Recurrence rates of Barrett’s esophagus and dysplasia in patients successfully treated with radiofrequency ablation vs. cryoballoon ablation: a comparative study

Abstract

Background

Endoscopic eradication therapy (EET) is recommended for the management of dysplastic Barrett’s esophagus (BE) and intramucosal adenocarcinoma. Both radiofrequency (RFA) and cryoballoon ablation (CBA) can induce complete remission of intestinal metaplasia (CRIM). We aimed to compare long-term durability following CRIM in patients treated with RFA vs. CBA.

Methods

We conducted a retrospective cohort study to analyze patient outcomes following EET with RFA or CBA at two referral centers. We measured and compared the recurrence rate of BE, with or without dysplasia, after CRIM, using Cox proportional hazard models and propensity score-matched analyses. Our secondary aim was to identify predictors of recurrence.

Results

681 patients who achieved CRIM were included (RFA 610; CBA 71), with median follow-up of 4.1 years and 4.4 years following RFA and CBA, respectively. A multivariable model revealed a higher risk of any recurrence with RFA (hazard ratio [HR] 2.19, 95%CI 1.18–4.06; P = 0.01), but not dysplastic recurrence (HR 0.88, 95%CI 0.39–1.97; P = 0.75). However, a comparable risk of “any recurrence” was observed after excluding intestinal metaplasia of the gastroesophageal junction (HR 1.18, 95%CI 0.61–2.30; P = 0.62). Propensity score-matched analysis (54 cases in each group) showed that the two groups had comparable recurrence risk. Baseline maximum BE length was associated with a higher risk of any recurrence (HR 1.07; P<0.001) and dysplastic recurrence (HR 1.11; P = 0.001).

Conclusions

The rates of any BE recurrence and dysplastic recurrence were comparable with the two modalities in this nonrandomized study.

Introduction

Endoscopic eradication therapy (EET) is recommended for the management of patients with dysplastic Barrett’s esophagus (BE) and intramucosal adenocarcinoma (IMC) and has been proven to be safe and effective [1]. EET typically involves endoscopic resection of all visible abnormalities, followed by ablation of residual BE mucosa with the goal being to achieve complete remission of dysplasia (CRD) and intestinal metaplasia (CRIM).

Radiofrequency ablation (RFA) is currently considered the preferred ablative modality based on evidence from randomized controlled trials [2] [3]. A meta-analysis reported pooled percentages for achieving CRIM of 78% and CRD of 91% of the patients treated with RFA. Recurrence was observed in 13% of patients, and progression to cancer was seen in 0.7% of those who achieved CRIM [4]. A previous study based on the US RFA registry showed a recurrence rate of 20% in patients after achieving CRIM, with recurrence associated with increasing age, BE segment length, and non-Caucasian race [5].

Cryoballoon ablation (CBA) uses a portable compact device with a rotating diffuser that sprays nitrous oxide to induce cryonecrosis. A US multicentric prospective clinical trial reported 1-year CRD and CRIM rates of 97% and 91%, respectively [6]. In a recent single-center study, the cumulative recurrence rate was 1.9% for dysplasia and 14.6% for intestinal metaplasia [7].

A prior propensity-matched cohort study evaluating the efficacy of CBA vs. RFA showed a comparable chance of achieving CRIM and CRD in patients undergoing CBA vs. RFA [8]. The CBA group had a higher stricture rate compared with the RFA group (10.4% vs. 4.4%). A multicenter Dutch study also showed comparable efficacy of CBA (88%) and RFA (90%) for endoscopic BE surface regression, but less post-procedural pain associated with CBA [9].

Despite endoscopic therapy being successful in achieving CRIM, there remains uncertainty regarding the durability of response and the recurrence rates of intestinal metaplasia, dysplasia, and carcinoma. Several studies have reported on recurrence rates of BE (+/− dysplasia) following successful EET [10] [11]; however, there are limited data comparing the long-term durability between RFA and CBA endoscopic ablation.

We aimed to compare the incidence of any BE recurrence (detection of intestinal metaplasia +/− dysplasia/carcinoma) and dysplastic recurrence (detection of intestinal metaplasia + dysplasia/carcinoma) following CRIM in a large two-center cohort of patients with BE-related dysplasia/neoplasia undergoing either CBA or RFA as the first-line ablative therapy. We also assessed predictors of BE recurrence after achievement of CRIM.

Methods

This study was approved by the Mayo Clinic and Northwell Health Institutional Review Boards.

Data collection

We retrospectively interrogated a prospectively maintained database of all patients who had undergone EET and were followed up for the management of dysplastic BE and IMC from October 2014 to April 2024 at these two academic tertiary referral centers.

Patient selection

We included patients who underwent RFA or CBA therapy as their primary ablation modality for the management of dysplastic BE and IMC. CRIM was defined as the endoscopic absence of columnar mucosa in the tubular esophagus and the absence of intestinal metaplasia on endoscopic surveillance biopsies obtained from the tubular esophagus and gastroesophageal junction (GEJ). Patients were required to have had at least one follow-up visit after achieving CRIM. Prior to receiving RFA/CBA ablation, all patients were eligible to undergo endoscopic resection (endoscopic mucosal resection [EMR] and/or endoscopic submucosal dissection [ESD]). Patients who did not have a follow-up visit after achieving CRIM and who received CBA or RFA as a rescue treatment after failure of their primary ablation modality were excluded from the analysis.

Method of endoscopic procedures

All endoscopic procedures were conducted by experienced endoscopists (P.G.I., C.L.L, A.J.T., or Dr. Kenneth K. Wang) who specialize in endoscopic ablation therapies following standard clinical practice guidelines. Standard diagnostic and therapeutic endoscopes from Olympus (Center Valley, Pennsylvania, USA) or Fujinon-Fujifilm Medical Systems USA (Lexington, Massachusetts, USA) were used. Esophagogastroduodenoscopy (EGD) was performed with the patient under sedation or monitored anesthesia care as clinically indicated. The BE mucosa was thoroughly evaluated using standard high definition white-light endoscopy, as well as electronic chromoendoscopy. Long-segment BE was defined as a BE length ≥3 cm.

Modality selection

Patients with BE segments >6 cm were typically treated with RFA as the primary ablation modality, while those with BE segments ≤6 cm were given the choice of RFA or CBA as their primary ablation method and shared decision-making was used to choose the ablation modality. This is owing to the fact that the CBA catheter had only focal ablation capacity, being able to treat a section of up to 2–2.5 cm with a single ablation, making it challenging to ablate circumferential segments >6 cm in one session. In contrast, RFA has a circumferential ablation catheter that can treat longer circumferential segments in an efficient manner. Patients were also informed of data showing reduced post-procedural pain after ablation with CBA compared with RFA, and a collaborative decision was made on which ablation modality to use.

Ablation technique

Radiofrequency ablation

RFA (Medtronic, Minneapolis, Minnesota, USA) is a thermal technique for mucosal ablation using bipolar electrodes arranged on circumferential or focal ablation devices ([Fig. 1] a). Circumferential BE segments >3 cm were ablated using a balloon-based device (HALO 360 or HALO 360 Express; Medtronic), and noncircumferential segments or segments <3 cm were ablated by the focal technique using focal devices available in multiple configurations (HALO 60 or HALO 90; Medtronic). RFA therapy was applied as previously described using standard techniques.

Fig. 1 Endoscopic images showing Barrett’s esophagus being treated using: a radiofrequency ablation; b cryoballoon ablation.

Endoscopy was repeated after 8–12 weeks to ablate any residual BE mucosa until endoscopic remission was achieved, with histology confirming the absence of intestinal metaplasia.

Cryoballoon ablation

CBA was performed with a portable, commercially available system (C2 focal cryoballoon; Pentax Medical Corporation, Montvale, New Jersey, USA) that consists of a small handheld controller housing a capsule containing liquid nitrous oxide that is connected to a disposable, low-pressure, compliant, transparent balloon catheter ([Fig. 1] b). The detailed method for this therapy was previously outlined [9] [12].

Briefly, the CBA catheter was advanced through the biopsy channel of a therapeutic endoscope. During CBA, the tip of the endoscope was placed at the proximal margin of the cryoballoon to enable complete visualization of the treatment area through the transparent balloon and to promote precise targeting and minimal ice patch overlap. The nitrous oxide was released from the controller down the catheter shaft to a small opening in the diffuser centered within the balloon, resulting in balloon inflation and a discrete, well-defined ice patch approximately 2–2.5 cm in diameter after 8–10 seconds (for Gen 2 and Gen 1, respectively) of cryogen application, cooling the tissue to −85°C.

The entire BE segment, from the most proximal extent of the BE to 1 cm into the gastric cardia, was treated circumferentially, moving from distal to proximal. The number of ablations needed to treat the entire BE segment was at the endoscopist’s discretion. Either the focal or pear balloon was used for CBA at the endoscopist’s discretion. Patients could receive four to five ablation sessions within 12 months, with each treatment session separated by 10–12 weeks.

Histopathologic analysis

The baseline histology was defined as the grade of dysplasia observed on initial EGD before any endoscopic therapy. In addition to the resection of any visible lesions, four-quadrant biopsy samples were obtained starting at the GEJ and then every 1 cm through the length of the BE segment. The highest grade of dysplasia detected was the overall histologic grade assigned to that patient on subsequent surveillance events.

Successful treatment was determined as the achievement of CRIM. Recurrence was classified as any recurrence (the histologic presence of intestinal metaplasia, with or without dysplasia) or dysplastic recurrence (the histologic presence of intestinal metaplasia with dysplasia) on biopsy specimens taken from the esophagus and/or the GEJ after CRIM had been achieved. All histologic assessments were performed by pathologists with specialist expertise in gastrointestinal pathology at each site.

Follow-up

All subjects who achieved endoscopic remission (i.e. CRIM on histology) underwent endoscopic surveillance with careful inspection of the BE segment. Biopsy specimens were obtained from the GEJ and the neosquamous mucosa. Flat residual columnar mucosa was treated with CBA, RFA, or other thermal ablation modalities, such as argon plasma coagulation or multipolar electrocoagulation at the treating endoscopist’s discretion.

The frequency of endoscopic surveillance varied based on the initial histologic grade of dysplasia. Patients with low grade dysplasia (LGD) before EET underwent EGD every 6 months for the first year and then every year thereafter. Those with high grade dysplasia (HGD)/IMC at baseline underwent surveillance endoscopies every 3 months for the first year, every 6 months for the following year, and then every year thereafter. All patients enrolled in the study were followed up until the end of the study period in April 2024.

Outcomes

The primary outcome was the incidence rate of any BE recurrence (detection of intestinal metaplasia +/− dysplasia/carcinoma) and dysplastic recurrence (detection of intestinal metaplasia + dysplasia/carcinoma) on biopsy specimens from the tubular esophagus and GEJ on a single surveillance endoscopy after successfully achieving CRIM.

Statistical analysis

Abstracted information included basic demographics, BE segment endoscopic characteristics, treatment details (including dates of endoscopic resection, initial ablation therapy, CRIM, and recurrence of BE), histology, and date of last follow-up. Data were summarized as mean (SD), median (interquartile range [IQR]), or proportions (%), as warranted.

Baseline variables were compared between treatment groups using the Kruskal–Wallis or Pearson’s chi-squared tests, as appropriate. Kaplan–Meier analysis was used to estimate the recurrence over time, starting from the date of achieving CRIM (the date of first pathologic documentation of CRIM). Cox proportional hazard models were used to assess the association of baseline covariates with the outcomes of any/dysplastic recurrence, with Firth adjustments as needed to avoid possible overfitting. Variables of interest in the models were age, sex, body mass index (BMI), history of ever smoking, length of BE at initial procedure, presence of a hiatal hernia, treatment groups (RFA vs. CBA), histology at baseline procedure (LGD vs. HGD/esophageal adenocarcinoma [EAC]), and method of endoscopic resection (EMR, ESD). The alpha level was set at 0.05 for statistical significance.

To decrease the risk of confounding bias, patients in the CBA and RFA groups were also matched using propensity score matching. Propensity scores were estimated by logistic regression analysis, with treatment strategy group (RFA vs. CBA) as the dependent variable and maximum BE length, prior endoscopic resection, BMI, presence of a hiatal hernia, and a random site effect as the independent variables because of imbalance according to standardized differences. Matching was performed in a 1 : 1 ratio (54 RFA cases and 54 CBA cases) using a greedy matching algorithm with a caliper of ±0.25 times the SD of the propensity score logit. Subsequent Cox models were fit to compare the post-CRIM recurrence outcomes between the CBA and RFA groups in the matched sets, which included a random site effect.

Sensitivity analysis performed using the multivariable and propensity-matched Cox models is presented in the subset with a BE length ≤6 cm and excluding recurrence of intestinal metaplasia of the GEJ during follow up.

All analyses were performed using SAS software v.9.4 (SAS Institute Inc., Cary, North Carolina, USA).

Results

We identified 681 patients who achieved CRIM post-EET, of whom 610 had undergone RFA (Mayo Clinic 591 [96.9%]; Northwell 19 [3.1%]) and 71 CBA (Mayo Clinic 35 [49.3%]; Northwell 36 [50.7%]) as their primary ablation modality. Baseline demographics were largely similar ([Table 1]), except for a higher proportion of patients with the presence of a diaphragmatic hernia and long-segment BE in the RFA group. The mean (SD) BE segment length was higher in the RFA group (4.5 [3.4] cm) compared with the CBA group (2.7 [2.0] cm). Long-segment BE was present in 64.3% of patients who underwent RFA compared with 46.5% in those subjected to CBA. The proportion of patients with HGD/IMC was comparable between the RFA (417 [68.4%]) and CBA groups (46 [64.8%]). Approximately two-thirds of patients in each group also underwent endoscopic resection before commencing ablation.

Table 1 Baseline clinical characteristics of the patients with dysplastic Barrett’s esophagus (BE) who underwent radiofrequency ablation (RFA) and cryoballoon ablation (CBA) as their primary ablation therapy.

	RFA (n = 610)	CBA (n = 71)
BMI, body mass index; EMR, endoscopic mucosal resection, ESD, endoscopic submucosal dissection, LGD, low grade dysplasia; HGD, high grade dysplasia; IMC, intramucosal carcinoma.
Age, mean (SD), years	65.2 (10.1)	67.1 (9.1)
Sex, male, n (%)	500 (82.0%)	61 (85.9%)
BMI, mean (SD), kg/m2	30.8 (5.7)	29.6 (5.1)
Ever smoker, n (%)	404 (66.2%)	41 (57.7%)
Diaphragmatic hernia, n (%)	496 (81.3%)	46 (64.8%)
• Hernia length, mean (SD), cm	3.7 (1.5)	2.6 (1.3)
BE length, mean (SD), cm	4.5 (3.4)	2.7 (2.0)
Long-segment BE, n (%)	392 (64.3%)	33 (46.5%)
Baseline BE dysplasia grade
LGD, n (%)	193 (31.6%)	25 (35.2%)
HGD + IMC, n (%)	417 (68.4%)	46 (64.8%)
Prior endoscopic resection, n (%)	420 (68.9%)	46 (64.8%)
Cap EMR, n (%)	390 (63.9%)	35 (49.3%)
ESD, n (%)	38 (6.2%)	11 (15.5%)

Post-CRIM recurrence

The median (IQR) follow-up from the date of achieving CRIM was 4.1 (1.7–7.2) years in the RFA group and 4.4 (3.2–5.1) years in the CBA group. The median number of surveillance endoscopic sessions was six (IQR 3–9) in the RFA group, compared with five (IQR 4–7) in the CBA group. The incidence of any recurrence was higher in the RFA group compared with the CBA group (11.2 vs. 4.4 per 100 patient-years; P = 0.001); however, the incidence of dysplastic recurrence in the RFA and CBA groups (3.75 vs. 2.83 per 100 patient-years; P = 0.66) was comparable.

The cumulative probability of any recurrence and dysplastic recurrence is shown in [Fig. 2]. The grade of dysplasia on recurrent histology is presented in [Table 2]. The majority of patients in the RFA group (66.7%) had nondysplastic BE on recurrence, while the CBA group had nondysplastic BE (36.4%) and indefinite/LGD (36.4%) as the most prevalent histologic types. Approximately 60% of nondysplastic BE recurrence originated from biopsies taken at the GEJ ([Table 2]).

Fig. 2 Kaplan-Meier curves for patients treated by cryoballoon ablation (CBA) and radiofrequency ablation (RFA) comparing the incidence over time (95%CIs marked by shaded areas) of: a any recurrence (4.4 vs. 11.2 per 100 patient-years); b dysplastic recurrence (2.8 vs. 3.7 per 100 patient-years).

Table 2 Distribution of recurrence histologic type and site on follow-up endoscopies after achieving complete remission of intestinal metaplasia following either radiofrequency ablation (RFA) and cryoballoon ablation (CBA).

Grade of dysplasia on recurrence	RFA (n = 219)		CBA (n = 11)
	Tubular esophagus (n = 100)	GEJ (n = 119)	Tubular esophagus (n = 6)	GEJ (n = 5)
EAC, esophageal adenocarcinoma; GEJ, gastroesophageal junction; HGD, high grade dysplasia; LGD, low grade dysplasia; NDBE, nondysplastic Barrett’s esophagus.
NDBE	58 (58.0%)	88 (73.9%)	3 (50.0%)	1 (20.0%)
Indefinite/LGD	25 (25.0%)	22 (18.5%)	1 (16.6%)	3 (60.0%)
HGD	12 (12.0%)	7 (5.9%)	1 (16.6%)	1 (20.0%)
EAC	5 (5.0%)	2 (1.7%)	1 (16.6%)	0

In a multivariable model that adjusted for baseline clinical and histologic variables, we found that ablation modality affected the chance of any recurrence (RFA vs. CBA: hazard ratio [HR] 2.19, 95%CI 1.18–4.06; P = 0.01) ([Table 3]); however, it did not affect the chance of dysplastic recurrence (HR 0.88, 95%CI 0.39–1.97; P = 0.75) ([Table 4]). The baseline length of BE was associated with an increased chance of any recurrence (HR 1.07; P < 0.001) and dysplastic recurrence (HR 1.11; P = 0.001). Dysplastic recurrence was higher in those who had undergone prior endoscopic mucosal resection (HR 2.02, 95%CI 1.10–3.71; P = 0.02) or had HGD/IMC on their baseline pathology (HR 1.95, 95%CI 1.07–3.56; P = 0.03).

Table 3 Multivariable model comparing radiofrequency ablation (RFA) vs. cryoballoon ablation (CBA) for any Barrett’s esophagus (BE) recurrence after achieving complete remission of intestinal metaplasia.

Covariate1	All BE patients (n = 658)		BE segment ≤6 cm (n = 508)
	Hazard ratio (95%CI)	P value	Hazard ratio (95%CI)	P value
EMR, endoscopic mucosal resection; ESD, endoscopic submucosal dissection; HGD, high grade dysplasia; IMC, intramucosal adenocarcinoma; LGD, low grade dysplasia. 1 Covariates included in the model were age, sex, body mass index (BMI), history of ever smoking, length of BE at initial procedure, presence of a hiatal hernia, treatment group (RFA vs. CBA), baseline histology (LGD vs. HGD/IMC), and prior endoscopic resection (EMR, ESD).
Age	1.01 (0.99–1.02)	0.24	1.01 (0.99–1.02)	0.52
Sex, male	1.11 (0.76–1.60)	0.60	1.02 (0.67–1.56)	0.93
BMI	0.99 (0.97–1.02)	0.61	0.999 (0.97–1.03)	0.95
Tobacco use	1.33 (0.99–1.78)	0.06	1.51 (1.05–2.17)	0.03
BE length	1.07 (1.03–1.11)	<0.001	1.23 (1.13–1.34)	<0.001
Diaphragmatic hernia	1.07 (0.74–1.54)	0.73	1.16 (0.76–1.77)	0.50
Prior EMR	1.04 (0.76–1.41)	0.83	1.23 (0.85–1.79)	0.28
Prior ESD	0.42 (0.18–0.97)	0.42	0.46 (0.16–1.28)	0.13
HGD/IMC	1.34 (0.98–1.84)	0.06	1.47 (1.01–2.13)	0.04
RFA	2.19 (1.18–4.06)	0.01	2.14 (1.15–3.99)	0.02

Table 4 Multivariable model comparing radiofrequency ablation (RFA) vs. cryoballoon ablation (CBA) for dysplastic Barrett’s esophagus (BE) recurrence after achieving complete remission of intestinal metaplasia.

Covariate1	All BE patients (n = 658)		BE segment ≤6 cm (n = 508)
	Hazard ratio (95%CI)	P value	Hazard ratio (95%CI)	P value
EMR, endoscopic mucosal resection; ESD, endoscopic submucosal dissection; HGD, high grade dysplasia; IMC, intramucosal adenocarcinoma; LGD, low grade dysplasia. 1 Covariates included in the model were age, sex, body mass index (BMI), history of ever smoking, length of BE at initial procedure, presence of a hiatal hernia, treatment group (RFA vs. CBA), baseline histology (LGD vs. HGD/IMC), and prior endoscopic resection (EMR, ESD).
Age	1.02 (1.00–1.05)	0.11	1.03 (1.00–1.06)	0.10
Sex, male	1.43 (0.79–2.59)	0.24	1.53 (0.78–3.00)	0.21
BMI	0.99 (0.94–1.03)	0.50	0.98 (0.93–1.03)	0.43
Tobacco use	1.50 (0.90–2.50)	0.12	1.54 (0.80–2.94)	0.19
BE length	1.11 (1.04–1.18)	0.001	1.17 (1.01–1.36)	0.04
Diaphragmatic hernia	1.04 (0.56–1.95)	0.90	1.34 (0.64–2.79)	0.44
Prior EMR	2.02 (1.10–3.71)	0.02	2.28 (1.06–4.92)	0.04
Prior ESD	0.81 (0.24–2.73)	0.73	1.23 (0.30–5.08)	0.77
HGD/IMC	1.95 (1.07–3.56)	0.03	1.90 (0.91–3.96)	0.09
RFA	0.88 (0.39–1.97)	0.75	0.80 (0.36–1.76)	0.57

Propensity matching

A propensity score-matched analysis was conducted using a 1:1 matching ratio based on maximum BE length, prior endoscopic resection, BMI, presence of hiatal hernia, and a random site effect. Baseline characteristics of the 54 CBA and 54 RFA cases are shown in Table 1s, see online-only Supplementary material. The propensity score-matched analysis revealed a similar risk of any BE recurrence post-CRIM in the RFA and CBA groups (HR 1.01, 95%CI 0.39–2.64; P = 0.98). A comparable risk of dysplastic BE recurrence post-CRIM was also observed in both groups (HR 0.18, 95%CI 0.02–1.52; P = 0.12). Fig. 1s shows a love plot depicting appropriate balancing between the CBA and RFA groups for all variables included in the propensity score model, with all standardized differences, including the length of the BE segment, being within the recommended ±0.25.

Sensitivity analysis in patients with BE ≤6 cm

In patients with a maximum BE length ≤6 cm (n = 527: RFA n = 458; CBA n = 69), the median (IQR) follow-up was 4.0 (1.6–7.2) years for RFA compared with 4.4 (3.2–5.1) years in the CBA group. Baseline characteristics are shown in Table 2s. Recurrence was seen in 11 CBA patients and 151 RFA patients (Table 3s). The incidence of any recurrence (10.11 vs. 4.51 per 100 patient-years; P = 0.004) was higher in the RFA group compared with the CBA group; however, the incidence of dysplastic recurrence was comparable in the RFA and CBA groups (3.01 vs 2.87 per 100 patient-years; P = 0.93). The cumulative probabilities of any recurrence and dysplastic recurrence are shown in Figs. 2s and 3s, respectively.

In a multivariable model among patients with a BE length ≤6 cm ([Table 3]), the ablation modality (RFA) was associated with any recurrence (HR 2.14, 95%CI 1.15–3.99; P = 0.02); however, no significant association was observed with dysplastic recurrence (HR 0.80, 95%CI 0.36–1.76; P = 0.57) ([Table 4]). Propensity score-matched analysis was performed for those with a BE length ≤6 cm (100 patients) (Fig. 4s), and showed a nonsignificantly higher chance of any BE recurrence post-CRIM in the RFA group vs. the CBA group (HR 2.10, 95%CI 0.82–5.36; P = 0.12) and comparable risk of dysplastic BE recurrence (HR 0.50, 95%CI 0.09–2.73; P = 0.42).

Sensitivity analysis with exclusion of nondysplastic BE recurrence at the GEJ

Approximately 60% of patients in the RFA group (n = 88) and 20% patients in the CBA group (n = 1) had nondysplastic recurrence at the GEJ. In a multivariable model treating these patients as not having recurrent events (i.e. defining recurrence as only those with nondysplastic recurrence in the tubular esophagus or a dysplastic recurrence at the tubular esophagus/GEJ), we observed a comparable risk of “any recurrence” (RFA vs. CBA: HR 1.18, 95%CI 0.61–2.30; P = 0.62). Propensity matched analysis showed similar results (HR 0.56, 95%CI 0.17–1.87; P = 0.35).

Discussion

In this study with a median length of follow-up of more than 4 years following CRIM, we observed that the incidence of any BE recurrence was not significantly higher in patients achieving CRIM following RFA compared with those treated by CBA on propensity score-matched analyses. The incidence of dysplastic recurrence was also comparable between RFA and CBA after successful EET. These results remained consistent on a priori planned sensitivity analyses in patients with maximum BE segment length ≤6 cm. The recurrence risk was also comparable between the RFA and CBA group when nondysplastic recurrences at the GEJ were excluded from the analysis. Increasing length of the BE segment and the presence of HGD or IMC prior to ablation were associated with a higher risk of dysplastic recurrence.

A previous meta-analysis reported post-RFA pooled recurrence rates of intestinal metaplasia, dysplastic BE, and HGD/EAC as 9.5%, 2.0%, and 1.2% per patient-year, respectively [13]. Increasing age and BE length were predictive of recurrence. Another study comparing recurrence rates between stepwise radical endoscopic resection and RFA demonstrated higher overall recurrence rates [8.6 vs. 5.1 per 100 patient-years; P = 0.01] but comparable rates of dysplasia recurrence (2.5 vs 1.2 per 100 patient-years; P = 0.11) [14]. A multicentric study reported an estimated annual incidence rate of 9.6% for any recurrence and 2.8% for dysplastic recurrences, which decreased to 4.3% for any recurrence after excluding nondysplastic BE recurrence at the GEJ [15]. We observed similar incidences of 11.2 per 100 patient-years for any recurrence and 3.75 per 100 patient-years for dysplastic recurrence in the RFA group in our study.

Dbouk et al. demonstrated an intestinal metaplasia recurrence rate of 5 per 100 person-years in patients treated with CBA. Recurrent intestinal metaplasia was seen in 14.6% of the patients (6 in esophagus and 1 in GEJ among a total of 59 patients) after a median of 20.7 months following CRIM [7]. Our study showed a similar recurrence rate (4.4 per 100 patient-years, overall 15% of patients).

Despite being an effective therapy, RFA has been associated with post-procedural pain, high cost, strictures, and recurrence risk in subsequent years requiring lifelong surveillance of patients. Consequently, other endoscopic ablative modalities, such as cryotherapy, have been developed. Liquid nitrogen spray cryotherapy is a noncontact method of BE eradication. A meta-analysis studying liquid nitrogen spray therapy showed pooled intestinal metaplasia recurrence rates of 14.8% after achieving CRIM [16]. Another single-center study assessing liquid nitrogen spray cryotherapy for BE-associated HGD/carcinoma showed recurrence rates of 12.2%, 4.0%, and 1.4% per person-year over a follow-up period of 5 years for intestinal metaplasia, dysplasia, and HGD/carcinoma, respectively [17].

CBA offers several unique features that make it an easy-to-use and efficient endoscopic ablation modality. Compared with spray cryotherapy, CBA is time efficient and eliminates the need for additional measures such as gas venting, abdominal compression, or repeating of the freeze–thaw cycle. The procedure is also facilitated by portable equipment, a small footprint controller, and disposable catheters. A novel 180° cryoballoon (CbAS180) has recently been developed, enabling semicircumferential treatment over a 3- cm length, offering a decreased risk of overlap between ablations, shortening the procedural time, and offering a high technical success rate of 96% [18].

Our study has several notable strengths. First, we commenced our analysis from the achievement of a clinically relevant histologic event (i.e. CRIM). We classified our outcomes as any recurrence and dysplastic recurrence, which further enhances the robustness and clinical applicability of our findings. Moreover, while our data were retrospectively analyzed, they were obtained from prospectively maintained and consistently updated databases. Our ability to examine patient charts allowed us to abstract in granular detail a variety of important variables that affect the clinical management of these patients, which ensures reliability and accuracy. Additionally, we conducted robust statistical analyses incorporating both multivariable regression analysis and propensity score matching to adjust for known confounders and reduce the risk of bias. Consistent results in both these analyses increase our confidence in the results.

We acknowledge certain limitations in our study. First, our patient population represents a cohort treated at two academic tertiary referral centers, which might limit the generalizability of our findings; however, current societal recommendations advocate that EET should be conducted at centers with sufficient volume and expertise in endoscopic resection and ablation [19] [20].

Another limitation is the absence of randomization between the treatment groups. Patients with BE segments >6 cm were treated with RFA and patients with BE segments ≤6 cm were provided with the option of choosing between CBA or RFA, which may be a source of indication bias. We attempted to mitigate the effects of these confounding variables using propensity score matching (Fig. 1s) and also by performing a subset analysis on patients with BE segments ≤6 cm, which demonstrated similar results. Although differences in the baseline characteristics between the RFA and CBA groups could potentially influence our results, it is notable that multivariable regression and propensity score-matched analyses did not demonstrate any significant difference in dysplastic BE recurrence between the modalities. Both approaches help address potential biases and strengthen the reliability of our results.

In conclusion, CBA is a comparably effective and durable alternative modality to RFA for EET in terms of both any BE recurrence and dysplastic BE recurrence. This study highlights the importance of surveillance endoscopic evaluation and treatment, particularly in those with long-segment BE undergoing successful EET. The choice of therapy should be individualized, considering the risks and benefits, and engaging in shared clinical decision-making. The recurrence rate of BE is low but, based on current results, it is imperative to understand that current endoscopic therapy provides a treatment rather than a definitive cure and this emphasizes the need for continued meticulous surveillance. Reassuringly, most recurrences can be treated endoscopically.

Conflict of Interest

A.J. Trindade has provided consultancy to Exact Sciences, Lucid Diagnostics, and Boston Scientific, and has received research funding from Exact Science and Lucid Diagnostics. P.G. Iyer has received research funding from Exact Sciences, Pentax Medical, and CDx Medical, and provided consultancy to Exact Sciences, Pentax Medical, CDx Medical, and Castle Biosciences. K. Sachdeva, P. Chandi, A. Verma, R. Dierkhising, D.C. Codipilly, and C.L. Leggett declare that they have no conflict of interest.

Acknowledgement

We wish to acknowledge the support of Dr. Kenneth K. Wang, an endoscopist at our institution who performed some of the procedures included in this study.

• References

• 1 Shaheen NJ, Falk GW, Iyer PG. et al. Diagnosis and management of Barrett's esophagus: an updated ACG guideline. Am J Gastroenterol 2022; 117: 559-587
• 2 Shaheen NJ, Sharma P, Overholt BF. et al. Radiofrequency ablation in Barrett's esophagus with dysplasia. NEJM 2009; 360: 2277-2288
• 3 Phoa KN, van Vilsteren FG, Weusten BL. et al. Radiofrequency ablation vs endoscopic surveillance for patients with Barrett esophagus and low-grade dysplasia: a randomized clinical trial. JAMA 2014; 311: 1209-1217
• 4 Orman ES, Li N, Shaheen NJ. Efficacy and durability of radiofrequency ablation for Barrett's Esophagus: systematic review and meta-analysis. Clin Gastroenterol Hepatol 2013; 11: 1245-1255
• 5 Pasricha S, Bulsiewicz WJ, Hathorn KE. et al. Durability and predictors of successful radiofrequency ablation for Barrett’s esophagus. Clin Gastroenterol Hepatol 2014; 12: 1840-1847
• 6 Canto MI, Trindade AJ, Abrams J. et al. Multifocal cryoballoon ablation for eradication of Barrett's esophagus-related neoplasia: A prospective multicenter clinical trial. Am J Gastroenterol 2020; 115: 1879-1890
• 7 Dbouk M, Simons M, Wang B. et al. Durability of cryoballoon ablation in neoplastic Barrett's esophagus. Techn Innov Gastrointest Endosc 2022; 24: 136-144
• 8 Agarwal S, Alshelleh M, Scott J. et al. Comparative outcomes of radiofrequency ablation and cryoballoon ablation in dysplastic Barrett’s esophagus: a propensity score− matched cohort study. Gastrointest Endosc 2022; 95: 422-431
• 9 van Munster SN, Overwater A, Haidry R. et al. Focal cryoballoon versus radiofrequency ablation of dysplastic Barrett’s esophagus: impact on treatment response and postprocedural pain. Gastrointest Endosc 2018; 88: 795-803
• 10 Sawas T, Alsawas M, Bazerbachi F. et al. Persistent intestinal metaplasia after endoscopic eradication therapy of neoplastic Barrett’s esophagus increases the risk of dysplasia recurrence: meta-analysis. Gastrointest Endosc 2019; 89: 913-925
• 11 van Vilsteren FG, Pouw RE, Seewald S. et al. Stepwise radical endoscopic resection versus radiofrequency ablation for Barrett's oesophagus with high-grade dysplasia or early cancer: a multicentre randomised trial. Gut 2011; 60: 765-773
• 12 Schölvinck DW, Künzli HT, Kestens C. et al. Treatment of Barrett’s esophagus with a novel focal cryoablation device: a safety and feasibility study. Endoscopy 2015; 47: 1106-1112
• 13 Krishnamoorthi R, Singh S, Ragunathan K. et al. Risk of recurrence of Barrett’s esophagus after successful endoscopic therapy. Gastrointest Endosc 2016; 83: 1090-1106
• 14 Fujii-Lau LL, Cinnor B, Shaheen N. et al. Recurrence of intestinal metaplasia and early neoplasia after endoscopic eradication therapy for Barrett’s esophagus: a systematic review and meta-analysis. Endosc Int Open 2017; 5: E430-E449
• 15 Sami SS, Ravindran A, Kahn A. et al. Timeline and location of recurrence following successful ablation in Barrett’s oesophagus: an international multicentre study. Gut 2019; 68: 1379-1385
• 16 Chandan S, Bapaye J, Khan SR. et al. Safety and efficacy of liquid nitrogen spray cryotherapy in Barrett’s neoplasia–a comprehensive review and meta-analysis. Endosc Int Open 2022; 10: E1462-E1473
• 17 Ramay FH, Cui Q, Greenwald BD. Outcomes after liquid nitrogen spray cryotherapy in Barrett's esophagus–associated high-grade dysplasia and intramucosal adenocarcinoma: 5-year follow-up. Gastrointest Endosc 2017; 86: 626-632
• 18 Overwater A, van Munster SN, Nagengast WB. et al. Novel cryoballoon 180° ablation system for treatment of Barrett’s esophagus-related neoplasia: a first-in-human study. Endoscopy 2022; 54: 64-70
• 19 Fudman DI, Lightdale CJ, Poneros JM. et al. Positive correlation between endoscopist radiofrequency ablation volume and response rates in Barrett's esophagus. Gastrointest Endosc 2014; 80: 71-77
• 20 Sharma P, Shaheen NJ, Katzka D. et al. AGA clinical practice update on endoscopic treatment of Barrett’s esophagus with dysplasia and/or early cancer: expert review. Gastroenterology 2020; 158: 760-769

Publikationsverlauf

Eingereicht: 19. November 2024

Angenommen nach Revision: 30. April 2025

Accepted Manuscript online:
30. April 2025

Artikel online veröffentlicht:
28. Mai 2025