1. Clinical Management of Colorectal Neoplasia in IBD Steven H. Itzkowitz, MD, FACP, FACG, AGAF Professor of Medicine and Oncological Sciences Icahn School of Medicine at Mount Sinai New York City, N.Y.

2. ACCME/Disclosures Dr. Steven Itzkowitz declares affiliation with Exact Sciences Corporation (Scientific Advisory Board; Research support)

3. Teaching Points 1.Risk of CRC in IBD is high (2x general pop’n.) 2.CRC incidence has been decreasing –Better colonoscopy? –Better medical management? –Earlier colectomy? 3.Colonoscopic surveillance reduces incidence of CRC 4.Risk factors have been identified –Clinical –Pathological 5.Dysplasia is the most important and actionable risk factor

4. Teaching Points 6.Dysplasia in IBD can be difficult to detect. 7. Better endoscopic detection of dysplasia –High-definition white light endoscopy (HD-WLE) –Chromoendoscopy 8. Better endoscopic resection techniques –Resectable vs. unresectable lesions 9. Better characterization of dysplastic lesions –Improved nomenclature (polypoid; nonpolypoid; invisible) 10. Better understanding of natural history of dysplasia 11. More refined colonoscopic surveillance guidelines

5. 1. Risk of CRC is High in IBD

6. Cumulative Risk of CRC in UC meta-analysis; older data (0.5-1.0% per year, after 10 years of disease) Eaden et al. Gut 48:526, 2001

7. Canavan, Aliment Pharm Ther, 2006 Cumulative Risk of CRC in Crohn’s Colitis older data

8. AuthorCountryDates Annual Incidence Rate Relative Risk (95% CI) PalliItaly1978-920.12%1.79 (0.9-3.3) BernsteinCanada1984-970.20%2.75 (1.9-4.0) LakatosHungary1974-20040.15%-- JessMinnesota, US1940-20010.10%1.1 (0.4-2.4) WintherDenmark1962-870.06%1.05 (0.6-1.8) Loftus E. Gastro Clin NA 35:517, 2006 2.Incidence Rates of CRC in IBD is Declining Population Based Studies

9. Cumulative Incidence of CRC AuthorCountry10 yrs20 yrs30 yrs Eaden ’01Meta-analysis1.6%8.3%18.4% Winther ’04Denmark0.4%1.1%2.1% Rutter ’06St. Mark’s; London0%2.5%7.6% Lakatos ’06Hungary0.6%5.4%7.5% Choi ’15St. Mark’s; London0.1%2.9%6.7%* *10% at 40 yrs 2.Incidence Rates of CRC in IBD is Declining Incidence over Time

10. DecadeNo of Studies Patient-YearsNo of CRC cases Incidence rate per 1000 py 1950’s34,750224.29 (0.95-7.64) 1960’s719,304804.18 (2.67-5.68) 1970’s412,909403.22 (0.67-5.77) 1980’s14123,8663102.58 (1.81-3.34) 1990’s1287,4991321.53 (1.06-2) 2000’s23369,8295251.29 (1-1.58) 2010-201318861,4781,1801.21 (0.95-1.48) Castano-Milla et al. Aliment Pharm Ther 39:645, 2014 Nonetheless, IBD patients are still considered to be at high risk for CRC. 2.Incidence Rates of CRC in IBD is Declining Temporal Trends

11. AuthorInterventionOR95% CI Eaden ’00No colonoscopy1.0-- 1-20.220.09-0.55 >20.420.16-1.10 Barium enema (ever)1.070.88-1.29 Velayos ’06No colonoscopy1.0-- 1-20.40.2-0.7 >20.30.1-0.8 Ananthakrishnan ’15No recent colonoscopy1.0-- C’scopy in past 36 mos0.560.39-0.80 3. Colonoscopy Reduces Incidence of CRC in IBD Eaden et al. Aliment Pharm Ther 14:145, 2000 Velayos et al. Gastroenterology 130;1941, 2006 Ananthakrishnan et al. Clin Gastro Hepatol 13:322, 2015

12. 4. Factors That Increase CRC Risk (Clinical) Risk FactorRelative Risk Duration >8-10 yrs2.4-2.8 Extent of colitis: Pancolitis14.8 Left sided2.8 Proctitis1.7 Primary sclerosing cholangitis4.8 Family history of colon cancer: Age >502.5 Age <509.2

13. Risk FactorRelative Risk Active inflammation: Histologic3.0-5.1 Colonoscopic2.5 Anatomic abnormalities: Shortened colon28.4 Stricture (UC)5.7 Inflammatory pseudopolyps2.1-2.5 History of dysplasia9 4. Factors That Increase CRC Risk (Pathological)

14. 4. Factors that Decrease CRC Risk Risk FactorEvidence Surveillance colonoscopyYes Regular doctor visitsYes (few studies) Chemoprevention: Mesalamine (5-ASA)Yes (referral centers) ThiopurinesYes (referral centers) Anti-TNFInsufficient data SteroidsMaybe Ursodeoxycholic acidMaybe Folic acidNo

15. 5. Dysplasia is the most important and actionable risk factor

16. Dysplasia-Carcinoma Sequence AdenomaNormal Dysplasia Colitis Cancer Sporadic: IBD:

17. Pathogenesis of Colorectal Carcinoma. Beaugerie L, Itzkowitz SH. N Engl J Med 2015;372:1441-1452

18. 6. Dysplasia in IBD can be difficult to detect.

19. White LightChromoendoscopy

20. 6. Dysplasia in IBD can be difficult to detect. White LightNarrow Band Imaging

21. 6. Dysplasia in IBD can be difficult to detect. No dysplasia

22. 6. Dysplasia in IBD can be difficult to detect. Visible (polypoid) dysplasia

23. 7. Better endoscopic detection of dysplasia

24. Prospective Studies: Chromo Detects More Dysplasia Than White Light Endoscopy AuthorYearNo. of PtsIncreased Yield of Lesions Kiesslich20031653-fold Hurlstone20041624-fold Rutter20041004.5-fold Hurlstone20057003-fold Kiesslich20071614.75-fold Marion20081021.5-fold (patients) Choi20151,0982-fold But, we don’t know the natural history of patients with chromo-detected dysplasia

25. SCENIC Consensus: Enhanced Imaging Techniques High-definition (HD-WLE) is recommended over standard white light endoscopy (S-WLE) Chromoendoscopy is recommended over S-WLE 2-3 fold increase in dysplastic lesions However, whether the additional lesions are assoc with increase in risk of CRC is not known. Chromoendoscopy is suggested over HD-WLE (conditional recommendation) SCENIC Consensus Statement. Gastrointest Endosc 81:489, 2015

26. 8. Better endoscopic resection techniques

27. 9. Better characterization of dysplastic lesions

28. Dysplasia: Macroscopic definitions SCENIC Consensus Statement. Gastrointest Endosc 81:489, 2015 TermDefinition VisibleIdentified on targeted biopsies from a lesion visualized at colonoscopy - PolypoidLesion protruding into the lumen >2.5 mm PedunculatedAttached to the mucosa by a stalk SessileNo stalk; entire base is contiguous with mucosa - NonpolypoidLittle (<2.5 mm) or no protrusion above the mucosa Superficial elevatedProtrusion <2.5 mm FlatNo protrusion DepressedAt least a portion of the lesion is below the mucosa InvisibleIdentified on random (non-targeted) biopsies; no visible lesion

29. Endoscopic polypectomy (without surgical resection) is adequate treatment for adenoma-like polyps in UC patients. Caveats: Polypectomy must be complete. The base of the polyp should be separately biopsied and found to have no dysplasia. There should be no dysplasia elsewhere in the colon. Management of Polypoid Dysplasia Engelsgjerd et al., Gastroenterology 117:1288, 1999 Rubin et al., Gastroenterology 117:1295, 1999 Odze et al. Clin Gastro Hepatol 2:534, 2004

30. Polypoid: After complete removal of endoscopically resectable polypoid dysplastic lesion, surveillance colonoscopy is recommended over colectomy. Non-polypoid: After complete removal of endoscopically resectable non-polypoid dysplasia, surveillance colonoscopy is suggested over colectomy. Invisible: If invisible dysplasia is detected, confirm by a GI pathologist, suggest referral to endoscopist with IBD surveillance expertise (chromo; HD- WLE). SCENIC Consensus Statement. Gastrointest Endosc 81:489, 2015 SCENIC Consensus: Management of Dysplastic Lesions

31. 10. Better understanding of natural history of dysplasia

32. What is the Natural History of High-Grade Dysplasia?

33. Probability of Finding Cancer at Colectomy Pathology that prompted colectomy If colectomy done immediately If colectomy done after some follow-up Bernstein ’94DALM17/40 (43%)-- Bernstein ‘94HGD10/24 (42%)15/47 (32%) Connell ’94HGD8/12 (67%)-- Rutter ’06HGD5/11 (46%)2/8 (25%) Choi ’15HGD16/29 (55%)4/19 (21%) Bernstein et al. Lancet 343:71, 1994 Connell et al. Gastroenterology 1994 Rutter et al. Gastroenterology 130:1030, 2006 Choi et al. Am J Gastroenterology 110:1022, 2015

34. What is the Natural History of Low-Grade Dysplasia?

35. Probability of Finding Cancer at Colectomy Pathology that prompted colectomy If colectomy done immediately If colectomy done after some follow-up Bernstein ’94DALM17/40 (43%)-- Bernstein ‘94HGD10/24 (42%)15/47 (32%) Connell ’94HGD8/12 (67%)-- Rutter ’06HGD5/11 (46%)2/8 (25%) Choi ’15HGD16/29 (55%)4/19 (21%) Bernstein ‘94LGD3/16 (19%)17/204 (8%) Ullman ’03LGD2/11 (19%)-- Rutter ’06LGD2/10 (20%)7/36 (19%) Choi ’15LGD12/47 (26%)18/113 (16%) Bernstein et al. Lancet 343:71, 1994 Connell et al. Gastroenterology 1994 Rutter et al. Gastroenterology 130:1030, 2006 Ullman et al. Gastroenterology 2003 Choi et al. Am J Gastroenterology 110:1022, 2015

36. Progression of LGD to HGD or Cancer StudyHospital No. of Pts with LGD Progression Rate Connell ’94St. Mark’s London954% @ 5 yrs Ullman ’03Mount Sinai4653% @ 5 yrs Ullman ’02Mayo Clinic; MN1833% @ 5 yrs Rutter ’06St. Mark’s London4723% @ 5 yrs Lindberg ‘96Huddinge Sweden3735% @ 20 yrs Lim ’03Leeds, UK2910% @ 10 yrs Befrits ’02Karolinska, Sweden602% @ ~10yrs Jess ’06Olmstead Cty, MN290% @ 18 yrs Note: Patients with IBD would only agree to colectomy if CRC risk “right now” was at least 73%* *(Siegel CA et al, Inflammatory Bowel Diseases, 2010)

37. Rate of Progression from LGD to HGD/CRC: Higher in the Distal Colon 5-year AN-free survivals for distal and proximal LGD: 75±7% and 95±3%, respectively p = 0.019 Distal LGD (n=13/68) Proximal LGD (n=2/53)

38. Risk Factors for LGD Progression to HGD/CRC St. Marks Hospital, London Extensive UC with LGD diagnosed 1993-2012. 172 pts, followed for 4 years (median) 33/172 (19%) developed HGD or CRC Choi et al. Am J Gastroenterol Sept, 2015 Risk FactorHR95% Conf. Interval Non-polypoid LGD8.63.0-24.8 Invisible LGD4.11.3-13.4 LGD size >1 cm3.81.5-13.4 Hx of “indefinite dysplasia”2.81.2-6.5

39. 122460367248 Dukes’ C1 Dukes’ C2 Dukes’ B2 Dukes’ C2 Dukes’ B1 Dukes’ A Months of follow-up ILNILL NI Ca LL H LLN N = No dysplasia I = Indefinite dysplasia L = Low-grade dysplasia H = High-grade dysplasia Ca = Carcinoma IILNNCa Flat LGD 7 6 5 4 3 2 1 Ullman et al. Gastroenterology 2003 LGD Can Progress to CRC without Going Through HGD

40. What is the Natural History of Indefinite Dysplasia?

41. IND Progression to HGD/CRC: Intermediate between LGD and NoD Ullman, et al. Clin Gastro Hepatol 6:1225, 2008

42. Choi et al. Am J Gastroenterol 110:1022, 2015 Dukes’ StageTotalSurveillance DetectedInterval CRCDefaulted/Transferred A28 (32%)2413 B18 (20%)1026 C24 (27%)1176 D7 (8%)--43 Unknown11 (13%)317 TOTAL88 (100%)48 (55%)15 (17%)25 (28%) St. Mark’s Surveillance Program (40-year experience: 1971-2012) 1,375 patients (4.1% had PSC) Median age of onset of UC: 30 Median follow-up duration: 11 years per patient Colonoscopies: 8,650 total 1,070 exams between 2003-2012 (55% had >1 chromo) Colorectal cancer diagnosed: 88/1,375 (6.4%) Survival worse with later stage CRC and interval CRC

43. Surveillance reduced the risk of advanced-stage CRC and interval CRC. The incidence of dysplasia increased over time. –not due to over-diagnosis by chromoendoscopy Post-colonoscopy CRC rate was lower following chromoendoscopy compared to (standard) WLE No significant change in the proportion of CRCs detected at colectomy for dysplasia over time. 38% of patients with CRC had synchronous dysplasia elsewhere in the colon. HGD would benefit from colectomy IND behavior is similar to LGD Sporadic adenomas: no additional CRC risk Choi et al. Am J Gastroenterol 110:1022, 2015 St. Mark’s Surveillance Program Conclusions

44. Recommendations for Managing Dysplasia in IBD (AGA; BSG Guidelines)

45. AGA: Recommended Surveillance Practice 2010 Begin at 8 years of colitis (all patients except proctitis) –for PSC: at time of PSC diagnosis Repeat colonoscopy every 1-2 years. –Consider shorter interval if: active inflammation, (+) family hx of CRC, PSC, stricture, pseudopolyps Representative bx’es from each colonic segment –chromoendoscopy if have expertise Attention to polyps/raised lesions –Biopsy flat mucosa adjacent to polypectomy sites Do for Crohn’s colitis what you would do for UC Farraye, Odze, Eaden, Itzkowitz. AGA Technical Review on the diagnosis and management of colorectal neoplasia in IBD. Gastroenterology 138:746-74, 2010

46. 5 Years BSG 2010 IBD Surveillance Guidelines Screening Colonoscopy at 10 Years (preferably in remission; pancolonic dye spray) Intermediate Risk Extensive UC; mildly active histo/endo inflamm. Post-inflammatory polyps FH CRC age >50 Higher Risk Extensive UC; mod-severe histo/endo inflamm. Stricture in past 5 yrs Dysplasia in past 5 yrs declining surgery PSC FH CRC age <50 3 Years 1 Year Cairns et al. Gut 59:666-690, 2010 Lower Risk Extensive UC; inactive histo/endo inflamm. Left-sided colitis Crohn’s colitis <50% of colon

47. British Society of Gastroenterology: 2010 IBD Surveillance Guidelines Biopsy Protocol –Pancolonic dye spray with targeted biopsy of abnormal areas (Grade A recommendation) –Otherwise, 2-4 random biopsies from every 10 cm. Other considerations –Patient preference –Multiple post-inflammatory polyps –Age –Co-morbidity –Accuracy and completeness of examination Cairns et al. Gut 59:666-690, 2010

48. Problems with BSG Guidelines Starting at 10 yrs will miss some cancers. –17% of CRC diagnosed before 10 yrs of disease* Interval cancers often occur approx. 2 yrs after the last colonoscopy Assumes that quiescent disease is low risk. Limited evidence to support changing the surveillance interval based on: – family history: very few studies stratified risk based on age of FDR with CRC – endoscopic appearance of the colon * Lutgens et al. Gut, 2008

49. High-grade Dysplasia Low-grade Dysplasia Indefinite Dysplasia No Dysplasia Repeat colonoscopy within 6 months: Repeated dysplasia confirmed? Repeat colonoscopy 6-12 months LGD Able to be completely removed endoscopically, with no dysplasia elsewhere? Repeat colonoscopy 3 years Repeat colonoscopy 1-2 years AGA Guidelines BSG/ECCO Guidelines Repeat colonoscopy 5 years Lower Risk: - Inactive inflamm. - Left-sided UC - Crohn’s colitis <50% of colon Intermediate Risk: - Mild inflamm. - Inflamm. polyps - Family history CRC age >50 Higher Risk: - Moderate-severe inflammation - Stricture - Previous dysplasia - PSC - Family history CRC age <50 YesNo Colonoscopy for cancer screening intent COLECTOMY Invasive cancer HGD YesNo Beaugerie L, Itzkowitz SH. N Engl J Med 2015;372:1441-1452

50. Endoscopically non-resectable lesion, regardless of grade of dysplasia Multifocal dysplasia Repeated dysplasia Stricture in UC (also Crohn’s if new change in symptoms) Inability to perform surveillance (stricture; risk of numerous bowel preps and endoscopies, ?pseudopolyps) Patient preference (or lack of adherence with surveillance) When to Consider Colectomy?

51. Patients with longstanding UC and Crohn’s colitis are at increased risk of CRC (despite trends towards decreasing incidence) High-definition WLE should replace standard definition WLE. Chromoendoscopy is preferred, especially in higher risk patients (refer to expert endoscopist if necessary) Polypoid dysplasia (LGD or HGD), if completely resected by endoscopy, can be followed with continued surveillance LGD has a variable rate of progression but warrants close surveillance (or colectomy on individualized basis) IND carries an increased risk of neoplastic progression compared to a history of no dysplasia When in doubt, refer to an IBD Center of Excellence! Summary