Esofago de Barret

PRIMER
Barrett oesophagus
Yonne Peters1,8, Ali Al-Kaabi1,8, Nicholas J. Shaheen2, Amitabh Chak3, Andrew Blum3,
Rhonda F. Souza4, Massimiliano Di Pietro5, Prasad G. Iyer6, Oliver Pech7,
Rebecca C. Fitzgerald5 and Peter D. Siersema1*
Abstract | Barrett oesophagus (BE), the only known histological precursor of oesophageal
adenocarcinoma (EAC), is a condition in which the squamous epithelium of the oesophagus is
replaced by columnar epithelium as an adaptive response to gastro-oesophageal reflux. EAC has
one of the fastest rising incidences of cancers in Western countries and has a dismal prognosis. BE
is usually detected during endoscopic examination, and diagnosis is confirmed by the histological
presence of intestinal metaplasia. Advances in genomics and transcriptomics have improved our
understanding of the pathogenesis and malignant progression of intestinal metaplasia. As the
majority of EAC cases are diagnosed in individuals without a known history of BE, screening for
BE could potentially decrease disease-related mortality. Owing to the pre-malignant nature of BE,
endoscopic surveillance of patients with BE is imperative for early detection and treatment of
dysplasia to prevent further progression to invasive EAC. Developments in endoscopic therapy
have resulted in a major shift in the treatment of patients with BE who have dysplasia or early
EAC, from surgical resection to endoscopic resection and ablation. In addition to symptom
control by optimization of lifestyle and pharmacological therapy with proton pump inhibitors,
chemopreventive strategies based on NSAIDs and statins are currently being investigated for
BE management.
Barrett oesophagus (BE) involves the formation of a with dysplasia who are at high risk of malignant pro-
metaplastic columnar epithelium with crypt architecture gression. The increasing knowledge of risk factors for
(which resembles the epithelium of the intestine) from both BE and EAC could help define clinical pathways
the squamous epithelium of the oesophagus, which is a to enable identification of individuals at high risk who
reparative response to reflux-induced damage (FIG. 1). It could undergo endoscopic or non-endoscopic screening
is now widely accepted that BE is a pre-malignant con- tests or be candidates for preventive strategies to halt BE
dition that predisposes patients to develop oesophageal development or progression8,9.
adenocarcinoma (EAC). BE is detected by endoscopy In this Primer, we provide an up-to-date overview
and histopathological assessment of biopsy samples and of the epidemiology and pathophysiology of BE, dis-
is diagnosed on the basis of the presence of intestinal cuss the criteria for the diagnosis of BE and explore
metaplasia (IM). However, ongoing debate about some developments in screening and surveillance. In addition,
diagnostic features stands in the way of a comprehensive we review the management of BE, including endoscopic
definition of BE1–4. eradication strategies, and the effect of the condition on
The prevalence of BE in the general population is dif- the quality of life (QOL) of patients.
ficult to determine, as accurate population-based esti-
mates are rare and the majority of individuals with BE Epidemiology
are not diagnosed, although the prevalence in Western Trends in incidence and prevalence
countries is ~1–2% in the general population and ~10% EAC is one of the cancers with the fastest rising prev-
in populations that report acid reflux symptoms5,6. alence in Western countries10. As BE is the precursor
Epidemiological and histopathological evidence indi- lesion in the majority of EAC cases, data suggest that the
cate that many cases of EAC arise in individuals with incidence and prevalence of BE are also rising. Studies
BE7 by the progression of IM to dysplasia and finally of the incidence and prevalence of BE are confounded
to neoplasia. Given the high mortality in patients with by the increasing use of gastrointestinal (GI) endoscopy;
*e-mail: peter.siersema@
radboudumc.nl EAC, improved diagnosis of individuals with BE or because BE is often associated with gastric reflux with no
https://doi.org/10.1038/ early stages of EAC could improve outcomes by ena- distinguishing symptoms, detection of BE is expected
s41572-019-0086-z bling clinical surveillance and treatment of patients to increase as upper endoscopy use increases, regardless
NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 1
0123456789();
PRIMER
Author addresses descent20,21. The effect of Hispanic ethnicity is less clear,

as some studies documented a reduced risk of develop-
1
Department of Gastroenterology and Hepatology, Radboud University Medical Center, ing BE20,21, whereas others estimate a risk similar to that
Nijmegen, Netherlands. in the white population22. Although BE is reportedly
2
Division of Gastroenterology and Hepatology, University of North Carolina at Chapel Hill, uncommon in East Asians, a meta-analysis of 28 studies
Chapel Hill, NC, USA.
showed that the prevalence of histologically confirmed
3
Division of Gastroenterology and Liver Diseases, Case Western Reserve University,
Cleveland, OH, USA. BE was 1.3% in East Asians23.
4
Department of Medicine and the Center for Esophageal Diseases, Baylor University The risk of developing BE is approximately twofold
Medical Center at Dallas and the Center for Esophageal Research, Baylor Scott and higher in men than in women24. Increasing age is an
White Research Institute, Dallas, TX, USA. additional strong risk factor — prevalence increases by
5
Medical Research Council Cancer Unit, University of Cambridge, Cambridge, UK. 50–100% for every decade of life in adults11,25.
6
Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN, USA.
7
Department of Gastroenterology, St John of God Hospital, Regensburg, Germany. Lifestyle factors. Cigarette smoking is a moderate risk fac-
8
These authors contributed equally: Yonne Peters, Ali Al-Kaabi tor for the development of BE26. Compared with a never
smoker, a current or past smoker has >50% increased
of true increases in underlying incidence11,12. To control risk of developing BE. Furthermore, a dose–response
for this detection bias, the proportion of upper endos- relationship exists, as greater cumulative pack-years of
copies that result in a diagnosis of BE has been assessed, exposure result in a greater risk of BE5,27.
which in some studies showed a steady increase in BE Diet is also a risk factor for the development of BE.
diagnoses12. The current best estimate of the prevalence After controlling for appropriate risk factors, increasing
of BE in adults is 1–2%5 (~10% in those with chronic intake of vegetables has a protective effect against BE28.
gastro-oesophageal reflux disease (GERD) symptoms) The intake of red meat or processed meat does not seem
in Western countries13–15. to be associated with increased risk of BE29. Increased
Accurate estimates of the annual risk of EAC among levels of nitric oxide from dietary sources can be detected
patients with BE are difficult to obtain, as incidence rates in the distal oesophagus, and cell culture experiments
vary between studies. Early studies before the year 2000 suggest that they may contribute to BE. High levels of
usually overestimated EAC incidence (2–4% per year) nitric oxide result in the formation of higher oxides
in patients with non-dysplastic BE owing to publication of nitrogen, which might contribute to mutagenesis in
bias and inclusion of high-risk patients16. Population- epithelial cells at the gastro-oesophageal junction30,31.
based studies and large meta- analyses in the past The association between alcohol use and BE risk
10 years reported lower progression rates of 0.1–0.5% is unclear, as some studies reported no association,
per patient-year17,18, although the lower estimate might whereas others reported that alcohol increased risk of
be too low, as it is possible that patients with IM of the BE32. Overall, alcohol use is not a strong risk factor,
gastric cardia or with ultra-short BE were included in although the type of alcohol that is consumed may mat-
these pathology registry-based studies. ter, as wine seems to be protective, whereas hard spirits
Although BE is a well- established risk factor for augment the risk of developing BE33.
EAC, the assumption that all patients who develop EAC Obesity is a strong risk factor for BE34. Overall, the
go through the same reflux-induced response leading risk of developing BE is approximately twofold higher
to adenocarcinoma was challenged by a retrospec- in patients who are obese (body mass index (BMI)
tive analysis that found that only 46% of patients with >30 kg/m2) than in those who are not34. Furthermore,
EAC presented with endoscopic confirmation of BE the type of obesity (that is, the distribution of fat) seems
and histopathological evidence of IM7. Furthermore, to be important; truncal obesity (the presence of large
comparison of patients with EAC who had confirmed amounts of intra-abdominal fat; also known as central
BE at presentation to those without BE suggested the obesity) seems to carry an especially high risk35. In fact,
existence of two EAC phenotypes with different tumour when waist circumference is controlled for, the associa-
behaviour and response to therapy7. These findings tion between obesity and BE disappears in some stud-
raise the question of whether EAC always develops ies35. The pathophysiological basis for this relationship
through the IM–dysplasia–EAC sequence. is unclear but may be due to trophic hormonal effects
that are associated with truncal obesity, such as increased
Risk factors serum levels of insulin and leptin, which have been
Risk factors for BE include demographic, lifestyle (such shown to be an independent risk factor for BE36. In addi-
as diet and weight), GERD-specific and miscellaneous tion, truncal obesity can mechanically influence GERD
risk factors (TABLE 1). GERD, obesity and smoking are through increased abdominal pressure and reduced
responsible for ~80% of the burden of EAC19. When lower oesophageal sphincter (LES) pressure37,38. However,
combined, these risk factors can be used to stratify risk because the association between obesity and BE persists
in populations to optimize screening programmes for even after controlling for GERD symptoms, mechanisms
BE and EAC. beyond obesity’s promotion of reflux may be present35.
Demographic factors. BE is a disease with a white male GERD-specific factors. As BE is considered to be a com-
predominance. After controlling for other relevant risk plication of chronic GERD, it is perhaps not surprising
factors, individuals of white ethnicity are at 2–3-fold that risk factors for gastric reflux are also strongly asso-
higher risk of developing BE than individuals of African ciated with BE. Reflux-induced injury has been linked
2 | Article citation ID: (2019) 5:35 www.nature.com/nrdp
0123456789();
PRIMER
Normal Barrett Dysplastic Oesophageal such as statins and proton pump inhibitors52 (PPIs), may
oesophageal oesophagus Barrett adenocarcinoma be associated with decreased risk of developing BE.
epithelium oesophagus
Oesophageal
Squamous Gland injury Columnar Risk factors for progression of BE
epithelium formation epithelium
In general, many of the risk factors for development of
BE also seem to be risk factors for progression of BE
to dysplasia or EAC (TABLE 1). Age is an important risk
factor for progression of BE. Among patients with non-
Basement membrane dysplastic BE, the risk of dysplasia increases by ~3.3%
per year of age53–55. White ethnicity and male sex53,55,56
Fig. 1 | Development and progression of Barrett oesophagus. Oesophageal injury, are also associated with increased risk of progression
mainly due to gastrointestinal reflux, might lead to Barrett oesophagus, a lesion that to EAC. The length of BE (measured from the gastro-
is characterized by replacement of the normal squamous epithelium by columnar oesophageal junction (GEJ)) is similarly directly corre-
epithelium, termed intestinal metaplasia. This pre-malignant condition might progress lated with risk of progression to dysplasia and EAC53–56.
to dysplasia and oesophageal adenocarcinoma, which may invade the submucosa.
Active reflux, smoking, BMI and waist circumference
Figure adapted from REF.219, Springer Nature Limited.
also increase the risk of progression55.
Debate exists about whether IM is a risk factor for
to cellular and molecular changes in the oesophagus39,40. malignant progression. A large, population-based cohort
Physiological alterations that predispose to GERD are study demonstrated a substantially higher annual EAC
also associated with BE, including LES hypotension41, risk in patients with IM than in those without IM (0.38%
increased gastric acid production42 and increased bile in versus 0.07%; HR 3.54, 95% CI 2.09–6.00, P < 0.001)57,
refluxate43. Symptoms of heartburn and regurgitation are although another study found that IM did not increase
strongly associated with the presence of BE, and dura- the risk of malignant progression58. The presence of dys-
tion of GERD symptoms may also be a risk factor for plasia is a major risk factor for progression to EAC59,60.
BE44. Although GERD is a strong risk factor for both BE Furthermore, there is uncertainty among pathologists
and EAC, 40–50% of patients with these disorders do not about the histopathological diagnosis of low-grade dys-
report chronic reflux symptoms, suggesting that silent plasia60 (LGD). The annual risk of malignant progression
reflux or other risk factors are important in the patho- is 5.2–9.1% in patients with LGD that was confirmed by
genesis of BE and EAC. Indeed, the prevalence of BE in an expert pathologist60,61. A meta-analysis estimated that
individuals without frequent, chronic reflux symptoms the risk of EAC is 6.6% per annum in patients with high-
is substantial and, in some studies, comparable to that in grade dysplasia (HGD), but considerably higher rates
individuals with frequent, chronic reflux45. (up to 28%) have been reported in therapeutic studies62–64.
Hiatal hernia size also correlates with risk of BE46,
although whether this is solely due to the increased sever- Mechanisms/pathophysiology
ity of GERD symptoms in those with bigger hiatal her- Pathogenesis
nias or is due to another, not-yet-elucidated mechanism, BE occurs as a result of epithelial injury in the distal
remains unknown. oesophagus, which is caused by reflux of acid, bile and
Gastric infection with Helicobacter pylori is negatively other noxious substances, and a subsequent reparative
correlated with the presence of BE, especially for the response. Reflux-induced damage is usually repaired by
most virulent forms of H. pylori47,48. The reduced gas- regeneration of squamous cells, but in some individu-
tric acid output that is associated with chronic H. pylori als, the squamous epithelium in the distal oesophagus is
infection might result in a lower prevalence of BE in replaced by a differentiated columnar epithelium (that
these patients. is, IM)65. Similarly to gastric and intestinal mucosa, BE
has a glandular structure comprising crypts that contain
Family history. Familial studies have implicated a stem cells that are located approximately one-third of the
genetic component in predisposition to BE and a family distance from the base of the crypts. There are various
history of BE or EAC may be a strong risk factor for cell lineages within the metaplastic columnar lining of
disease49,50. The prevalence of BE in individuals with BE, including columnar cells that strongly express the
a family history of BE is estimated to be 6–7.3%49,50. mucins MUC1 and MUC5AC and the trefoil family
Familial aggregation must be interpreted with caution, member TFF1; mucus-secreting cells that express MUC6,
as it could also be due to common environmental expo- TFF2 and TFF3; and goblet cells that express MUC2 and
Celcalilit.
sures, such as diet and smoking, or a genetic suscep- MUC3, which are similar to the epithelial cell types that
tibility to other risk factors, such as obesity or GERD. are present in the intestine66,67.
In addition, studies of heritability of these traits may be Exposure to gastric refluxate, which
MSSA
th-frequently con-
confounded by detection bias in family members, given tains bile acids from duodeno-gastro-oesophageal reflux,
the increased use of endoscopy in this group compared is associated with increased- -
oxidative stress, activation
with the general population49. of inflammatory
- mediators and
- DNA damage in cells of
BE68. The resulting inflammatory infiltrate is accompa-
Medication. Whereas the use of NSAIDs has been -
nied by changes in the expression of key genes that are
strongly inversely associated with the risk of EAC, NSAID involved in cell fate and development, including BMP4,-
use is not associated with decreased risk of BE51. However, PTGS2, SHH, the caudal-type homeobox genes CDX1
some studies suggest that use of other medications, and CDX2, Notch and SOX9 (REFS69–73).
0123456789();
PRIMER
Table 1 | Selected risk factors for BE and EAC contained glands that were positive for phosphorylated
SMAD1 (pSMAD1), pSMAD5 and/or pSMAD8 but
Risk factor Association with Association with Refsb lacked expression of the IM markers CDX2 and MUC2
BE (OR (95% CI))a EAC (OR (95% CI))a
(REF.69). In another mouse model of BE (caused by surgi-
Male sex 2.0 (1.8–2.2) 2.2 (1.8–2.5) 24,53–55
cally induced reflux), high levels of pSMAD1, pSMAD5
White ethnicity + + 20,22,33
and/or pSMAD8, CDX2 and MUC2 were observed in
Increased age 1.0 (1.0–1.1) 1.0 (1.0–1.1) c 11,25,53–55 columnar epithelial cells69. This study demonstrated
that a non- specialized columnar epithelium may be
Presence of GERD symptoms 2.9 (1.9–4.5) 7.7 (5.3–11.4) 14,44
an intermediate stage in the formation of a specialized
Hiatal hernia size 3.9 (3.0–5.1) Unknown 46
IM, in which the downstream BMP targets pSMAD1,
BMI (per unit increase) 1.0 (0.9–1.0) 1.0 (0.9–1.2) c 34,35,55
pSMAD5 and/or pSMAD8 are required as cofactors to
Waist circumference ratio 1.2 (1.0–1.3) 2.1 (1.3–3.2) 35 activate the transcriptional activity of CDX2. However,
(per 5 cm increase) other signalling pathways may also be involved in this
Cigarette smoking 1.4 (1.2–1.7) 1.5 (1.1–2.0)c 26,55,56,229 process (FIG. 2b).
Alcohol intake 1.1 (0.6–1.8) 1.1 (0.8–1.5)c 55,230
Transcommitment of progenitor cells. Despite some
Helicobacter pylori infection 0.7 (0.6–0.8) 0.5 (0.4–0.7) 47,48,231
evidence for transdifferentiation, reprogramming of
NSAID use 1.0 (0.8–1.3) 0.7 (0.5–1.0)c 51,55
immature pluripotent stem cells (that is, transcom-
Family history of GERD, + + 49,50 mitment) is perhaps a more likely explanation for the
BE or EAC multiple different cell types observed in BE. BE was
BE, Barrett oesophagus; BMI, body mass index; EAC, oesophageal adenocarcinoma; GERD, initially thought to result from migration of columnar
gastro-oesophageal reflux disease. + indicates positive for risk factor. aOdds ratios for family epithelial cells from the gastric cardia in response to
history are not available and in the case of white ethnicity, they depend on the population that
is used for comparison. bIf a recent meta-analysis has been published, it is preferentially listed. tissue damage76. However, in animal studies, a colum-
c
Odds ratio for factors associated with progression of BE to EAC. nar epithelium still develops in a damaged oesophageal
mucosa that is separated from the gastric mucosa by a
Several hypotheses exist about the cellular origins normal squamous epithelium77, which is not consist-
of BE, although consensus is lacking65 (FIG. 2a). For exam- ent with this hypothesis. Evidence exists for migration
ple, BE may develop by transdifferentiation of mature of progenitor cells (marked by expression of the stem
oesophageal squamous cells to columnar epithelial cells, cell marker LGR5) from the gastric cardia to the distal
or IM may result from differentiation of immature pro- oesophagus in a Notch1-dependent manner in a mouse
genitor cells (transcommitment). Although the various model of BE (induced by chronic inflammation due
models seem to be distinct, they may not be mutu- to oesophageal overexpression of IL-1β)78. IL-1β over-
ally exclusive and the formation of BE may involve a expression was sufficient to induce IM, dysplasia and
combination of these processes. cancer, which was accelerated by dietary bile acids. The
molecular similarity between EAC and chromosomally
Transdifferentiation of oesophageal squamous cells. BE unstable gastric cancer, and the resemblance between
may develop by direct transdifferentation of oesophageal BE-associated EAC and normal gastric mucosa in their
squamous cells into metaplastic columnar cells74. Indirect chromatin structure, is further (indirect) evidence for
transdifferentiation, in which the squamous cell first -
the gastric origin of BE79,80.
dedifferentiates into a transitional cell that then differ- Another hypothesis is that stem cells in oesophageal
·
entiates into a columnar cell, has also been proposed in submucosal glands undergo transcommitment to form
the setting of GERD65. In transdifferentiation, the reflux- a columnar epithelium. In this model, the columnar epi-
induced inflammatory environment (mediated by prosta- thelium and the submucosal glands are a morphologi-
glandin E2 (PGE2), nuclear factor-κB (NF-κB), TNF cal continuum and different clones of immature stem
and other molecules) leads to increased sonic hedgehog cells are reprogrammed to differentiate into BE81,82. In
signalling and decreased Notch signalling (FIG. 2b). the porcine oesophagus, submucosal glands contain
The transcription factor CDX2 is an important con- a population of pluripotent stem cells that can differ-
tributor to the transdifferentiation of squamous cells entiate into both squamous epithelium and columnar
into IM (FIG. 2b). Selective overexpression of Cdx2 from
- epithelium. Interestingly, these pluripotent stem cells
the Krt14 promoter in the mouse oesophagus and fore- proliferate in response to injury, which is also observed
stomach induces a glandular phenotype in the squamous- in oesophageal submucosal glands in humans83,84. In
cell-lined oesophagus, which is characterized by reduced support of this model, single- cell transcriptomics of
basal keratinocyte proliferation, reduced barrier func- oesophageal cells from biopsy samples of patients with
tion and the presence of transitional cells with secretory BE and healthy individuals identified a cell population
features72. In patients with reflux- related symptoms, in BE that expresses the stem cell factors olfactomedin 4
CDX2 expression was detected in oesophageal biopsy
-
(OLFM4) and LEFTY1 and had a transcription profile
samples with IM but not in those without IM75. However, that substantially overlaps with that of oesophageal
other factors probably contribute to IM, as it does not submucosal gland cells85.
-
occur in transgenic mice overexpressing Cdx2 (REF.72). Studies in mice suggest that BE may develop from
Overexpression of the- TGFβ family member Bmp4 the proximal migration of dormant stem cells or transi-
in mice induced a columnar phenotype in the epithe- tional epithelium progenitor cells that are present at the
lium at the squamocolumnar junction (SCJ), which SCJ. In this model, BE develops through competition
0123456789();
PRIMER
between stem cells in the squamous epithelium and contains KRT5+KRT7+p63+ basal progenitor cells, over-
-
residual embryonic stem cells at the SCJ for stem cell laid by luminal KRT7+ cells71. A surgically induced reflux
>
niches. In p63-deficient adult mice that lack stratified model and a Cdx2-overexpression mouse model with
epithelia, embryonic KRT7+p63− stem cells from the
-
lineage tracing showed that BE-type metaplasia arises
SCJ
.. proliferate in response to damage to the squamous from this transitional epithelium71.
epithelium, which leads to oesophageal metaplasia86. Other sources of these progenitor cells include res-
In fact, a histologically distinct multilayered transition ident stem cells in the basal layer of the oesophagus
zone at the SCJ is present in both mice and humans and or multipotent bone- marrow-derived stem cells87,88,
a Oesophageal stratified Intestinal metaplasia with Squamocolumnar Gastric columnar

squamous epithelium columnar epithelium junction epithelium
Transitional cell
Indirect
Restoration Oesophageal
LGR5+
of squamous injury stem cell
epithelium
Direct KR7+p63– KR7+p63+
embryonic transitional basal
stem cell progenitor cell
Squamous Stem cell Columnar cell

epithelium
Submucosal
gland
Stem cell
Crypt base
columnar
cell
Blood Paneth Bone-marrow-derived
vessel cell progenitor cell
b ↓ Notch
signalling
↓ Squamous
↑ MYC cytokeratins
↑ CDX1 KLF4 ↑ Columnar
Pro-inflammatory cytokeratins
cytokines ↑ Mucins
• NF-κB BMP–SMAD Intestinal
Reflux ↑ CDX2 (e.g. MUC2)
signalling metaplasia
• PGE2
• TNF
↑ Hedgehog SOX9
signalling FOXA2
Fig. 2 | Pathogenesis of Barrett oesophagus. a | Pathogenetic mechanisms in Barrett oesophagus (BE). Various cell
types have been proposed to give rise to intestinal metaplasia (the replacement of oesophageal squamous epithelium
by intestinal columnar epithelium in response to oesophageal injury, typically by gastric reflux) in BE, which can progress
to oesophageal adenocarcinoma. Intestinal metaplasia can arise by transdifferentiation of squamous cells, directly or
via a transitional cell, or by migration of bone-marrow-derived progenitor cells to the mucosa. Furthermore, various
-
stem cells, including residual embryonic stem cells, submucosal gland stem cells or gastric cardia stem cells, or progenitor
-
cells, including gastro-oesophageal junction or basal squamous progenitor cells, can undergo transcommitment.
The pathogenetic mechanisms might act separately or in combination to give rise to the BE lesion. b | Molecular mediators
of transdifferentiation in> BE. Reflux-induced oesophageal injury leads to inhibition of Notch signalling and increased
expression of MYC and the homeobox protein CDX1, which promote transdifferentiation by decreasing levels of squamous
--
-
epithelium cytokeratins and increasing levels of columnar epithelium cytokeratins and mucins73. The transcription
pathways mediating these changes are unclear, although KLF4 expression seems to be upregulated, which is linked to
increased transcription of CDX2 and MUC2 (REF.227). Furthermore, reflux activates Hedgehog signalling in squamous
->
epithelial cells, leading to secretion of sonic hedgehog (SHH). SHH induces BMP4 expression in stromal fibroblasts, which
in turn activates BMP4–SMAD1/SMAD5/SMAD8 (BMP–SMAD) signalling in squamous cells and thereby increases CDX2
and MUC2 expression (REF.228). BMP and SHH also upregulate the transcription factors SOX9 and FOXA2, which in turn
can upregulate the levels of CDX2 and MUC2 (REF.228). NF-κB, nuclear factor-κB; PGE2, prostaglandin E2.
0123456789();
PRIMER
although robust evidence from clinical studies for either and whole-genome doubling), somatic driver and pas-
of these sources is very limited. senger mutations and a tumour-promoting, inflamma-
These models of BE pathogenesis have mostly been tory microenvironment (that is, oxidative stress)97,98.
studied using surgically induced reflux or transgenic This conceptual framework can be useful to classify the
mouse models of BE, which have multiple limitations numerous acquired genetic alterations described during
that must be considered when translating these results malignant progression of BE into the major carcinogenic
to humans89. capabilities99.
Germline susceptibility and environmental factors. Pathways for malignant progression. Whole- exome
Susceptibility to BE is probably a composite effect of sequencing of DNA extracted from areas of BE adjacent to
environmental risk factors, various germline and somatic EAC has revealed a somatic mutation frequency of 1.3–5.4
genetic variants and epigenetic variants. Genetic and mutations per Mb of DNA in non-neoplastic metapla-
other analyses have identified a multitude of genetic sia100–102, which is higher than that in prostate or breast car-
polymorphisms and humoral factors (such as circulat- cinomas. The mutation pattern was indicative of genomic
ing levels of various cytokines) that are associated with damage caused by oxidative stress, most likely due to
risk of development and progression of BE90–92. Genome- GERD. Mutations in TP53 are early shared mutations
wide association studies (GWASs) showed that germline in tumour development, as they are found in both EAC
genetic susceptibility factors constitute ~35% of the and adjacent (high-risk) non-dysplastic BE100. Although
heritability of BE and ~25% of EAC93. Furthermore, GWASs the mutational load in EAC is high, many somatic muta-
showed that a polygenic component underlies disease risk tions, such as those in ARID1A and SMARCA4, were also
in unrelated individuals with BE or EAC and that a sub- present in biopsy samples from non-dysplastic BE that
stantial proportion of single-nucleotide polymorphisms did not progress to HGD or EAC (low risk). These early-
(SNPs) overlap in individuals with BE or EAC, which occurring passenger mutations contribute to increased
suggests a shared genetic basis of susceptibility93. These clonal expansion and changes in clonal diversity102,
SNPs are in or near genes that regulate oesophageal devel- whereas the acquisition of a driver mutation pushes the
opment (for example, FOXF1 and FOXP1), ion transport cell towards cancer103. Analysis of BE without dyspla-
(cystic fibrosis transmembrane conductance regulator sia suggests that the level of genetic diversity at baseline
(CFTR)) and immune regulation (the major histocompat- remains constant over time and is associated with an
ibility complex genes and the antioxidant microsomal glu- increased risk of malignant progression in BE101,102,104.
tathione S-transferase 1)94,95. Another GWAS explored the Only a minority of tumours progress along this tra-
association between several well-known epidemiological ditional pathway of stepwise loss of function of TP53
risk factors for BE (GERD, cigarette smoking and BMI) and other tumour suppressor genes (such as CDKN2A and
and seven SNPs that are implicated as risk factors for BE96. SMAD4) followed by oncogene amplification and the
Only a SNP in FOXP1 combined with at least weekly development of genomic instability (FIG. 3).
reflux symptoms modified the risk of developing BE and Whole- exome sequencing and whole- genome
EAC96. An analysis of a large number of SNPs identified sequencing studies have identified other pathways to
multiple risk-modifying polymorphisms. Individuals with malignancy, involving alterations in large regions of
obesity (BMI ≥30 kg/m2) and an rs491603-AA genotype the genome, which can arise at any stage and accelerate
had a threefold increased likelihood of BE compared with malignant progression100,102,105,106. In fact, the frequency
those with BMI <25 kg/m2. Risk of BE in individuals of whole-genome duplication suggests that a substan-
with a history of heavy smoking and an rs11631094-AA tial proportion of tumours in Barrett metaplasia might
genotype at chromosome 15p14 was 50% that of light develop through the genome duplication pathway.
smokers90. Smoking status and rs13429103, and recurrent In this pathway, TP53 mutations occur first, followed
GERD symptoms and three SNPs (rs12465911, rs2341926 by whole-genome duplication in dysplastic tissues and
and rs13396805), affected the risk of EAC (OR ~2.0)90. then genomic instability and oncogene amplification in
These findings need to be validated in future studies. cancerous tissues98,100 (FIG. 3).
Other genomic events, such as chromothripsis
Malignant progression in BE (chromosome shattering), kataegis (localized regions of
The transformation of BE to dysplasia and then to EAC hypermutation) and breakage–fusion–bridge (chromo-
requires accumulation of genetic and epigenetic alter- some breakage followed by fusion and bridge formation),
ations at an early stage, which enable metaplastic cells may also be mechanisms of malignant progression in BE.
to acquire the core physiological capabilities of tumour Although the understanding of their role in progression
cells97. These acquired features include generation of is incomplete, the presence of these genomic catastro-
their own mitogenic signals (generally through onco- phes in BE with HGD and in EAC suggests that multiple
gene activation), resistance to growth inhibitory signals mechanisms are involved in the rapid progression105,106.
(generally through inactivation of tumour suppressor
genes), avoidance of apoptosis, immortalization, vascu- Diagnosis, screening and prevention
larization, invasion of adjacent structures and metastasis Diagnosis
to distant structures, reprogramming of energy meta- Endoscopy is the gold-standard test for a diagnosis of
bolism and evasion of tumour-destroying immune cells97. BE. Upper endoscopy is indicated in patients >50 years
The acquisition of these core capabilities is facilitated by of age who have new-onset symptoms of reflux or dys-
genomic instability (which can be reflected by aneuploidy pepsia and at any age if these symptoms persist despite
0123456789();
PRIMER
Neoplastic BE cell
Oncogene
amplification
Traditional and genomic
pathway instability
Loss of other
tumour suppressor
genes
Non-dysplastic
BE cell Mutation in tumour suppressor gene
TP53 Chromothripsis
mutation Rapid pathways
p53 Genomic catastrophes
inactivation Break–fusion–bridge
• Chromothripsis
• Kataegis
• Break–fusion–bridge
Genome duplication Kataegis
Genomic
instability
and oncogene
amplification
Fig. 3 | Mechanisms of malignant progression in Barrett oesophagus. Non-dysplastic cells in Barrett oesophagus (BE)
first acquire mutations in TP53 that inactivate p53. In the traditional pathway, stepwise accumulation of loss-of-function
mutations in other tumour suppressor genes occurs next, followed by oncogene amplification and genomic instability,
which eventually lead to cancer formation. Conversely, in the genome-doubling pathway, TP53-mutant non-dysplastic
cells in BE undergo whole-genome duplication to form dysplastic tissues, which progress to oesophageal adenocarcinoma
by genomic instability and oncogene amplification. Rapid progression to malignancy can also occur by chromothripsis
(chromosome shattering), kataegis (localized regions of hypermutation) and breakage–fusion–bridge.
initial medical management or in the presence of alarm European Society of Gastrointestinal Endoscopy require
symptoms (that is, dysphagia, persistent vomiting, a columnar epithelium that is clearly visible endoscopi-
weight loss, upper GI bleeding or anaemia)107,108. A BE cally and ≥1 cm in length (measured from the GEJ) for a
diagnostic yield of 3–12% was obtained in patients who diagnosis of BE2,3,109. The minimum length cut-off stems
were referred with GERD or dysphagia13–15. The final mostly from the high interobserver variability in diagno-
diagnosis of BE is made by synthesis of endoscopic and sis of sub-centimetric BE and the very low cancer risk
histopathological evidence of columnar-lined epithelium associated with ultra-short BE2,112,113.
(FIG. 4). Although diagnosis seems straightforward, there Quality indicators have been published to guide
continues to be debate about the precise endoscopic and physicians in endoscopic diagnosis and in judging the
histopathological criteria (FIG. 5; TABLE 2). appropriateness of diagnosis and management of BE114.
These indicators include careful delineation of endo-
Endoscopic diagnosis. During endoscopic examination, scopic landmarks (diaphragmatic pinch and GEJ) and
the BE segment is inspected with high-definition white- use of the Prague classification for measuring the length
light (HD-WL) endoscopy and a targeted biopsy of visi- of BE115 (FIG. 4). Adherence to quality indicators and to
ble mucosal lesions is performed. In the Seattle protocol, the Seattle biopsy protocol led to a 2–3-fold higher rate
random four- quadrant biopsy samples are obtained of dysplasia diagnosis116,117.
every 2 cm, starting from the upper end of the gastric
folds1–3,109. The minimum length of a columnar-lined epi- Pathological diagnosis. Debate also exists about histo-
thelium (circumferential or maximum length of the BE pathological criteria for a BE diagnosis — in particu-
segment) for an endoscopic diagnosis of BE is the sub- lar, whether the presence of specialized IM with goblet
ject of debate (TABLE 3). The guidelines of the American cells is required. The guidelines of all the major socie-
Gastroenterology Association, Cancer Council Australia ties agree that IM is a diagnostic prerequisite, although
and the Japanese Society of Gastroenterology do not some guidelines accept that short BE segments may
indicate a minimum segment length 1,110,111, whereas be columnar-lined epithelium without the presence of
those of the American College of Gastroenterology IM (TABLE 3). Furthermore, IM can be missed owing
(ACG), the British Society of Gastroenterology and the to insufficient sampling, as the yield of IM correlates
0123456789();
PRIMER
directly with the number of endoscopic biopsy sam- of dysplasia (TABLE 2). The Vienna classification has
ples obtained118. However, although IM might not be improved the agreement among pathologists in staging
essential for a BE diagnosis in all guidelines, routine early neoplasia in the oesophagus, although the cyto-
surveillance for short- segment BE (<3 cm in length) logical and architectural changes in LGD can be subtle
without IM (confirmed by two endoscopies) is not and agreement among pathologists remains very low
recommended by any Western society owing to a very (κ = 0.11)121. The Seattle biopsy protocol clearly detects
low cancer risk in the absence of IM57. In addition, in more dysplasia than ad hoc random biopsies122, although
some guidelines, cases with tongues of columnar-lined sampling error is inevitable as this protocol samples <5%
oesophagus <1 cm in length (that is, an irregular SCJ of the BE epithelium118. The role of computer-assisted
or Z-line) should not be routinely biopsied and an inci- or brush biopsies for increasing the yield of dysplasia is
dental finding of IM within an irregular Z- line is not currently being assessed123.
an indication for routine surveillance and should be Most LGD cases and up to one-third of HGD cases
diagnosed as IM of the gastric cardia, not BE. IM of the are invisible by standard white-light endoscopy.
gastric cardia is a common finding in routine endoscopy Consequently, considerable effort has been dedicated
(up to 18% of all upper endoscopies)119, has a different to developing novel imaging modalities to improve
epidemiology from BE and EAC (a lower male:female detection of early neoplasia, although convincing evi-
ratio and an association with H. pylori infection) and has dence that these modalities can improve diagnosis of
a very low cancer risk11,119. BE is still lacking. A multicentre, randomized, crossover
trial compared narrow-band imaging (in which light of
Staging. Progression from BE to EAC occurs through a specific wavelengths is used to improve the resolution
morphological continuum of progressive derangement of of the surface mucosa) with targeted biopsies to standard
cytological features (such as variably sized and enlarged high-definition endoscopy with targeted and random
nuclei, rounded nuclei, loss of polarity and prominent biopsies124. Although narrow-band imaging diagnosed
nucleoli) and the development of glandular architecture a higher proportion of areas with dysplasia than did stan-
(complex budding or branching of glands and back-to- dard endoscopy (30% versus 21%, P = 0.01), this did
back glands). According to the Vienna classification, GI not translate into a diagnostic benefit in the per-patient
epithelial neoplasia is classified as indefinite for dyspla- analysis124. Other imaging techniques, such as acetic acid
sia (that is, dysplasia that cannot be confidently identi- chromoendoscopy, confocal laser endomicroscopy and
fied or ruled out), LGD, HGD, non-invasive carcinoma volumetric laser endomicroscopy, are being evaluated
and suspicion of invasive carcinoma120. To date, dyspla- (reviewed elsewhere125). In summary, a thorough endo-
sia remains the best available marker of cancer risk in scopic examination with high- definition endoscopy,
patients with BE, although there is considerable inter- strict adherence to basic quality indicators and allow-
observer and intraobserver variability in the interpretation ing sufficient time for mucosal cleaning and inspection
a
b
Top of BE tongue 36 cm
Maximum Gastro-oesophageal
Squamocolumnar junction
End of circular extent junction 40 cm
BE segment
39 cm
Circumferential extent
40 cm Maximum
Gastro-oesophageal extent Circumferential
junction Hiatal 36 cm extent
hernia 39 cm
42 cm
Diaphragm
c d
Fig. 4 | Endoscopic diagnosis of Barrett oesophagus. Endoscopic criteria for a diagnosis of Barrett oesophagus (BE)
are described in the Prague classification (REF.115). a,b | The lower measurement boundary is formed by the most proximal
extent of the gastric folds, at the gastro-oesophageal junction. The two upper measurement boundaries are marked by
the proximal margin of the circumferential BE segment (circumferential extent) and the proximal margin of the longest
tongue-like BE segment (maximum extent). In these panels, the Prague classification is C1M4. c | A short segment of
columnar-epithelium-lined oesophagus can be difficult to size endoscopically. d | Measurement of BE segment length
using open biopsy forceps (usually 6 mm maximum open size). In this case, the segment length was measured as 1 cm
(Prague classification: C0M1). Image in part b courtesy of I. Koutroubakis, on behalf of the Annals of Gastroenterology,
Annals of Gastroenterology, Greece.
0123456789();
PRIMER
a seem to be the key elements for maximizing detection

of dysplasia in routine practice.
Despite these advances in endoscopy and stan-
dardization of diagnostic approaches, EAC is detected
in 20–25% of patients with non-dysplastic or LGD BE
within 1 year of a BE diagnosis, indicating that initial
endoscopy frequently misses dysplasia and EAC126.
The number of patients with EAC who present at the
time of or shortly after an initial diagnosis of BE greatly
exceeds the number of incident cancers, which demon-
strates the importance of early diagnosis of BE and
related neoplasias.
Prevention
b c Modifiable risk factors. Strategies to minimize the
effect of risk factors, such as GERD, obesity and smok-
ing, might reduce the incidence of BE and related
neoplasias and are likely be more effective than phar-
macological chemoprevention9. In 2018, the World
Cancer Research Fund (WCRF), in collaboration with
the American Institute of Cancer Research, published
a series of recommendations on diet, physical activ-
ity and weight management for cancer prevention127.
Although recreational and occupational physical
activity are associated with ~35% lower risk of BE and
EAC128,129, this association varies greatly among studies
and the lack of a dose–response relationship warrants
careful interpretation of these findings. Diet also seems
to have a role in the prevention of BE development and
d e progression. In a study comparing patients with GERD,
non-dysplastic BE, dysplastic BE or EAC, adherence
to the WCRF recommendations for high fruit and
low processed meat consumption was 5–6-fold lower
in the advanced disease group (dysplasia and EAC)
than in the less advanced disease group (GERD and
BE)130. In two case–control studies of the population
of Northern Ireland, high dietary levels of vitamin D
directly correlated with increased risk of EAC (OR 1.99,
95% CI 1.3–3.86)131, whereas high magnesium intake
protected against reflux oesophagitis (OR 0.31, 95%
CI 0.11–0.87) and BE (OR 0.29, 95% CI 0.12–0.71)132.
A vitamin-D-restricted diet is not a viable chemopre-
ventive strategy given the protective effect of vitamin D
against other types of cancer, including colorectal,
breast, prostate and some haematological malignan-
Fig. 5 | Histopathological features of Barrett oesophagus. Progression from cies133. Conversely, a diet rich in magnesium might be
squamous epithelium to intestinal metaplasia, dysplasia and oesophageal
an interesting component of a preventive strategy, as
adenocarcinoma (see TABLE 2 for detailed description of these stages). a | Normal
oesophageal squamous epithelium. Magnification ×4. b | Intestinal metaplasia magnesium has well- established anti- inflammatory
containing goblet cells: intestinalized columnar-lined epithelium, normal cytology properties and is associated with a reduced risk of
and normal surface maturation (nuclear:cytoplasmic ratio of surface epithelium is oesophagitis and BE in a retrospective study132,134.
lower than basal cells). Magnification ×10. c | Low-grade dysplasia: mild architectural
changes, surface maturation disrupted (stratification) and mild, diffuse cytological Primary chemoprevention. PPIs, NSAIDs and statins are
abnormalities (for example, enlarged hyperchromatic nuclei, mitotic activity the most promising classes of drugs for primary chemo-
and normal cell polarity). Magnification ×20. d | High-grade dysplasia: marked prevention of BE. Acid-suppression therapy using PPIs
architectural changes, no surface maturation and marked cytological abnormalities is commonly prescribed for patients with BE to reduce
(for example, enlarged nuclei, loss of cell polarity, mitotic activity and atypical heartburn symptoms, although whether these drugs
mitoses). Magnification ×20. e | Oesophageal adenocarcinoma: marked architectural
reduce cancer risk in BE is debated and most GI soci-
changes with crowded irregular glands showing marked cytological abnormalities
with enlarged nuclei, loss of cell polarity, mitotic activity and atypical mitoses. Lamina eties currently recommend acid-suppression therapy
propria and submucosa invasion is present. Magnification ×10. Staining in all panels for symptom control but not for primary chemopreven-
is with haemotoxylin and eosin. Image in part a courtesy of R . S. van der Post, tion. Long-term adverse effects have been increasingly
Radboud University Medical Centre, Netherlands. Images in parts b–e courtesy of reported in low-quality retrospective studies (reviewed
M. O’Donovan, Cambridge University Hospital, UK. elsewhere135), whereas long-term follow-up data from
0123456789();
PRIMER
Table 2 | Histopathological features of BE

Stage Histopathology featuresa Interobserver HGD and EAC risk per
agreement (κ value) 100 person-years (95% CI)
NDBE • Intestinalized columnar-lined epithelium 0.58–0.73 (REFS232,233) • EAC: 0.33 (0.28–0.38)234,b
• Normal cytology • HGD/EAC: 0.26 (0.22–0.31)17;
• Normal surface maturation (nuclear:cytoplasmic 0.68 (0.61–0.74)18
ratio of surface epithelium is lower than
basal cells)
IND • Normal architecture 0.15–0.21 • EAC: 0.8 (0.5–0.12)236
• Normal surface maturation (REFS232,233,235) • HGD/EAC: 1.4 (1.0–1.9)236
• Mild cytological abnormalities
LGD • Mild architectural changes 0.11–0.32 • EAC: 0.54 (0.32–0.76)237,b; 2.51
• Surface maturation disrupted (stratification) (REFS121,232,233) (1.46–3.99)61
• Mild, diffuse cytological abnormalities (such as • HGD/EAC: 1.73 (0.99–2.47)237,b;
enlarged, hyperchromatic nuclei, mitotic activity 5.18 (4.32–8.10)61
and normal cell polarity)
HGD • Marked architectural changes 0.43–0.76 (REFS121,232) • EAC: 6.6 (5.0–8.5)62,b
• No surface maturation
• Marked cytological abnormalities (such as
enlarged nuclei, loss of cell polarity, mitotic
activity and atypical mitoses)
EAC • Marked architectural changes: carcinoma 0.61–0.82 (REFS235,238) NA
invades through basement membrane
• No surface maturation
• Marked cytological abnormalities (such as
enlarged nuclei, loss of cell polarity, mitotic
activity and atypical mitoses)
BE, Barrett oesophagus; EAC, oesophageal adenocarcinoma; HGD, high-grade dysplasia; IND, indefinite for dysplasia; LGD,
low-grade dysplasia; NA , not applicable; NDBE, non-dysplastic Barrett oesophagus. aHistopathology features are from REF.239.
b
Meta-analysis.
the SOPRAN and LOTUS prospective trials confirm that with EAC and >5,000 population-based controls, users of
PPIs are generally safe as maintenance therapy136. The aspirin and non-aspirin NSAIDs had a 32% lower risk
two main rationales for using potent acid suppression of EAC than controls (OR 0.68, 95% CI 0.56–0.83) and
for chemoprevention are the inhibition of carcinogenesis there was a dose–response relationship142. In a meta-
by reduction in inflammatory pathways137 and the sup- analysis of nine observational studies that included
pression of the pro-proliferative effects of intermittent 5,446 participants, aspirin and non-aspirin NSAID use
acid exposure, which is supported by in vitro data138,139. was associated with a reduced risk of EAC and HGD
In a meta-analysis of seven observational studies that among individuals with BE (relative risk (RR) 0.64, 95%
included 2,813 patients with BE, PPI use was associ- CI 0.53–0.77)143. In the AspECT trial, aspirin alone did
ated with a 71% reduction in cancer risk (adjusted OR not delay death compared with no aspirin (TR 1.25, 95%
0.29, 95% CI 0.12–0.79)140 and a weak dose–response CI 0.92–1.70), although there was a small effect in delay-
relationship was present, despite considerable hetero- ing development of HGD (TR 1.51, 95% CI 1.00–2.29)
geneity among studies. However, a subsequent nation- but not EAC (TR 1.02, 95% CI 0.64–1.64). The combi-
wide study of the Danish population assessing the effect nation of aspirin and high-dose PPI had the strongest
of PPIs in 9,883 patients with BE did not show any effect in delaying all- cause mortality, EAC and HGD
cancer-protective effect. In the multicentre, two-by-two, (TR 1.59, 95% CI 1.14–2.23)141, although this benefit
factorial, randomized AspECT trial, 2,557 patients with required compliance with high-dose PPI and aspirin
BE received either low-dose or high-dose esomeprazole, treatment for >10 years.
with or without 300 mg aspirin141. Whereas high-dose Statins can be chemopreventive independently of
PPI did not affect the occurrence of dysplasia or EAC, their lipid- lowering activity, as they reduce cell pro-
it delayed all-cause mortality (time ratio (TR) 1.36, 95% liferation and induce apoptosis in oesophageal cells
CI 1.01–1.82). Although the effect on all-cause mortality by inhibition of RAS farnesylation and ERK and AKT
is difficult to explain, the similar rate of adverse events signalling pathways144. In a case–control study of 303
for the two PPI doses provides additional evidence for individuals with BE and 909 control individuals, statin
the safety of high-dose PPI. The association between PPI use was associated with 43% lower risk of BE (OR 0.57,
use and a decreased risk of progression in BE, and the 95% CI 0.38–0.87)145 and the effect was much stronger
possibility that inflammation may promote progression among obese individuals (OR 0.26, 95% CI 0.09–0.71).
to neoplasia, validates once-daily PPI therapy, even in These results were confirmed in another smaller case–
patients without reflux symptoms. control study of 123 individuals with BE and 268 con-
Aspirin and cyclooxygenase 2 (COX2) inhibitors trol individuals in eastern England, which showed a
have also been evaluated for potential chemoprevention, similar protective effect of statins against BE (OR 0.62,
given the role of COX2 in inflammation and neoplastic 95% CI 0.37–0.93) and a greater protective effect with
progression. In a pooled analysis of >1,200 individuals the combination of statins and aspirin (OR 0.43, 95%
0123456789();
PRIMER
CI 0.21–0.89)52. Similarly, in another case–control study major risk factor for BE, they fail to include other easily
of 311 individuals with EAC and 856 matched controls, identifiable predisposing conditions (such as smoking,
statin use was inversely related to EAC development poor diet and obesity) that are available for risk stratifi-
(OR 0.65, 95% CI 0.47–0.91)146. Interestingly, the lack of cation. Furthermore, the substantial prevalence of BE in
a protective effect with other lipid-lowering drugs in two patients without GERD casts doubt on whether a GERD-
of these studies145,146 indicates that this protective effect centric screening strategy can substantially influence
is specific to statins. persistently poor EAC outcomes6,45.
Prospective studies are required to confirm these In the past decade, professional societies have devel-
findings. Despite these promising results, chemopreven- oped a more multifaceted approach to screening rec-
tion is not currently recommended by society guidelines, ommendations, such that the target population for BE
but its use should probably be re-considered in future screening parallels the characteristics of the population
revisions of guidelines1–3,109. at highest risk of developing EAC (TABLE 4). Although
guidelines vary, current recommendations are to con-
Screening sider screening by conventional upper endoscopy in
Patients with EAC have a dismal prognosis (5-year patients with chronic, frequent gastro- oesophageal
survival = 15%), and >40% of patients are diagnosed reflux symptoms and the presence of several other risk
with EAC after it has metastasized147. However, a prior factors1–4,109. After an initial negative endoscopy, the
diagnosis of BE before a cancer diagnosis is associated risk of developing BE in patients with chronic GERD
with detection of the tumour at an earlier stage and is low150. Therefore, one-time endoscopy could reassure
improved survival148. Although BE is a known precursor of patients with GERD without endoscopic evidence of BE
EAC, in daily practice >90% of patients with EAC never that the risk of developing BE in the future is extremely
had prior endoscopy and EAC is usually diagnosed low. Although an absolute age cut-off for BE screening is
when symptoms develop, outside a programme of BE not recommended, it is important to consider the overall
surveillance148. life expectancy of the patient and explain the implica-
The fact that endoscopic therapy has become the tions of a diagnosis of BE before screening is conducted.
standard of care for dysplasia or EAC in patients with General population screening for BE is currently not
BE, together with the recent developments in mini- recommended in any society guidelines.
mally invasive (non)endoscopic screening techniques
for BE, means that screening for EAC and its precur- Risk models. A risk model was created from the US
sor BE potentially fits the WHO criteria for screening Veteran’s Administration database, using cigarette
(Supplementary Box 1) and might decrease EAC-related smoking, age, waist:hip ratio and GERD symptoms as
mortality. To date, upper endoscopy combined with predictive variables for BE, and substantially improved
biopsy sampling is the only screening method for BE prediction of the presence of BE compared with a model
that is approved by most societies. using GERD symptoms alone (area under the curve
(AUC) = 0.61 versus 0.72)151. This score was validated
Target populations. Before 2010, chronic GERD was in a case–control study in four independent data sets and
the only criterion for entry in endoscopic screening showed similar results152. Attempts to improve predic-
programmes. For example, the first guideline from the tion models using genetic variants that are known to be
ACG recommended endoscopic screening for anyone associated with BE or other humoral biomarkers, such
with >5 years of reflux symptoms149. Although these as serum leptin or interleukin levels, are underway but
guidelines correctly recognize GERD symptoms as a are not yet ready for clinical use90,153.
Table 3 | Criteria for diagnosis of BE

Society Year Endoscopic criterion Histological criterion Ref.
AGA 2011 Columnar epithelium extending for any length Intestinal metaplasia 1
above the GEJ into the tubular oesophagus

ASGE 2012 Salmon or pink colour, in contrast to the Specialized intestinal metaplasia 4
light-grey appearance of the oesophageal

squamous mucosa
BSG 2014 Distal oesophageal mucosa replaced by Columnar metaplasia (columnar-lined 3
metaplastic columnar epithelium, which is clearly epithelium, regardless of the presence

visible endoscopically (≥1 cm) above the GEJ or absence of intestinal metaplasia)
ACG 2016 Extension of salmon-coloured mucosa into the Specialized intestinal metaplasia 2
tubular oesophagus for ≥1 cm proximal to the GEJ

JSG 2016 Columnar epithelium that continues from the Columnar metaplasia (columnar-lined 110
stomach to oesophagus for any length epithelium, regardless of the presence

or absence of intestinal metaplasia)
ESGE 2017 Distal oesophagus lined with columnar epithelium Specialized intestinal metaplasia 109
with a minimum length of 1 cm (tongues or circular)

ACG, American College of Gastroenterology ; AGA , American Gastroenterological Association; ASGE, American Society for
Gastrointestinal Endoscopy ; BSG, British Society of Gastroenterology ; ESGE, European Society of Gastrointestinal Endoscopy ;
GEJ, gastro-oesophageal junction; JSG, Japanese Society of Gastroenterology.
0123456789();
PRIMER
Overall, tools that incorporate risk factors for analysis reported conflicting results about the diagnostic
BE and EAC could be used to identify high-risk patients ability of ECE.
and improve population screening. However, the devel- Collecting cells from the oesophagus for histo-
opment of minimally invasive screening tools may pathological analysis without acquiring images offers
reduce the need for these BE risk scores by enabling an alternative approach to endoscopy to screen for BE.
more widespread screening. The most studied non-endoscopic cytological device
for BE screening is an encapsulated sponge device
Minimally invasive screening tools. Upper endoscopy is attached to a string (Cytosponge)8. This device, com-
invasive, expensive and not suitable for widespread use. bined with measuring trefoil factor 3 (TFF3; a cellular
Therefore, a need exists for a safe, effective, minimally marker of intestinalization), has shown promising accu-
invasive screening method that is acceptable to patients. racy and acceptability in clinical studies8,13 and is now
Several alternatives are currently being investigated154,155 being evaluated in a cluster-randomized trial of 13,000
(FIG. 6), although only the ACG guidelines suggest an patients in primary care; further biomarkers can be
alternative to upper endoscopy, unsedated transnasal used to assess for the presence of dysplasia163. A sim-
endoscopy (TNE)2. ilar approach using non- endoscopic cell collection
Despite a narrower working channel and smaller devices (for example, inflatable balloons and capsule
biopsy sample size, TNE has comparable clinical effec- sponges) combined with a panel of methylated DNA or
tiveness (in terms of participation rates, yield, safety, microRNA (miRNA) markers also showed promise for
tolerability and patient preference) to upper endoscopy non-endoscopic detection of BE, but the results of larger
and has lower direct and indirect costs156–158. Disposable studies are awaited164–167. Although copious quantities of
silicone sheaths to enable reuse of the endoscope with- cells are collected, a drawback of these non-endoscopic
out reprocessing (Endosheath) or completely disposable cell collection devices is the limited control over their
transnasal capsules (EG Scan) provide more options passage through the oesophagus and the possible failure
for TNE, although biopsies are not always possible159. to retrieve cells from the GEJ.
In addition, TNE can be performed by non-physicians Biomarker detection in the breath (using an elec-
and outside of hospitals, which could further reduce tronic nose device to detect volatile organic compounds)
costs and increase access160. or in blood (detection of miRNAs) is an attractive
Novel minimally invasive techniques are in devel- method for screening, as it is non- invasive, provides
opment and showed promising results in case–control results quickly and is fairly inexpensive168, although few
studies but are not yet ready for clinical application studies of this method have been published to date.
(Supplementary Table 1). An imaging-based method,
oesophageal capsule endoscopy (ECE), enables direct Cost-effectiveness. All published modelling studies have
non-invasive visualization of the oesophagus, although assessed the cost-effectiveness of BE screening in white
biopsies are not possible161. In a meta-analysis assessing men >50 years of age who have chronic reflux symp-
the accuracy of ECE for BE detection in 618 patients, toms169 and have assumed a BE prevalence of 10% and a
pooled sensitivity and specificity were 77% and 86%, rate of progression to EAC of 0.5% (which may be higher
respectively162, although the studies included in the than current estimates) in this population. These studies
have mostly found that screening with standard upper
endoscopy is cost-effective compared with no screening
Table 4 | Recommended populations for BE screening and decreases EAC incidence by 15–25% in this popu-
Society Year Screening population Ref. lation (incremental cost-effectiveness ratios (ICERs)
were less than US$50,000 in most studies). Data on cost-
AGA 2011 Multiple risk factors (≥50 years of age, male sex, white 1
effectiveness of screening with TNE are somewhat limi-
ethnicity, chronic GERD symptoms, hiatal hernia and
obesity (elevated body mass index and intra-abdominal ted, but ICERs of US$55,000 and US$28,000 have been
distribution of body fat)) reported169,170. Other (mostly older) studies did not use
endoscopic therapy for the treatment of BE-related dys-
ASGE 2012 Multiple risk factors (male sex, white ethnicity, >50 years 4
of age, increased duration of reflux symptoms, smoking, plasia or mucosal (stage T1a) EAC in the models, which
obesity and family history of BE) may underestimate the positive effect of screening171,172.
In addition, participation rates and direct and indirect
BSG 2014 Chronic GERD symptoms with ≥3 risk factors (50 years of 3
age, white ethnicity, male sex, obesity and first-degree costs were not included in estimates, and unreliable
relative with BE or EAC) assumptions (such as 100% participation) were made.
Newer, minimally invasive screening technologies (such
ACG 2016 Men with either >5 years GERD and/or with more than 2
weekly symptoms and ≥2 risk factors (>50 years of age, as capsule sponge-based techniques173) are reported to
central obesity (waist circumference >102 cm or WHR >0.9), be cost-effective in individuals with reflux symptoms
white ethnicity, smoking and first-degree relative with BE (Supplementary Table 1), but cost-effectiveness at the
or EAC) population level has not been assessed to date.
ESGE 2017 Longstanding GERD symptoms (>5 years) and multiple risk 109
factors (≥50 years of age, white ethnicity, male sex, obesity Management
and first-degree relative with BE or EAC) Strategies for the management of BE depend on the stag-
ACG, American College of Gastroenterology ; AGA , American Gastroenterological Association; ing of the disease. Endoscopic treatment is indicated for
ASGE, American Society of Gastrointestinal Endoscopy ; BE, Barrett oesophagus; BSG, British
Society of Gastroenterology ; EAC, oesophageal adenocarcinoma; ESGE, European Society of patients with BE who have LGD, HGD, mucosal EAC
Gastrointestinal Endoscopy ; GERD, gastro-oesophageal reflux disease; WHR , waist:hip ratio. and ‘low-risk’ submucosal EAC (FIG. 7) and is preferred
0123456789();
PRIMER
a b c
5.4 mm
9.8 mm
d e f
Cardiac-type cell
Goblet cell
0 cm
1 cm
Squamous cell
TFF3+ goblet cell
2 cm
Fig. 6 | Screening techniques for Barrett oesophagus. a | Standard upper endoscope (Pentax EG29 (bottom)) and
transnasal endoscope (Pentax EG16 (top)). b | Electronic nose. c | Capsule endoscopy. d | Cytosponge sampling device
(expanded (left) and unexpanded (right)). e,f | A positive Cytosponge test shown by haemotoxylin and eosin staining
(part e) and trefoil factor 3 (TFF3) immunohistochemistry (part f), with positive staining of the gland groups with goblet
cells but not of cardiac-type mucosa and squamous epithelial cells. Magnification ×20 in parts e and f.
over surgical treatment. Management begins with careful Endoscopic surveillance with HD- WL endoscopy
inspection of the dysplastic BE segment and endoscopic and histopathological evaluation of biopsy samples is
resection of all visible lesions to enable adequate staging, now considered the standard of care. In some centres,
followed by ablative therapy of the remaining BE3,109,174. several other advanced imaging methods (described
Ablation should not be used as a primary treatment of earlier) are used for surveillance in clinical practice,
early cancer in BE or endoscopically visible lesions, as although their routine use in patients with BE is not
the target mucosa is destroyed and histological staging supported by scientific evidence or recommended by
is not possible after this treatment. In the absence of vis- clinical guidelines.
ible mucosal irregularities, endoscopic ablation without The cost-effectiveness of surveillance in all patients
resection is sufficient. with BE is the subject of debate. Many gastroentero-
Endoscopic treatment of non-dysplastic BE is not logists argue that only patients at high risk of malignant
indicated, as risk of progression to HGD or EAC is progression should be included in a surveillance pro-
too low to justify treatment. Any ablative therapy in gramme. An easy scoring system was developed that
patients with non-dysplastic BE is associated with costs can identify patients at high risk of malignant progres-
and potential complications and therefore should not sion56, which identified male sex, smoking, BE length
be performed outside controlled studies. Instead, endo- and baseline-confirmed LGD as significantly associated
scopic surveillance of patients with non-dysplastic BE is with progression to EAC. A similar model combining
recommended in all available guidelines3,109,174. age, male sex and BE length could predict 71% of HGD
and EAC cases176. Further prospective studies are needed
Surveillance to validate these models.
The aim of surveillance in patients with BE is to
improve patient outcomes by detection of dysplasia Endoscopic resection
or EAC at an early stage to ensure effective treatment. Endoscopic mucosal resection. Patients with BE lesions
The efficacy of current surveillance strategies in redu- containing dysplasia or superficial early EAC should
cing mortality in patients with BE compared with the undergo endoscopic mucosal resection (EMR) as the
general population is debated148,175. A meta- analysis initial diagnostic and therapeutic procedure (FIG. 8).
demonstrated that surveillance was associated with EMR provides a tissue specimen that can be evaluated
diagnosis of earlier-stage EAC59. Furthermore, all-cause for prognostic factors, such as dysplasia grade, differen-
mortality was lower for cancers that are detected during tiation grade, infiltration depth, vascular invasion and
surveillance than for cancers that were detected out- completeness of the resection. In a retrospective study,
side surveillance programmes. However, the benefit histopathological assessment of EMR tissue samples
of lower mortality was eliminated after adjustment for changed the diagnosis in ~50% of patients owing to the
lead-time and length-time biases. large tissue sample obtained with EMR177.
The surveillance intervals are risk stratified by the The most common EMR technique involves the use
presence and grade of dysplasia and, in European soci- of a multiband ligation device and a dedicated snare that
ety guidelines, by the length of the BE segment (TABLE 5). can be advanced through the working channel with the
0123456789();
PRIMER
Diagnosis of BE
• EGD
• Histopathology of biopsy sample
Non-dysplastic Indefinite for HGD Submucosal

LGD
BE dysplasia Mucosal EAC EAC
• Repeat EGD at 3–6 months Visible lesion at EGD

• Acid-suppression therapy
No Yes
EMR
Dysplasia confirmed by Histopathological ESD

histopathology confirmation of EAC
No No
Yes
Yes
Low-risk features in resection specimen
• Submucosal infiltration <500 μm
• Good and/or moderate differentiation
Yes • No lymphovascular infiltration
• Tumour-free deep margin
Surveillance Follow LGD Follow-up Endoscopic ablation No
EGD every and/or HGD EGD at in case of residual
3–5 years flowchart 3–6 months BE segment Surgery
Fig. 7 | Algorithm for management of non-dysplastic and dysplastic Barrett oesophagus. In cases of Barrett
oesophagus (BE) confirmed by oesophagogastroduodenoscopy (EGD) and histopathological assessment of biopsy
samples, BE is staged as non-dysplastic, indefinite for dysplasia, low-grade dysplasia (LGD), high-grade dysplasia (HGD) or
mucosal oesophageal adenocarcinoma (EAC), or submucosal EAC. For lower-grade disease (non-dysplastic or indefinite
for dysplasia), surveillance at various intervals is recommended. For higher-grade disease with lesions visible by EGD,
endoscopic mucosal resection (EMR) followed by ablation of the residual BE segment is recommended. In case of no visible
lesion, endoscopic ablation is recommended in patients without life-limiting comorbidity, otherwise strict surveillance
is warranted. Endoscopic submucosal dissection (ESD) may be considered in selected cases of EAC (with poorly lifting
tumours, >15 mm or lesions at risk of submucosal invasion). If the endoscopic resection does not fulfil low-risk criteria,
surgery is needed.
ligation device in place. EMR is an advanced endoscopic experienced hands, it is technically demanding and
procedure that requires training by an experienced requires intensive training. In several prospective case
endoscopist in high-volume centres. A case series from series, complete resection (R0) rates of BE-associated
the UK demonstrated that EMR in the upper GI tract EAC were 38.5–79% and stricture rates were up to
was associated with a substantially higher mortality and 60%180,181. A prospective randomized series comparing
perforation rate when performed by less experienced EMR and ESD in 40 patients with mucosal BE-associated
endoscopists178. Therefore, current society guidelines EAC found no significant difference in complete remis-
recommend that BE-associated EAC should be treated sion rates182. However, EMR is still the treatment of
only in expert centres (centres with ten or more new choice in current society guidelines181. Indications for
cases with early neoplasia per year, access to experi- ESD include bulky lesions that cannot be removed by
enced oesophageal surgery teams and prospective data EMR and suspicion of submucosal infiltration.
collection for all BE cases)109.
The largest case series of EMR in 1,000 patients with Complications of endoscopic therapy. The major com-
BE who have mucosal EAC demonstrated excellent long- p lications of endoscopic resection include severe
term complete remission rates of neoplasia (93.8%) after bleeding, perforations and strictures. Although ESD
follow-up of almost 5 years179. The reported complica- is technically more challenging than EMR and has a
tion rate was only 1.5%, although neoplastic recurrence steeper learning curve, the complication rates are low
rate is a problem in endoscopic treatment (14.5% in this in experienced hands. Bleeding is one of the most
series). As all patients were in a follow-up programme, common complications of endoscopic therapy (<0.5%
recurrences were diagnosed at an early stage, making for EMR and 0.9–6.7% for ESD)179,183. Perforations after
endoscopic re-treatment feasible in almost all cases. EMR and ESD occur in up to 5% of treated patients179,182.
Stricture formation after EMR is related to the extent
Endoscopic submucosal dissection. Endoscopic sub- of resection and is minimal if <50% of the circum-
mucosal dissection (ESD) enables en bloc resection ference is resected184. Complications of endoscopic
of lesions of any size that invade the mucosa and therapy can usually be successfully treated during
submucosa. Although ESD is safe and effective in the same endoscopic procedure using endoscopic
0123456789();
PRIMER
bleeding control techniques, clip or stent placement resection followed by RFA in 132 patients with BE and
or dilatation. HGD or early EAC, complete remission rates for neopla-
sia and IM were 98% and 93%, respectively187,188. Most
Endoscopic ablation current society guidelines therefore recommend endo-
Radiofrequency ablation. The high rate of confirmed scopic resection combined with RFA as the treatment
LGD progression to EAC is the reason why endoscopic of choice in patients with HGD and early BE-associated
ablation is recommended as an alternative to frequent EAC3,109,174.
endoscopic follow-up (FIG. 8). In radiofrequency ablation
(RFA), thermal ablation of the mucosa is performed Cryoablation. Cryoablation is one of the newest tech-
using an electromagnetic current. In a multicentre, niques for ablation of BE, and two approaches are avail-
sham-controlled study (the AIM Dysplasia trial), com- able. Endoscopic spray cryotherapy involves spraying
plete eradication of BE was achieved in 91% of patients either liquid nitrogen or rapidly expanding carbon
and of dysplasia in 98% of patients after 3-year follow- dioxide gas over the BE segment. Cryoballoon ablation
up63,185. These results were confirmed in a randomized involves expanding a balloon at the level of the BE seg-
clinical trial of RFA versus PPIs and endoscopic follow- ment and then a focal spray ablation is performed. Both
up186; ablation reduced the risk of progression to HGD methods destroy the target mucosa by rapid freezing of
by 25% and to EAC by 7.4% over 3-year follow-up. the tissue. Data regarding the treatment of dysplasia and
Complete ablation of the residual BE epithelium after EAC in patients with BE using cryoablation are limited.
endoscopic resection of neoplastic lesions can signifi- However, complete eradication of dysplasia and complete
cantly reduce recurrence rates178. RFA of the residual ablation of BE mucosa have been reported in 87–96% and
BE epithelium is the current treatment standard and 57–96% of patients, respectively188,189. Cryoablation was
has been evaluated extensively (>100 peer- reviewed also effective in patients with BE and early EAC, with
articles and >250,000 patients treated worldwide). In complete remission in 75% of these patients, including
the Euro-2 prospective multicentre trial of endoscopic those in whom other endoscopic treatments failed188,189.
Table 5 | Recommendations for BE surveillance and management

Society Non-dysplastic BE Indefinite for dysplasia LGD HGD Ref.
(year)
AGA EGD every 3–5 years Not specified • EGD every 6–12 months • EGD every 3 months 1
(2011) • Consider endoscopic • Endoscopic eradication therapy

eradication therapy rather than surveillance or surgery
ASGE • EGD every 3–5 years Repeat EGD with maximal acid • Repeat EGD in 6 months • EGD every 3 months 4
(2012) • Consider no surveillance suppression to confirm LGD (only patients who are not
• Consider ablation in • EGD every year candidates for endoscopic or
select cases • Consider endoscopic surgical treatment)
therapy • Consider endoscopic treatment
• Consider surgical consultation
BSG • Irregular Z-line: no Repeat EGD at 6 months with • Surveillance: EGD every • Mucosal irregularity: EMR 3
(2014) surveillance maximal acid suppression 6 months • Endoscopic therapy is preferred

• BE <3 cm without IM: • Ablation cannot be over oesophagectomy or
no surveillance recommended routinely surveillance
• BE <3 cm with IM: EGD every
3–5 years
• BE ≥3 cm: EGD every
2–3 years
• Consider no surveillance on
the basis of patient’s fitness
and risk of progression
ACG EGD every 3–5 years • Repeat EGD at 3–6 months • Endoscopic treatment Endoscopic treatment (patients 2
(2016) after optimization of acid (patients without life- without life-limiting comorbidity)
suppression limiting comorbidity)
• Persistent indefinite for • EGD every 12 months
dysplasia: EGD after 1 year
ESGE • BE <1 cm: no surveillance Repeat EGD at 6 months with • Repeat EGD at 6 • Repeat EGD 109
(2017) • BE 1–3 cm: EGD every 5 years optimization of anti-reflux months • Visible irregularity: EMR
• BE 3–10 cm: EGD every 3 years medication • If persistent LGD: • Persistent HGD: ablation
• BE ≥10 cm: referral to BE endoscopic ablation • No dysplasia: repeat EGD
expert centre 3 months
• Consider discharge for
patients with limited
life expectancy and
advanced age
ACG, American College of Gastroenterology ; AGA , American Gastroenterological Association; ASGE, American Society of Gastrointestinal Endoscopy ;
BE, Barrett oesophagus; BSG, British Society of Gastroenterology ; EGD, oesophagogastroduodenoscopy ; EMR , endoscopic mucosal resection; ESGE, European
Society of Gastrointestinal Endoscopy ; HGD, high-grade dysplasia; IM, intestinal metaplasia; LGD, low-grade dysplasia.
0123456789();
PRIMER
a b c
*
EMR
d e f
RFA
Fig. 8 | Endoscopic approaches for treatment of Barrett oesophagus. Endoscopic mucosal resection (EMR) and
radiofrequency ablation (RFA) of early Barrett oesophagus (BE)-associated oesophageal adenocarcinoma. a | BE with visible
lesion (asterisk). b | EMR, starting with suction of the lesion into the cap. c | Complete endoscopic resection (white arrow).
d | A circumferential RFA balloon is positioned in the BE segment. e | Whitish discoloration (arrowhead) is the immediate
effect after treatment with circumferential RFA. f | Application of focal RFA (black arrow) to residual islands of BE.
Argon plasma coagulation. Argon plasma coagula- Stepwise endoscopic resection

tion (APC), one of the earliest thermal ablation tech- Recurrence of EAC can also be prevented by complete
niques used for BE eradication, involves passing a stepwise endoscopic resection of the neoplastic lesion
high-frequency electric current through ionized argon and the entire BE segment. In a study of stepwise
gas applied to a lesion. APC can achieve remission of radical resection with long-term follow- up (median
BE in most patients after initial APC treatment190. In a 76 months), complete remission rates at the last follow-
prospective randomized trial, APC ablation of the BE up endoscopy were 95% for IM and 97% for neoplasia195.
epithelium that remained after successful endoscopic However, the main disadvantage of this approach is the
resection of mucosal EAC significantly reduced the rate high stricture rate (>50%). In a meta-analysis compar-
of recurrent or metachronous neoplasia compared with ing a two-step approach (focal EMR followed by RFA)
surveillance187. Hybrid APC combines submucosal fluid and stepwise radical resection, complete remission rates
injection and APC with a higher-frequency electric cur- were similar but the complication rate was significantly
rent than for conventional APC, which seems to result in higher for the stepwise approach (strictures 33.5% ver-
a more effective ablation with fewer strictures owing to sus 10.2%; bleeding 7.5% versus 1.1%; perforation 1.3%
the protective effect of the submucosal fluid cushion191. versus 0.2%)196. Owing to these excellent results and
A major downside of APC is the operator dependency, the lower rate of adverse events, the two-step approach
the larger number of sessions (than other ablative ther- is the recommended treatment strategy in all society
apies) that are needed to achieve complete eradication guidelines3,109.
of the BE epithelium and the fairly high risk of residual
islands of metaplasia. Nonetheless, APC is substantially Managing submucosal EAC in BE
less expensive than RFA and cryoablation. EAC that infiltrates the submucosa is associated with
a substantially increased risk of lymph node meta-
Complications of endoscopic ablation. Patients often stasis197. Tumours that infiltrate the upper third of the
experience chest pain after thermal ablation. Stricturing submucosa (TNM (tumour, node, metastasis) stage
is the most common complication of RFA (occurring in pT1sm1; invasion depth ≤500 μm) have a lymph node
5–14% of patients)192. EMR before RFA may increase this metastasis risk of 0–21%, whereas risk increases to
stricture rate. In the AIM Dysplasia trial, bleeding after 36–54% when the tumour invades deeper layers of the
RFA occurred in one patient (<1%) and was probably a submucosa (pT1sm2 or pT1sm3)198,199.
result of anticoagulant therapy185. If ablative therapy does Endoscopic therapy can be safely performed in so-
not destroy all metaplastic epithelium, the partially ablated called low-risk submucosal EAC (that is, Tsm1 tumour,
mucosa may heal with an overlying layer of neosquamous invasion depth ≤500 μm, good to moderate differentiation
epithelium to produce subsquamous IM (SSIM). Most with no lymphovascular invasion and diameter <20 mm)
patients with BE and neoplasia seem to have SSIM193, in patients with BE198,199. In a case series of 67 patients
as the overlying squamous epithelium hides this SSIM with submucosal BE-associated EAC who were treated
from endoscopic detection and might protect it from by endoscopic resection, only one patient developed a
RFA. The clinical relevance of SSIM is still uncertain, but lymph node metastasis199, representing a risk of 1.5%,
these buried glands may have malignant potential194. which is below the usual mortality for oesophagectomy.
0123456789();
PRIMER
Thus, these data suggest that endoscopic treatment of (59%) reported the procedure as burdensome and expe-
low-risk T1sm1 BE-associated EAC can be recommended rienced distress beforehand212. A follow-up study showed
as an alternative to oesophagectomy. that patients with BE with and without nonspecific
upper GI symptoms experienced less discomfort, pain
Follow-up after endoscopic therapy and overall burden during a surveillance endoscopy than
Endoscopic follow- up after endoscopic treatment is patients without BE who were undergoing a diagnostic
mandatory owing to the frequent recurrence of BE and endoscopy owing to nonspecific upper GI symptoms209.
dysplasia3,109,174. However, evidence- based data about Patients who interpreted their risk of developing EAC
surveillance intervals are very limited and recommen- as high had higher levels of procedural discomfort and
dations are typically based on expert opinions2,4. The seemed to have worse strategies for coping with sur-
incidence of neoplastic recurrence after complete erad- veillance endoscopies212. Thus, elevated impact of event
ication of IM has been modelled, which has yielded scores may be associated with anxiety concerning the
evidence-based suggested surveillance intervals200. For physical burden of the upper endoscopy but also with
patients with LGD, surveillance endoscopy at 1-year and the worry about cancer risk, which is usually followed
3-year follow-up are recommended. For patients with by subsequent relief from a negative test.
HGD or mucosal adenocarcinoma, surveillance endo-
scopies at 3-month, 6-month and 1-year follow-up after Cancer risk perception and cancer worries. Patients with
complete remission should be performed, and annu- BE may feel psychologically burdened by the threat of
ally thereafter. The implementation of this surveillance developing EAC, which may negatively affect their
schedule seems to be feasible and protects well against QOL213. Misperception of cancer risk can have important
invasive adenocarcinoma200. psychological consequences and may affect screening-
related health behaviour214. Perceptions of cancer risk
Quality of life vary widely in patients with BE. In two studies, ~60%
A diagnosis of BE affects various domains of QOL, of patients with BE underestimated their numerical
increases health-care costs, affects health-care utilization annual risk of EAC (annual risk 0.1–0.2%)215,216. By
and health behaviour and carries the potential for mor- contrast, two studies reported that the majority of the
bidity. Decreased QOL is usually due to GERD symptoms 134 patients with BE overestimated their 1-year can-
and concern about carrying a pre-malignant condition201. cer risk (mean perceived risk 6% and 13.6%)217,218 and
were willing to accept low success rates and high risks
QOL in non-dysplastic BE of complications to undergo endoscopic therapy218.
Symptoms. A validated QOL score specifically for Overestimating EAC risk is associated with more reflux
patients with BE does not exist. However, in numerous symptoms, lower QOL scores and worse illness per-
studies, the severity and frequency of GERD symp- ceptions216,217. Risk perceptions did not correlate with
toms are associated with physical pain and reduced endoscopic surveillance attendance215,217. Nonetheless,
social, emotional and physical functioning202,203. Both patients’ knowledge of cancer risk seems to be insuf-
generic and GERD-specific QOL scores are substan- ficient and it is questionable whether patients with
tially reduced in individuals with BE compared with BE are provided with enough information about their
the general population but are comparable to those of diagnosis, their cancer risk and the role of surveillance
individuals with GERD204–206. As patients with BE are a and treatment.
subset of those with GERD in most studies, they likely
had similar symptoms and thus may not represent the QOL in dysplastic BE or early EAC
total BE population204–206. Therefore, symptom control A diagnosis and (surgical) treatment of dysplasia and
by PPIs seems to be an important factor in maximizing EAC, regardless of the stage, have a major impact on
QOL but alone is not sufficient205. patients and their QOL219,220. Endoscopic therapy for
dysplasia and early EAC is less invasive and beneficial
Labelling effect and psychological burden. A diagnosis in terms of procedural risk and long- term symptoms
of a potentially life-threatening disease may adversely compared with oesophagectomy63. As endoscopic treat-
affect patient QOL independent of the disease’s biologi- ment has become standard of care, the risk of disease
cal effects (labelling effect)207,208. Studies assessing the recurrence and the burden of subsequent surveillance
potential psychological consequences of living with must also be considered. Two consecutive studies
BE did not show differences among patients with BE showed that patients with early Barrett neoplasia treated
and those with GERD206. By contrast, increased anxi- endoscopically have better QOL in both physical and
ety and depression scores (using the Hospital Anxiety mental domains of the 36-Item Short Form Survey
and Depression Scale), both before and after endoscopic (SF-36) and fewer EAC- related symptoms compared
surveillance, have been shown in patients with BE com- with those treated surgically221,222. By contrast, fear
pared with patients with GERD symptoms undergoing of cancer recurrence seems to be higher in patients
endoscopy209,210. endoscopically treated for BE-associated EAC, despite
excellent 5-year survival222,223.
Burden of endoscopic surveillance. Although it has been QOL following RFA of dysplastic BE was assessed
shown to be safe, upper endoscopy is an invasive proce- in the AIM Dysplasia trial63,220. Eradication of dysplas-
dure that is not well tolerated by all individuals211. In a tic BE reduced worry of cancer recurrence, depres-
questionnaire study of 180 patients with BE, the majority sion scores and impact on daily work and family life
0123456789();
PRIMER
compared with persistent BE (sham procedure)220. QOL factors for BE and EAC are available to enrich the popu-
was particularly improved in patients with complete lation suitable for screening, but they have important
eradication of dysplasia and BE. These improvements limitations. Importantly, although strong risk factors
seem to be secondary to a decrease in patients’ perceived for BE, including GERD symptoms, age, white ethni-
risk of cancer220. city, male sex, waist circumference and smoking, have
been identified, chronic GERD symptoms are still most
Outlook commonly used to select individuals for endoscopic
Since the initial description of BE in 1950, progress has screening. A more sophisticated risk stratification
been made in understanding of BE pathogenesis and model for clinical practice and a cost-effective, accu-
extensive research has yielded improvements in endo- rate, minimally invasive or serum-based screening test
scopic diagnosis and management of BE and the iden- that is acceptable to both patients and providers are
tification of dysplasia in BE. However, many challenges needed. Although several promising modalities have
and opportunities in clinical practice and research still been evaluated, randomized, well-powered studies of
remain (Supplementary Box 2). the appropriate populations are needed to confirm their
ability to identify BE and reduce EAC-related burden.
Pathophysiology Another important question is whether screening
To improve understanding of BE pathogenesis, further should be offered to individuals at high risk of BE or to
research is needed on the different inflammatory pro- the general population.
cesses that lead to IM and malignant progression. The
majority of patients with chronic GERD do not develop Surveillance
BE. Thus, the molecular factors that contribute to the Screening for BE will be beneficial only if it is coupled
development of IM need to be further investigated. with effective surveillance. However, current endoscopic
Although various models of BE pathogenesis speculate surveillance of BE has numerous limitations. Although
about the cellular origin of metaplastic cells, none of the risk of developing EAC is substantially increased
these models explains all aspects of BE; thus, alterna- in patients with BE, the vast majority of these patients
tive pathogenic mechanisms or even a combination of never progress to HGD or EAC18. Therefore, in addi-
mechanisms may more fully explain BE pathogenesis. tion to identifying high-risk patients, patients with BE
Another outstanding question is whether BE is an inter- at low risk of malignant progression must be identified
mediate stage in the development of all EACs7, which to spare them the risks and costs of unnecessary sur-
could be addressed by studies in experimental models veillance and intervention. However, despite extensive
and of Barrett metaplasia in patients. research to establish risk factors for malignant progres-
sion, most guidelines base surveillance intervals on
Diagnosis the presence and extent of dysplasia. Stratification of
Although upper endoscopy and histopathology are the patients with BE based on age, segment length, a history
gold standard for diagnosis, consensus about various cri- of persistent non-dysplastic BE or additional factors may
teria, such as the minimum length of the oesophageal be a first step towards a more personalized surveillance
columnar epithelium and the presence of IM, is neces- approach18,56,224. The efficacy and cost-effectiveness of
sary to obtain a more precise, universal definition of BE. surveillance is currently being assessed in an ongoing
However, this consensus cannot occur independently of large, randomized trial comparing standard surveillance
efforts to establish an accurate and reliable cancer risk and ‘at need’ endoscopy225.
stratification model for BE.
In addition, the optimal management strategy in Prevention
patients with dysplasia is unclear owing to difficulties Substantial progress has been made in the prevention
in detecting and diagnosing dysplasia. Novel imaging and management of BE. The available evidence suggests
modalities might improve the detection of dysplasia, but that maintenance therapy with PPIs for reflux disease is
none seem to be ready for clinical application at present. safe, when clinically appropriate. The protective effect
As histological diagnosis of dysplasia is affected by intra- of aspirin and high- dose PPI seems to be small and
observer and interobserver variability121, more objective mostly related to a reduction in nonspecific causes of
markers are needed to determine the risk of malignant mortality (such as cardiovascular events), whereas their
progression in BE. Artificial intelligence may help in independent chemopreventive effect on EAC unfor-
optimizing endoscopic and histopathological evaluation tunately remains far less certain. Statins are promising
of BE and dysplasia in BE in the future. agents, and the retrospective data seem to support future
prospective trials.
Screening
Early detection and prevention of BE may be the best Management
strategy to combat the increasing incidence of EAC, Endoscopic resection and RFA are effective treatments
as BE is a well-defined precursor lesion and effective that have high eradication rates for BE and dyspla-
endoscopic treatment for dysplasia and early EAC is sia. Endoscopic follow-up after endoscopic treatment
available. Although conventional endoscopy has been is mandatory to detect recurrence of BE or dyspla-
the focus of past research on screening methods, it sia. As many studies have probably overestimated
is not feasible as a primary screening method owing the risk of recurrence by including IM at the GEJ as
to cost and invasiveness. Tools that incorporate risk a BE diagnosis223, the priority in research is to define
0123456789();
PRIMER
recurrent disease after treatment to support the devel- QOL

opment of evidence-based recommendations about the Patients with BE may not be sufficiently knowledgeable
surveillance intervals. about their diagnosis, associated cancer (recurrence)
Promising alternative ablative therapies are in devel- risk and the benefits and drawbacks of surveillance to
opment; for example, success rates with cryoablation make fully informed decisions about surveillance strat-
(cryospray or cryoballoon) are encouraging, with a egies and treatment options. Providing information
safety profile comparable to RFA. However, additional to all newly diagnosed patients with BE may improve
prospective trials of cryoablation are needed before it can disease- specific knowledge and may be helpful for
substitute for RFA. Direct comparison of these ablation engaging patients in shared decision-making to affect
modalities in multicentre trials is ongoing. Nevertheless, health behaviour and QOL3.
cryoablation will likely be useful as a second-line
treatment for patients who fail to respond to RFA226. Published online xx xx xxxx
1. Spechler, S. J., Sharma, P., Souza, R. F., Inadomi, J. M. 19. Olsen, C. M. et al. Population attributable fractions 36. Chandar, A. K. et al. Association of serum levels of
& Shaheen, N. J. American Gastroenterological of adenocarcinoma of the esophagus and adipokines and insulin with risk of Barrett’s esophagus:
Association medical position statement on the gastroesophageal junction. Am. J. Epidemiol. 174, a systematic review and meta-analysis. Clin.
management of Barrett’s esophagus. Gastroenterology 582–590 (2011). Gastroenterol. Hepatol. 13, 2241–2255 (2015).
140, 1084–1091 (2011). 20. Abrams, J. A., Fields, S., Lightdale, C. J. & Neugut, A. I. 37. Ayazi, S. et al. Obesity and gastroesophageal reflux:
2. Shaheen, N. J., Falk, G. W., Iyer, P. G. & Gerson, L. B. Racial and ethnic disparities in the prevalence of quantifying the association between body mass index,
ACG clinical guideline: diagnosis and management Barrett’s esophagus among patients who undergo esophageal acid exposure, and lower esophageal
of Barrett’s esophagus. Am. J. Gastroenterol. 111, upper endoscopy. Clin. Gastroenterol. Hepatol. 6, sphincter status in a large series of patients with reflux
30–50 (2016). 30–34 (2008). symptoms. J. Gastrointest. Surg. 13, 1440–1447
3. Fitzgerald, R. C. et al. British Society of 21. Wang, A., Mattek, N. C., Holub, J. L., Lieberman, D. A. (2009).
Gastroenterology guidelines on the diagnosis and & Eisen, G. M. Prevalence of complicated 38. Wu, J. C., Mui, L. M., Cheung, C. M., Chan, Y. &
management of Barrett’s oesophagus. Gut 63, gastroesophageal reflux disease and Barrett’s Sung, J. J. Obesity is associated with increased
7–42 (2014). esophagus among racial groups in a multi-center transient lower esophageal sphincter relaxation.
4. Evans, J. A. et al. The role of endoscopy in Barrett’s consortium. Dig. Dis. Sci. 54, 964–971 (2009). Gastroenterology 132, 883–889 (2007).
esophagus and other premalignant conditions of 22. Keyashian, K. et al. Barrett’s esophagus in Latinos 39. Leodolter, A. et al. Progression of specialized
the esophagus. Gastrointest. Endosc. 76, 1087–1094 undergoing endoscopy for gastroesophageal reflux intestinal metaplasia at the cardia to macroscopically
(2012). disease symptoms. Dis. Esophagus 26, 44–49 evident Barrett’s esophagus: an entity of concern
5. Ronkainen, J. et al. Prevalence of Barrett’s esophagus (2013). in the ProGERD study. Scand. J. Gastroenterol. 47,
in the general population: an endoscopic study. 23. Shiota, S., Singh, S., Anshasi, A. & El-Serag, H. B. 1429–1435 (2012).
Gastroenterology 129, 1825–1831 (2005). Prevalence of Barrett’s esophagus in Asian countries: 40. Dunbar, K. B. et al. Association of acute
6. Rex, D. K. et al. Screening for Barrett’s esophagus a systematic review and meta-analysis. Clin. gastroesophageal reflux disease with esophageal
in colonoscopy patients with and without heartburn. Gastroenterol. Hepatol. 13, 1907–1918 (2015). histologic changes. JAMA 315, 2104–2112 (2016).
Gastroenterology 125, 1670–1677 (2003). 24. Cook, M. B., Wild, C. P. & Forman, D. A systematic 41. Iascone, C., DeMeester, T. R., Little, A. G. &
7. Sawas, T. et al. Identification of prognostic phenotypes review and meta-analysis of the sex ratio for Barrett’s Skinner, D. B. Barrett’s esophagus. Functional
of esophageal adenocarcinoma in two independent esophagus, erosive reflux disease, and nonerosive reflux assessment, proposed pathogenesis, and surgical
cohorts. Gastroenterology 155, 1720–1728 (2018). disease. Am. J. Epidemiol. 162, 1050–1061 (2005). therapy. Arch. Surg. 118, 543–549 (1983).
8. Ross-Innes, C. S. et al. Evaluation of a minimally 25. Edelstein, Z. R., Bronner, M. P., Rosen, S. N. & 42. Collen, M. J., Lewis, J. H. & Benjamin, S. B. Gastric acid
invasive cell sampling device coupled with assessment Vaughan, T. L. Risk factors for Barrett’s esophagus hypersecretion in refractory gastroesophageal reflux
of trefoil factor 3 expression for diagnosing Barrett’s among patients with gastroesophageal reflux disease. Gastroenterology 98, 654–661 (1990).
esophagus: a multi-center case-control study. disease: a community clinic-based case-control study. 43. Mulholland, M. W., Reid, B. J., Levine, D. S. &
PLOS Med. 12, e1001780 (2015). Am. J. Gastroenterol. 104, 834–842 (2009). Rubin, C. E. Elevated gastric acid secretion in patients
9. Vaughan, T. L. & Fitzgerald, R. C. Precision prevention of 26. Westra, W. M. et al. Smokeless tobacco and cigar with Barrett’s metaplastic epithelium. Dig. Dis. Sci.
oesophageal adenocarcinoma. Nat. Rev. Gastroenterol. and/or pipe are risk factors for Barrett esophagus in 34, 1329–1334 (1989).
Hepatol. 12, 243–248 (2015). male patients with gastroesophageal reflux disease. 44. Taylor, J. B. & Rubenstein, J. H. Meta-analyses of the
10. Arnold, M., Laversanne, M., Brown, L. M., Devesa, S. S. Mayo Clin. Proc. 93, 1282–1289 (2018). effect of symptoms of gastroesophageal reflux on
& Bray, F. Predicting the future burden of esophageal 27. Cook, M. B. et al. Cigarette smoking increases risk the risk of Barrett’s esophagus. Am. J. Gastroenterol.
cancer by histological subtype: international trends of Barrett’s esophagus: an analysis of the Barrett’s 105, 1730–1737 (2010).
in incidence up to 2030. Am. J. Gastroenterol. 112, and Esophageal Adenocarcinoma Consortium. 45. Ward, E. M. et al. Barrett’s esophagus is common
1247–1255 (2017). Gastroenterology 142, 744–753 (2012). in older men and women undergoing screening
11. Jung, K. W. et al. Epidemiology and natural history of 28. Zhao, Z. et al. Dietary fruit, vegetable, fat, and red and colonoscopy regardless of reflux symptoms. Am. J.
intestinal metaplasia of the gastroesophageal junction processed meat intakes and Barrett’s esophagus risk: Gastroenterol. 101, 12–17 (2006).
and Barrett’s esophagus: a population-based study. a systematic review and meta-analysis. Sci. Rep. 6, 46. Andrici, J., Tio, M., Cox, M. R. & Eslick, G. D.
Am. J. Gastroenterol. 106, 1447–1455 (2011). 27334 (2016). Hiatal hernia and the risk of Barrett’s esophagus.
12. Alcedo, J. et al. Trends in Barrett’s esophagus 29. Keszei, A. P. et al. Meat consumption and the risk J. Gastroenterol. Hepatol. 28, 415–431 (2013).
diagnosis in Southern Europe: implications for of Barrett’s esophagus in a large Dutch cohort. 47. Wang, Z. et al. Helicobacter pylori infection is
surveillance. Dis. Esophagus 22, 239–248 (2009). Cancer Epidemiol. Biomarkers Prev. 22, 1162–1166 associated with reduced risk of Barrett’s esophagus:
13. Kadri, S. R. et al. Acceptability and accuracy of a non- (2013). an analysis of the Barrett’s and Esophageal
endoscopic screening test for Barrett’s oesophagus in 30. Iijima, K. et al. Dietary nitrate generates potentially Adenocarcinoma Consortium. Am. J. Gastroenterol.
primary care: cohort study. BMJ 341, c4372 (2010). mutagenic concentrations of nitric oxide at the 113, 1148–1155 (2018).
This prospective study screens individuals with gastroesophageal junction. Gastroenterology 122, 48. Eross, B. et al. Helicobacter pylori infection reduces
long-term PPI use for BE using the minimally 1248–1257 (2002). the risk of Barrett’s esophagus: a meta-analysis and
invasive Cytosponge. 31. Moriya, A. et al. In vitro studies indicate that systematic review. Helicobacter 23, e12504 (2018).
14. Lagergren, J., Bergstrom, R., Lindgren, A. & Nyren, O. acid catalysed generation of N-nitrosocompounds 49. Chak, A. et al. Familiality in Barrett’s esophagus,
Symptomatic gastroesophageal reflux as a risk factor from dietary nitrate will be maximal at the gastro- adenocarcinoma of the esophagus, and adenocarcinoma
for esophageal adenocarcinoma. N. Engl. J. Med. 340, oesophageal junction and cardia. Scand. J. of the gastroesophageal junction. Cancer Epidemiol.
825–831 (1999). Gastroenterol. 37, 253–261 (2002). Biomarkers Prev. 15, 1668–1673 (2006).
15. Lin, E. C., Holub, J., Lieberman, D. & Hur, C. Low 32. Thrift, A. P. et al. Lifetime alcohol consumption and 50. Verbeek, R. E. et al. Familial clustering of Barrett’s
prevalence of suspected Barrett’s esophagus in patients risk of Barrett’s esophagus. Am. J. Gastroenterol. esophagus and esophageal adenocarcinoma in a
with gastroesophageal reflux disease without alarm 106, 1220–1230 (2011). European cohort. Clin. Gastroenterol. Hepatol. 12,
symptoms. Clin. Gastroenterol. Hepatol. 17, 857–863 33. Kubo, A. et al. Alcohol types and sociodemographic 1656–1663 (2014).
(2018). characteristics as risk factors for Barrett’s esophagus. 51. Thrift, A. P. et al. The use of nonsteroidal anti-
16. Shaheen, N. J., Crosby, M. A., Bozymski, E. M. & Gastroenterology 136, 806–815 (2009). inflammatory drugs and the risk of Barrett’s
Sandler, R. S. Is there publication bias in the reporting 34. Singh, S. et al. Central adiposity is associated with oesophagus. Aliment. Pharmacol. Ther. 34,
of cancer risk in Barrett’s esophagus? Gastroenterology increased risk of esophageal inflammation, metaplasia, 1235–1244 (2011).
119, 333–338 (2000). and adenocarcinoma: a systematic review and meta- 52. Beales, I. L., Dearman, L., Vardi, I. & Loke, Y. Reduced
17. Hvid-Jensen, F., Pedersen, L., Drewes, A. M., analysis. Clin. Gastroenterol. Hepatol. 11, 1399–1412 risk of Barrett’s esophagus in statin users: case-
Sorensen, H. T. & Funch-Jensen, P. Incidence of (2013). control study and meta-analysis. Dig. Dis. Sci. 61,
adenocarcinoma among patients with Barrett’s 35. Kubo, A. et al. Sex-specific associations between 238–246 (2016).
esophagus. N. Engl. J. Med. 365, 1375–1383 (2011). body mass index, waist circumference and the risk 53. Brown, C. S. et al. Predicting progression in Barrett’s
18. Peters, Y. et al. Incidence of progression of persistent of Barrett’s oesophagus: a pooled analysis from the esophagus: development and validation of the
non-dysplastic Barrett’s esophagus to malignancy. International BEACON Consortium. Gut 62, Barrett’s Esophagus Assessment of Risk Score (BEAR
Clin. Gastroenterol. Hepatol. 17, 869–877 (2018). 1684–1691 (2013). Score). Ann. Surg. 267, 716–720 (2018).
0123456789();
PRIMER
54. Gatenby, P., Bhattacharjee, S., Wall, C., Caygill, C. & 74. Slack, J. M. Metaplasia and transdifferentiation: from 98. Sikkema, M. et al. Aneuploidy and overexpression of
Watson, A. Risk stratification for malignant progression pure biology to the clinic. Nature reviews. Mol. Cell Biol. Ki67 and p53 as markers for neoplastic progression
in Barrett’s esophagus: gender, age, duration and 8, 369–378 (2007). in Barrett’s esophagus: a case-control study. Am. J.
year of surveillance. World J. Gastroenterol. 22, 75. Groisman, G. M., Amar, M. & Meir, A. Expression Gastroenterol. 104, 2673–2680 (2009).
10592–10600 (2016). of the intestinal marker Cdx2 in the columnar-lined 99. Souza, R. F., Morales, C. P. & Spechler, S. J. Review
55. Krishnamoorthi, R. et al. Factors associated with esophagus with and without intestinal (Barrett’s) article: a conceptual approach to understanding the
progression of Barrett’s esophagus: a systematic metaplasia. Mod. Pathol. 17, 1282–1288 (2004). molecular mechanisms of cancer development in
review and meta-analysis. Clin. Gastroenterol. Hepatol. 76. Barbera, M. & Fitzgerald, R. C. Cellular origin of Barrett’s oesophagus. Aliment. Pharmacol. Ther. 15,
16, 1046–1055 (2018). Barrett’s metaplasia and oesophageal stem cells. 1087–1100 (2001).
56. Parasa, S. et al. Development and validation of a Biochem. Soc. Trans. 38, 370–373 (2010). 100. Stachler, M. D. et al. Paired exome analysis of
model to determine risk of progression of Barrett’s 77. Li, H. et al. Mechanisms of columnar metaplasia and Barrett’s esophagus and adenocarcinoma. Nat. Genet.
esophagus to neoplasia. Gastroenterology 154, squamous regeneration in experimental Barrett’s 47, 1047–1055 (2015).
1282–1289 (2018). esophagus. Surgery 115, 176–181 (1994). This paper investigates the malignant progression
This study of 2,697 patients with BE develops a 78. Quante, M. et al. Bile acid and inflammation activate of BE to EAC by whole-exome sequencing and
scoring system based on male sex, smoking, length gastric cardia stem cells in a mouse model of shows that the emergence of EAC follows a more
of BE and baseline LGD and identifies patients Barrett-like metaplasia. Cancer Cell 21, 36–51 rapid pathway that involves genome duplication.
with BE at low risk, intermediate risk and high risk (2012). 101. Ross-Innes, C. S. et al. Whole-genome sequencing
of EAC. 79. Kübler, K. et al. Tumor mutational landscape is a provides new insights into the clonal architecture of
57. Bhat, S. et al. Risk of malignant progression in record of the pre-malignant state. Preprint at bioRxiv Barrett’s esophagus and esophageal adenocarcinoma.
Barrett’s esophagus patients: results from a large https://doi.org/10.1101/517565 (2019). Nat. Genet. 47, 1038–1046 (2015).
population-based study. J. Natl Cancer Inst. 103, 80. The Cancer Genome Atlas Research Network. 102. Weaver, J. M. J. et al. Ordering of mutations in
1049–1057 (2011). Integrated genomic characterization of oesophageal preinvasive disease stages of esophageal carcinogenesis.
58. Kelty, C. J., Gough, M. D., Van Wyk, Q., Stephenson, T. J. carcinoma. Nature 541, 169–175 (2017). Nat. Genet. 46, 837–843 (2014).
& Ackroyd, R. Barrett’s oesophagus: intestinal 81. Coad, R. A. et al. On the histogenesis of Barrett’s 103. Frankell, A. M. et al. The landscape of selection in
metaplasia is not essential for cancer risk. Scand. oesophagus and its associated squamous islands: 551 esophageal adenocarcinomas defines genomic
J. Gastroenterol. 42, 1271–1274 (2007). a three-dimensional study of their morphological biomarkers for the clinic. Nat. Genet. 51, 506–516
59. Codipilly, D. C. et al. The effect of endoscopic relationship with native oesophageal gland ducts. (2019).
surveillance in patients with Barrett’s esophagus: a J. Pathol. 206, 388–394 (2005). 104. Martinez, P. et al. Dynamic clonal equilibrium and
systematic review and meta-analysis. Gastroenterology 82. Leedham, S. J. et al. Individual crypt genetic predetermined cancer risk in Barrett’s oesophagus.
154, 2068–2086 (2018). heterogeneity and the origin of metaplastic glandular Nat. Commun. 7, 12158 (2016).
This systematic review and meta-analysis shows epithelium in human Barrett’s oesophagus. Gut 57, 105. Newell, F. et al. Complex structural rearrangements
that surveillance of patients with BE is associated 1041–1048 (2008). are present in high-grade dysplastic Barrett’s
with detection of earlier-stage EAC and a small 83. Kruger, L. et al. Ductular and proliferative response oesophagus samples. BMC Med. Genomics 12, 31
survival benefit. of esophageal submucosal glands in a porcine (2019).
60. Duits, L. C. et al. Patients with Barrett’s esophagus model of esophageal injury and repair. Am. J. Physiol. 106. Nones, K. et al. Genomic catastrophes frequently
and confirmed persistent low-grade dysplasia are Gastrointest. Liver Physiol. 313, G180–G191 (2017). arise in esophageal adenocarcinoma and drive
at increased risk for progression to neoplasia. 84. von Furstenberg, R. J. et al. Porcine esophageal tumorigenesis. Nat. Commun. 5, 5224 (2014).
Gastroenterology 152, 993–1001 (2017). submucosal gland culture model shows capacity 107. Moayyedi, P. M. et al. ACG and CAG clinical guideline:
61. Kestens, C., Offerhaus, G. J., van Baal, J. W. & for proliferation and differentiation. Cell. Mol. management of dyspepsia. Am. J. Gastroenterol. 112,
Siersema, P. D. Patients with Barrett’s esophagus and Gastroenterol. Hepatol. 4, 385–404 (2017). 988–1013 (2017).
persistent low-grade dysplasia have an increased risk 85. Owen, R. P. et al. Single cell RNA-seq reveals profound 108. Muthusamy, V. R. et al. The role of endoscopy in the
for high-grade dysplasia and cancer. Clin. Gastroenterol. transcriptional similarity between Barrett’s oesophagus management of GERD. Gastrointest. Endosc. 81,
Hepatol. 14, 956–962 (2016). and oesophageal submucosal glands. Nat. Commun. 9, 1305–1310 (2015).
62. Rastogi, A. et al. Incidence of esophageal 4261 (2018). 109. Weusten, B. et al. Endoscopic management of Barrett’s
adenocarcinoma in patients with Barrett’s esophagus 86. Wang, X. et al. Residual embryonic cells as precursors esophagus: European Society of Gastrointestinal
and high-grade dysplasia: a meta-analysis. of a Barrett’s-like metaplasia. Cell 145, 1023–1035 Endoscopy (ESGE) position statement. Endoscopy 49,
Gastrointest. Endosc. 67, 394–398 (2008). (2011). 191–198 (2017).
63. Shaheen, N. J. et al. Radiofrequency ablation in 87. Seery, J. P. Stem cells of the oesophageal epithelium. 110. Iwakiri, K. et al. Evidence-based clinical practice
Barrett’s esophagus with dysplasia. N. Engl. J. Med. J. Cell Sci. 115, 1783–1789 (2002). guidelines for gastroesophageal reflux disease 2015.
360, 2277–2288 (2009). 88. Hutchinson, L. et al. Human Barrett’s adenocarcinoma J. Gastroenterol. 51, 751–767 (2016).
This sham-controlled randomized trial analyses of the esophagus, associated myofibroblasts, and 111. Whiteman, D. C. et al. Australian clinical practice
whether endoscopic RFA could eradicate dysplastic endothelium can arise from bone marrow-derived cells guidelines for the diagnosis and management
BE and decrease the rate of malignant progression. after allogeneic stem cell transplant. Stem Cells Dev. of Barrett’s esophagus and early esophageal
64. Overholt, B. F. et al. Photodynamic therapy with 20, 11–17 (2011). adenocarcinoma. J. Gastroenterol. Hepatol. 30,
porfimer sodium for ablation of high-grade dysplasia 89. Kapoor, H., Lohani, K. R., Lee, T. H., Agrawal, D. K. 804–820 (2015).
in Barrett’s esophagus: international, partially & Mittal, S. K. Animal models of Barrett’s esophagus 112. Pohl, H. et al. Length of Barrett’s oesophagus and
blinded, randomized phase III trial. Gastrointest. and esophageal adenocarcinoma-past, present, and cancer risk: implications from a large sample of
Endosc. 62, 488–498 (2005). future. Clin. Transl Sci. 8, 841–847 (2015). patients with early oesophageal adenocarcinoma.
65. Souza, R. F. Reflux esophagitis and its role in the 90. Dong, J. et al. Interactions between genetic variants Gut 65, 196–201 (2016).
pathogenesis of Barrett’s metaplasia. J. Gastroenterol. and environmental factors affect risk of esophageal 113. Thota, P. N. et al. Low risk of high-grade dysplasia or
52, 767–776 (2017). adenocarcinoma and Barrett’s esophagus. Clin. esophageal adenocarcinoma among patients with
66. Glickman, J. N. et al. Mucin core polypeptide Gastroenterol. Hepatol. 16, 1598–1606 (2018). Barrett’s esophagus less than 1cm (irregular Z line)
expression in the progression of neoplasia in Barrett’s 91. Dong, J. et al. Determining risk of Barrett’s within 5 years of index endoscopy. Gastroenterology
esophagus. Hum. Pathol. 37, 1304–1315 (2006). esophagus and esophageal adenocarcinoma based 152, 987–992 (2017).
67. Lavery, D. L. et al. The stem cell organisation, and the on epidemiologic factors and genetic variants. 114. Sharma, P. et al. Quality indicators for the
proliferative and gene expression profile of Barrett’s Gastroenterology 154, 1273–1281 (2018). management of Barrett’s esophagus, dysplasia, and
epithelium, replicates pyloric-type gastric glands. 92. Sun, X. et al. Genomic regions associated with esophageal adenocarcinoma: international consensus
Gut 63, 1854–1863 (2014). susceptibility to Barrett’s esophagus and esophageal recommendations from the American
68. McQuaid, K. R., Laine, L., Fennerty, M. B., Souza, R. adenocarcinoma in African Americans: the cross Gastroenterological Association Symposium.
& Spechler, S. J. Systematic review: the role of bile BETRNet admixture study. PLOS ONE 12, e0184962 Gastroenterology 149, 1599–1606 (2015).
acids in the pathogenesis of gastro-oesophageal reflux (2017). 115. Sharma, P. et al. The development and validation of
disease and related neoplasia. Aliment. Pharmacol. 93. Ek, W. E. et al. Germline genetic contributions to risk an endoscopic grading system for Barrett’s esophagus:
Ther. 34, 146–165 (2011). for esophageal adenocarcinoma, Barrett’s esophagus, the Prague C & M criteria. Gastroenterology 131,
69. Mari, L. et al. A pSMAD/CDX2 complex is essential for and gastroesophageal reflux. J. Natl Cancer Inst. 105, 1392–1399 (2006).
the intestinalization of epithelial metaplasia. Cell Rep. 1711–1718 (2013). 116. Antony, A. et al. Adherence to quality indicators in
7, 1197–1210 (2014). 94. Gharahkhani, P. et al. Genome-wide association endoscopic surveillance of Barrett’s esophagus and
70. Wang, D. H. et al. Aberrant epithelial-mesenchymal studies in oesophageal adenocarcinoma and Barrett’s correlation to dysplasia detection rates. Clin. Res.
Hedgehog signaling characterizes Barrett’s metaplasia. oesophagus: a large-scale meta-analysis. Lancet. Oncol. Hepatol. Gastroenterol. 42, 591–596 (2018).
Gastroenterology 138, 1810–1822 (2010). 17, 1363–1373 (2016). 117. Ooi, J. et al. Dedicated Barrett’s surveillance sessions
71. Jiang, M. et al. Transitional basal cells at the squamous- This meta-analysis of all GWASs of BE and EAC managed by trained endoscopists improve dysplasia
columnar junction generate Barrett’s oesophagus. investigates novel genetic risk variants for BE detection rate. Endoscopy 49, C1 (2017).
Nature 550, 529–533 (2017). development and malignant progression. 118. Harrison, R. et al. Detection of intestinal metaplasia in
72. Kong, J., Crissey, M. A., Funakoshi, S., Kreindler, J. L. 95. Buas, M. F. et al. Germline variation in inflammation- Barrett’s esophagus: an observational comparator study
& Lynch, J. P. Ectopic Cdx2 expression in murine related pathways and risk of Barrett’s oesophagus and suggests the need for a minimum of eight biopsies.
esophagus models an intermediate stage in the oesophageal adenocarcinoma. Gut 66, 1739–1747 Am. J. Gastroenterol. 102, 1154–1161 (2007).
emergence of Barrett’s esophagus. PLOS ONE 6, (2017). 119. Hirota, W. K. et al. Specialized intestinal metaplasia,
e18280 (2011). 96. Dai, J. Y. et al. A newly identified susceptibility locus near dysplasia, and cancer of the esophagus and
73. Vega, M. E. et al. Inhibition of Notch signaling FOXP1 modifies the association of gastroesophageal esophagogastric junction: prevalence and clinical
enhances transdifferentiation of the esophageal reflux with Barrett’s esophagus. Cancer Epidemiol. data. Gastroenterology 116, 277–285 (1999).
squamous epithelium towards a Barrett’s-like Biomarkers Prev. 24, 1739–1747 (2015). 120. Schlemper, R. J. et al. The Vienna classification of
metaplasia via KLF4. Cell Cycle 13, 3857–3866 97. Hanahan, D. & Weinberg, R. A. Hallmarks of cancer: gastrointestinal epithelial neoplasia. Gut 47, 251–255
(2014). the next generation. Cell 144, 646–674 (2011). (2000).
0123456789();
PRIMER
121. Vennalaganti, P. et al. Discordance among pathologists 142. Liao, L. M. et al. Nonsteroidal anti-inflammatory drug 162. Bhardwaj, A., Hollenbeak, C. S., Pooran, N. &
in the United States and Europe in diagnosis of low- use reduces risk of adenocarcinomas of the esophagus Mathew, A. A meta-analysis of the diagnostic accuracy of
grade dysplasia for patients with Barrett’s esophagus. and esophagogastric junction in a pooled analysis. esophageal capsule endoscopy for Barrett’s esophagus
Gastroenterology 152, 564–570 (2017). Gastroenterology 142, 442–452 (2012). in patients with gastroesophageal reflux disease. Am. J.
122. Abela, J. E. et al. Systematic four-quadrant biopsy 143. Zhang, S. et al. Cyclooxygenase inhibitors use is Gastroenterol. 104, 1533–1539 (2009).
detects Barrett’s dysplasia in more patients than associated with reduced risk of esophageal 163. Offman, J. et al. Barrett’s oESophagus trial 3 (BEST3):
nonsystematic biopsy. Am. J. Gastroenterol. 103, adenocarcinoma in patients with Barrett’s esophagus: study protocol for a randomised controlled trial
850–855 (2008). a meta-analysis. Br. J. Cancer 110, 2378–2388 comparing the Cytosponge-TFF3 test with usual care
123. Vennalaganti, P. R. et al. Increased detection of Barrett’s (2014). to facilitate the diagnosis of oesophageal pre-cancer
esophagus-associated neoplasia using wide-area 144. Ogunwobi, O. O. & Beales, I. L. Statins inhibit in primary care patients with chronic acid reflux. BMC
trans-epithelial sampling: a multicenter, prospective, proliferation and induce apoptosis in Barrett’s Cancer 18, 784 (2018).
randomized trial. Gastrointest. Endosc. 87, 348–355 esophageal adenocarcinoma cells. Am. J. Gastroenterol. 164. Iyer, P. G. et al. Highly discriminant methylated DNA
(2018). 103, 825–837 (2008). markers for the non-endoscopic detection of Barrett’s
124. Sharma, P. et al. Standard endoscopy with random 145. Nguyen, T., Khalaf, N., Ramsey, D. & El-Serag, H. B. esophagus. Am. J. Gastroenterol. 113, 1156–1166
biopsies versus narrow band imaging targeted biopsies Statin use is associated with a decreased risk of (2018).
in Barrett’s oesophagus: a prospective, international, Barrett’s esophagus. Gastroenterology 147, 314–323 165. Moinova, H. R. et al. Identifying DNA methylation
randomised controlled trial. Gut 62, 15–21 (2013). (2014). biomarkers for non-endoscopic detection of Barrett’s
125. Waterhouse, D. J., Fitzpatrick, C. R. M., di Pietro, M. 146. Nguyen, T., Duan, Z., Naik, A. D., Kramer, J. R. & esophagus. Sci. Transl Med. 10, eaao5848 (2018).
& Bohndiek, S. E. Emerging optical methods for El-Serag, H. B. Statin use reduces risk of esophageal 166. Chettouh, H. et al. Methylation panel is a diagnostic
endoscopic surveillance of Barrett’s oesophagus. adenocarcinoma in US veterans with Barrett’s biomarker for Barrett’s oesophagus in endoscopic
Lancet Gastroenterol. Hepatol. 3, 349–362 (2018). esophagus: a nested case-control study. biopsies and non-endoscopic cytology specimens.
126. Visrodia, K. et al. Magnitude of missed esophageal Gastroenterology 149, 1392–1398 (2015). Gut 67, 1942–1949 (2018).
adenocarcinoma after Barrett’s esophagus diagnosis: 147. Siegel, R., Naishadham, D. & Jemal, A. Cancer 167. Li, X. et al. Selection and application of tissue
a systematic review and meta-analysis. Gastroenterology statistics, 2013. CA Cancer J. Clin. 63, 11–30 (2013). microRNAs for nonendoscopic diagnosis of Barrett’s
150, 599–607 (2016). 148. Verbeek, R. E. et al. Surveillance of Barrett’s esophagus esophagus. Gastroenterology 155, 771–783 (2018).
127. World Cancer Research Fund/American Institute for and mortality from esophageal adenocarcinoma: 168. Chan, D. K. et al. Breath testing for Barrett’s esophagus
Cancer Research. Diet, Nutrition, Physical Activity a population-based cohort study. Am. J. Gastroenterol. using exhaled volatile organic compound profiling
and Cancer: a Global Perspective, Continuous Update 109, 1215–1222 (2014). with an electronic nose device. Gastroenterology 152,
Project Expert Report (World Cancer Research Fund 149. Sampliner, R. E. Practice guidelines on the diagnosis, 24–26 (2017).
International, 2018). surveillance, and therapy of Barrett’s esophagus. 169. Honing, J. et al. Endosheath ultrathin transnasal
128. Lam, S., Alexandre, L., Luben, R. & Hart, A. R. The Practice Parameters Committee of the American endoscopy is a cost-effective method for screening for
The association between physical activity and the College of Gastroenterology. Am. J. Gastroenterol. 93, Barrett’s esophagus in patients with GERD symptoms.
risk of symptomatic Barrett’s oesophagus: a UK 1028–1032 (1998). Gastrointest. Endosc. 89, 712–722 (2018).
prospective cohort study. Eur. J. Gastroenterol. 150. Stoltey, J., Reeba, H., Ullah, N., Sabhaie, P. & Gerson, L. 170. Nietert, P. J. et al. Cost-effectiveness of screening a
Hepatol. 30, 71–75 (2018). Does Barrett’s oesophagus develop over time in population with chronic gastroesophageal reflux.
129. Kunzmann, A. T. et al. Physical activity, sedentary patients with chronic gastro-oesophageal reflux Gastrointest. Endosc. 57, 311–318 (2003).
behaviour and risk of oesophago-gastric cancer: disease? Aliment. Pharmacol. Ther. 25, 83–91 (2007). 171. Inadomi, J. M. et al. Screening and surveillance for
a prospective cohort study within UK Biobank. United 151. Rubenstein, J. H. et al. Prediction of Barrett’s Barrett esophagus in high-risk groups: a cost-utility
European Gastroenterol. J. 6, 1144–1154 (2018). esophagus among men. Am. J. Gastroenterol. 108, analysis. Ann. Intern. Med. 138, 176–186 (2003).
130. Realdon, S. et al. Adherence to WCRF/AICR lifestyle 353–362 (2013). 172. Gerson, L. B., Groeneveld, P. W. & Triadafilopoulos, G.
recommendations for cancer prevention and the risk of This article presents a study in which a prediction Cost-effectiveness model of endoscopic screening and
Barrett’s esophagus onset and evolution to esophageal model for BE based on GERD, age, abdominal surveillance in patients with gastroesophageal reflux
adenocarcinoma: results from a pilot study in a high-risk obesity and cigarette use with an AUC of 0.72 is disease. Clin. Gastroenterol. Hepatol. 2, 868–879
population. Eur. J. Nutr. 55, 1563–1571 (2016). developed in 822 male patients with colorectal (2004).
131. Mulholland, H. G., Murray, L. J., Anderson, L. A., cancer who are screened using upper endoscopy. 173. Heberle, C. R. et al. Cost effectiveness of screening
Cantwell, M. M. & FINBAR study group. Vitamin D, 152. Thrift, A. P., Vaughan, T. L., Anderson, L. A., patients with gastroesophageal reflux disease for
calcium and dairy intake, and risk of oesophageal Whiteman, D. C. & El-Serag, H. B. External validation Barrett’s esophagus with a minimally invasive cell
adenocarcinoma and its precursor conditions. of the Michigan Barrett’s esophagus prediction tool. sampling device. Clin. Gastroenterol. Hepatol. 15,
Br. J. Nutr. 106, 732–741 (2011). Clin. Gastroenterol. Hepatol. 15, 1124–1126 (2017). 1397–1404 (2017).
132. Dai, Q. et al. Dietary magnesium, calcium:magnesium 153. Thrift, A. P., Garcia, J. M. & El-Serag, H. B. A 174. Wani, S. et al. Endoscopic eradication therapy for
ratio and risk of reflux oesophagitis, Barrett’s multibiomarker risk score helps predict risk for patients with Barrett’s esophagus-associated dysplasia
oesophagus and oesophageal adenocarcinoma: Barrett’s esophagus. Clin. Gastroenterol. Hepatol. 12, and intramucosal cancer. Gastrointest. Endosc. 87,
a population-based case-control study. Br. J. Nutr. 1267–1271 (2014). 907–931 (2018).
115, 342–350 (2016). 154. Spechler, S. J., Katzka, D. A. & Fitzgerald, R. C. New 175. Corley, D. A. et al. Impact of endoscopic surveillance
133. Garland, C. F. et al. The role of vitamin D in cancer screening techniques in Barrett’s esophagus: great on mortality from Barrett’s esophagus-associated
prevention. Am. J. Public Health 96, 252–261 (2006). ideas or great practice? Gastroenterology 154, esophageal adenocarcinomas. Gastroenterology 145,
134. Dibaba, D. T., Xun, P. & He, K. Dietary magnesium 1594–1601 (2018). 312–319 (2013).
intake is inversely associated with serum C-reactive 155. Sami, S. S. & Iyer, P. G. Recent advances in screening 176. Holmberg, D., Ness-Jensen, E., Mattsson, F. &
protein levels: meta-analysis and systematic review. for Barrett’s esophagus. Curr. Treat. Options Lagergren, J. Clinical prediction model for tumor
Eur. J. Clin. Nutr. 68, 510–516 (2014). Gastroenterol. 16, 1–14 (2018). progression in Barrett’s esophagus. Surg. Endosc.
135. Vaezi, M. F., Yang, Y. X. & Howden, C. W. Complications 156. Sami, S. S. et al. Performance characteristics https://doi.org/10.1007/s00464-018-6590-5
of proton pump inhibitor therapy. Gastroenterology of unsedated ultrathin video endoscopy in the (2018).
153, 35–48 (2017). assessment of the upper GI tract: systematic review 177. Peters, F. P. et al. Histologic evaluation of resection
136. Attwood, S. E. et al. Long-term safety of proton and meta-analysis. Gastrointest. Endosc. 82, specimens obtained at 293 endoscopic resections
pump inhibitor therapy assessed under controlled, 782–792 (2015). in Barrett’s esophagus. Gastrointest. Endosc. 67,
randomised clinical trial conditions: data from the 157. Sami, S. S. et al. A randomized comparative 604–609 (2008).
SOPRAN and LOTUS studies. Aliment. Pharmacol. effectiveness trial of novel endoscopic techniques and 178. Markar, S. R. et al. The influence of procedural
Ther. 41, 1162–1174 (2015). approaches for Barrett’s esophagus screening in the volume and proficiency gain on mortality from upper
137. O’Sullivan, K. E. et al. The role of inflammation in community. Am. J. Gastroenterol. 110, 148–158 GI endoscopic mucosal resection. Gut 67, 79–85
cancer of the esophagus. Expert Rev. Gastroenterol. (2015). (2018).
Hepatol. 8, 749–760 (2014). This randomized controlled trial shows that mobile 179. Pech, O. et al. Long-term efficacy and safety of
138. Ouatu-Lascar, R., Fitzgerald, R. C. & Triadafilopoulos, G. and outpatient unsedated TNE may be an effective endoscopic resection for patients with mucosal
Differentiation and proliferation in Barrett’s esophagus alternative to standard endoscopy to screen for BE. adenocarcinoma of the esophagus. Gastroenterology
and the effects of acid suppression. Gastroenterology 158. Moriarty, J. P. et al. Costs associated with Barrett’s 146, 652–660 (2014).
117, 327–335 (1999). esophagus screening in the community: an economic 180. Subramaniam, S. et al. Complex early Barrett’s
139. Abu-Sneineh, A. et al. The effects of high-dose analysis of a prospective randomized controlled trial neoplasia at 3 Western centers: European Barrett’s
esomeprazole on gastric and oesophageal acid of sedated versus hospital unsedated versus mobile Endoscopic Submucosal Dissection Trial (E-BEST).
exposure and molecular markers in Barrett’s community unsedated endoscopy. Gastrointest. Gastrointest. Endosc. 86, 608–618 (2017).
oesophagus. Aliment. Pharmacol. Ther. 32, Endosc. 87, 88–94 (2018). 181. Pimentel-Nunes, P. et al. Endoscopic submucosal
1023–1030 (2010). 159. Sami, S. S. et al. Acceptability, accuracy, and safety of dissection: European Society of Gastrointestinal
140. Singh, S., Garg, S. K., Singh, P. P., Iyer, P. G. & disposable transnasal capsule endoscopy for Barrett’s Endoscopy (ESGE) Guideline. Endoscopy 47, 829–854
El-Serag, H. B. Acid-suppressive medications and esophagus screening. Clin. Gastroenterol. Hepatol. (2015).
risk of oesophageal adenocarcinoma in patients 17, 638–646 (2018). 182. Terheggen, G. et al. A randomised trial of endoscopic
with Barrett’s oesophagus: a systematic review and 160. Alashkar, B. et al. Development of a program to submucosal dissection versus endoscopic mucosal
meta-analysis. Gut 63, 1229–1237 (2014). train physician extenders to perform transnasal resection for early Barrett’s neoplasia. Gut 66,
141. Jankowski, J. A. Z. et al. Esomeprazole and aspirin in esophagoscopy and screen for Barrett’s esophagus. 783–793 (2017).
Barrett’s oesophagus (AspECT): a randomised factorial Clin. Gastroenterol. Hepatol. 12, 785–792 (2014). 183. Yang, D. et al. Endoscopic submucosal dissection for
trial. Lancet 392, 400–408 (2018). 161. Sharma, P. et al. The diagnostic accuracy of esophageal early Barrett’s neoplasia: a meta-analysis. Gastrointest.
This randomized trial of 2,557 patients with BE capsule endoscopy in patients with gastroesophageal Endosc. 87, 1383–1393 (2018).
shows that the combination of high-dose PPI and reflux disease and Barrett’s esophagus: a blinded, 184. Tomizawa, Y. et al. Safety of endoscopic mucosal
aspirin is the best strategy to reduce the risk of prospective study. Am. J. Gastroenterol. 103, 525–532 resection for Barrett’s esophagus. Am. J. Gastroenterol.
EAC and HGD. (2008). 108, 1440–1447 (2013).
0123456789();
PRIMER
185. Shaheen, N. J. et al. Durability of radiofrequency esophagus is related to gender and manifestation review and meta-analysis. J. Gastroenterol. Hepatol.
ablation in Barrett’s esophagus with dysplasia. of disease. Am. J. Gastroenterol. 104, 2695–2703 28, 1258–1273 (2013).
Gastroenterology 141, 460–468 (2011). (2009). 230. Anderson, L. A. et al. The association between alcohol
186. Phoa, K. N. et al. Radiofrequency ablation versus 207. Rodger, A. J., Jolley, D., Thompson, S. C., Lanigan, A. and reflux esophagitis, Barrett’s esophagus, and
endoscopic surveillance for patients with Barrett & Crofts, N. The impact of diagnosis of hepatitis C virus esophageal adenocarcinoma. Gastroenterology 136,
esophagus and low-grade dysplasia: a randomized on quality of life. Hepatology 30, 1299–1301 (1999). 799–805 (2009).
clinical trial. JAMA 311, 1209–1217 (2014). 208. Rubenstein, J. H. & Inadomi, J. M. Defining a clinically 231. Rokkas, T., Pistiolas, D., Sechopoulos, P., Robotis, I. &
187. Manner, H. et al. Ablation of residual Barrett’s significant adverse impact of diagnosing Barrett’s Margantinis, G. Relationship between Helicobacter
epithelium after endoscopic resection: a randomized esophagus. J. Clin. Gastroenterol. 40, 109–115 (2006). pylori infection and esophageal neoplasia: a meta-
long-term follow-up study of argon plasma coagulation 209. Essink-Bot, M. L. et al. Different perceptions of the analysis. Clin. Gastroenterol. Hepatol. 5, 1413–1417
versus surveillance (APE study). Endoscopy 46, 6–12 burden of upper GI endoscopy: an empirical study in (2007).
(2014). three patient groups. Qual. Life Res. 16, 1309–1318 232. Salomao, M. A., Lam-Himlin, D. & Pai, R. K.
188. Thota, P. N. et al. Cryotherapy and radiofrequency (2007). Substantial interobserver agreement in the diagnosis
ablation for eradication of Barrett’s esophagus with 210. Cooper, S. C. et al. Endoscopic surveillance for of dysplasia in Barrett esophagus upon review of a
dysplasia or intramucosal cancer. Dig. Dis. Sci. 63, Barrett’s oesophagus: the patients’ perspective. patient’s entire set of biopsies. Am. J. Surg. Pathol.
1311–1319 (2018). Eur. J. Gastroenterol. Hepatol. 21, 850–854 (2009). 42, 376–381 (2018).
189. Dumot, J. A. et al. An open-label, prospective trial of 211. Levine, D. S., Blount, P. L., Rudolph, R. E. & Reid, B. J. 233. Montgomery, E. et al. Reproducibility of the diagnosis
cryospray ablation for Barrett’s esophagus high-grade Safety of a systematic endoscopic biopsy protocol in of dysplasia in Barrett esophagus: a reaffirmation.
dysplasia and early esophageal cancer in high-risk patients with Barrett’s esophagus. Am. J. Gastroenterol. Hum. Pathol. 32, 368–378 (2001).
patients. Gastrointest. Endosc. 70, 635–644 (2009). 95, 1152–1157 (2000). 234. Desai, T. K. et al. The incidence of oesophageal
190. Milashka, M. et al. Sixteen-year follow-up of Barrett’s 212. Kruijshaar, M. E. et al. The burden of upper adenocarcinoma in non-dysplastic Barrett’s
esophagus, endoscopically treated with argon plasma gastrointestinal endoscopy in patients with Barrett’s oesophagus: a meta-analysis. Gut 61, 970–976 (2012).
coagulation. United European Gastroenterol. J. 2, esophagus. Endoscopy 38, 873–878 (2006). 235. Sonwalkar, S. A. et al. A study of indefinite for
367–373 (2014). 213. Gerson, L. B., Ullah, N., Hastie, T. & Goldstein, M. K. dysplasia in Barrett’s oesophagus: reproducibility of
191. Manner, H. et al. The tissue effect of argon-plasma Does cancer risk affect health-related quality of life diagnosis, clinical outcomes and predicting progression
coagulation with prior submucosal injection in patients with Barrett’s esophagus? Gastrointest. with AMACR (alpha-methylacyl-CoA-racemase).
(Hybrid-APC) versus standard APC: a randomized Endosc. 65, 16–25 (2007). Histopathology 56, 900–907 (2010).
ex-vivo study. United European Gastroenterol. J. 2, 214. Macrae, F. A., Hill, D. J., St John, D. J., Ambikapathy, A. 236. Kestens, C., Leenders, M., Offerhaus, G. J.,
383–390 (2014). & Garner, J. F. Predicting colon cancer screening van Baal, J. W. & Siersema, P. D. Risk of neoplastic
192. Orman, E. S., Li, N. & Shaheen, N. J. Efficacy and behavior from health beliefs. Prev. Med. 13, 115–126 progression in Barrett’s esophagus diagnosed as
durability of radiofrequency ablation for Barrett’s (1984). indefinite for dysplasia: a nationwide cohort study.
Esophagus: systematic review and meta-analysis. 215. Kruijshaar, M. E. et al. Patients with Barrett’s Endoscopy 47, 409–414 (2015).
Clin. Gastroenterol. Hepatol. 11, 1245–1255 (2013). esophagus perceive their risk of developing esophageal 237. Singh, S. et al. Incidence of esophageal adenocarcinoma
193. Anders, M. et al. Subsquamous extension of intestinal adenocarcinoma as low. Gastrointest. Endosc. 65, in Barrett’s esophagus with low-grade dysplasia:
metaplasia is detected in 98% of cases of neoplastic 26–30 (2007). a systematic review and meta-analysis. Gastrointest.
Barrett’s esophagus. Clin. Gastroenterol. Hepatol. 12, 216. van der Ende-van Loon, M. C. M., Rosmolen, W. D., Endosc. 79, 897–909 (2014).
405–410 (2014). Houterman, S., Schoon, E. J. & Curvers, W. L. Cancer 238. Grotenhuis, B. A. et al. Inter- and intraobserver
194. Gray, N. A., Odze, R. D. & Spechler, S. J. Buried risk perception in relation to associated symptoms in variation in the histopathological evaluation of early
metaplasia after endoscopic ablation of Barrett’s Barrett’s patients: a cross sectional study on quality of oesophageal adenocarcinoma. J. Clin. Pathol. 63,
esophagus: a systematic review. Am. J. Gastroenterol. life. United European Gastroenterol. J. 6, 1316–1322 978–981 (2010).
106, 1899–1908 (2011). (2018). 239. Kerkhof, M. et al. Grading of dysplasia in Barrett’s
195. Belghazi, K. et al. Long-term follow-up results of 217. Shaheen, N. J. et al. The perception of cancer risk in oesophagus: substantial interobserver variation
stepwise radical endoscopic resection for Barrett’s patients with prevalent Barrett’s esophagus enrolled in between general and gastrointestinal pathologists.
esophagus with early neoplasia. Gastrointest. Endosc. an endoscopic surveillance program. Gastroenterology Histopathology 50, 920–927 (2007).
87, 77–84 (2018). 129, 429–436 (2005).
196. Desai, M. et al. Efficacy and safety outcomes of 218. Stier, M. W. et al. Perceptions of risk and therapy Acknowledgements
multimodal endoscopic eradication therapy in among patients with Barrett’s esophagus: a patient The authors thank R. S. van der Post and M. O’Donovan for
Barrett’s esophagus-related neoplasia: a systematic survey study. Dis. Esophagus 31, dox109 (2018). providing the histopathological images.
review and pooled analysis. Gastrointest. Endosc. 85, 219. Smyth, E. C. et al. Oesophageal cancer. Nat. Rev. Dis.
482–495 (2017). Primers 3, 17048 (2017). Author contributions
197. Cho, J. W. et al. Lymph node metastases in esophageal 220. Shaheen, N. J. et al. Quality of life following Introduction (Y.P. and A.A.-K.); Epidemiology (Y.P. and N.J.S.);
carcinoma: an endoscopist’s view. Clin. Endosc. 47, radiofrequency ablation of dysplastic Barrett’s Mechanisms/pathophysiology (A.A.-K., A.C., A.B. and R.F.S.);
523–529 (2014). esophagus. Endoscopy 42, 790–799 (2010). Diagnosis, screening and prevention (Y.P., M.D.P., P.G.I. and
198. Manner, H. et al. Early Barrett’s carcinoma with 221. Rosmolen, W. D. et al. Quality of life and fear of R.C.F.); Management (Y.P., A.A.-K. and O.P.); Quality of life
“low-risk” submucosal invasion: long-term results of cancer recurrence after endoscopic treatment for early (Y.P.); Outlook (Y.P., A.A.-K. and P.D.S.); Overview of Primer
endoscopic resection with a curative intent. Am. J. Barrett’s neoplasia: a prospective study. Dis. Esophagus (Y.P., A.A.-K. and P.D.S.).
Gastroenterol. 103, 2589–2597 (2008). 30, 1–9 (2017).
199. Manner, H. et al. Efficacy, safety, and long-term results 222. Rosmolen, W. D. et al. Quality of life and fear of cancer Competing interests
of endoscopic treatment for early stage adenocarcinoma recurrence after endoscopic and surgical treatment for N.J.S. has received research grants from Medtronic, CSA
of the esophagus with low-risk sm1 invasion. Clin. early neoplasia in Barrett’s esophagus. Endoscopy 42, Medical, C2 Therapeutics, CDx Medical, EndoStim and
Gastroenterol. Hepatol. 11, 630–635 (2013). 525–531 (2010). Interpace Diagnostics and has served as a consultant for
200. Cotton, C. C., Haidry, R., Thrift, A. P., Lovat, L. & 223. Haidry, R. J. et al. Improvement over time in outcomes Pfizer, Boston Scientific and Shire. A.C. has equity interest in
Shaheen, N. J. Development of evidence-based for patients undergoing endoscopic therapy for and is a consultant for Lucid Diagnostics. R.F.S. has served as
surveillance intervals after radiofrequency ablation Barrett’s oesophagus-related neoplasia: 6-year a consultant and receives research support from Ironwood
of Barrett’s esophagus. Gastroenterology 155, experience from the first 500 patients treated in the Pharmaceuticals. M.D.P. has received educational grants
316–326 (2018). UK patient registry. Gut 64, 1192–1199 (2015). from Olympus and Medtronic and has served as consultant
This article provides evidence-based surveillance 224. Nguyen, T., Thrift, A. P., Yu, X., Duan, Z. & for Medtronic. P.G.I. has received research funding from Exact
intervals for patients with endoscopically El-Serag, H. B. The annual risk of esophageal Sciences, Pentax Medical, Symple Surgical, Nine Point
treated BE. adenocarcinoma does not decrease over time in Medical and Medtronic and has served as a consultant for
201. Britton, J. et al. Effect of diagnosis, surveillance, and patients with Barrett’s esophagus. Am. J. Gastroenterol. Medtronic, Symple Surgical and CSA Medical. O.P. has a
treatment of Barrett’s oesophagus on health-related 112, 1049–1055 (2017). speaker honorarium from Olympus, Fujifilm, Medtronic,
quality of life. Lancet Gastroenterol. Hepatol. 3, 225. Old, O. et al. Barrett’s Oesophagus Surveillance Boston Scientific and Cook. R.C.F. is named on patents
57–65 (2018). versus endoscopy at need Study (BOSS): protocol and related to the Cytosponge and associated assays, which have
This paper presents a comprehensive review of the analysis plan for a multicentre randomized controlled been licensed to Covidien (now Medtronic), and has served
QOL in patients with BE. trial. J. Med. Screen. 22, 158–164 (2015). as a consultant to Medtronic. P.D.S. has received research
202. Kamolz, T., Pointner, R. & Velanovich, V. The impact This article presents a study protocol for a funding from Pentax Medical, EndoStim, Yakult, OncoDNA
of gastroesophageal reflux disease on quality of life. randomized study that will provide data to evaluate and Ella- CS. Y.P., A.A.-K. and A.B. declare no competing
Surg. Endosc. 17, 1193–1199 (2003). the efficacy and cost-effectiveness of screening interests.
203. Ofman, J. J. The economic and quality-of-life impact of patients with BE for the presence of EAC.
symptomatic gastroesophageal reflux disease. Am. J. 226. Visrodia, K. et al. Cryotherapy for persistent Barrett’s Publisher’s note
Gastroenterol. 98, S8–S14 (2003). esophagus after radiofrequency ablation: a systematic Springer Nature remains neutral with regard to jurisdictional
204. Eloubeidi, M. A. & Provenzale, D. Health-related review and meta-analysis. Gastrointest. Endosc. 87, claims in published maps and institutional affiliations.
quality of life and severity of symptoms in patients 1396–1404 (2018).
with Barrett’s esophagus and gastroesophageal reflux 227. Kazumori, H., Ishihara, S., Takahashi, Y., Amano, Y. Reviewer information
disease patients without Barrett’s esophagus. Am. J. & Kinoshita, Y. Roles of Kruppel-like factor 4 in Nature Reviews Disease Primers thanks S. McDonald,
Gastroenterol. 95, 1881–1887 (2000). oesophageal epithelial cells in Barrett’s epithelium A. Bansal, M. Jansen, R. Odze, K. Krishnadath, H. Barr and
205. Kulig, M. et al. Quality of life in relation to symptoms development. Gut 60, 608–617 (2011). the other anonymous reviewer(s), for their contribution to the
in patients with gastro-oesophageal reflux disease — 228. Wang, D. H. et al. Hedgehog signaling regulates peer review of this work.
an analysis based on the ProGERD initiative. Aliment. FOXA2 in esophageal embryogenesis and Barrett’s
Pharmacol. Ther. 18, 767–776 (2003). metaplasia. J. Clin. Invest. 124, 3767–3780 (2014). Supplementary information
206. Lippmann, Q. K., Crockett, S. D., Dellon, E. S. & 229. Andrici, J., Cox, M. R. & Eslick, G. D. Cigarette smoking Supplementary information is available for this paper at
Shaheen, N. J. Quality of life in GERD and Barrett’s and the risk of Barrett’s esophagus: a systematic https://doi.org/10.1038/s41572-019-0086-z.
0123456789();

Esofago de Barret

Uploaded by

Copyright:

Available Formats

Esofago de Barret

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Esofago de Barret

Uploaded by

Copyright:

Available Formats

PRIMER

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 1

Author addresses descent20,21. The effect of Hispanic ethnicity is less clear,

2 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 3

4 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

a Oesophageal stratiﬁed Intestinal metaplasia with Squamocolumnar Gastric columnar

Squamous Stem cell Columnar cell

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 5

6 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

Genome duplication Kataegis

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 7

8 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

a seem to be the key elements for maximizing detection

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 9

Table 2 | Histopathological features of BE

10 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

Table 3 | Criteria for diagnosis of BE

above the GEJ into the tubular oesophagus

light-grey appearance of the oesophageal

metaplastic columnar epithelium, which is clearly epithelium, regardless of the presence

tubular oesophagus for ≥1 cm proximal to the GEJ

stomach to oesophagus for any length epithelium, regardless of the presence

with a minimum length of 1 cm (tongues or circular)

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 11

12 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 13

Non-dysplastic Indeﬁnite for HGD Submucosal

• Repeat EGD at 3–6 months Visible lesion at EGD

Dysplasia conﬁrmed by Histopathological ESD

14 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

Table 5 | Recommendations for BE surveillance and management

(2011) • Consider endoscopic • Endoscopic eradication therapy

(2014) surveillance maximal acid suppression 6 months • Endoscopic therapy is preferred

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 15

Argon plasma coagulation. Argon plasma coagula- Stepwise endoscopic resection

16 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 17

18 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

recurrent disease after treatment to support the devel- QOL

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 19

20 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

NATURE REVIEWS | DISEASE PRIMERS | Article citation ID: (2019) 5:35 21

22 | Article citation ID: (2019) 5:35 www.nature.com/nrdp

You might also like