Long-term follow-up after RFA with or without prior EMR for early Barrett's neoplasia

A relatively new form of endoscopic ablation therapy for Barrett's epithelium,
is radiofrequency ablation (RFA), by using radiofrequency energy the mucosa can
be ablated up to the superficial submucosa. Afterwards the mucosa regenerates
with neo-squamous epithelium (NSE). In patients with non-dysplastic Barrett's
esophagus (BE), RFA has been demonstrated to be safe and effective in the
short- and long term (5 years) for eradication of IM, when applied for BE
containing low-grade intraepithelial neoplasia (LGIN) results are similar, with
favourable rates for complete eradication of dysplasia and IM. Furthermore, RFA
is a modality with a low complication rate, low rate of esophageal stenosis and
pre-existing genetic abnormalities in BE are absent in the NSE that regenerates
after RFA.

To investigate whether RFA can be used effectively to treat patients with
high-grade intra-epithelial neoplasia (HGIN) or early carcinoma (EC), in
combination with endoscopic resection, two clinical trial studies (AMC-I and
AMC-II) including 23 patients, were conducted at the Academic Medical Centre
(AMC) in 2005/2006. Complete eradication of dysplasia and IM was achieved in
all patients. After a median follow-up of 14 months no recurrence of dysplasia
or IM was observed. There were no adverse events or strictures and in none of
the biopsies obtained from NSE buried Barrett glands were found.

Buried Barrett*s, a condition in which residual areas with IM are hidden
underneath NSE is an issue to address when using ablation therapy. Ablative
modalities like photodynamic therapy (PDT) or argon plasma coagulation (APC)
are associated with high rates of buried Barrett*s (0-56%). Of concern is
occult malignant progression of SSIM hidden underneath NSE and therefore not
detectable during routine endoscopic inspection. This has been described in
incidental case reports in patients who underwent treatment with APC and PDT.
To assess whether or not RFA is associated with SSIM a separate study has been
conducted in the AMC including the previously reported 23 patients at 2 years
follow-up. Rigorous evaluation of the NSE after RFA was performed by taking
biopsies, keyhole biopsies and ER specimens from the NSE to assess the presence
of SSIM. In addition, pre-RFA cytology brushes from the BE segment and post-RFA
cytology brushes from the NSE were used to evaluate eradication of genetic
abnormalities pre- and post-RFA. No SSIM nor persisting genetic abnormalities
were detected in any of the study subjects.

Another issue to address is the presence of IM below the neosquamocolumnar
junction (neo-SCJ). So far its clinical relevance remains debatable since it is
unknown whether IM is a physiological finding or if it is the first sign of
recurrence of Barrett. In a normal population IM can be detected in the cardia
in up to 25% of patients and is not considered a premalignant condition.
Studies and case reports have suggested that after thermal ablation therapy,
the mucosal histology of the cardia undergoes a change, one study noting a rise
in presence of IM from 8.5% pre-ablation to 28% post-ablation, and a rise of
5.3% for dysplasia. A previous study conducted in the AMC showed focal
non-dysplastic IM in the cardia in two-thirds of patients with IM during
follow-up evaluation after treatment with RFA. Several theories arise
concerning development of IM. The first concerns an association with H. Pylori
(HP) and IM elsewhere in the stomach. The second describes that repeated injury
caused by the ablative method to the cardia, might generate a favourable
condition in which cardiac dysplasia may arise. Another possibility is that the
degree of damage to the cardia i.e. by acid and bile refluxate is already
higher in patients treated for HGIN or cancer, accounting for a higher
incidence of dysplasia in the cardia in dysplastic than non-dysplastic ablated
Barrett. The risk of malignant progression of IM in the cardia seems lower than
that of IM in (short-segment) BE. However the number of patients investigated
so far remains small. Long-term data are necessary to evaluate behaviour of IM
in the cardia. At the moment we believe that only dysplastic IM in the cardia
should be treated.

The aim of this study is to answer the question if RFA is an effective
treatment modality in the long term, for patients with Barrett's esophagus
containing HGIN or EC.

Endoscopic evaluation
All patients will be evaluated by an expert gastroenterologist at the AMC, who
will record any visible abnormalities on a case record form (CRF); specific
attention will be paid to the appearance of the neo-SCJ. Subsequently all
patients will undergo an EUS and if necessary FNA, to exclude any abnormal
regional lymph nodes, subsquamous growth or invasive growth of recurrent
neoplasia, in patients who were previously treated for a (pre)cancerous
condition.

Histopathological evaluation of neosquamous epithelium
At the follow-up endoscopy, 4Q/2cm biopsies will be taken from neosquamous
epithelium. The neosquamous biopsies will be evaluated for the presence of
intestinal metaplasia, dysplasia and subsquamous IM, by an expert
GI-pathologist.

Histopathological evaluation of endoscopic resection specimens
Next to the biopsies, an endoscopic resection specimen from neosquamous
epithelium will be obtained, by using the multi-band mucosectomy technique
(MBM). ER has proved before to be an adequate method for detection of buried
Barrett. 18 By using the multi-band mucosectomy (MBM) technique, a specimen can
be obtained safely, and it can easily be used on areas previously treated with
RFA.
The specimens will be evaluated for SSIM by an expert GI-pathologist, who will
also score penetration depth of all ER-specimens on a standardized CRF.
The ER-specimen will be taken from an area with previous Barrett*s mucosa.
Prior to the MBM, 1-2 biopsies will be taken at the target site, to relate
presence of SSIM in biopsies to the findings of the ER specimens.

Histopathological evaluation of biopsy samples below the neo-squamocolumnar
junction
At least 4 quadrant biopsies will be taken from the area immediately distal to
the neo-SCJ (within 5 mm); these biopsies will be evaluated for presence of
intestinal metaplasia by an expert pathologist.
Furthermore at least 2 additional biopsies from antrum and corpus will be
obtained to evaluate presence of HP and IM.

In this study a regular follow-up endoscopy will be performed, with standard
biopsy sampling. General risks associated with a gastroscopy are: mild
irritation of the throat due to introduction of the endoscope, difficulties
swallowing and retrosternal pain. During follow-up endoscopy an endoscopic
resection will be performed by using the multiband mucosectomy technique, minor
bleeding may occur in 6% of the cases, usually easily managed with endoscopic
hemostatic techniques. The additional risk of an EUS is rarely a perforation of
the esophagus, which can be treated in a conservative manner or by surgical
intervention.