1. SCOTT SANDERS, CHRISTIAN ELL, S. MICHAEL GRIFFIN, STEPHEN ATTWOOD
Consensus Statements for Management of Barrett’s Dysplasia and Early- Stage Esophageal Adenocarcinoma, Based on a Delphi Process
CATHY BENNETT, NIMISH VAKIL, JACQUES BERGMAN, REBECCA HARRISON, ROBERT ODZE, MICHAEL VIETH, STUART GITTENS et al.
SCOTT SANDERS, CHRISTIAN ELL, S. MICHAEL GRIFFIN, STEPHEN ATTWOOD
HUGH BARR, JOHN ALLEN, MARK K. FERGUSON, PAUL MOAYYEDI, and JANUSZ A. Z. JANKOWSKI
Gastroenterology 2012; 143:336-46

3. International Funding
International Society of Diseases of the Esophagus
British Society of Gastroenterology
American College of Gastroenterology
American Gastroenterological Association
American Society for Gastrointestinal Endoscopy
Association of Upper Gastrointestinal Surgeons
FORT - Fight Oesophageal Reflux Together
German Society of Endoscopy
Netherlands Association of Hepatogastroenterologists
Oesophageal Cancer Research Fund

4. BADCAT consensus group
104 experts agreed to participate
25 core group developed statements
Chosen for expertise
Methodology, diagnosis, therapy, endoscopy, surgery
Chosen for international representation
14 countries (North America, Europe, Asia)
95 voted, 92 authors

5. BAD CAT methodology 1
Stage 1. Selection of experts, creation of sub section panels and endorsement of key stakeholders UEGW – Nov 2009
Stage 2. Formulation of questions, collection of evidence and voting rounds 1-5 BSG March 2010, AGA May 2010 and ISDE 2010
Stage 3. Final analysis and conclusions
Stage 4. Research into key areas

6. BADCAT methodology 2 In depth literature search (MEDLINE, EMBASE)
Additional literature from consensus group
Modified Delphi consensus methodology
Core group developed statements
All reviewed and added literature
Iterative voting
5 point adjectival scale
80% agreement considered consensus
Evidence graded according to GRADE system

7. Delphi and BAD CAT What was unique about our process? Delphi
Custom web-based system
Large group
Rapid feedback, commenting
Transparency and control
Literature search reviewed by all
Reproducibly systematic

8. Adjectival scale Disagree strongly Disagree with major reservation
Neither agree nor disagree
Agree with major reservation
Agree strongly
80% agreement threshold
aggregated agreement scores

9. Literature search Exhaustive search of electronic databases
Formulated by information specialists, included all terms (keywords and MeSH headings)
Peer reviewed
Over 11,000 Barrett’s and HGD records in data base, updated
Each participant reviewed hits for inclusion
Tagged to each statement online
Systematic

10. Voting rounds 2009-2010 4 rounds of web based anonymized voting
4 face to face meetings (WCG 2009, BSG 2010, DDW 2010), editorial
Statements reworded/changed at each round
New evidence added (tagged)
Grading evidence
92 final statements

11. Agreement over time
% reached consensus

12. BADCAT statement 92 statements on HGD in Barrett’s Definitions
Epidemiology
Prevention
Diagnosis
Medical therapy
Endoscopic therapy
Surgical therapy
Ethics

13. Highlights 6 key statements: Pathology Risk of progression
Routine endoscopy
Surveillance
Endoscopic treatment
Surgery

14. Pathology
At least 2 experienced gastrointestinal pathologists should evaluate all Barrett’s biopsies when a diagnosis of dysplasia is considered.
94% agree with statement

15. 2 pathologists should evaluate HGD in Barrett’s
Inter-observer variability between pathologists in differentiating HGD from intramucosal EA as described here.
Five studies showed the prediction of progression of esophageal dysplasia is improved if at least 2 expert pathologists agree on a diagnosis of dysplasia, and increases when more pathologists concur with the diagnosis.
Evidence: Moderate.

16. Risk of progression
Non-goblet columnar metaplasia of the esophagus may progress to cancer, but the magnitude of risk is unknown.
92% agree with statement

17. Risk of cancer in non-goblet cell columnar metaplasia
The US definition of Barrett’s requires that intestinal metaplasia is present (i.e. + goblet cells).
Non-goblet columnar metaplasia of the distal esophagus shows biological features of intestinal differentiation, and possesses molecular abnormalities consistent with a risk of malignancy of neoplasia precursor lesions.
Two retrospective studies evaluated the risk of neoplasia in patients with columnar metaplasia of the esophagus either with or without goblet cells. There were 991 patients with intestinal metaplasia and 631 without intestinal metaplasia.
The incidence of cancer progression from BE was similar in the 2 patient groups (4.5% vs 3.6% in one study and 3.1% vs 3.2% in the other).
Non-goblet cell columnar metaplasia has malignant potential, although the relative risk is unclear.
Evidence: Low.

18. Endoscopy: lumps and bumps
Visible lumps or nodules consisting of HGD suggest a more advanced lesion with invasion might be present.
(ulcers we know less about, no case series), but 90% agreed if not healed with PPI they are suspicious and should be monitored)
99% agree with statement

19. Visible lumps or nodules consisting of HGD = higher risk of cancer
Endoscopic mucosal resection (EMR/ER) of visible lumps with HGD on endoscopic biopsy results in upgrading the final diagnosis to cancer in 40% of cases.
Oesophagectomy series performed for presumed HGD identified by endoscopic biopsies, coexisting OA was found in 7 of 9 patients (78%) with a visible lesion and 7 of 22 patients (32%) without a visible lesion (P =0.02).
Evidence: Low.

20. Surveillance protocols
For evaluation of patients with BE, the use of high-resolution endoscopes and targeted biopsies of every suspicious lesion followed by 4-quadrant biopsies every 2 cm are recommended.
98% agree with statement

21. Surveillance protocols
High-resolution endoscopes preferred
There was no superiority to chromoendoscopy over standard endoscopy, although acetic acid spraying can improve visualization of lesions.
4-quadrant biopsies are still necessary at 1–2-cm intervals throughout the entire BE segment.
There are no data demonstrating superiority of 1-cm intervals compared with 2-cm
Evidence: Very low.

22. Endoscopic treatment or surgical?
Endoscopic treatment should be preferred over surgical treatment for management of most patients with HGD in BE.
93% agree with statement

23. Endoscopic or surgical treatment?
HGD in BE is rarely associated with lymph node involvement, deeper invasion has been ruled out by EMR .
Two case series reported that survival after EMR was high, similar to that expected in a surgical cohort.
One cohort study reported that disease-specific survival rate after endoscopic treatment was not different from surgical therapy. The case series reported a lower morbidity than might be expected in surgical patients.
Endoscopic treatment is associated with a higher rate of HGD recurrence, although this can usually be re-treated endoscopically.
Severe complications (such as bleeding, perforation, or stricture) are uncommon with endoscopic treatment.
Surgical resection is still possible and curative.
Evidence: Low - no surgery vs endo RCTs - see Cochrane review

24. Surgery for HGD
Reported operative mortality rate for esophagectomy for HGD and T1m generally ranges from 0% to 4%, with a mean overall operative mortality of 2%.
95% agree with statement

25. Low surgical mortality rates
30% mortality sometimes quoted.
Retrospective data are primarily from self-selected high-volume centres and analysis is retrospective.
Ten case series evaluating a total of 567 HGD or early EA patients.
Operative mortality rate for esophagectomy for HGD and early EA ranges from 0% to 4%, with an overall operative 30-day mortality rate of approximately 2%.
Specialist centres.
Evidence: Moderate.

26. BAD CAT Clinical pathways

27. BAD CAT Main conclusions
Pathology 1. At least 2 experienced gastrointestinal pathologists should evaluate all Barrett’s biopsies when a diagnosis of dysplasia is considered. Endoscopy 1. The Prague C&M Criteria is the best available tool for grading the endoscopic extent of BE. 2. Visible lumps in nodules consisting of HGD suggest a more advanced lesion with invasion might be present. Populations at risk 1. Men have approximately twice the rate of developing HGD or esophageal cancer compared with women, and the rate at which OA is increasing in Western populations is twice as high in men as it is in women. 2. Non-Hispanic white patients with BE are at higher risk for development of HGD/cancer compared with other racial/ethnic groups with BE. 3. Obesity is an independent risk factor for development of EA. Therapy 1. Endoscopic treatment should be preferred over endoscopic surveillance for management of most patients with HGD/T1m BE. 2. RFA is currently the best available ablation technique for treatment of flat HGD and for eradication of residual BE after focal EMR. 3. The operative mortality is improved if surgery is undertaken in specialist surgical centres.

28. Does it end there? Developing countries
Free text commenting – template better?
Low quality evidence
No meta-analysis
Some areas remain uncertain (lack of data)

29. Uncertainties
It is uncertain if proton pump inhibitors prevent the development of HGD/cancer.
91% agree it’s uncertain

30. PPIs and progression to HGD or early cancer
BAD CAT Data from retrospective observational studies
Evaluate BE patients on PPI and not on PPI
Three studies suggest PPI protective
One study suggests no benefit
Problems of patient selection/confounding
We agreed it’s uncertain if PPIs prevent progression
Where did that leave us?
El-Serag HB et al. AJG 2004; 99:

31. BOB CAT
Benign Barrett's and CAncer Taskforce 'BoB CAT' consensus group
To use a systematic evidence base review in iterative steps to produce guidelines for best clinical and cost effective management of benign Barrett’s esophagus in prevention of esophageal adenocarcinoma.

32. PPIs and progression BOB CAT
The use of PPI’s (compared to no therapy or H2RAs) is associated with a decrease in progression from benign Barrett’s metaplasia to Barrett’s neoplasia (dysplasia and cancer)
There is no evidence from high quality prospective trials (RCTs) that PPIs use prevents progression of BE to neoplasia but there is scientific plausibility and observational/epidemiological evidence for this inverse association and protective effect.

33. Conclusions BADCAT consensus group International Multidisciplinary
Modified Delphi consensus methodology
Inform GI community how to manage HGD and early cancer in BE
BOB CAT Benign Barrett's and CAncer Taskforce 'BoB CAT' consensus group