activity and biomarker measures: ulcerative colitis...• icc screen, wk 6, wk 10 site vs. central:...
TRANSCRIPT
Activity and Biomarker
Measures: Ulcerative
Colitis
Mark T. Osterman, MD MSCE
Assistant Professor of Medicine
University of Pennsylvania
Outline
• Activity Measures
–Clinical activity
–Endoscopic activity
–Histological activity
• Biomarker Measures
–Serum biomarkers
–Stool biomarkers
Index Variables Score Remis
Truelove-Witts Freq, bleed, T, HR, Hgb, ESR N/A N/A
Powell-Tuck Well-being, pain, freq, consist, T,
bleed, anorexia, N/V, tender, EIM
0-20 0
Rachmilewitz Freq, bleed, PGA, pain, T, EIM, lab 0-29 <4
Seo Bleed, freq, ESR, Hgb, alb 50-250 N/A
PGA Sx relief (Likert) 1-6 1
Lichtiger Freq, noct, bleed, incont, pain,
well-being, tender, α-diarrheal
0-21 <3
Invest Glob Eval Sx score (Likert) 0-4 0
SCCAI Freq, noct, urge, bleed, well-
being, EIM
0-19 N/A
UC Clin Score Freq, bleed, PFA, PGA 0-12 <1
Pt-Def Remis In remission: yes/no Yes/no Yes
Purely Clinical Indices
Patient-Defined Remission • Survey
–56 outpatients
–Are you in remission: yes / no
• 7-point Likert score
–Has your condition improved /
worsened since last visit
• Remission defined as absence of rectal
bleeding and modified Baron Score 0-2:
sens 86% / spec 76%
Higgins PD et al, Gut 2005
SCCAI Bowel frequency (day): 1-3
4-6
7-9
>9
0
1
2
3
Bowel frequency (night): 0
1-3
4-6
0
1
2
Urgency: None
Hurry
Immediately
Incontinence
0
1
2
3
Bleeding: None
Trace
Occasionally frank
Usually frank
0
1
2
3
Well-being: Very well
Slightly below par
Poor
Very poor
Terrible
0
1
2
3
4
EIM: Uveitis
Pyoderma gangrenosum
Erythema nodosum
Arthropathy
1
1
1
1
Remission
• <2 correlates with PDR
Response
• >2-pt decrease correlates
with patient-defined
significant improvement
Relapse
• >5 highly predictive per
Seo Index
Walmsley RS et al, Gut 1998
Jowett SL et al, Scand J Gastroenterol 2003
Higgins PD et al, Gut 2005
Partial Mayo Score Stool frequency: Normal
1-2 / day above normal
3-4 / day above normal
>5 / day above normal
0
1
2
3
Bleeding: None
Streaks of blood in <50% of BMs
Obvious blood in >50% of BMs
Blood alone
0
1
2
3
PGA: Normal (no symptoms)
Mild disease
Moderate disease
Severe disease
0
1
2
3
Schroeder KW et al, N Engl J Med 1987
Rutgeerts P et al, N Engl J Med 2005
Lewis JD et al, Gastroenterology 2008
Remission
• <2 with no individual subscore >1
Response
• Decrease of >3 pts and >30% from baseline
+ Rectal bleeding decrease >1 pt or absolute subscore 0-1
Partial Mayo Score
Lewis JD et al, Inflamm Bowel Dis 2008
Remission Response
6-pt PMS
Remission
• <1
Response
• Decrease of >2 pts from baseline
How Many Days of Data • Issues: Sx recall, day-to-day variability of Sx
• Recall: 6-pt PMS (n=28, 61% inactive)
– Questionnaires: baseline (all), 1 of next 7 d
– No difference in ΔPMS from baseline for 1-2
vs. 3-5 vs. 6-8 d groups
• Variability: 6-pt PMS (n=31, 68% inactive)
– 7d electronic diary, only 58% completed all
– Mean max ΔPMS higher with active
– ρ = 0.82, 0.86, 0.91 for avg of all 7d vs. last 1,
avg of last 2, avg of last 3 d
Henao MP et al, Inflamm Bowel Dis 2015
The Problem with Clinical Indices “It is evident that placebos have a high degree of
therapeutic effectiveness…being produced in 35.2 +
2.2% of cases.” ̶ Henry Knowles Beecher, 1955
Beecher HK, JAMA 1955
Hrobjartsson A et al, N Engl J Med 2001
Barsky AJ et al, JAMA 2002
PBO Type Pts Trials Pooled Estimate
Binary 3099 21 RR 0.97 (0.88-1.07)
Continuous 2363 24 SMD -0.20 (-0.37 to -0.04)
Nocebo phenomenon
• Reporting of AEs to PBO
• 25% freely report AEs to PBO but 71% report AEs
when actively asked about them
Placebo Response in UC Pooled remission rate 13% (range 0-40%)
Su C et al, Gastroenterology 2007
Predictors OR (95% CI)
Study location (Europe vs. N. America) 2.9 (1.5-5.4)
F/u time (per 1 wk) 2.0 (1.2-3.1)
# F/u visits (per 1 visit) 2.2 (1.3-3.7)
Disease duration (per 1 yr) 1.4 (1.1-1.7)
Baseline UCDAI (per 1-pt) 0.1 (0.04-0.49)
Strict outcome (DAI 0 vs. <3) Rate 5% vs. 17%
Rectal bleeding subscore (per 1-pt) 0.04 (<0.001-0.25)
Endoscopic mucosal healing 0.4 (0.2-0.7)
Pooled response rate 28% (range 0-67%)
Other Predictors OR (95% CI)
Publication yr (per 1 yr) 1.04 (1.002-1.1)
↓ UCDAI >3 vs. >2 Rate 30% vs. 52%
No bleeding
Bleeding on light touch
Spontaneous bleeding
0
1
2
Endoscopic Indices Powell-Tuck
Baron Score Normal
Hypovascular, granular, hyperemia
Bleeding on light touch
Spontaneous bleeding
0
1
2
3
Modified Baron Score
Normal
Hypovascular, granular, hyperemia
Bleeding on light touch
Spontaneous bleeding
Clear ulceration, denuded mucosa
0
1
2
3
4
Vascular pattern: Normal
Patchy obliteration
Obliteration
0
1
2
Bleeding: None
Mucosal
Luminal mild
Luminal moderate / severe
0
1
2
3
Erosions / ulcers: None
Erosions
Superficial ulcer
Deep ulcer
0
1
2
3
Normal
Mild: erythema, hypovascular, mild friability
Moderate: marked erythema, avascular, friability, erosions
Severe: spontaneous bleeding, ulceration
0
1
2
3
Endoscopic Indices Mayo Endoscopic Score: 0-3
UC Endoscopic Index of Severity (UCEIS): 0-8
Schroeder KW et al, N Engl J Med 1987
Travis SP et al, Gut 2012
Mayo Endoscopic Score
Vascular pattern: Normal
Partially visible
Complete loss
0
1
2
Granularity: None
Fine
Coarse
0
1
2
Ulceration: None
Erosions / pinpoint ulceration
Numerous shallow ulcers w/ mucopus
Deep excavated ulcerations
Diffusely ulceration w/ >30% involvement
0
1
2
3
4
Bleeding / friability: Normal
Friable, bleeding to touch
Spontaneous bleeding
0
1
2
Endoscopic Indices UC Colonoscopic Index of Severity (UCCIS): 0-162
Score = 3.1 x (vascular pattern sum) + 3.6 x (granularity sum) +
3.5 x (ulceration sum) + 2.5 x (bleeding/friability sum)
Thia KT et al, Inflamm Bowel Dis 2011
Samuel S et al, Clin Gastroenterol Hepatol 2013
Flex Sig vs. Colonoscopy
Endoscopy Status # Discordant r
Active (n=239)
Mayo >2
Mayo >1
UCEIS
9
1
N/A
0.84
0.96
0.92
Mucosal Healing (n=139)
Mayo <1
Mayo = 0
7
1
0.85
0.95
Colombel J-F et al, Gastroenterology 2016
• Analysis of 239 / 331 videos from
EUCALYPTUS (Etrolizumab phase 2 RCT)
• Only 4.5-7.5% had worse disease proximally
The Problem With Endoscopy Descriptor Intrarater κ
(2 pairs per rater)
Interrater κ
(21 per rater)
Vascular pattern 0.51 0.34
Erythema 0.37 0.25
Mucosal surface 0.37 0.26
Edema 0.33 0.23
Mucopus 0.38 0.32
Bleeding 0.51 0.29
Incidental friability 0.37 0.30
Contact friability 0.33 0.23
Erosions / ulcers 0.56 0.36
Erosion / ulcer extent 0.51 0.32
Travis SP et al, Gut 2012
Central Reading of Endoscopy
• Analysis of high-dose mesalamine tab (800 mg) RCT
• Significance not reached (UCDAI endo = 0) with site
read vs. central read due to PBO rate 21% vs. 14%
• 7 central readers evaluated
• ICC screen, wk 6, wk 10 site vs. central: 0.11, 0.31, 0.44
• Site reader scores generally higher than central reader
• Central: unbiased, expert, no interobserver variability
Feagan BG et al, Gastroenterology 2013
Central UCDAI Mod Baron UCEIS VAS
Intra ICC 0.89 0.88 0.89 0.91
Inter ICC 0.79 0.78 0.83 0.78
Stricter Outcome Measures
• Corticosteroid-free remission
• Rectal bleeding
– ACT / GEMINI 1: response rectal bleeding
decrease >1 pt or absolute subscore 0-1
– Tofacitinib: rectal bleeding subscore = 0
• Endoscopic mucosal healing / remission
– Even stricter but yet to be incorporated as
primary outcome
• ?Histological healing
Why Histology as Outcome
• Logical as UC is mucosal disease
• Histological healing is true mucosal
healing
• Up to 24-90% of endoscopically normal
may have microscopic inflammation
• May be better predictor of relapse, CS
use, hospitalization than endoscopy
Osterman MT, J Clin Gastroenterol 2013
Bryant RV et al, Gut 2016
Riley SA et al, Gut 1991
Zenlea T et al, Am J Gastroenterol 2016
Semiquant:
0 = none
1 = mild
2 = mod
3 = severe
Crypt abscesses 0-3
PMNs in lamina propria 0-3
Surface epithelial integrity 0-3
Mucin depletion 0-3
Round cells in lamina propria 0-3
Crypt architecture irregularity 0-3
PMNs in epithelium: None
<25% crypts
25-75% crypts
>75% crypts
0
1
2
3
PMNs in lamina propria: Mild increase
Mod increase
Marked increase
4
5
6
Erosion or ulceration 7
Riley SA et al, Gut 1991
Feagan BG et al, N Engl J Med 2005
Riley Index
Modified Riley Score
Geboes Index
Geboes K et al, Gut 2000
0.0
0.1
0.2
0.3
Architectural changes: None
Mild
Mild or mod diffuse or multifocal
Severe diffuse or multifocal
1.0
1.1
1.2
1.3
Chronic inflammatory infiltrate: No increase
Mild or equivocal increase
Moderate increase
Marked increase
2.0 (2A for Eos, 2B for PMNs)
2.1
2.2
2.3
Lamina propria Eos and PMNs: No increase
Mild or equivocal increase
Moderate increase
Marked increase
3.0
3.1
3.2
3.3
Epithelial PMNs: None
<5% crypts involved
<50% crypts involved
>50% crypts involved
4.0
4.1
4.2
4.3
Crypt destruction: None
Probable – local excess of PMNs in part of crypt
Probable – marked attenuation
Unequivocal crypt destruction
5.0
5.1
5.2
5.3
5.4
Erosion or ulceration: No erosion, ulceration, granulation tissue
Recovering epithelium + adjacent inflammation
Probable erosion – focally stripped
Unequivocal erosion
Ulcer or granulation tissue
Proposed histologic healing
Predictive of clinical relapse
Bressenot A et al, Gut 2015
Index / Components Intrarater κ Interrater α
Riley
PMNs in lamina propria
Crypt abscesses
Mucin depletion
Surface epithelial integrity
Round cells in lamina propria
Crypt architecture irregularity
0.75-0.89
0.67-1.00
0.40-0.57
0.73-0.93
0.40-0.73
0.53-0.71
0.85
0.67
0.73
0.75
0.67
0.64
Geboes
Architectural changes
Chronic inflammatory infiltrate
Lamina propria Eos
Lamina propria PMNs
Epithelial PMNs
Crypt destruction
Erosion or ulceration
0.58-0.70
0.47-0.73
0.46-0.56
0.66-0.93
0.68-0.70
0.75-0.84
0.85-0.92
0.65
0.73
0.40
0.82
0.74
0.63
0.82
The Problem With Histology
Bressenot A et al, Gut 2015
Index / Components Intrarater κ Interrater α
Riley
PMNs in lamina propria
Crypt abscesses
Mucin depletion
Surface epithelial integrity
Round cells in lamina propria
Crypt architecture irregularity
0.75-0.89
0.67-1.00
0.40-0.57
0.73-0.93
0.40-0.73
0.53-0.71
0.85
0.67
0.73
0.75
0.67
0.64
Geboes
Architectural changes
Chronic inflammatory infiltrate
Lamina propria Eos
Lamina propria PMNs
Epithelial PMNs
Crypt destruction
Erosion or ulceration
0.58-0.70
0.47-0.73
0.46-0.56
0.66-0.93
0.68-0.70
0.75-0.84
0.85-0.92
0.65
0.73
0.40
0.82
0.74
0.63
0.82
The Problem With Histology
Basal Plasmacytosis
• Dense infiltration of plasma cells in
lower third of mucosa
• Predictive of relapse in quiescent UC
– Modified Baron 0-1: HR 4.5 (1.7-11.9)
– Mayo = 0: OR 5.13 (1.32-19.99)
Goldman H, Pathology of GI Tract 1998
Bitton A et al, Gastroenterology 2001
Bessissow T et al, Am J Gastroenterol 2012
MARQUEE Study: CCFA CRA
• Prevalence of endoscopic (Mayo, UCEIS, UCCIS)
and histological disease (Riley, basal plasmacy)
in clinically quiescent UC
• Correlate endoscopic and histological scores
• Determine risk of relapse by endoscopic and
histological score to see which most predictive
• Basis for possible RCT of therapy escalation to
biologic for patients in clinical remission on
optimized 5-ASA but with active mucosal
disease
Serum Biomarkers • 451 pediatric patients
– PUCAI: sens / spec for cutoffs CRP 71-86% / 77-91%,
ESR 67-81% / 69-79%
– Beattie endo: r CRP 0.55, ESR 0.41
– 36% of mod-severe had CRP < 5 mg/L but useful as
f/u indicator if elevated when active
• 722 endoscopies in 552 patients
– Mayo: r CRP 0.50, ESR 0.40
– Modified Baron: r CRP 0.52, ESR 0.43
• Limited data on predicting clinical relapse
• CRP short ½-life 19h, ESR affected by Hct
• CRP: mesenteric adipocytes (CD, severe UC) Turner D et al, J Crohns Colitis 2011
Yoon JY et al, Dig Dis Sci 2014
Sands BE Gastroenterology 2015
Stool Biomarkers
• Lactoferrin
– Iron-binding protein in PMN granules and
serum, secreted by mucosal membranes
• Calprotectin
– Calcium- and zinc-binding protein
– 60% of cytosolic proteins in PMNs
• Fecal immunochemical test (FIT)
– Quantification of hemoglobin concentration
Kane SV et al, Am J Gastroenterol 2003
Roseth AG et al, Scand J Gastroenterol 1992
Levi Z et al, Ann Intern Med 2007
Stool vs. Serum: Endo Activity
Marker / Disease Pooled
Sens (%)
Pooled
Spec (%)
CRP: IBD Range 5-10 mg/dL
49 (34-64) 92 (72-98)
Fecal Calprotectin Range 6-280 µg/g
IBD
UC
88 (84-90)
88 (84-92)
73 (66-79)
79 (68-87)
Fecal Lactoferrin: IBD Range 0.07-7.25 µg/mL
82 (73-88) 79 (62-89)
Mosli MH et al, Am J Gastroenterol 2015
Meta-analysis: 19 studies, 2499 patients
FC Level: Endo Activity
FC cutoff
(µg/g)
Pooled
Sens (%)
Pooled
Spec (%)
50 92 (90-94) 60 (52-67)
100 84 (80-88) 66 (59-73)
250 80 (76-84) 82 (77-86)
Lin J-F et al, Inflamm Bowel Dis 2014
Meta-analysis: 8 studies, 744 patients
FC Level: Outcomes in RCT FC cutoff (mg/kg) Sens (%) Spec (%) Sum (%)
Clinical Remission
50
100
150
200
250
43
54
68
70
71
91
83
79
73
71
134
137
147
143
142
Endoscopic Remission
50
100
150
200
250
52
67
79
82
82
88
81
75
70
67
140
147
153
151
149
Mucosal Healing
50
100
150
200
250
29
44
54
57
58
92
89
85
79
71
121
133
139
136
134
Sandborn WJ et al, Gastroenterology 2016
FC Level: Outcomes in RCT FC cutoff (mg/kg) Sens (%) Spec (%) Sum (%)
Clinical Remission
50
100
150
200
250
43
54
68
70
71
91
83
79
73
71
134
137
147
143
142
Endoscopic Remission
50
100
150
200
250
52
67
79
82
82
88
81
75
70
67
140
147
153
151
149
Mucosal Healing
50
100
150
200
250
29
44
54
57
58
92
89
85
79
71
121
133
139
136
134
Sandborn WJ et al, Gastroenterology 2016
FC: Predicting Relapse Study Details N FC Cutoff
(µg/g)
Sens
(%)
Spec
(%)
Mao Meta-analysis
5 studies, 12 mo
318 120-164 77 71
Yamamoto 5-ASA maint
12 mo
80 170 76 76
De Vos IFX maint
↑ as early as 3 mo
87 300
2 in 1 mo
58
62
93
100
Molander IFX induction
12 mo
26
60
139 72
all
80
All
De Vos IFX ind, FC wk 2
Wk 10 endo remis
53 <50
80% ↓
54 67
FC: Histological Inflammation Study Details N Findings
Guardiola Prospective
Clin and endo remis
PMS, Mayo endo
Histo: epithelial PMN
59 Active: FC 278 (136-
696) vs. 68 (30-172)
FC 155: sens 78,
spec 71
Theede Active and inactive
PMS, Mayo / UCEIS
Histo: mod Harpaz
120 FC by histo 0-3: 237,
1398, 1080, 1530
FC 171: sens 75,
spec 90
Zittan Active and inactive
Mayo, UC/colonic CD
Histo: Geboes / basal
plasmacytosis
27 FC 100: r Geboes
0.77, basal plas 0.80
FC 100: sens 100,
spec 77
The Problem With FC • Correlates with disease activity / extent
– Severely active short vs. mildly active long segment
• Associated with lifestyle / diet
– Higher with higher age, obesity, physical inactivity
– Lower with fiber and vegetable consumption
• Much variability during bowel prep
– Can go from very high to very low so best to get
sample before or 2-5 days after
• ?Varies with amount of blood / mucus in BMs
• Affected by assay
– 2 ELISA, 1 enz fluoroassay, 2 quant immunochrom:
r 0.65, 0.65, 0.69, 0.70, 0.75, 0.77, 0.82, 0.85, 0.91, 0.93 Kolho K-L et al, J Pediatr Gastroenterol Nutr 2012
Labaere D et al, United Eur Gastroenterol J 2014
Kawashima K et al, BMC Gastroenterology 2016
Poullis A et al, Cancer Epidemiol Biomarkers Prev 2004
FC: Intra-Individual Variability Study Details Within-BM Within-day Btwn-day
Lasson 18 pt, 287 sample
Q-BM x 2d
2 sample / ea BM
Mild-mod UC
ICC 0.79
(0.48-0.90)
Med CV 0.52
↑ with ↑ time
btwn BM,
looser BMs
1st on d1 vs.
1st on d2:
med CV 0.41
Calafat 18 pt, 56 sample
1st, 1-3 more BM
in 1d
Severe UC
Med CV 0.40
FC highest
in 1st, 2nd BM
Kristensten 50 pt, 150 sample
BM AM, PM, AM
2 sample / 1 BM
Mixed IBD activity
Mean CV
0.03
UC: κ 0.69
CD: κ 0.84
UC: κ 0.68
CD: κ 0.84
Du 45 pt, 312 sample
3 sample / 1st, 1-2
more BM in 1d
50% inactive IBD
Clin sig
var 8-25%
Lowest:
1st, FC 250
Clin sig var
13-26%
Lowest: FC
250
DEAR
Red circles denote primary outcome
compared in the RCT
Randomized
Randomized
Measure FC
FC<50 mcg/g FC>=50 mcg/g
Week 6
Baseline
Measure FC
Week 12
Week 48
Remission
F/U to 48 weeks
NOT taking mesalamine
Measure FC Measure FC
Randomized
Measure FC
Taking 2.4g
MMX-M
Measure FC
<50 mcg/g >=50 mcg/g <50 mcg/g >=50 mcg/g
Measure FC
Taking mesalamine*
Switch to 2.4g
MMX-M
<50 mcg/g >=50 mcg/g <50 mcg/g >=50 mcg/g
Measure FC Measure FC Measure FC Measure FC Measure FC Measure FC
Group 1
Group 3 Group 4 Group 2
* 3gm/day or less of mesalamine other than MMX
mesalamine (MMX-M)
<50 mcg/g
2b
>=50 mcg/g
2a
2.4g
MMX-M
2aSD
4.8g
MMX-M
2aID
2.4g MMX-M
3SD
4.8g
MMX-M
4.8g
MMX-M
2.4g
MMX-M
4.8g
MMX-M
No MMX-
M
2.4g
MMX-M
4.8g
MMX-M
4.8g MMX-M
3ID
No MMX-M
4SD
2.4g MMX-M
4ID
Osterman MT et al, Clin Gastroenterol Hepatol 2014
n=58 n=61
ITT
PP
Osterman MT et al, Clin Gastroenterol Hepatol 2014
Stool Trek: The Next Generation
Vinding KK et al, Inflamm Bowel Dis 2016
• Dip 3-5X in BM, shake, settle 10 min, 2 drops supernat at S
• Lat flow device (LFD): nitrocellulose memb, gold-labeled
α-FC (T) and control (C) Abs, incubate 10 min, take pic
• Ratio of T:C lines staining intensity determines FC
• Dots: green (FC <200), yellow (200-600), red (>600)
Fecal Immunochemical Test (FIT)
Nakarai A, Am J Gastroenterol 2013
Takashima S et al, Am J Gastroenterol 2015
Nakarai A et al, World J Gastroenterol 2016
Outcome N / Endo Sens (%) Spec (%)
Endo (Hgb <100 ng/mL)
Mayo 0
Mayo 0-1
Mayo 2-3
152 / 310
92
60
87
71
87
60
MH: FIT (100) vs. FC (250)
Mayo 0
Mayo 0-1
92 / 105
95 vs. 82
81 vs. 70
62 vs. 62
68 vs. 66
Clinical relapse (n=194)
• FIT Hgb <100: HR 0.17 (0.10-0.28)
• Clinical remission >1y and neg FIT: HR 0.11 (0.04-0.23)