Controversies in Surveillance and Therapy
for Colorectal Dysplasia in IBD:
Case Studies
Advances in Inflammatory Bowel Disease
Hollywood, Florida
December 13, 2013
Thomas Ullman MD
Mount Sinai, New York
Fernando Velayos MD MPH
University of California, San Francisco
Risk of CRC in IBD is elevated
Inflammation of the colon is the key factor
Site
All CD
Colon
Ileum
*
*
*
RR
2.5
4.5
1.1
95% CI
1.3-4.7
1.3-14.9
0.8-1.5
Ulcerative colitis
Crohn’s disease
General population
Canavan C et. al.Aliment Pharmacol Ther 2006: 23; 1097
Known risk factors are almost all non-modifiable
• Non-modifiable risk factors:
–
–
–
–
–
Duration (increases after 10 years)
Extent (15X greater in pancolitis)
PSC (5X greater)2
Family history of CRC (2.5X greater) 1
Inflammatory polyps (“pseudopolyps”-2.5X) 3,4
• Potentially modifiable risk factor:
– Histologic inflammation at surveillance colonoscopy3
Normal
Epithelium
1Askling
Inflamed
Epithelium
Indefinite
Dysplasia
Low-Grade
Dysplasia
High-Grade
Dysplasia
Cancer
J, et al. Gastroenterology. 2001
BU, et al. Dis Colon Rectum. 2001
3Rutter, et al. Gastroenterology. 2004. Bansal, et al. Presented at ACG 2005, Honolulu. Rubin et al. Presented at DDW
2006, Los Angeles.
4Velayos et. al . Gastroenterology. 2006
2Lindberg
Controversies to cover today
1. Surveillance: Is it effective, when to start, in
whom, how frequent to repeat colonoscopy?
2. Vocabulary of dysplasia: time to simplify?
3. What to do when dysplasia in detected:
polypectomy, proctocolectomy, partial
resection?
4. Performance of surveillance and role of
chromoendoscopy: what is standard of care?
5. New algorithm for thinking and managing
dysplasia in IBD: Can we mimic what we are
doing in non-IBD patients?
Controversy 1
Surveillance: Is it effective, when to start, in
whom, how frequent to repeat colonoscopy?
45 year old man with L sided ulcerative colitis
diagnosed 5 years ago. Based on 2010 AGA
guidelines what strategy is recommended?
A. Begin screening at 15 years, then every 5
years
B. Begin screening at 8 years, and then every 12 years
C. Begin screening at 8 years, then every 1-5
years
D. Average risk screening, not at increased risk
based on his limited extent
Is there sufficient rationale for
performing surveillance
colonoscopy in patients with IBD?
•
•
Grade B: There is moderate certainty that
surveillance colonoscopy results in at least
moderate reduction of CRC risk in patients with
IBD.
Despite the lack of randomized controlled trials,
surveillance colonoscopy is recommended for
patients with IBD at increased risk for
developing CRC.
Patients with extensive UC or CD of the colon
are most likely to benefit from surveillance.
Farraye FA, Odze R, Eaden J, Itzkowitz S. Diagnosis and management of colorectal neoplasia in
inflammatory bowel disease. Gastroenterology 2010; 138:746-774.
Most recent GI society surveillance guidelines
-which to choose?
Society
First colonoscopy
(Screening)
Interval subsequent colonoscopy
ACG (2004) and ASGE
(2006)
All patients 8-10 years
after diagnosis
Immediately in PSC
All patients 8-10 years
after diagnosis
Immediately in PSC
Every 1-2 years
Crohn’s and Colitis
Foundation (2006)
AGA (2010)
All patients 8 years after
symptom onset (except
proctitis and
procotosigmoiditis)
British Society
Gastroenterology (2010)
All patients 10 years after
diagnosis to determine
extent and endoscopic risk
factors
- Next 2 in 1-2 years
-Then every 1-3 years until 20 years of disease,
then return to every 1-2 years
- Yearly in PSC
-Every 1-2 years after screening
-Every 1-3 years after 2 negative examinations
- Yearly in pancolitis with active/moderate
inflammation or stricture or PSC or history of
dysplasia or FH CRC age <50
-Every 3 years in pancolitis with mild inflammation
or inflammatory polyps or FH CRC >50 years
- Every 5 years in quiescent pancolitis or left sided
colitis
Controversy 2
Vocabulary of dysplasia: time to simplify?
You are performing surveillance in pt with UC and
biopsies of lesion in area inflammation-path shows
tubular adenoma. Assuming area around lesion shows
no dysplasia, what would you call this lesion?
A.
B.
C.
D.
E.
Sporadic adenoma
Adenoma-like lesion or mass (ALM)
Dysplasia-associated lesion or mass (DALM)
Raised Dysplasia
Flat Dysplasia
You are performing surveillance in pt with UC and
biopsies of lesion in area inflammation-path shows
tubular adenoma. Assuming area around lesion shows
no dysplasia, what would you call this lesion?
A.
B.
C.
D.
E.
Sporadic adenoma
Adenoma-like lesion or mass (ALM)
Dysplasia-associated lesion or mass (DALM)
Raised Dysplasia
Flat Dysplasia
You are performing surveillance in pt with UC and
path shows tubular adenoma. What would you
call this lesion?
A.
B.
C.
D.
E.
Adenoma-like lesion or mass (ALM)
Dysplasia-associated lesion or mass (DALM)
Raised Dysplasia
Flat Dysplasia
Occult dysplasia
Pathologist cannot decideimportance of dysplasia is given by
endoscopic context
• Tubular adenoma= low-grade dysplasia
Indefinite
Low-Grade
High-Grade
Vocabulary for dysplasia in IBD
• Traditional: Macroscopic classification
“sporadic”
“Flat”
“Invisible?”
“DALM”
“ALM”
“Elevated”
• Better:
• How detected (Non-targeted vs. targeted biopsies)
• Can borders be defined
Itzkowitz S. and Harpaz N.
Gastroenterology 126:1634, 2004
Controversy 3
What to do when dysplasia in detected:
polypectomy, proctocolectomy, partial
resection?
Normal
Epithelium
Inflamed
Epithelium
Indefinite
Dysplasia
Low-Grade
Dysplasia
High-Grade
Dysplasia
Cancer
You are performing surveillance in pt with UC and
path shows dysplasia. Based on 2010 AGA
Guidelines, what is the recommended action
A. Ongoing surveillance with white light
endoscopy
B. Ongoing surveillance with chromoendoscopy
C. Proctocolectomy
D. Segmental resection
E. No recommendation
2010 AGA Guidelines for management
dysplasia-mostly grade A
Farraye Gastroenterology 2010; 138: 738
Perspective: What proportion of dysplasia fall into
the “flat” category
• Rutter 2006
– 25/110 (22.7%) LGD “invisible” or flat
• Rubin 2007
– 29/75 LGD invisible (38.7%)
• Velayos 2009
– 16/61 (26.2%) LGD invisible
• Marion 2008
– 3/12 LGD invisible (25%)
Rutter MD et. al.. GI Endoscopy 2004: 60(3):334
Rubin DT et. al.. GI Endoscopy 2007: 65 (7): 998
Velayos FS et al ACG 2009
Marion JF et al AJG 2008: 103: 2342
Perspective: What proportion of dysplasia
fall into this category
~25%
~75%
Gastroenterology 2010; 138: 738
AGA Guidelines-management of
dysplasia
Questions and
parameters to
decide
“nonadenoma like
dysplasia
lesion or
mass”
“adenoma-like “flat highlesion or mass grade
and no flat
dysplasia”
dysplasia
elsewhere”
“flat lowgrade
dysplasia”
Treatment?
Surgery
(grade A)
Polypectomy
(grade A)
Insufficient
(grade I)
Surgery
(grade A)
* Further adenoma 50%-need close surveillance
Farraye F Gastroenterology 2010; 138: 738
Bernstein C Lancet 1994
Controversy 4
Performance of surveillance and role of
chromoendoscopy: what is standard of care?
You are planning to perform surveillance colonoscopy on
patient with IBD and are deciding on what is the current
standard of care with regard to enhanced dysplasia detection
technique. Which of the following statements is true based
on 2010 AGA Guidelines?
A. Chromoendoscopy is superior to white light
colonoscopy for detecting dysplasia and should be
performed for every surveillance
B. NBI/iScan (virtual chromoendoscopy) is superior to
white light colonoscopy for detecting dysplasia and is
an easier alternative to chromoendoscopy
C. Chromoendoscopy is an acceptable alternative to
white light colonoscopy in those experienced in the
technique
D. Chromoendoscopy does not eliminate the need for
random biopsies
Surveillance Technique
• Based on expert opinion
• Technique: 4-quadrant biopsies every 10 cm of
mucosa; at least 33 biopsies; extra focus on
nodules, masses, strictures; every 5 cm in
rectosigmoid
Kornbluth and Sachar, Am J Gastro, 2004.
Itzkowitz and Present, Inflammatory Bowel Diseases, 2005.
Itzkowitz and Harpaz, Gastroenterology 126:1634, 2004.
23
Chromoendoscopy proposed as means
of improving sensitivity of colonoscopy
• Two main uses in IBD Surveillance
– Improve detection of subtle colonic lesions
(increase sensitivity of surveillance)
– Once lesion detected-to aid in differentiating
between neoplastic and non-neoplastic based on
crypt architecture and modified pit pattern
“Invisible” dysplasia happens in IBDReason for “enhanced” surveillance techniques
Rutter MD et. al.. GI Endoscopy 2004: 60(3):334
Toruner et. al.. Inflamm Bowel Dis 2005: 11:428
Significance of Pit Patterns
Type I/II predict non-neoplastic
lesions
Kudo S et al. Endoscopy 1993
Type III/IV/V predict neoplastic lesions
Difference Between Chromoendoscopy and
Virtual chromoendoscopy
• Chromoendoscopy
– Dye spray through catheter
– Absorptive dye: (stain taken up by noninflammed mucosa but poorly taken
up by active inflammation and dysplasia): methylene blue
– Contrast dye (coats surface to highlight subtle disruptions of normal
contours): indigo carmine
• Virtual chromoendoscopy
– Rotating color filters the R-G-B bands while increasing the relative intensity
of blue bands
– Post-processing techniques (i-Scan/Fujinon) to achieve pseudocolored
image
– Enhance tissue vasculature (differential optical absorption of light by Hb
associated with dysplasia (blue band)) or mucosal contours
SURFACE guidelines for
chromoendoscopy
• Strict patient selection
– Avoid active disease
• Unmask the mucosal surface
– Excellent bowel prep; remove mucus and debris
• Reduce peristaltic waves
• Full-staining length of the colon
• Augmented detection with dyes
– 0.4% indigo carmine; 0.1% methylene blue
• Crypt architecture analysis
– Pit pattern III/IV of concern
• Endoscopic targeted biopsies
– Biopsy all mucosal alterations, especially pit pattern III/IV
Chromoendoscopy Finds More Dysplasia than
Conventional Exams
Number of Dysplastic Lesions
Author
(Year)
Institution
# of UC
Patients
Type of
Imaging
Chromo
Conventional
Sensitivity /
Specificity
Kiesslich
(2003)
University of
Mainz, Germany
263
Methylene
blue
32
10
93% sens.
93% spec.
Rutter
(2004)
St. Mark’s
Hospital,
Harrow, UK
100
Indigo
carmine
7
0
Not given
Hurlstone
(2005)
The Royal
Hallamshire
Hospital,
Sheffield, UK
350
Indigo
Carmine-and
Magnification
69
24
93% sens.
88% spec.
19
4
94.7% sens.
98.3% spec.
97.8%
accuracy
8
7
Not given
17
9
Not given
Kiesslich
(2007)
University of
Mainz, Germany
161
Confocal
endomicrosco
py
Dekker
(2007)
Academic
Medical Center,
Amsterdam, The
Netherlands
42
Narrow-band
imaging
Marion
(2008)
Mount Sinai,
New York, USA
102
Methylene
Blue
Role of chromoendoscopy in surveillance
• Not yet standard of care
• Chromoendoscopy (not virtual chromo)-is an
alternative surveillance technique mentioned in
guidelines from Crohn’s and Colitis Foundation
of America (2006) and AGA (2010) and British
Society of Gastroenterology Guidelines (2010)
Controversy 5
Can we create a new/unified algorithm for
thinking and managing dysplasia in IBD: Can we
mimic what we are doing in non-IBD patients?
You are performing colonoscopy on a non-IBD patient and
come across the following lesion in the ascending colon. You
are able to define borders and lifts with saline. What would
you do?
A. Biopsy, if no cancer, schedule
colonoscopy later to remove
endoscopically (yourself or refer)
B. Attempt complete endoscopic
removal at the time of
procedure, if no cancer
confirmed, continue surveillance
C. Biopsy, if no cancer, refer to
surgeon for segmental resection
D. Biopsy, if no cancer, refer to
surgeon for proctocolectomy
Proposal-three parameters relevant for preventing
CRC and CRC mortality in IBD once any type of
dysplasia is detectedNOTE: it is what you are already doing in non-IBD
patients
1. Rate of progression of dysplasia to
advanced dysplasia or CRC (metachronous)
2. Rate of occult cancer in patients diagnosed
with dysplasia (synchronous)
3. Resectability of the dysplastic lesion
Proposal-three parameters relevant for preventing
CRC and CRC mortality in IBD once any type of
dysplasia is detectedNOTE: it is what you are already doing in non-IBD
patients
1. Rate of progression of dysplasia to
advanced dysplasia or CRC (metachronous)
2. Rate of occult cancer in patients diagnosed
with dysplasia (synchronous)
3. Resectability of the dysplastic lesion
Can I
see it?
Is it
discreet?
Can I
resect it?
3 questions to ask in this case
1. Rate of progression of
dysplasia to advanced
dysplasia or CRC
(metachronous)
2. Rate of occult cancer
in patients diagnosed
with dysplasia
(synchronous)
3. Resectability of the
dysplastic lesion
Controversy regarding progression
of “flat” LGD to HGD or Cancer
Study
Setting
LGD (n)
Rate
Connell 1994
St Mark’s
9
54% @5y
Ullman 2002
Mayo Clinic
18
33% @5y
Ullman 2003
Mount Sinai
46
53% @5y
Rutter 2006
St Mark’s
36
25% @5y
Lindberg 1996
Huddinge
37
35% @20y
Befrits 2002
Karolinska
60
2% @10y
Lim 2003
Leeds, UK
29
10% @10y
Van Schaik 2010
6 Dutch centers
70
12% @5y
Controversy regarding progression
of “flat” LGD to HGD or Cancer
Study
Setting
LGD (n)
Rate
Connell 1994
St Mark’s
9
54% @5y
Ullman 2002
Mayo Clinic
18
33% @5y
Ullman 2003
Mount Sinai
46
53% @5y
Rutter 2006
St Mark’s
36
25% @5y
Van Schaik 2010
6 Dutch centers
21
37% @5y
Lindberg 1996
Huddinge
37
35% @20y
Befrits 2002
Karolinska
60
2% @10y
Lim 2003
Leeds, UK
29
10% @10y
Controversy in the agreement of
dysplasia
GI Pathologists
Kappa statistic indicates how
much greater observer
agreement exists than would be
expected by chance
Range -1.0 to +1.0
Value 0= pure chance only
Value 1.0= perfect agreement
Value >0.75 =excellent agreement
Value 0.4-0.74= fair to good agreement
Value <0.4= poor agreement
- Eaden J J of Pathol 2001; 194:152
P1
P1 P2
P3
P4
P5
P6
P7
P8
P9
P10
P11
P12
P13
P2 P3 P4 P5
0.43 0.25 0.12 0.15
0.12 0.16 0.24
0.44 0.38
0.44
-
General Pathologists
P6
0.59
0.40
0.27
0.18
0.27
-
P7
0.48
0.36
0.39
0.17
0.26
0.51
-
P8
0.2
0.24
0.18
0.25
0.29
0.14
0.13
-
P9
0.22
0.15
0.24
0.17
0.14
0.35
0.32
0.13
-
P10
0.37
0.28
0.47
0.20
0.29
0.36
0.39
0.21
0.32
-
P11
0.19
0.19
0.33
0.27
0.2
0.24
0.34
0.13
0.28
0.21
-
P12
0.23
0.27
0.52
0.31
0.48
0.38
0.43
0.33
0.25
0.48
0.39
-
P13
0.33
0.26
0.35
0.17
0.12
0.43
0.40
0.11
0.26
0.3
0.43
0.29
-
Very few kappa values over 0.5
All pathologists agreed only on 4 of 51 (7.8% agreement (all HGD))
GI pathologists agreed only on 6 slides (11.7% agreement (4 HGD, 2 reactive atypia))
General pathologists agreed on 8 slides ( 15.7 % agreement (5HGD,2LGD,1 atypia))
3 questions to ask in this case
1. Rate of progression of
dysplasia to advanced
dysplasia or CRC
(metachronous)
2. Rate of occult cancer
in patients diagnosed
with dysplasia
(synchronous)
3. Resectability of the
dysplastic lesion
What is the probability of finding
occult (synchronous) cancer after a
diagnosis fLGD?
Study
If colectomy done
immediately
Bernstein 1994
3/16 (19%)
Ullman 2003
2/11 (19%)
Rutter 2006
2/10 (20%)
3 questions to ask in this case
1. Rate of progression of
dysplasia to advanced
dysplasia or CRC
(metachronous)
2. Rate of occult cancer
in patients diagnosed
with dysplasia
(synchronous)
3. Resectability of the
dysplastic lesion
Characteristics to resectability
You already ask yourself this when you do screening
and surveillance in patients without IBD
Can I
see it?
Is it
discreet?
Can I
resect it?
Fact: Non-resectable colonic
dysplasia is managed with surgery
• Concern in IBD is typically the type of surgery
– Colectomy in IBD vs. limited resection in non-IBD
Proposal: 3 parameters relevant for
managing dysplasia
Questions and
parameters to
decide
“nonadenoma like
dysplasia
lesion or
mass”
Progression
No info
Occult Cancer
43%
Resectability
No
Treatment?
Surgery
(grade A)
“adenoma-like “flat highlesion or mass grade
and no flat
dysplasia”
dysplasia
elsewhere”
* Further adenoma 50%-need close surveillance
Farraye F Gastroenterology 2010; 138: 738
Bernstein C Lancet 1994
“flat lowgrade
dysplasia”
Proposal: 3 parameters relevant for
managing dysplasia
Questions and
parameters to
decide
“nonadenoma like
dysplasia
lesion or
mass”
“adenoma-like “flat highlesion or mass grade
and no flat
dysplasia”
dysplasia
elsewhere”
Progression
No info
<5%*
Occult Cancer
43%
<5%
Resectability
No
Yes
Treatment?
Surgery
(grade A)
Polypectomy
(grade A)
* Further adenoma 50%-need close surveillance
Farraye F Gastroenterology 2010; 138: 738
Bernstein C Lancet 1994
“flat lowgrade
dysplasia”
Proposal: 3 parameters relevant for
managing dysplasia
Questions and
parameters to
decide
“nonadenoma like
dysplasia
lesion or
mass”
“adenoma-like “flat highlesion or mass grade
and no flat
dysplasia”
dysplasia
elsewhere”
Progression
No info
<5%*
High
Occult Cancer
43%
<5%
42%
Resectability
No
Yes
No
Treatment?
Surgery
(grade A)
Polypectomy
(grade A)
Surgery
(grade A)
* Further adenoma 50%-need close surveillance
Farraye F Gastroenterology 2010; 138: 738
Bernstein C Lancet 1994
“flat lowgrade
dysplasia”
Proposal: 3 parameters relevant for
managing dysplasia
Questions and
parameters to
decide
“nonadenoma like
dysplasia
lesion or
mass”
“adenoma-like “flat highlesion or mass grade
and no flat
dysplasia”
dysplasia
elsewhere”
“flat lowgrade
dysplasia”
Progression
No info
<5%*
High
1-12% vs 2555%
Occult Cancer
43%
<5%
42%
19%
Resectability
No
Yes
No
No
Treatment?
Surgery
(grade A)
Polypectomy
(grade A)
Surgery
(grade A)
Insufficient
(grade I)
* Further adenoma 50%-need close surveillance
Farraye F Gastroenterology 2010; 138: 738
Bernstein C Lancet 1994
Our approach to these controversies
1. Grade B evidence for surveillance in IBD. GI society
guidelines share first exam 8-10 yrs/PSC at diagnosis
– Next exam varies (1-3 years)
2. Simplified approach to dysplasia-based on how found:
targeted vs. non-targeted biopsy and if can define borders
3. Dysplasia mngmt: polypectomy-ALM; surgery-HGD/DALM;
not clear-flat LGD
4. Follow either surveillance technique based on expert
opinion or chromo, no role virtual chromo
– More likely to come across raised lesions or subtle
abnormalities (75%)-don’t just focus on 33 biopsies/dye spray
– No need random biopsy with chromo after training
5. Proposal: the 3 parameters we use to manage non-IBD
dysplasia can be applied to IBD-dysplasia (to be tested)