Introduction
Barrett’s esophagus (BE) is an important risk factor for development of esophageal
adenocarcinoma (EAC), incidence of which continues to rise out of proportion to other
malignancies [1]
[2]
[3]. The condition is characterized by a sequential progression from intestinal metaplasia
(IM) to low-grade dysplasia (LGD) to high-grade dysplasia (HGD), and eventually to
EAC. Annual risk of EAC is approximately 0.5 % in patients with LGD [4] and 4 % to 8 % in patients with HGD [5]
[6]. Because more than 50 % of cases of invasive EAC present with incurable locally
advanced or metastatic disease, therapy for BE presents an opportunity to halt neoplastic
progression before cancer develops. Consequently, clinical practice guidelines [7]
[8]
[9]
[10]
[11] recommend endoscopic surveillance of patients with known BE and endoscopic eradication
therapy (EET) for those who are found to have confirmed HGD or intramucosal carcinoma
(IMC). In addition, some data support use of EET for LGD to prevent progression to
more advanced lesions [12].
Nodular BE requiring treatment is commonly eradicated with endoscopic mucosal resection
(EMR) whereas flat disease is amenable to radiofrequency ablation (RFA). RFA uses
thermal energy to destroy tissue to a depth of 1000 microns. Since results of the
Ablation of Intestinal Metaplasia (AIM)-dysplasia trial were published in 2009 [13], RFA has been extensively studied and used for treatment of dysplastic BE. RFA’s
efficacy and durability have made it the most evidence-based, and often the first-line
treatment for dysplastic BE [14]
[15]. In contrast, cryotherapy is a more novel treatment option which results in mucosal
ablation by delivery of cryogen that causes tissue destruction as a result of extremely
cold temperatures. Spray cryotherapy utilizes a spray catheter through which cryogen
is applied directly to the esophageal mucosa. Alternatively, balloon cryotherapy uses
a self-contained balloon-delivery system which is inflated in the esophagus and cryogen
is sprayed into the inside of the balloon in a targeted four-quadrant fashion. Originally
applied in the fields of dermatology, urology, and gynecology, cryotherapy has been
found to be safe and effective in treatment of BE [16]
[17]
[18]
[19]. However, data remain limited and rigorous comparative efficacy studies relative
to RFA have not been published.
Lacking large placebo-controlled trials and robust comparative data, endoscopists
have been reluctant to use cryotherapy as a first-line treatment. However, failure
of RFA has often been attributed to the fact that some areas of BE may be too thick.
Therefore, cryotherapy, which is thought to produce a deeper treatment effect, has
been increasingly used as salvage therapy in patients who have had an incomplete response
to RFA or combined EMR and RFA. Hypothesizing that cryotherapy is safe and effective
in a salvage capacity, we aimed to assess clinical outcomes among a large cohort of
patients who had failed conventional treatments.
Patients and methods
Study design
This was a retrospective cohort study of consecutive cases of dysplastic BE or IMC
treated with EET at a single tertiary care academic medical center (Medical University
of South Carolina in Charleston, South Carolina, United States) over a 12-year period
from 2007 to 2018. Cryotherapy was first used at our center in 2012, but prior endoscopic
interventions from as early as 2007 were included. Our local institutional review
board approved the study. Patients who declined consent for their medical records
to be used for research were excluded.
Patients
Potentially eligible patients were identified using a clinical database and electronic
medical record query. Records were manually reviewed to identify patients who met
the following inclusion criteria: 1) pathology reviewed at our center showing BE with
dysplasia or IMC arising from BE; and 2) underwent EET using cryotherapy as salvage
therapy at our center. Two investigators (BJE, PE) independently reviewed procedure
reports for each potentially eligible subject to adjudicate whether cryoablation was
implemented primarily as salvage therapy due to refractoriness to EMR and/or RFA.
A subject was deemed to have undergone salvage cryoablation if one or more of the
following criteria were satisfied: (1) the procedure report explicitly stated that
cryoablation was for salvage purposes; (2) the procedure report explicitly stated
that cryoablation was selected because the BE was refractory to EMR and/or RFA; (3)
nodularity or IMC was present, but EMR was not feasible due to non-lifting or inability
to adequately capture target tissue; or (4) cryoablation was used after both EMR and
RFA were performed, but prior to achieving remission. Cases in which cryoablation
was used even though EMR or RFA were suitable and feasible were not considered salvage.
For example, patients who underwent cryoablation as the first modality or those who
underwent cryoablation of flat BE after EMR of nodularity were not considered eligible.
Cases in which cryoablation was used as salvage therapy, including those in which
other treatment modalities (RFA, EMR) were subsequently used after the initial cryotherapy
session, were included.
Treatment and follow-up
EET was performed by one of four endoscopists who had at least 5 years of experience
with advanced BE treatment. In some cases, advanced endoscopy fellows assisted under
the direct supervision of the attending endoscopists. Consistent with standard practice,
EMR was typically attempted for nodular lesions, those with known IMC, and those that
were deemed by the endoscopist to have a morphologically concerning appearance. RFA
was generally reserved for flat BE with HGD using the treatment protocol described
by Shaheen et al [13]. Cryotherapy was performed using both the spray and balloon liquid nitrogen delivery
systems. Liquid nitrogen spray cryotherapy (truFreeze Spray Cryotherapy, Lexington,
Massachusetts, United States) was the mainstay of cryotherapy treatment at our center
until early 2017. Patients who received treatment with cryotherapy after this date
received either spray cryotherapy or balloon cryotherapy (C2 CryoBalloon, Pentax Medical,
Redwood City, California, United States). If patients were responding to spray cryotherapy,
the delivery modality was not altered. Ablation sessions generally occurred every
1 to 3 months. In some cases, argon plasma coagulation (APC) was used in limited capacity
as an adjunct therapy at the discretion of the endoscopist. The primary goal of cryotherapy
treatment was eradication of all neoplasia. The secondary goal was eradication of
all intestinal metaplasia.
Follow-up endoscopy typically occurred 2 to 3 months after the last treatment session.
Once remission was achieved, targeted biopsies were obtained from any visible lesions
concerning for recurrence. In addition, random four-quadrant biopsies were obtained
every 1 to 2 cm from the gastroesophageal junction and the prior treatment area (neosquamous-lined
esophagus) according to the Seattle protocol [20]. Patients received twice-daily proton pump inhibitors as maintenance therapy. Biopsy
and EMR specimens were evaluated by two expert gastrointestinal pathologists for presence
of BE, dysplasia, and cancer according to standard definitions. Pathologists were
not blinded to endoscopic treatments.
Outcomes
We aimed to determine the proportion of patients undergoing salvage cryotherapy who
achieved CE-D and CE-IM, defined according to accepted pathologic standards. Recurrence
was defined as intestinal metaplasia or dysplasia identified on post-remission surveillance
biopsies prompting reinitiation of EET. Adverse events (AEs) related to EET were defined
according to published data [21].
Data collection and analysis
Data collected included age, gender, race and ethnicity, body mass index (BMI), tobacco
use history, alcohol use history, personal and family cancer history, gastrointestinal
surgical history, prior gastroesophageal reflux disease (GERD), presence and characteristics
of hiatal hernia, BE segment length, Prague classification, presence of nodules, intervention
type (RFA, cryotherapy, EMR), location and extent of treatment, and pathology results.
The index endoscopy, defined as the first treatment session at our institution, was
used to calculate time to CE-D and CE-IM. Results for the primary analysis are presented
using descriptive statistics. In an exploratory analysis, stepwise logistic regression
was used to identify factors associated with CE-D in patients undergoing salvage cryotherapy.
Results
Over a 12-year period, 352 patients underwent EET at our center. Eighty-one of them
underwent cryotherapy as part of their treatment regimen. Of that group, 46 patients
received salvage cryotherapy ([Fig. 1]). The majority of these patients (40/46) received spray cryotherapy as salvage therapy,
while six received balloon cryotherapy. Six patients in the cohort had previously
failed to respond to RFA at another institution. The majority of patients were white
(98 %), male (91 %), and the median age at the time of index endoscopy at our center
was 66 years ([Table 1]). Most had a significant smoking history (52 % had at least a 20-pack-year history)
and the median BMI was 27. A majority (87 %) had a history of symptomatic GERD. Six
patients had undergone previous fundoplication. Most (98 %) had a hiatal hernia, with
median size of 3 cm. A majority of patients (54%) had HGD at referral, while 33 %
had LGD, and 13 % had IMC. All patients had long-segment BE (≥ 3 cm) with 59 % of
patients having nodularity noted. All patients underwent RFA treatment prior to cryotherapy,
either at our center or prior to referral, and 50 % of patients underwent previous
EMR at our institution. Pathology immediately prior to treatment with cryotherapy
is reported in [Fig. 2].
Fig. 1 Patient flow diagram.
Table 1
Cohort characteristics.
|
Treatment cohort
(n = 46)
|
Achieved CE-D
(n = 38)
|
No CE-D
(n = 8)
|
|
Age, median
|
66
|
66
|
63
|
|
Male %, (n/N)
|
91 (42/46)
|
89 (34/38)
|
100 (8/8)
|
|
Tobacco history %, (n/N)
|
|
|
37 (17/46)
|
42 (16/38)
|
13 (1/8)
|
|
|
11 (5/46)
|
8 (3/38)
|
25 (2/8)
|
|
|
52 (24/46)
|
50 (19/38)
|
63 (5/8)
|
|
Alcohol %, (n/N)
|
|
|
57 (26/46)
|
58 (22/38)
|
50 (4/8)
|
|
|
11 (5/46)
|
5 (2/38)
|
38 (3/8)
|
|
|
33 (15/46)
|
37 (14/38)
|
13 (1/8)
|
|
BMI, median
|
27
|
27
|
26
|
|
History of symptomatic GERD, %, (n/N)
|
87 (40/46)
|
95 (36/38)
|
50 (4/8)
|
|
Hiatal hernia > 3 cm %, (n/N)
|
26 (12/46)
|
29 (11/38)
|
13 (1/8)
|
|
Nodularity %, (n/N)
|
|
|
41 (19/46)
|
39 (15/38)
|
50 (4/8)
|
|
|
35 (16/46)
|
37 (14/38)
|
25 (2/8)
|
|
|
20 (9/46)
|
21 (8/38)
|
13 (1/8)
|
|
|
4 (2/46)
|
3 (1/38)
|
13 (1/8)
|
|
Initial pathology %, (n/N)
|
|
|
33 (15/46)
|
34 (13/38)
|
25 (2/8)
|
|
|
54 (25/46)
|
50 (19/38)
|
75 (6/8)
|
|
|
13 (6/46)
|
16 (6/38)
|
0 (0/8)
|
|
Prior RFA %, (n/N)
|
100 (46/46)
|
100 (38/38)
|
100 (8/8)
|
|
Prior EMR %, (n/N)
|
50 (23/46)
|
47 (18/38)
|
63 (5/8)
|
CE-D, complete eradication of dysplasia; PY, pack years; BMI, body mass index; LGD,
low-grade dysplasia; HGD, high-grade dysplasia; IMC, intramucosal carcinoma; RFA,
radiofrequency ablation; EMR, endoscopic mucosal resection
Fig. 2 Initial pathology, nodularity, and pre-cryotherapy treatments.
Of the 46 patients who underwent salvage cryotherapy, 38 (83 %) reached CE-D and 21
(46 %) reached CE-IM ([Fig. 3]). In 15 of the 17 patient who reached CE-D, but not CE-IM, additional treatment
was intentionally discontinued because competing illness rendered the risk-benefit
ratio of ongoing treatment unfavorable. The remaining two patients are undergoing
ongoing treatment with a goal of CE-IM. Of the eight patients who had not achieved
CE-D or CE-IM, five are still undergoing treatment, two were lost to follow-up, and
one developed endoscopically untreatable esophageal cancer.
Fig. 3 Pre-cryotherapy pathology and outcomes.
Among the 38 patients who achieved CE-D, median time to CE-D was 18 months, median
number of total interventions (RFA, cryotherapy, and EMR) at our center was five and
median number of cryoablation sessions was two ([Table 2]). Among the 21 patients who achieved CE-IM, median time to CE-IM was 22 months,
median number of total interventions at our center was six, and median number of cryoablation
sessions was two. In four cases, patients reached CE-D with RFA and/or EMR, and cryotherapy
was first used in an attempt to reach CE-IM. In each of these cases, CE-IM was obtained
with a median number of two cryotherapy sessions. All patients with IMC in our study
sample had T1a lesions. Of the six patients, five had lesions that were resected endoscopically,
in accordance with clinical practice guidelines. Cryotherapy in these patients was
used to eradicate residual BE after EMR. A single patient with IMC (T1a lesion) had
disease not amenable to EMR and underwent cryotherapy as salvage therapy and eventually
reached CE-IM.
Table 2
Results.
|
Achieved CE-D
(n = 38)
|
Achieved CE-IM
(n = 21)
|
No CE-IM
(n = 17)
|
|
Time to CE-D, median months (range)
|
18 (4 – 59)
|
15 (5 – 59)
|
26 (4 – 59)
|
|
Median number of RFA treatments at our center to achieve CE-D (range)
|
2 (0 – 5)
|
2 (0 – 5)
|
2 (0 – 5)
|
|
Median number of EMR treatments at our center to achieve CE-D (range)
|
0 (0 – 4)
|
0 (0 – 3)
|
1 (0 – 4)
|
|
Median number of cryotherapy treatments at our center to achieve CE-D (range)
|
2 (0 – 10)
|
2 (0 – 8)
|
2 (1 – 10)
|
|
Median number of total interventions at our center to achieve CE-D (range)
|
5 (1 – 14)
|
4 (1 – 10)
|
5 (2 – 14)
|
|
Time to CE-IM, median months (range)
|
n/a
|
22 (7 – 65)
|
n/a
|
|
Median number of RFA treatments at our center to achieve CE-IM (range)
|
n/a
|
2 (0 – 8)
|
n/a
|
|
Median number of EMR treatments at our center to achieve CE-IM (range)
|
n/a
|
0 (0 – 3)
|
n/a
|
|
Median number of cryotherapy treatments at our center to achieve CE-IM (range)
|
n/a
|
2 (1 – 8)
|
n/a
|
|
Median number of total interventions at our center to achieve CE-IM (range)
|
n/a
|
6 (1 – 12)
|
n/a
|
CE-D, complete eradication of dysplasia; RFA, radiofrequency ablation; EMR, endoscopic
mucosal resection; CE-IM, complete eradication of intestinal metaplasia.
Two patients who achieved remission had recurrence during follow-up. In one patient
who reached CE-D at our center after two RFA sessions, recurrence with LGD was seen
after 48 months. The patient was treated with six additional RFA treatments, again
reached CE-D, then received one cryotherapy to reach CE-IM. Another patient was found
to have a recurrence of HGD 25 months after CE-IM, which was successfully eradicated
by EMR. That patient has since undergone one additional surveillance exam with no
evidence of recurrence.
Three patients developed treatment-related strictures requiring dilation. One patient
developed a stricture 1 month after the third cryotherapy treatment and required seven
dilations, although cryotherapy was able to be resumed and CE-IM was obtained in 19
months following six cryotherapy sessions. Another developed a stricture 1 month after
initial cryotherapy treatment and required four dilations. Again, cryotherapy was
able to be resumed and CE-IM was obtained in 29 months following two cryotherapy sessions.
A third patient developed a stricture following a second session of cryotherapy which
had CE-D. One dilation was performed and further treatment was not pursued.
Discussion
We present a 12-year retrospective cohort study of consecutive cases of refractory
dysplastic BE or IMC treated with salvage cryotherapy at a single tertiary care academic
medical center. Our study demonstrated CE-D and CE-IM rates of 82.6 % and 45.6 %,
respectively, in this cohort of patients who failed conventional treatment. Among
patients who achieved CE-D, median time to CE-D was 18 months and median number of
cryoablation sessions was two. In addition, there was a very low progression to invasive
EAC (one case; 2.1 %) in this high-risk cohort, despite six patients having IMC at
referral. The stricture rate was 6.5 %; there were no other serious complications.
Since the development and initial clinical use of esophageal cryotherapy in 1999 [22], studies in treatment-naive patients have demonstrated encouraging efficacy, safety,
and durability [18]
[23]
[24]
[25]. In addition, because it can be applied to non-lifting and nodular tissue, cryotherapy
has become increasingly popular as a salvage modality when patients are not candidates
for, or have experienced suboptimal response to EMR, RFA, or the combination. Given
its molecular effects, cryoablation allows for deeper tissue ablation, which makes
it an attractive treatment option for RFA-refractory BE, which may be the result of
increased mucosal thickness [26]. However, absence of level I data has limited its widespread use. A recent review
and meta-analysis of 11 studies comprising 148 BE patients treated with cryotherapy
for persistent IM or dysplasia after RFA [19] demonstrated CE-D and CE-IM in 76.0 % and 45.9 % of patients, respectively. Our
study, which to our knowledge is the largest single-center report on this topic, demonstrated
similar CE-D and CE-IM rates, affirming the efficacy of cryoablation in the salvage
capacity. Despite the high-risk nature of our cohort—54 % with HGD and 13 % with IMC—there
was a very low progression to invasive EAC (2.1 %). This is similar to the 2.9 % progression
reported by Canto et al [16].
The median number of cryotherapy sessions required per patient to achieve CE-D was
two, which is fewer than has been reported previously. Canto et al [16] reported a median of four CO2 cryotherapy sessions to achieve a complete response for HGD. Previously studies have
cited comparable data for liquid nitrogen cryotherapy (mean 3.9) [18], APC (mean 4) [27], and for RFA (mean 3.5) [13]. The most likely explanation is that we allowed for continued treatment with RFA,
EMR, and APC concurrently with cryotherapy treatment at the discretion of the endoscopist.
In patients who reached CE-D, the median number of total interventions (RFA, cryotherapy,
and EMR) at our center was five, which is comparable to published data. Based on our
data, although cryotherapy played an important role in salvage therapy, outcomes may
not have been solely due to this intervention. The additive impact of continued RFA,
EMR, and APC may have contributed to the CE-D and CE-IM rates observed in this study.
On this basis, the precise role of cryotherapy in the dysplastic BE treatment algorithm
remains debatable. Some experts favor a more limited role for cryotherapy, arguing
that the population of patients whose disease is truly RFA-refractory should be much
smaller than commonly reported due to the lack of stringent adherence to a strict
RFA treatment algorithm [26]. According to this philosophy, only those whose disease meets the strict definition
of RFA-refractory should be considered for cryotherapy. In contrast, although impossible
to fully infer intent in a retrospective cohort, the approach to cryotherapy in our
study was more liberal. Future studies will clarify the precise place for cryotherapy
in the treatment algorithm and whether strict adherence to the definition of RFA refractoriness
is clinically important.
Our study adds to the excellent safety profile reported for cryotherapy. Aside from
stricture formation, no other AEs were noted, which is comparable to the low (2.9 %)
overall serious AE rate reported in CO2 cryotherapy by Canto et al [16] and published AE data for RFA (3.4 %) [28]. Treatment was complicated by stricture formation requiring dilation in three patients
(6.5%), which is comparable to RFA (6.0 % to 7.6 %) [13]
[28]. Of these three patients in our study, two were able to resume treatment and reach
CE-IM. Furthermore, our study is an appraisal of real-world treatment of unselected
patients by four endoscopists, adding to generalizability compared to prior studies
that involved only one or two expert endoscopists.
The results of this study should be considered in the context of several important
limitations. First, retrospective collection of data could not ensure definitive and
systematic determination of whether cryotherapy was used in a salvage capacity. This
eligibility criterion was adjudicated post hoc by manual review of procedure reports
using aforementioned criteria that are based on clinical assumptions aiming to determine
endoscopist intent. While we believe these assumptions are reasonable, cases may have
been misclassified if the intent was misinterpreted, potentially biasing study results.
Second, the small sample size, retrospective nature, non-randomized design, and lack
of a comparison arm limit any definitive conclusions. In addition, pathologic interpretation
of BE samples was not performed in a blinded manner. Lastly, because treatments were
not limited to cryotherapy and included alternative endoscopic therapies, the study
allowed for a more real-world experience, but that makes the multimodal treatment
data more difficult to interpret.
Conclusion
In summary, this study showed that cryotherapy appears effective for salvage treatment
of patients with refractory dysplastic BE and IMC, successfully achieving CE-D and
CE-IM in of 82.6 % and 45.6 % of patients, respectively. Higher-quality studies, ideally
including randomized trials, are needed to confirm these findings and establish the
exact role of cryotherapy in the treatment armamentarium for BE.
