation pass moving from proximal to distal is performed.

●After the second ablation pass, the catheter and endoscope can be removed from the patient and the procedure is completed.

While using the Barrx360 Express, the standard ablation regimen (ie, one ablation at 10 J/cm², followed by catheter cleaning, followed by one ablation at 10 J/cm²) has been typically used for treating BE rather than alternative ablation regimens. In a trial including 104 patients with BE and dysplasia, three ablation regimens using Barrx360 Express were evaluated: standard regimen (one ablation at 10 J/cm², followed by catheter cleaning, followed by one ablation at 10 J/cm²), a simple-single regimen (one ablation at 10 J/cm² without catheter cleaning), and a simple-double regimen (two consecutive ablations at 10 J/cm² without catheter cleaning) [22]. Enrollment in the simple-double regimen arm was stopped early because 6 of 28 patients (21 percent) developed severe esophageal stenosis. The simple-single regimen resulted in lower rates of eradication of BE compared with the standard regimen after three months (73 versus 85 percent; mean difference 13 percent, 95% CI 5-23 percent). Of note, procedure time was longer with standard ablation regimen (31 versus 17 minutes).

Follow-up endoscopy and subsequent ablation regimens include:

●Follow-up circumferential ablation – Twelve weeks after the first circumferential ablation treatment, patients undergo a follow-up endoscopy and additional therapy is carried out if needed. A circumferential ablation is performed if:

•There is residual circumferential BE measuring 2 cm or more

•There are multiple islands or tongues of BE

●Follow-up focal ablation – At the 12-week follow-up endoscopy, patients are treated with secondary focal ablation using the Barrx90 catheter for (see 'Focal ablation' below):

•Residual BE with a circumferential extent of less than 2 cm

•Circular treatment at the upper end of the gastric folds (at least once)

•Small tongues of BE

•Scattered islands of BE

Focal ablation — Focal RFA with the Barrx90 catheter also uses radiofrequency energy to ablate small areas of BE. For focal ablation, the electric current is delivered through an electrode array attached to the end of the endoscope (picture 3).

The electrode array is mounted on an articulated platform (figure 2), allowing the electrode to move front-to-back and left-to-right, ensuring optimal tissue contact. It can be attached with a flexible strap to the distal end of any endoscope with a diameter of 8.6 mm to 12.8 mm without impairing endoscopic view or function. The electrode array is 20.6 mm long and 13.2 mm wide, with an active electrode surface of 20 mm × 13 mm.

The FLEX generator is used for both circumferential (Barrx360 Express) and focal RFA. The first-generation HALO90 energy generator is used only with the Barrx90 catheter.

●Electrode introduction – The Barrx90 electrode array fits on the tip of the endoscope and is placed at the 12 o'clock position in the endoscopic video image.

●The device and endoscope are then introduced under visual guidance. When the laryngeal cavity is seen, the tip of the endoscope is deflected slightly downward. The endoscope is gently advanced into the esophagus, passing the leading edge of the catheter behind the arytenoids.

In approximately 10 percent of cases, introducing the electrode array may prove difficult. In those cases, a Zenker diverticulum should be excluded. Introduction of the device should never be forced, due to the risk of perforation. In these cases, we will sometimes blindly pass a biopsy forceps or the spraying catheter into the esophagus to guide the endoscope into the proximal esophagus. In difficult cases, a CRE balloon may be used to open the upper esophageal sphincter by manually inflating the balloon to a low pressure and then advancing the endoscope and Barrx90 device along with the balloon.

●Residual Barrett's epithelium is positioned at the 12 o'clock position in the endoscopic video image, corresponding to the position of the electrode. The electrode is brought into close contact with the mucosa, deflected upward, and activated.

The simplified ablation regimen for focal RFA consists of three successive ablations at 12 J/cm², without cleaning the ablation zone or electrode. This simplified method has been effective but with shorter procedure times and possibly less patient discomfort because it requires a single introduction of the endoscope and ablation catheter [23,24]. In comparison, the previous regimen for focal ablation was a double-clean-double method consisting of two applications of 15 J/cm², followed by catheter cleaning and then two applications of 15 J/cm², resulting in a "double" application of radiofrequency energy. In a trial including 84 patients with dysplastic BE, rates of eradication of BE were not significantly different for a regimen of three successive ablations at 12 J/cm² without catheter cleaning compared with a double-clean-double regimen [24]. In another trial including 41 patients with 46 pairs of residual islands of BE, rates of eradication of BE were not significantly different for the simplified regimen (three applications of radiofrequency energy at 15 J/cm² without cleaning) compared with the double-clean-double regimen (73 versus 67 percent) [23].

In addition to treatment of any visually apparent BE, ablation of the entire Z-line is recommended (even if no clear tongues of BE are observed) to ensure eradication of all BE at the gastroesophageal junction (picture 4) [25].

Other ablation devices — Three ablation catheters have been added to the Barrx FLEX system: the Barrx90 Ultra catheter, the Barrx60 catheter, and a channel RFA device that can be advanced through the scope. All three catheters can be used alongside the Barrx90 catheter for secondary focal RFA of BE; however, no studies have yet evaluated the use of these devices in clinical practice. Recommendations for the use of these catheters are therefore based on previous experiences with the Barrx90 device:

●Barrx90 Ultra device – The electrode array is mounted on an articulated platform in a similar way as the Barrx90 device, allowing the electrode to move front-to-back and left-to-right, ensuring optimal tissue contact. It can be attached with a flexible strap to the distal end of any endoscope with a recommended diameter of 8.6 mm to 9.8 mm without impairing endoscopic view or function. The electrode array is 40 mm long and 13 mm wide with an active electrode surface area of 520 mm², resulting in a 200 percent larger electrode surface as compared with the regular Barrx90 device. The Barrx90 Ultra is less well evaluated in terms of energy settings and safety when compared with the Barrx90 device.

To prevent potential stenosis after focal ablation, the recommended treatment regimen consists of two double applications of radiofrequency energy at 12 J/cm², which has been studied extensively since the introduction of the Barrx90 device [26]. An alternative regimen for the Barrx90 Ultra device consists of three applications of 12 J/cm². Patients can be treated with secondary focal ablation using the Barrx90 Ultra device if there are large tongues of residual BE or if there is short-segment BE.

●Barrx60 device – The electrode array is mounted on an articulated platform in a similar way as the Barrx90 device, allowing the electrode to move front-to-back and left-to-right, ensuring optimal tissue contact. It can be attached with a flexible strap to the distal end of any endoscope with a recommended diameter of 8.6 mm to 9.8 mm. The electrode array is 15 mm long and 10 mm wide; as a result, the active electrode surface area is 60 percent of the surface area of the Barrx90 device.

The recommended treatment regimen consists of two double applications of energy at 15 J/cm² or three applications of 12 J/cm². Patients can be treated with the Barrx60 device for small islands of BE in the presence of a stenosis.

●Channel RFA device – The channel device is a through-the-scope device and fits through the working channel of an endoscope with a recommended diameter of 2.8 mm or larger. The design of the shaft provides catheter maneuverability, and the translucence of the device provides visibility. The electrode array is 15.7 mm long and 7.5 mm wide and has approximately the same active electrode surface area as the Barrx60 device. Since the channel device has only been tested in animal studies, the recommended treatment regimen consists of two double applications of energy at 15 J/cm² with cleaning or three applications of 12 J/cm².

Post-treatment care — After RFA, acid suppressive therapy is important, not only to minimize patient discomfort but also to allow the esophagus to heal optimally and regenerate with squamous epithelium. Studies suggest that ongoing gastroesophageal reflux has an adverse effect on treatment outcome [27,28]. All patients should therefore receive high-dose proton pump inhibitors as maintenance therapy. In addition, extra acid suppression after each treatment session is advisable. We prescribe all patients esomeprazole 40 mg twice daily, supplemented with a histamine-2 receptor antagonist at bedtime, and 1 gram sucralfate suspension (eg, 5 mL of a 200 mg/mL or 10 mL of a 100 mg/mL suspension) four times a day for two weeks after each ablation session [27,28]. The proton pump inhibitor is continued as maintenance therapy. (See "Antiulcer medications: Mechanism of action, pharmacology, and side effects", section on 'Histamine-2 receptor antagonists'.)

After RFA, patients should adhere to a liquid diet for 24 hours. After 24 hours, patients may gradually advance to a soft and then normal diet at their own discretion, generally guided by their symptoms. Patients may experience symptoms of chest discomfort, sore throat, difficulty or pain with swallowing, and/or nausea, which usually improve daily.

For patients who have pain following the procedure, acetaminophen 500 to 1000 mg up to four times per day may be given as needed. Acetaminophen can be supplemented with diclofenac supplements 50 mg up to twice daily if the acetaminophen does not provide adequate relief. Other analgesic regimens include an antacid/lidocaine slurry and liquid acetaminophen with or without codeine. Some patients may also require antiemetic medication.

For patients presenting with severe chest pain and fever following their procedures, observation and conservative management with maximal acid suppression and an analgesic regimen will usually suffice. Only in rare cases, when there is a clear suspicion of severe complications, is additional testing (eg, computed tomography) required.

EFFICACY

Eradication of intestinal metaplasia and dysplasia — Evidence from a number of well-designed studies, including a randomized, sham-controlled trial, suggests that RFA is highly effective at removing all BE at both the endoscopic and histologic level with a favorable safety profile [1,2,7,8,10,25,26,29-48]. Overall, studies report complete eradication of intestinal metaplasia (CE-IM) rates of 54 to 100 percent and complete eradication of dysplasia (CE-D) rates of 80 to 100 percent.

One factor that may account for some of the variability in CE-IM and CE-D rates seen in studies is the use of different RFA protocols (circumferential versus circumferential plus focal RFA, whether standard ablation of the Z-line was performed, whether endoscopic mucosal resection [EMR] was performed for visible lesions). Other factors may include stricter adherence to RFA protocols in prospective studies, variable baseline histology (nondysplastic BE, low-grade dysplasia [LGD], high-grade dysplasia [HGD], or intramucosal carcinoma [IMC]), differences in the length of BE segments included in the studies, and differences in follow-up protocols.

Long-term outcomes — Studies on long-term outcomes suggest that rates of recurrent dysplasia after RFA are low and range from 0.8 to 3 percent [25,33,37,48-53]. Additional factors that inform an evidence-based approach to surveillance protocols include timing of recurrence and endoscopic and histologic findings (eg, visible abnormality, dysplasia detected by random biopsy). In a cohort study including 1154 patients with dysplastic BE who had RFA at a referral center, 38 patients (3 percent) developed recurrent dysplasia (ie, LGD, HGD or cancer) after a median follow-up of 43 months [48]. All patients with HGD or cancer had endoscopically visible abnormalities.

Based on data from cohort studies, a model was developed to predict recurrent disease that may guide surveillance protocols [48,50,54]. Factors that were associated with increased risk for developing recurrence included new visible lesions detected during RFA, higher number of endoscopic resection procedures, male sex, increasing BE length, HGD or cancer at baseline, and younger age [54]. The model was validated with an area under the curve of 0.91 with good calibration.

Whether the risk of recurrence remains constant over time is uncertain [49,50]. In a study including 594 patients with dysplastic BE who were treated with RFA with median follow-up of nearly three years, the rate of recurrence was 2.8 percent annually without significant variation in recurrence rate during the total duration of follow-up [49]. However, in a study including 807 patients with dysplastic BE who had RFA with median follow-up of 2317 person-years, the risk of recurrence peaked at 1.6 years after CE-IM was achieved [50].

Combining endoscopic mucosal resection and RFA — Patients with visible abnormalities in a segment of BE that contains IMC or HGD may be treated with RFA, but only after endoscopic resection of the IMC or visible lesion. (See 'Following removal of visible lesions containing HGD or IMC' above and "Barrett's esophagus: Treatment of high-grade dysplasia or early cancer with endoscopic resection" and 'Adverse events' below.)

There is concern that patients undergoing RFA following endoscopic resection may be at increased risk for complications. However, one study with 90 patients (44 who underwent RFA following endoscopic resection and 46 who only underwent RFA) did not find a difference in the rate of stricture formation between the groups [55]. Patients who underwent RFA following endoscopic resection had a stricture rate of 14 percent, compared with 9 percent for those who underwent RFA alone (odds ratio [OR] 1.53, 95% CI 0.26-9.74).

An advantage of RFA over other ablation methods is that it does not appear to interfere with subsequent endoscopic resection for residual lesions. Studies suggest that it is possible to resect areas of Barrett's mucosa that persist after multiple RFA sessions using the ligate-and-cut technique, without the need for submucosal lifting [2,8]. Other endoscopic ablation techniques typically result in submucosal scarring, which complicates subsequent treatment with endoscopic resection.

The ability to perform endoscopic resection for residual dysplastic tissue may explain the higher success rates in studies that incorporated endoscopic resection for residual BE (83 to 100 percent) compared with other RFA studies that did not [30,31,56].

Effect on quality of life — Patients in the ablation of intestinal metaplasia-dysplasia trial were also evaluated to see if RFA had an effect on their quality of life [57]. At baseline, similar numbers of patients assigned to RFA and sham procedures reported worry about esophageal cancer (71 versus 85 percent, respectively) and worry about esophagectomy (61 versus 68 percent, respectively). Patients were also similar with regard to the presence/severity of depression, impaired quality of life, worry, stress, and dissatisfaction with the condition of their esophagus.

After undergoing an endoscopy 12 months following RFA or a sham procedure, patients were informed of their biopsy results (but not of their randomization group) and completed a second questionnaire. At follow-up, patients in the RFA group were less likely to be worried about esophageal cancer compared with those in the sham group (22 versus 66 percent) and were less likely to be worried about esophagectomy (17 versus 47 percent). They also showed significantly more improvement with regard to the presence/severity of depression, impaired quality of life, worry, stress, and dissatisfaction with the condition of their esophagus.

Interestingly, patients in the RFA group who did not achieve eradication of their BE still showed improvement in many of the quality-of-life domains, including worry about esophageal cancer and worry about esophagectomy, compared with those who underwent the sham procedure.

Subsequent cancer risk — Patients who undergo RFA for BE are at low risk of subsequently developing esophageal adenocarcinoma (EAC) [58,59]. In a registry study, 4982 patients who underwent RFA for BE were followed for a mean of 2.7 years [58]. EAC developed in 100 patients (2 percent; incidence of 7.8 per 1000 patient-years). Most of the cancers developed in patients with baseline HGD (83 patients). Cancers were also seen in patients with baseline LGD (12 patients) and nondysplastic BE (3 patients). Factors associated with cancer development included male sex, older age, longer BE segment length, and higher pathologic grade at baseline. These data should be cautiously interpreted since this is a registry study in which the appropriateness of indications, standardized treatment, and follow-up are not ensured. The strongest predictors for developing esophageal cancer after RFA are the indication for treatment and the complete remission rates for dysplasia and IM.

BURIED BARRETT'S ESOPHAGUS — The possibility of occult malignant progression of the buried glands has been suggested by cases of adenocarcinoma arising underneath neosquamous epithelium after ablation therapy with RFA, photodynamic therapy, or argon plasma coagulation [60-62]. However, because the Barrett's mucosa is protected from gastroesophageal refluxate by the neosquamous epithelium, the malignant potential of the buried glands may be lower than that seen with normal Barrett's mucosa [63,64]. In a cohort study including 376 patients with dysplastic BE who had RFA followed by random surveillance biopsies from neosquamous epithelium during follow-up, buried BE was found in 10 patients (2.7 percent of patients and 0.1 percent of biopsies) [48]. None of the patients with buried BE progressed to dysplasia during a median follow-up of four years.

Buried Barrett's prior to RFA — Some of the clinical uncertainty is due to the observation that buried Barrett's glands may be seen in up to 28 percent of patients in the absence of treatment, suggesting that buried glands found following RFA may have been present prior to treatment [65].

In a retrospective study of 112 patients who completed treatment with RFA, 15 percent of the subjects had evidence of buried glands before or during RFA treatment [66]. Importantly, 71 percent of cases showed buried glands during RFA treatment. A finding of buried glands was always concomitant with the presence of Barrett's mucosa visualized on endoscopy or biopsy. At the final evaluation, none of the patients showed evidence of buried glands and all patients were classified as being in complete remission.

In a sham-controlled study of 127 patients, 25 percent of the subjects had buried glands at baseline [7]. At follow-up, only 5 percent of those treated with RFA were found to have buried glands, compared with 40 percent of subjects in the sham group (p<0.001). Overall, in a systematic review that included 1004 patients who underwent RFA, buried metaplasia was found in only 1 percent [65].

False-negative results — Another concern is that buried Barrett's following RFA is underrecognized because the biopsies do not sample the neosquamous epithelium deeply enough [67,68]. A study of 16 patients examined the sampling depth and presence of buried glands in biopsies and endoscopic resection specimens from neosquamous epithelium after RFA [67]. Four-quadrant routine biopsies were obtained every 2 cm from the neosquamous epithelium. Immediately after each routine biopsy, a second "keyhole" biopsy was taken from the same biopsy site to obtain a deeper sample. In addition, one set of four-quadrant routine biopsies was obtained from the untreated squamous epithelium of the proximal esophagus. Finally, a tissue sample from the neosquamous epithelium was obtained using endoscopic resection.

The study showed:

●No difference in primary biopsy depth when comparing specimens obtained from the post-RFA neosquamous epithelium with specimens obtained from the untreated squamous epithelium (lamina propria present in 37 versus 36 percent of samples), suggesting that the post-RFA neosquamous epithelium is not more resistant to biopsy than untreated tissue.

●Keyhole biopsies and endoscopic resection sampled more deeply than routine biopsies (55 and 100 percent contained lamina propria, respectively).

●No buried Barrett's glands were detected in any of the primary, keyhole, or endoscopic resection specimens.

This study shows that biopsies obtained from squamous mucosa following RFA are similar to those obtained from normal squamous mucosa and that the low rate of buried glands reported after RFA compared with other ablation techniques is not a reflection of a sampling error specific to RFA. The absence of buried glands in all the specimens obtained in this study (including endoscopic resection specimens that sampled deep into the submucosa) is consistent with most clinical studies, in which buried glands are a rare finding. This suggests that RFA may be associated with complete eradication of all Barrett's epithelium.

False-positive results — Tissue artifacts and residual BE may also lead to an incorrect diagnosis of buried Barrett's glands. Biopsies from neosquamous epithelium near the neosquamocolumnar junction may lead to sampling of the transition from neosquamous to columnar epithelium, leading to a histologic finding of glandular mucosa beneath the neosquamous epithelium, which may mistakenly be interpreted as buried Barrett's glands. In the case of a residual island of BE that was not treated with RFA, tangential sampling of the island combined with tangential sectioning of the biopsy may result in an erroneous finding of buried Barrett's [69].

A diagnosis of buried Barrett's glands should only be made if the endoscopist is positive that there were no residual BE islands after detailed inspection with narrow band imaging and if the biopsies were not obtained at the level of the neosquamocolumnar junction [70]. (See "Barrett's esophagus: Evaluation with optical chromoscopy".)

INTESTINAL METAPLASIA OF THE CARDIA — There are limited data on the natural history of recurrent intestinal metaplasia (IM) of the cardia/gastroesophageal junction. After RFA treatment, the gastroesophageal junction is frequently biopsied, given the known increased risk of recurrences in this area, as endoscopic differentiation between gastric mucosa and IM is nearly impossible [32]. However, the clinical relevance of IM when detected in this area is uncertain because focal IM in this area may reflect insufficient treatment, recurrence of disease, or an irrelevant normal finding. Several studies have reflected on the relevance of IM at the cardia/gastroesophageal junction:

●In a cohort study including 1121 patients with dysplastic BE who had random biopsies of the cardia during a total of 2733 endoscopies and with median follow-up of 43 months, IM was detected in 150 patients (14 percent) during 198 endoscopies (7 percent) [48]. During subsequent endoscopic surveillance over a median follow-up of three years, IM was detected in 43 of 129 patients (33 percent). Three patients (2.3 percent) progressed to low-grade dysplasia (LGD), however, no patients developed high-grade dysplasia (HGD) or cancer. Random biopsies of the cardia showed LGD in less than 1 percent of patients. For patients with LGD, this finding was detected again in 75 percent of patients undergoing endoscopy during a median follow-up of two years. No patients with LGD progressed to HGD or cancer.

●In an earlier series that included 448 patients with BE with variable degrees of dysplasia, 17 patients developed recurrence of BE at the gastroesophageal junction only, after having achieved complete eradication after RFA [33]. Of these, 72 percent of patients had IM. Forty-seven percent of patients with IM or dysplasia at the cardia had a normal endoscopic appearance of the neosquamocolumnar junction. Follow-up results after detection of focal IM are not available.

These studies show that the management of focal IM of the gastroesophageal junction varies widely as formal guidelines are lacking. The gastroesophageal junction remains an area at risk after endoscopic treatment and should be carefully inspected during follow-up with high-resolution endoscopy combined with advanced imaging techniques. However, IM without dysplasia does not appear to warrant treatment as the majority of patients do not develop dysplasia during follow-up. If dysplasia is detected at the gastroesophageal junction, this can usually be managed endoscopically.

ADVERSE EVENTS — Adverse events reported with RFA include esophageal strictures, upper gastrointestinal hemorrhage, and chest pain [7,8]. Stricture rates of 0 to 56 percent have been described with other endoscopic ablation techniques [71]. Overall, studies of RFA for BE have shown lower rates of stricturing (0 to 6 percent) [7,8,10,26,34,48,72,73]. In a meta-analysis of 18 studies, the most common adverse events were stricture formation (pooled estimate 5 percent), pain (3 percent), and bleeding (1 percent) [36]. The stricturing seen with RFA has generally been associated with either prior endoscopic resection or a narrow esophagus at baseline due to underlying reflux disease.

●A study of 12 patients that compared measurements of esophageal inner diameter, motility, and compliance before RFA treatment and two months after the last ablation session showed no significant differences, suggesting that RFA does not impair the functional integrity of the esophagus [72].

●In a sham-controlled study of 127 patients, 6 percent of the RFA cohort experienced a stricture, but all resolved with a mean of 2.6 dilations [7]. There were no perforations or deaths.

●In a randomized trial with 136 patients, 12 percent of the ablation cohort (8 of 68 patients) experienced a stricture, but these appeared generally mild in nature as all resolved with a median of one dilation [10].

●In a community-based registry of 429 patients, nine strictures occurred following 788 procedures (1 percent of cases, 2 percent of patients) [34]. All of the strictures resolved after a median of three dilations. There were no cases of bleeding (other than one patient who vomited blood-tinged mucus), perforation, or death.

Performing endoscopic resection prior to RFA may increase the risk of complications. In a study of 65 patients, there were no complications in the 18 patients who had not previously undergone endoscopic resection [74]. In the 47 patients who had undergone endoscopic resection prior to circumferential RFA, mucosal lacerations were observed in patients who underwent RFA with a catheter that exceeded the smallest measured inner esophageal diameter and in patients whose endoscopic resection involved more than one-third of the esophageal circumference and was greater than 2.5 cm in length. Five cases of esophageal stenosis after RFA occurred, all in patients whose endoscopic resection involved more than 50 percent of the esophageal circumference and was longer than 2 cm in length.

Based on these observations, it is advisable to choose the size of the ablation catheters carefully in cases of prior endoscopic mucosal resection (EMR). (See 'Combining endoscopic mucosal resection and RFA' above and 'Circumferential ablation' above.)

Poor healing and poor squamous regeneration — Following RFA, some patients have poor mucosal healing and/or poor squamous regeneration [75,76]. Poor mucosal healing is defined as active inflammatory change with mucosal swelling and exudates and/or ulcerations ≥ 3 months post-RFA. Poor squamous regeneration (PSR) is defined as < 50 percent BE regression three months after RFA, provided that the esophagus was completely healed.

Data suggest that poor mucosal healing improves with medical management, whereas poor squamous regeneration is associated with disease progression and lack of treatment response. In a cohort study including 1386 patients with dysplastic BE who were treated with RFA, 134 patients (10 percent) had poor mucosal healing that resolved with increased acid suppression and additional follow-up time [75,76]. Of these patients, 67 patients (50 percent) had normal squamous regeneration and most achieved complete eradication of BE (97 percent). However, 67 patients (50 percent) had poor squamous regeneration, which was associated with higher risk of not achieving complete eradication compared with normal squamous regeneration (64 versus 2 percent). In addition, poor squamous regeneration was associated with higher rates of disease progression (15 versus <1 percent). Risk factors for poor squamous regeneration included higher body mass index, longer BE segment, reflux esophagitis, and < 50 percent squamous regeneration after baseline endoscopic resection.

Identifying risk factors for a complex RFA treatment course may inform management and improve outcomes. In a study including 1386 patients with BE who had endoscopic eradication therapy, 78 patients (6 percent) experienced a complex treatment course, defined as disease progression, treatment failure, or development of new visible lesion(s) [77]. Risk factors included BE length of ≥9 cm with a visible lesion containing high-grade dysplasia/cancer and <50 percent squamous conversion after RFA. A prognostic model was developed using these risk factors, and the model performed well with validation (area under the curve 0.84).

FOLLOW-UP ENDOSCOPY

Techniques — The cornerstone of endoscopic follow-up should consist of meticulous endoscopic inspection of the neosquamous mucosa and the neosquamocolumnar junction to rule out the presence of residual columnar mucosa. Techniques available to detect residual BE include:

●High-definition white light endoscopy for detailed endoscopic inspection of the treatment area

●Narrow band imaging or comparable technologies (eg, Fujifilm intelligent chromoendoscopy, i-scan) (see "Barrett's esophagus: Evaluation with optical chromoscopy")

Detailed inspection of the neosquamous mucosa after RFA by an endoscopist with a trained eye is important for two reasons: first, to detect even small areas of BE that can be additionally treated as any residual BE puts the patient at risk of developing esophageal adenocarcinoma (EAC) (picture 3); second, if random biopsies are obtained and, accidentally, small residual islands of Barrett's are sampled, this can result in a histologic finding of buried Barrett's, causing doubts on the efficacy of the treatment and resulting in a missed opportunity to treat endoscopically visible remnants of Barrett's mucosa.

In a study evaluating the incidence of buried Barrett's in biopsies obtained after RFA, buried glands were found in 0.1 percent of biopsies from endoscopically normal neosquamous epithelium. However, when small islands of columnar mucosa were intentionally biopsied, buried glands were detected in 21 percent of biopsies [70]. The low percentage of buried Barrett's in biopsies from neosquamous mucosa found in this study is similar to the 0 to 5 percent found in over 1000 patients from other studies on RFA for BE [7,26,34,35,65]. (See 'Buried Barrett's esophagus' above.)

We perform follow-up with high-resolution endoscopy and narrow band imaging, or a comparable technique, to carefully inspect the neosquamocolumnar junction in antegrade and retroflexed positions to rule out the presence of small islands of Barrett's mucosa. The neosquamocolumnar junction is the area most at risk for recurrence of neoplasia [44-46]. A reliable endoscopic tool to predict if all Barrett's mucosa has been eradicated at this level is not available [47]. Even endoscopic detection techniques such as narrow band imaging have not been able to aid the endoscopist in the differentiation between gastric mucosa and intestinal metaplasia (IM) [32]. Therefore, we always obtain biopsies immediately distal (<5 mm) to the neosquamocolumnar junction as an objective endpoint for eradication of IM. The downside of this biopsy protocol is that it can lead to overestimation of nondysplastic IM in the presence of a normal-appearing neosquamocolumnar junction on endoscopy if the cardia is sampled. The presence of IM in the cardia during follow-up examination appears to be of limited clinical relevance (see 'Intestinal metaplasia of the cardia' above). If IM is found in this region, we advocate repeating focal ablation only if the IM is detected at the first follow-up endoscopy. Touch-up ablation during subsequent follow-up endoscopies is unnecessary unless dysplasia is present.

Given the very low incidence of buried Barrett's reported in multiple studies, extensive biopsies from the neosquamous mucosa are, in our opinion, not necessary, provided that detailed inspection with high-resolution endoscopy with narrow band imaging did not show any columnar mucosa or mucosal irregularities.

If residual BE is found, ablation can be repeated every 12 weeks until it has been eradicated both visually and histologically. Most patients will need one circumferential ablation session and one to two focal ablation sessions to eradicate all dysplasia and Barrett's mucosa. We suggest a maximum number of two circumferential and three focal ablation sessions, which should be sufficient in most patients.

Surveillance intervals — Follow-up protocols are generally based on consensus opinion and vary by geographic region. However, emerging data will help inform follow up strategies in the future. The follow-up interval after complete eradication of IM (CE-IM) depends on the initial grade of dysplasia:

●For patients with intramucosal carcinoma (IMC)/high-grade dysplasia (HGD), we recommend gradually increasing the interval of follow-up endoscopies if there has been complete eradication of dysplasia (CE-D) and CE-IM to 3, 6, and 12 months after the last treatment session, then annually during the first five years. If there is sustained eradication of IM at that stage, surveillance can be stopped or intervals prolonged to every three to five years.

●For patients with low-grade dysplasia (LGD)/nondysplastic BE, we recommend follow-up endoscopy at 3 and 12 months after the last treatment session. If there is sustained eradication of IM at that stage, we stop surveillance. The published literature on the durability of RFA shows that the risk of progression in these patients is small [36]. For patients with LGD who have completed endoscopic eradication therapy and have no evidence of IM, professional societies recommend surveillance in one year, in three years, and every two years thereafter [78,79]. (See 'Efficacy' above.)

If dysplasia is eradicated but IM persists, patients should be followed every six months for one year, then annually for two years, and then every three years thereafter. (See "Barrett's esophagus: Surveillance and management", section on 'Surveillance'.)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Barrett's esophagus".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topics (see "Patient education: Barrett's esophagus (The Basics)")

●Beyond the Basics topics (see "Patient education: Barrett's esophagus (Beyond the Basics)")

SUMMARY AND RECOMMENDATIONS

●Indications – Radiofrequency ablation (RFA) is an endoscopic treatment for Barrett's esophagus (BE). RFA is an option for the treatment of patients with dysplastic (low-grade dysplasia [LGD] and high-grade dysplasia [HGD]) BE since RFA decreases their risk of malignant progression. Data on short- and long-term outcomes suggest that RFA is effective and safe treatment. (See 'Indications for radiofrequency ablation' above and 'Efficacy' above.)

For most patients with nondysplastic BE, the net health benefit of RFA may be too low to justify its use. However, RFA can be considered for selected patients (eg, <50 years old, a long BE segment, and a positive family history for esophageal adenocarcinoma [EAC]). (See 'Nondysplastic BE' above.)

●Ablation procedures – RFA is performed with the Barrx FLEX system, which uses a balloon-based catheter for primary circumferential RFA and focal catheters for secondary focal RFA of BE. Circumferential RFA involves inflation of a 4 cm-long self-sizing balloon ablation catheter (Barrx360 Express device) within the esophagus at the site of the BE. The balloon contains an electrode array through which an electric current is applied, ablating the Barrett's mucosa. (See 'Ablation procedures' above.)

Secondary focal RFA to ablate small areas of BE also uses electrical energy to ablate the tissue, and the electrode array is attached to the end of the endoscope (Barrx90 device). Catheters have been developed that may serve as an adjunct to the Barrx90 device. (See 'Ablation procedures' above.)

●Adverse events – Adverse events reported with RFA include esophageal strictures, upper gastrointestinal hemorrhage, and chest pain. These complications are generally mild and severe complications are rare. (See 'Adverse events' above.)

●Post-treatment care – After RFA, acid suppression is important, not only to minimize patient discomfort but also to allow the esophagus to heal optimally and regenerate with squamous epithelium. All patients should receive high-dose proton pump inhibitors as maintenance therapy. In addition, extra acid suppression after each treatment session is advisable. We prescribe all patients esomeprazole 40 mg twice daily, supplemented with a histamine-2 receptor antagonist at bedtime, and sucralfate suspension (5 mL of a 200 mg/mL or 10 mL of a 100 mg/mL suspension) four times a day for two weeks after each ablation session. The proton pump inhibitor is continued as maintenance therapy. (See 'Post-treatment care' above.)

●Endoscopic surveillance – Three months after the last treatment, the absence of residual Barrett's epithelium is confirmed by endoscopic inspection. The use of high-resolution endoscopy with narrow band imaging is important to detect potential recurrences in the tubular esophagus or the gastroesophageal junction. (See "Barrett's esophagus: Evaluation with optical chromoscopy" and 'Follow-up endoscopy' above.)

The recommended follow-up interval after successful RFA with complete eradication of IM (CE-IM) is based on consensus expert opinion, and guidance varies by geographic region. Evidence-based intervals are being developed. Existing protocols are informed by the initial grade of dysplasia (see 'Follow-up endoscopy' above):

•For patients with IMC/HGD, we perform follow-up endoscopy at 3, 6, and 12 months after the last treatment session, then annually during the first five years. If there is sustained eradication of IM at that stage, surveillance can be stopped or intervals prolonged to every three to five years.

•For patients with LGD/nondysplastic BE, we perform follow-up endoscopy at 3 and 12 months after the last treatment session. If there is sustained eradication of IM at that stage, surveillance should, in our opinion, be stopped. The risk of progression in these patients is small.

If dysplasia is eradicated but IM persists, we follow patients every six months for one year, then annually for two years, and then every three years thereafter. (See "Barrett's esophagus: Surveillance and management", section on 'Surveillance'.)