Results
Role of Video Capsule Endoscopy in Inflammatory Bowel Disease
Video Capsule Endoscopy in Suspected Crohn's Disease
Study Selection and Study Characteristics
We have identified 30 original articles (13 prospective) which evaluated the role
of VCE in suspected CD[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36] ([Table 1]).
Table 1
Summary of studies evaluating role of video capsule endoscopy in suspected Crohn's
diseases
Author
|
Indication
|
Study type
|
N
|
Diagnostic yield
|
Sensitivity
|
Specificity
|
PPV
|
NPV
|
Adverse events
|
Fireman et al 2003[7]
|
Suspected CD, negative conventional imaging
|
Prospective
|
17
|
71%
|
–
|
–
|
–
|
–
|
|
Herrerías et al 2003[8]
|
Suspected CD, negative conventional imaging
|
Prospective
|
21
|
43%
|
–
|
–
|
–
|
–
|
None
|
Ge et al 2004[10]
|
Suspected CD, negative conventional imaging
|
Prospective
|
20
|
65%
|
–
|
–
|
–
|
–
|
–
|
De Bona et al 2006[12]
|
Suspected CD, negative conventional imaging
|
Prospective
|
38
|
39.5% (46.2%with symptoms+ biochemical markers of inflammation)
|
–
|
–
|
–
|
–
|
2.6%
|
van Tuyl et al 2006[13]
|
Suspected small bowel disorders including CD
|
Retrospective
|
57
|
49%
|
–
|
–
|
61%
|
92%
|
–
|
May et al, 2007[15]
|
Abdominal pain+ diarrhea weight loss/anemia/elevated inflammatory markers
|
Prospective
|
50
|
54%
Additional symptom increased diagnostic yield
|
–
|
–
|
–
|
–
|
Retention (4%)
|
GIrelli et al 2007[14]
|
Suspected CD: pain and diarrhea >3 mo + fever/weight loss/anemia/EIM
|
Prospective
|
27
|
59%
|
93%
|
84%
|
–
|
–
|
Retention requiring surgery (11.1%)
|
Tukey et al 2009[18]
|
Suspected CD or pain and/or diarrhea
|
Retrospective
|
102
|
37% (13% final diagnosis of CD on follow up)
|
77%
|
89%
|
50% (depends on selection criteria)
|
96%
|
–
|
Figueiredo et al 2010[19]
|
Suspected CD
|
Retrospective
|
78
|
39.8%
(56% for those with negative ileoscopy)
|
93%
|
80%
|
77%
|
94%
|
Retention (4%)
|
Adler et al 2012[21]
|
Perianal disease, normal ileo-colonoscopy/SBFT/CTE/MRE
|
Prospective
|
26
|
24%
|
–
|
–
|
–
|
–
|
0%
|
Kalla et al 2013[24]
|
Suspected CD
|
Retrospective
|
265
|
12%
|
–
|
–
|
–
|
–
|
–
|
Egnatios et al 2015[27]
|
Chronic abdominal pain
|
Retrospective
|
90
|
24.4% (27.1% with additional symptoms, 19.4% only pain)
|
–
|
–
|
–
|
–
|
None
|
Mitselos et al 2016[29]
|
Chronic abdominal pain and/or diarrhea
|
Retrospective
|
91
|
17.6%
|
63.6%
|
92.5%
|
–
|
–
|
–
|
Lee and Lim 2016[28]
|
Symptomatic patients with Isolated ileitis
|
Retrospective
|
137
|
85.4% (high with ileal ulcer/erosion and high ESR)
|
–
|
–
|
–
|
–
|
–
|
Song et al 2017[30]
|
Chronic diarrhea
|
Retrospective
|
91
|
42.9% (hematochezia and hypoalbuminemia were predictors)
|
–
|
–
|
–
|
–
|
1%
|
Huang et al 2018[32]
|
Chronic abdominal pain >3 mo
|
Retrospective
|
341
|
28.15% (33.3% for abdominal pain + associated symptoms) (half had inflammatory pathology)
|
–
|
–
|
–
|
–
|
0%
|
Magalhaes et al 2019[33]
|
Suspected CD
|
Prospective
|
220
|
44.5% (high CRP, low iron increased yield)
|
–
|
–
|
–
|
–
|
–
|
Min et al 2013[25]
|
Suspected CD (pediatric)
|
Retrospective
|
17
|
6%
|
–
|
–
|
–
|
–
|
0%
|
Gralnek et al 2012[23]
|
Suspected CD (pediatric)
|
Prospective
|
10
|
80%
|
–
|
–
|
–
|
–
|
0%
|
Argüelles-Arias et al 2004[9]
|
Suspected CD (pediatric, ≥12–16 y)
|
Prospective
|
12
|
58.3%
|
–
|
–
|
–
|
–
|
0%
|
Wu et al 2020[34]
|
Symptomatic patients (abdominal pain, obscure GI bleed, diarrhea etc.) (Pediatric)
|
Retrospective
|
825
|
19.9% CD
|
–
|
–
|
–
|
–
|
Retention requiring surgery (0.4%)
|
Nuutinen et al 2011[20]
|
Suspected CD (pediatric) (8–188 mo)
|
Retrospective
|
26
|
62%
|
–
|
–
|
–
|
–
|
0%
|
Moy and Levine 2009[17]
|
Suspected CD (pediatric) (growth failure)
|
Retrospective
|
7
|
57.1%
Improvement in height after small bowel CD treatment
|
–
|
–
|
–
|
–
|
1 retained in stomach
|
Cohen et al 2012[22]
|
Suspected CD (pediatric)
|
Retrospective
|
184
|
15%
|
–
|
–
|
–
|
–
|
1 retention
|
Esaki et al 2014[26]
|
Suspected CD
|
Retrospective
|
80
|
72.5%
|
–
|
–
|
–
|
–
|
6.3% retention
|
van Tuyl et al 2007[16]
|
Suspected CD
|
Retrospective
|
22
|
71% definitive diagnosis, 14% probable diagnosis
|
–
|
–
|
–
|
–
|
–
|
Mow et al 2004[11]
|
Suspected CD
|
Retrospective
|
8
|
37.5%
|
–
|
–
|
–
|
–
|
12.5%
|
Eliakim et al 2018[31]
|
Suspected CD
(panenteric capsule)
|
Prospective
|
7
|
57.1%
|
–
|
–
|
–
|
–
|
None
|
Broderson et al 2022[36]
|
Suspected CD (panenteric capsule)
|
Prospective
|
59
|
44.8%
Better image quality with increased volume of PEG but no change in diagnostic yield
(43.9% versus 47.1%)
|
–
|
–
|
–
|
–
|
Not mentioned clearly
|
Tai et al 2020[35]
|
Suspected CD (panenteric capsule)
|
Prospective
|
22
|
13.6%
|
–
|
–
|
–
|
–
|
0%
|
Abbreviations: CD, Crohn's disease; CRP, c-reactive protein; ESR, erythrocyte sedimentation
rate; NPV, negative predictive values; PEG, polyethylene glycol; PPV, positive predictive
values.
Results
Suspicion of CD is the most common indication of VCE in IBD as per a Spanish physician
survey.[37] The diagnostic yield was highly variable (6–80%) across these studies. This wide
variability can be explained by the heterogeneity of study design, variable definitions
of suspected CD, pretest probability in the subjects studied, and variable age groups
(includes pediatric). The probability was higher with an increasing number of symptoms,
elevated biochemical markers of inflammation, anemia, and hypoalbuminemia.
Three studies evaluated the sensitivity and specificity of VCE which were excellent
(77–93% and 80–89%, respectively). The positive predictive values (PPVs) and negative
predictive values (NPVs) varied from 50 to 77% and 92 to 96%, respectively. VCE retention
rates varied from 0 to 12.5%. Two meta-analyses have shown that VCE was superior to
small bowel follow through (SBFT) and ileo-colonoscopy (IC) and comparable to computed
tomography enterography (one meta-analysis showed inferiority to VCE and another comparable
to VCE)/MRE in the evaluation of suspected CD (see comparative yield section).[38]
[39]
3.1.1. Capsule Endoscopy Differentiating Crohn's Disease from Mimics (e.g., Small
Bowel Tuberculosis)
Capsule endoscopy findings should be interpreted in correlation with other findings
to differentiate from other similar appearing lesions like small bowel tuberculosis
(SBTB), nonsteroidal anti-inflammatory drugs enteropathy, Behcet's disease, vasculitis,
and also normal variation (10%). In a prospective study, out of 37 suspected CD patients
on VCE, only 13% were subsequently diagnosed to have CD on 1 year follow-up.[18] On the contrary, 19% patients with nonspecific enteritis develop CD on follow-up.
High baseline LS (>135) and clinical suspicion were predictors of the subsequent development
of CD.[40]
To address the aforementioned issue, a prospective study from India in which VCE was
done in 26 patients after proving bowel patency showed that ileo-cecal valve involvement
and aphthous ulceration were universal in SBTB (100% compared with 33% in CD) and
CD (100% compared with 25% in SBTB), respectively. Large ulcers were more common in
SBTB as compared with CD (75 vs. 47%).[41]
3.1.2. Factors Affecting Yield of Video Capsule Endoscopy in Suspected Crohn's Disease
Based on clinical symptoms, the combination of abdominal pain and diarrhea was shown
to be highly predictive of CD on VCE. Nearly, one-third with the combination of symptoms
had CD in this retrospective analysis.[42] The diagnostic yield with only chronic abdominal pain was 20.9% based on a systematic
review.[43] The independent predictive factors of proximal small bowel involvement were ileal
involvement, stricturing behavior, and significant weight loss.[44] In suspected CD with negative IC and SBFT, the combination of anemia and increased
platelet count was a significant predictor of CD on VCE.[45] Fecal calprotectin as a predictor of the lesion in VCE has shown variable results
in several studies with variable cutoffs.[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
Fecal calprotectin level >194 μg/g had a sensitivity and specificity of 47 and 90%,
respectively, for diagnosing CD on VCE.[57] On the contrary, fecal calprotectin <50 μg/g had a negative predictive value of
91.8% of having CD on VCE based on a systematic review.[52]
Two studies have evaluated multiple parameters for the prediction of CD. A Spanish
multicenter study developed and validated a scoring index based on fecal calprotectin
(score 10), c-reactive protein (CRP; score 6), thrombocytosis (score 3), anemia (score
2), leukocytosis (score 2), and high erythrocyte sedimentation rate.[1] Score ranges 0 to 5, 6 to 15, and ≥16 predicted low, intermediate, and high risk
of inflammatory lesions on VCE, respectively.[61]
Another elegant study has shown the incremental yield of VCE with an increasing number
of International Conference on Capsule Endoscopy (ICCE) criteria. Suspected CD was
defined as clinical symptoms (chronic abdominal pain/diarrhea, weight loss, and growth
failure) plus any one or more extraintestinal manifestations, inflammatory markers,
and abnormal imaging (SBFT/CTE). The prevalence of CD in those with suspected CD not
supported by ICCE criteria, two criteria, and three criteria was 21.4, 52.6, and 77.8%,
respectively. In those with LS ≥135, 82.6% had CD.[62]
3.1.3. Value of Repeat Video Capsule Endoscopy in Suspected Crohn's Disease
In a letter to the editor, Robertson et al have reported findings of 18 patients with
suspected CD who underwent repeat VCE on follow-up. Those with nonspecific inflammation
on initial VCE (33%) were more likely to have repeat VCE suggestive of CD (33%) along
with those with higher fecal calprotectin levels.[63]
3.1.4. Role of Video Capsule Endoscopy in Presymptomatic Patients
VCE has been used in first-degree relatives of CD patients to identify those with
subclinical small bowel inflammation. A cross-sectional study in 2017 by Teshima et
al showed increased intestinal permeability, and small bowel ulceration (≥3) was seen
in 30 and 24% of the first-degree relatives (n = 223) with CD. However, increased intestinal permeability did not correlate with
small bowel inflammation.[64] Later, another study by Taylor et al in 2019 involving 480 asymptomatic first-degree
relatives of CD has shown a risk tool comprising family history of CD, genetic variants
associated with CD, and high level of fecal calprotectin predicted risk of presymptomatic
small bowel inflammation.[65]
Role of Video Capsule Endoscopy in Spondyloarthropathy
Three prospective studies have shown a high yield of VCE (12.5–42.2%) to diagnose
CD in the established case of spondyloarthropathy (SpA) with bowel symptoms.[66]
[67]
[68] Elevated fecal calprotectin (>100 µg/g) was the predictor of small bowel CD (odds
ratio = 4.5).[68] In a case series of three juvenile idiopathic arthritis, all cases were diagnosed
to have CD.[69]
Video Capsule Endoscopy in Known Crohn's Disease
Study Selection and Study Characteristics
A total of 29 original articles were identified (12 prospective) evaluating the role
of VCE in known CD.
Results
Overall diagnostic yield varied from 52 to 88.3%. The yield was high for symptomatic
CD (highest for diarrhea—73%), whereas it was 21.1 and 4.7% only in those with clinical
remission and clinico-biochemical remission, respectively.[16]
[20]
[22]
[23]
[24]
[25]
[26]
[31]
[35]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80]
[81]
[82]
[83]
[84]
[85]
[86]
[87]
[88] The incremental yield over SBFT was 32 to 32.7% and 7% over IC. In a prospective
study in pediatric CD, the diagnostic yield of VCE (41%) was higher than magnetic
resonance imaging (MRI)/small intestinal contrast ultrasonography (SICUS; 18.2%).[75] The incremental yield in the proximal small bowel was 28 to 50%.[23]
[83]
[85] The overall change in management after VCE varied from 21 to 71% which included
treatment escalation, deescalation, initiation of new medications (biologics/immunomodulators),
and decision for surgery.[16]
[24]
[85]
[87] Other implications of VCE in known CD are reclassification of disease phenotype
(11%), assessment of mucosal healing posttherapy, and prediction of relapse (higher
with jejunal disease).[78]
[82]
[84] Clinical and biochemical improvement can predict mucosal healing in less than half
(none at 2 weeks, 42% at 52 weeks) of the patients.[89]
[90]
[91] The rate of retention varied from 2.1 to 18.6% ([Table 2]).
Table 2
Summary of studies evaluating role of video capsule endoscopy in known Crohn's diseases
Author
|
Indication
|
Study type
|
N
|
Yield/Incremental yield
|
Impact on management
|
Adverse events/retention
|
Cotter et al 2014
|
Known CD
|
Retrospective
|
50
|
Incremental yield- 66%
|
Initiation of immunomodulators/biologics increased by 26%
|
6% retention
|
Dussault et al 2013
|
Known CD
|
Retrospective
|
77
|
Overall yield - 62%
|
53.5%
60% in unexplained anemia
58% when performed for assessing disease location
20% when performed for discordance between symptoms and morphology
|
4.2% transient retention after negative patency testing
|
Elosua et al 2022
|
Known CD
|
Retrospective
|
432
|
Overall yield - 63.7%
|
51.4%, Escalation- 46.1%, De-escalation-5.3%
Escalation:89.5% in moderate-severe disease, 57.8% mild disease
|
2.5% retention, all managed nonsurgically
|
Flamant et al 2013
|
Known CD
|
Retrospective
|
108
|
Jejunal lesions -56%
|
Increased risk of relapse with jejunal lesions
|
5.5% retention
|
Hansel et al 2018
|
Known CD
|
Prospective
|
50
|
Proximal small bowel incremental yield- 28%
|
Altered management- 34%
New medication initiated- 29%
Exclusion of active small bowel disease- 24%
|
None reported except dysphagia in one
|
Kopylov et al 2015
|
Known CD
|
Retrospective
|
187
|
Overall yield- 71.6%
|
52.3%
|
Retention 2.1%
|
Mehdizadeh et al 2010
|
Known CD
|
Retrospective
|
134
|
Overall yield - 52% (highest for diarrhea -73%)
Incremental yield- 32% (to SBFT)
7% (to IC)
|
Not evaluated
|
None
|
Melmed et al 2018
|
Known CD
|
Prospective
|
53
|
Proximal small bowel -85%
|
Not evaluated, high correlation with IC
No correlation with clinical severity indices (CDAI)
|
None were capsule related
|
Niv et al 2014
|
Known CD
|
Prospective
|
19
|
78.9% at week 0
84.6% at week 4
|
Not evaluated, No correlation with sequential clinical severity indices (CDAI)
|
No retention
|
Park et al 2007
|
Known CD
|
Retrospective
|
52
|
32.7% (over-SBFT)
|
28.8%
|
Retention-9.6%
Surgery- 3.8%
|
Petruzziello et al 2010
|
Known CD
|
Prospective
|
32
|
50% (in CD involving distal ileum)
|
Not evaluated
|
3% retention
|
Santos-Antunes et al 2015
|
Known CD
|
Retrospective
|
106
|
Proximal small bowel -46% (incremental yield)
|
40%
Immunomodulator therapy post-VCE 44 vs. 21% pre VCE
|
None
|
Lorenzo-Zu ́n ̃iga et al 2010
|
Known CD
|
Prospective
|
14
|
85.7%
|
64%
|
None
|
Long M et al 2011
|
Known CD, indeterminate colitis and pouchitis
|
Retrospective
|
86 (CD)
|
77.9%
|
Change in medication: 51.1%
New IBD medication: 39.5%
Surgery: 12.8%
|
16 cases of retention
8 required operative intervention
|
Nardo et al 2011
|
Known pediatric CD
|
Prospective
|
44
|
41% with VCE vs. 18.2% with MRI/SICUS
|
Not evaluated
|
None
|
Kalla et al 2013
|
Known CD
|
Retrospective
|
50
|
66%
|
Management was altered in 48%
|
None
|
Greener et al 2016
|
Known CD
|
Prospective
|
56
|
51%
|
Reclassification of disease phenotype in 11%
|
1 patient had temporary patency capsule retention
|
Min et al 2013
|
Known pediatric CD
|
Retrospective
|
50
|
70% (43% extensive disease compared with other imaging)
|
75%
Improved growth, BMI, HBI, and ESR on follow-up
|
None
|
Gralnek et al 2012
|
Known pediatric CD
|
Prospective
|
4
|
50% more proximal involvement detected
|
75% change in management
|
None
|
Oliva et al 2019
|
Known pediatric CD
|
Prospective
|
48
|
71% (panenteric capsule endoscopy)
|
71% change in management
Mucosal healing achieved at 24 and 52 wk were 54 and 58% respectively compared with
21% at baseline
|
No serious event except for nausea and vomiting in 3 patients
|
Nuutinen et al 2011
|
Known pediatric CD
|
Retrospective
|
9
|
56%
|
–
|
1 had retention requiring elective surgery
|
Cohen et al 2012
|
Known pediatric CD
|
Retrospective
|
61
|
57.3%
|
–
|
9.8%
|
Cohen et al 2008
|
Known pediatric CD
|
Retrospective
|
21
|
62%
|
23.8% change in management
|
1 had retention
|
Esaki et al 2014
|
Known CD
|
Retrospective
|
94
|
88.3%
|
75% (8/12) of colonic disease reclassified as ileo-colonic disease
|
7.4% retention
|
Tuyl et al 2007
|
Known CD
|
Retrospective
|
14
|
71% definitive diagnosis
14% probable diagnosis
|
21% changed management
|
None
|
Mow et al 2004
|
Known CD
|
Retrospective
|
20
|
70%
|
60%
|
5% retention
|
Eliakim et al 2018
|
Known CD
|
Prospective
|
29
|
55.17%
(panenteric capsule)
31% proximal disease
|
Not evaluated
|
None
|
Tai et al 2020
|
Known CD
|
Prospective
|
71
|
67.6%
|
38.7% had change in management
Upstaging of Montreal classification 33.8%
Mucosal healing 15.5%
|
2.8%
|
Kopylov et al 2015
|
Known CD (clinical remission or mild symptoms)
|
Prospective
|
56
|
21.1% in those in clinical remission
4.7% in clinico-biochemical remission
|
Mucosal healing and deep remission are rare in CD in clinical remission and hence
may require escalation of therapy
|
0%
|
Abbreviations: BMI, body mass index; CD, Crohn's disease; CRP, c-reactive protein;
ESR, erythrocyte sedimentation rate; HBI, harvey bradshaw index; MRI, magnetic resonance
imaging; NPV, negative predictive values; PEG, polyethylene glycol; PPV, positive
predictive values; SBFT, small bowel follow through; SICUS, small intestinal contrast
ultrasonography; VCE, video capsule endoscopy.
3.3. Video Capsule Endoscopy Compared with Other Imaging Technologies in Suspected
or Known Crohn's Disease
Study Selection and Study Characteristics
In total, 9 studies including 378 patients compared the diagnostic yield of VCE with
various other modalities (IC, SBFT, CTE, MRE, SICUS) in suspected CD.
Results
VCE was better than all the other modalities with regard to diagnostic yield except
MRE ([Table 3])[29]
[92]
[93]
[94]
[95]
[96]
[97]
[98]. An earlier meta-analysis and a recent meta-analysis have also shown the same for
both suspected and established CD.
Table 3
Video capsule endoscopy compared with other diagnostic modalities in suspected and
known Crohn's disease
Author
|
Study type
|
N
|
Indication
|
Diagnostic yield
|
Comparator
|
Diagnostic yield of competing technology
|
Eliakim et al 2002
|
Prospective
|
20
|
Suspected CD
|
70%
|
Barium meal follow through (BMFT)
Entero-CT
|
BMFT: 37%
EnteroCT: 50%
|
Di Nardo eta 2010
|
Prospective
|
18
|
Suspected IBD
|
50%
|
MRI and/ or SICUS
|
22.2%
|
Eliakim et al 2004
|
Prospective
|
35
|
Suspected CD
|
77%
|
Barium meal follow through (BMFT)
Entero-CT
|
BMFT: 23%
EnteroCT: 20%
|
Voderholzer et al 2004
|
Prospective
|
41
|
Known CD
|
60.9
|
CTE
|
29.2%
|
Albert et al. 2005
|
Prospective
|
27
|
Suspected CD (n = 14) + known CD (n = 13)
|
93%
|
MRI
Fluoroscopic enteroclysis
|
MRI: 78%
Enteroclysis: 33%
|
Chong et al 2005
|
Prospective
|
43
|
Suspected CD (n = 21) + known CD (n = 22)
|
Suspected CD:19%
Known CD 77%
Change in management 70%
|
Push enteroscopy
Enteroclysis
|
Enteroclysis (19% known CD, 6% suspected CD)
Push enteroscopy: 16% known CD, 0% suspected CD)
|
Marmo et al 2005
|
Prospective
|
31
|
Known CD
|
71%
89% (with terminal ileal involvement)
46% (proximal small bowel)
|
Enteroclysis
|
25.8%
37% (with terminal ileal involvement)
13% (proximal small bowel)
|
Hara et al 2006
|
Prospective
|
17
|
Suspected CD
|
71%
|
IC
CTE
SBFT
|
IC: 65%
CTE: 53%
SBFT: 24%
|
Efthymiou et al 2008
|
Prospective
|
55
|
Suspected CD (n = 26) + known CD (n = 29)
|
Suspected CD:65%
Known CD 74.1%
|
Enteroclysis
|
Enteroclysis (3% suspected CD, 40.7% known CD) (p < 0.05)
|
Solem et al 2008
|
Prospective
|
28
|
Suspected CD + known CD
|
Sensitivity:83%
Specificity:53%
|
CTE
IC
SBFT
|
CTE
Sensitivity: 67%
Specificity:100% (p = 0.02)
IC
Sensitivity:67%
Specificity:100% (p = 0.03)
SBFT
Sensitivity:50%
Specificity:100% (p = 0.2)
|
Bocker et al 2010
|
Prospective
|
21
|
Suspected or known CD
|
42.9%
|
MRI
|
28.6%
|
Petruzziello et al 2010
|
Prospective
|
30
|
Suspected CD
|
50%
Incremental yield: 33%
|
IC
SICUS
SBFT
|
IC: 63%
SICUS: 40%
SBFT: 50%
|
Casciani E et al 2011
|
Prospective
|
37
|
Pediatric suspected CD
|
Sensitivity: 97.6%
Specificity: 92.3%
Accuracy: 98.3%
|
MRE
|
Sensitivity: 91.9%
Specificity: 90.9%
Accuracy: 100%
|
Jensen et al 2011
|
Prospective
|
93
|
Suspected or newly diagnosed CD
|
Sensitivity: 100%
Specificity: 91%
(terminal ileum)
|
CTE
MRE
|
CTE
Sensitivity: 76%
Specificity: 85%
MRE
Sensitivity: 81%
Specificity: 86%
|
Wiarda et al 2011
|
Prospective
|
38
|
Suspected CD (n = 20) + known CD (n = 18)
|
Sensitivity: 57%
Specificity: 89%
PPV: 67%
NPV: 84%
|
MRE
|
Sensitivity: 73%
Specificity: 90%
PPV: 88%
NPV: 78%
|
Kovanlikaya et al 2013
|
Retrospective
|
23
|
Children with suspected or known IBD
|
Sensitivity 77.8%
|
MRE
|
Sensitivity: 75%
|
Aloi M et al, 2014
|
Prospective blinded, comparative
|
25
|
Pediatric suspected CD (n = 6) + known CD (n = 28)
|
Sensitivity in jejunum, proximal /midileum and terminal ileum were 92, 100, and 81%,
respectively
|
MRE
SICUS
|
Sensitivity:
MRE: jejunum: 75%, Proximal/mid ileum: 100%, terminal ileum: 81%
SICUS: jejunum: 92%, Proximal/mid ileum: 80%, terminal ileum: 94%
Specificity of capsule lower in jejunum (61%) and proximal/mid ileum (74%) but higher
in terminal ileum (90%)
|
Leighton et al 2014
|
Prospective
|
80
|
Suspected small bowel CD
|
VCE + IC (small bowel + colon)- 97.3%
VCE (terminal ileum+ cecum) - 49.2%
VCE (small bowel)- 93%
|
1. IC (terminal ileum + cecum)
2. SBFT + IC (small bowel + colon)
3. SBFT (small bowel)
|
1. IC (terminal ileum + cecum)
−70.5% (p = 0.09)
2. SBFT+ IC (vs. IC + VCE) − 57.3% (p < 0.001)
3. SBFT (small bowel) (vs. VCE) − 25.6%
|
Oliva et al 2015
|
Prospective
|
40
|
Pediatric known CD
|
Colon
Sensitivity: 89%
Specificity:100%
PPV: 100%
NPV: 91%
Small bowel
Sensitivity: 90%
Specificity: 94%
PPV: 95%
NPV: 90%
|
SICUS
MRE
|
SICUS (small bowel)
Sensitivity: 90%
Specificity: 93%
MRE (small bowel)
Sensitivity: 85%
Specificity: 89%
|
Leighton et al 2016
|
Prospective
|
114
|
Known active CD with proven bowel patency
|
Panenteric capsule endoscopy: 83.3%
|
IC
|
69.7%
|
Mitselos et al 2016
|
Retrospective
|
91
|
Suspected CD
|
Sensitivity: 81.82%
Specificity: 77.50%
PPV: 53.85%
NPV: 94.87%
AUC: 0.781
|
IC
|
Sensitivity: 63.64%
Specificity: 92.50%
PPV: 33.33%
NPV: 96.88%
AUC: 0.797
|
Carter D et al, 2018
|
Prospective
|
50
|
Suspected CD- negative ileocolonoscopy
|
38%
|
IUS
|
38%, Sensitivity: 72%, specificity: 84% compared with capsule endoscopy which was
considered gold standard
|
Gonzalez-Suarez et al 2018
|
Prospective
|
47
|
Suspected CD (n = 15) + known CD (n = 32)
|
76.6% (higher in jejunal, ileal and terminal ileal lesions)
|
MRE
|
44.7% (capsule significantly better for superficial and proximal lesions)
|
Hijaz et al 2019
|
Prospective
|
27
|
Children with CD or indeterminate colitis
|
Sensitivity: 83%
Specificity: 78.6%
|
MRE
|
Sensitivity: 100%
Specificity: 57.14%
Capsule has lower sensitivity but high specificity
|
Bruining DH et al 2020
|
Prospective
|
99
|
Known CD (Panenteric capsule)
|
Sensitivity 94% (proximal small bowel 97%)
Specificity: 74%
|
MRI + IC
|
Sensitivity: 100%
Specificity: 22% (p = 0.021; similar specificity in terminal ileum and colon)
|
Yamada et al 2021
|
Prospective
|
20
|
Known CD, Colon capsule endoscopy
|
Diagnostic accuracy (detecting ulcers)
Small bowel-88.3%
Large bowel-78.1%
|
Bidirectional double balloon enteroscopy
|
Reference standard
|
Dubsenco et al 2005
|
Prospective
|
39
|
Known and suspected CD
|
Sensitivity: 89.6% Specificity-100.0%, Positive predictive value:100%
Negative predictive value: 76.9
|
Small bowel series
|
Sensitivity: 27.6%
Specificity: 100.0%
PPV: 100.0%
NPV: 32.3%.
|
D′Haens et al 2015
|
Prospective
|
40
|
Colonic CD (paneneteric capsule PCCE-2)
|
Sensitivity: 86%
Specificity: 40%
|
Colonoscopy
|
Reference standard
Better estimated the disease severity compared with PCCE-2
PCCE-2 better tolerated
|
Papalia et al 2021
|
Prospective
|
47
|
Known ileo-colonic, nonstricturing CD for mucosal healing
|
Strong correlation with SES-CD scores in colonoscopy (r = 0.77), strongest in terminal ileum
|
Colonoscopy
|
PCCE-2 identified additional ulcers
PCCE-2 was complete in 68% cases compared with 89% colonoscopy
PCCE-2 noninvasive modality for monitoring
|
Ben Horin et al 2019
|
Prospective
|
61
|
Clinically quiescent known CD for predicting flare
|
Lewis score >350 predicted risk of flare with AUC 0.79
|
Fecal calprotectin
MRE
|
Fecal calprotectin (AUC)
2 y flare: 0,62
6 mo flare: 0.81
MRE risk prediction (AUC)
2 y:0.71
|
Abbreviations: AUC, area under the curve; CD, Crohn's disease; IC, ileo-colonoscopy;
CTMRI, magnetic resonance imaging; IUS, intestinal ultrasound; MRE, magnetic resonance
enterography; PCCE, panenteric colon capsule endoscopy; SES- CD, simple endoscopic
score Crohn's disease; SICUS, small intestinal contrast ultrasonography.
Similar results were found in known CD (10 articles and 2 meta-analysis)[3]
[99]
[100]
[101]
[102]
[103]
[104]
[105]
[106]
[107]. However, SICUS had comparable sensitivity (90%) and specificity (93%) to VCE for
small bowel involvement in pediatric CD.[102] Panenteric colon capsule endoscopy (PCCE) has been compared with MRE plus IC and
was found that MRE + IC had 100% sensitivity (94% with VCE) but low specificity (22%
compared with 74% with PCCE).[106] PCCE has been compared with reference endoscopic standard (bidirectional double
balloon enteroscopy), and diagnostic accuracy was 88.3 and 77.1% for small and large
bowel, respectively.[3] Compared with colonoscopy, the sensitivity and specificity of CCE were 86 and 40%,
respectively.[101] Moreover, risk prediction for future flare was better with the VCE LS (AUC—area
under the curve: 0.79) compared with MRE risk prediction (AUC: 0.71). Among 11 studies
(n = 439) comprising both suspected and known CD, VCE had higher sensitivity compared
with CTE, IC, and SBFT but lower specificity (53% compared with 100% with others).[108]
[109]
[110]
[111]
[112]
[113]
[114]
[115]
[116]
[117]
[118] Comparing the specificity of VCE according to the regions of small intestine, the
specificity of VCE was lower in jejunum (61%) and proximal/mid-ileum (74%), but higher
in terminal ileum (90%) compared with MRE and SICUS ([Table 3]).[117] VCE was better compared with MRE for proximal and superficial lesions.
3.4. Risk of Retention in Suspected or Known Crohn's Disease
The risk of retention of VCE is 13% in established CD and 1.6% in suspected CD based
on the earlier retrospective study by Cheifetz et al.[119] The overall retention rate in CD is 3.32% (2.35% suspected CD and 4.63% known CD)
based on an updated systematic review and meta-analysis.[120] This is the reason why VCE is not generally preferred in established CD except for
patients with anemia, obscure GI bleed, or assessment of mucosal healing. Retention
rates were lower in pediatric CD (1.64%) compared with adult CD (3.49%).[120]
3.4.1. Predictors of Capsule Retention
Negative patency capsule testing and negative CT/MRE are negative predictors of capsule
retention. The retention rates of patency capsule varied between 15.2 and 28% in known
CD.[121]
[122] A retrospective study showed that the retention rates after positive and negative
patency capsule testing were 11 and 2.1%, respectively.[123] Similarly a prospective study showed 28.9% retention rates with active inflammation
on MRE.[122] Based on meta-analysis, the retention rates in established CD remained 2.88% even
after patency capsule testing and 2.32% after CT/MRE. The risk of retention is reduced
after CT/MRE by 50%. MRE has a sensitivity of 92.3% and specificity of 59% to evaluate
capsule retention as compared with 97 and 83%, respectively, for patency capsules.
MRE tends to overdiagnose the risk of capsule retention. If only MRE is used to predict
retention instead of patency capsule, nearly 40% patients (specificity—59%) would
not undergo VCE due to fear of capsule retention.[122]
Other predictors were obstructive symptoms, stricturing/penetrating disease, BMI≤
5th percentile, suspected stenosis on SBFT, restricted diffusion on diffusion-weighted
MRI, extensive small bowel thickening on small bowel ultrasound, longer stricture
length and higher number of prestenotic dilatations, high CRP, and history of abdominal
surgery according to several studies[119]
[120]
[121]
[122]
[123]
[124]
[125]
[126]
[127]
[128]
[129]
[130]
[131]
[132]
[133]
[134]
[135]
[136] ([Supplementary Table S1], available in the online version).
3.4.2. Management of Retained Capsule
Capsule retention is asymptomatic in 85% which can be managed conservatively with
enteroscopy-guided removal electively. Partial air complete small bowel obstruction
occurs in the rest which requires endoscopic retrieval with or without balloon dilation.
There are several reports of retrieval by double-balloon enteroscope (success rate
80–92%) and recently novel motorized spiral enteroscopy which can avoid surgery in
the majority.[137]
[138]
[139] Moreover, it can help take surgical decisions as some of the patients may require
surgery even after capsule removal for treating the underlying disease.
3.4.3. Other Complications of Video Capsule Endoscopy
Although known CD substantially increases the risk of capsule retention, other complications
of VCE like swallow disorder, aspiration, and technique failure can be substantial
and clinically important. However, we could not find specific citations pertaining
to IBD with regard to this. Meta-analysis, which included IBD patients, showed that
the pooled rates of aspiration, technical failure, and swallow disorders were 0, 0.94,
and 0.75%, respectively.[140]
3.4. Video Capsule Endoscopy in postoperative Crohn's Disease
Study selection and study characteristics
Five prospective and one retrospective study including 313 patients have compared
VCE with colonoscopy in postoperative CD ([Supplementary Table S2], available in the online version).
Results
Except for the one retrospective study,[141] all the studies concluded that the yield of VCE was higher than colonoscopy in detecting
postoperative recurrence especially for proximal involvement out of the reach of the
colonoscopy.[142]
[143]
[144]
[145]
[146]
[147] Another prospective study has compared capsule endoscopy or no capsule endoscopy
in postoperative settings and has shown that VCE arm had lower hospitalization or
surgery[147] ([Supplementary Table S2], available in the online version).
3.5. Video Capsule Endoscopy in UC
Study Selection and Study Characteristics
In total, 14 prospective and 3 retrospective studies involving 612 patients have evaluated
the role of VCE in ulcerative colitis (UC) ([Table 4]).[11]
[20]
[23]
[31]
[75]
[148]
[149]
[150]
[151]
[152]
[153]
[154]
[155]
[156]
[157]
[158]
[159]
Table 4
Summary of studies evaluating role of video capsule endoscopy in ulcerative colitis
Author
|
N
|
Study type
|
Small bowel involvement
|
Correlation with colonoscopy
|
Cleanliness
|
Adverse events
|
Ye et al 2013
|
26
|
Prospective
|
|
Excellent correlation with extent and severity of UC
|
80%
|
None
|
Higurashi et al 2011
|
23 UC and 23 healthy volunteers
|
Prospective
|
57%, correlated with disease activity
|
Not studies
|
Not evaluated
|
None
|
Hisabe et al 2011
|
30
|
Prospective
|
36.6% (40% in active UC and 33% in postproctocolectomy) (extensive colitis, pouchitis
and age <20 y are predictors)
|
Not evaluated
|
–
|
None
|
Hosoe et al 2013
|
42
|
Prospective
|
Not evaluated
|
Strong correlation with colonoscopy
|
<50% good or excellent cleansing level
|
None
|
Juan Acosta et al 2014
|
42
|
Prospective
|
3 out of 42 patients
|
Good correlation for severity and extent of inflammation
|
80%
|
None
|
Matsubayashi et al 2020
|
41
|
Prospective
|
Not evaluated
|
Capsule Scoring of Ulcerative Colitis (CSUC) better than fecal biomarkers for predicting
relapse
|
Not mentioned
|
None
|
Meister et al 2013
|
13
|
Prospective
|
Not evaluated
|
Colonoscopy detected vessel vulnerability, granulated mucosa, mucosal damage and disease
extension better than capsule endoscopy
|
90% good or fair
|
None
|
Okabayashi et al 2018
|
33
|
Prospective
|
Not evaluated
|
Good correlation with endoscopic indices of severity. Active disease had longer transit
time with resultant poor acceptability
|
77.2% acceptable
|
5.1% from laxatives, 7.7% delayed excretion (> 24 hours)
|
Oliva et al 2014
|
30, Pediatric
|
Prospective
|
Not evaluated
|
High sensitivity (96%), specificity (100%), positive predictive value (100%), negative
predictive value (85%)
|
62% adequate, 24% fair
|
None
|
Shi et al 2017
|
150
|
Prospective
|
Not evaluated
|
Good correlation (R = 0.64–0,67) for severity of mucosal inflammation (sensitivity: 97%)
|
66%
|
21% mainly related to bowel preparation, one serious adverse event due to retention
by unexpected rectal tumor
|
Sung et al 2012
|
100
|
Prospective
|
Not evaluated
|
High sensitivity (89%), specificity (75%), positive predictive value (93%), negative
predictive value (65%)
|
64% acceptable, 31% fair
|
No serious adverse events, all related to bowel preparation
|
Nardo et al 2011
|
29
|
Prospective
|
No incremental yield
|
Not evaluated
|
Not evaluated
|
None
|
Gralnek et al 2012
|
2
|
Prospective
|
50%, diagnosis changed to CD
|
Not evaluated
|
Not evaluated
|
None
|
Nuutinen et al 2011
|
21
|
Retrospective
|
4.8%
|
–
|
–
|
None
|
Cohen et al 2011
|
5
|
Retrospective
|
80% (UC with suspicion of CD)
|
–
|
–
|
None
|
Mow et al 2004
|
20
|
Retrospective
|
59%
|
Not evaluated
|
Not evaluated
|
None
|
Eliakim et al 2018
|
5
|
Prospective
|
40% (panenteric capsule)
|
Not evaluated
|
Not mentioned separately
|
None
|
Abbreviation: CD, Crohn's disease.
Results
Overall results indicate that VCE had an excellent correlation with colonoscopy for
severity/extent of inflammation and is better than fecal biomarkers. Patient acceptability
was better than colonoscopy. Small bowel involvement in UC with VCE is variable (4.8–80%)
and is dependent on the pretest probability (80% for those with suspicion of CD in
a small series).[11]
[23]
[31]
[149]
[150]
[156] In postproctocolectomy cases, extensive colitis, pouchitis, and age less than 20
years were predictors of small bowel involvement.[150] Active UC was also a predictor of small bowel involvement (40% compared with overall
36.6%).[150] No adverse events are reported except those related to bowel preparation and one
case of retention due to unexpected rectal tumor ([Table 4]).[157] Bowel preparation was acceptable in 62 to 90% cases ([Table 4]).
3.5.1. Role of Video Capsule Endoscopy in Pouchitis
A single-center prospective study has shown that all of the patients with chronic
antibiotic refractory pouchitis have small bowel lesions from duodenum to ileum detectable
on VCE which ranges from aphthous to deep, fissuring ulcers. None of the patients
have any prior evidence of CD on review of surgical biopsy. These patients need to
be followed up further, and the significance of such lesions is still unknown.[160] In a retrospective study, small bowel capsule endoscopy in pouchitis showed positive
findings in 65.2%. Initiation of new IBD medications was noted in 56.5%, and small
bowel resection was done in 4.4% following VCE.[76]
3.6. Video Capsule Endoscopy in Inflammatory Bowel Disease-Unclassified
Study Selection and Study Characteristics
VCE could be particularly helpful in the IBD unclassified subgroup, where up to 16.7
to 50% patients can be diagnosed with CD after undergoing VCE as per four retrospective
and five prospective studies in 177 adult and pediatric patients ([Supplementary Table S3], available in the online version).[23]
[24]
[25]
[75]
[148]
[161]
[162]
[163]
[164]
Results
However, VCE can miss a diagnosis of CD as five of the aforementioned studies have
shown that 0 to 16.7% patients develop CD on follow-up.[23]
[148]
[161]
[162]
[164] A change in existing treatment after CD diagnosis may not be necessary in all the
patients as the reported change in treatment after CD diagnosis was seen in 0 to 100%.[24]
[163]
[164] A confirmed diagnosis of UC after exclusion of small bowel involvement in IBD-U
can occur in 5.5 to 59.3%.[23]
[161]
[162]
[163]
[164] No change in diagnosis of IBD-U can occur in 0 to 75% cases ([Supplementary Table S3], available in the online version). LS >135 is a predictor of CD diagnosis with a
sensitivity and specificity of 90 and 100%, respectively.[164]
3.7. Scoring Systems for Video Capsule Endoscopy
Scoring systems in VCE may help predict disease severity and disease course similar
to conventional endoscopic scoring. There are various validated scores for small bowel,
colon, and panenteric evaluation.
3.7.1. Lewis Score
Study Selection and Study Characteristics
Total 9 studies (2 prospective) including 811 patients evaluated the role of the LS
alone.[62]
[165]
[166]
[167]
[168]
[169]
[170]
[171]
[172]
Results
LS was initially developed by Gralnek et al based on edema, ulceration, and stenosis
in three tertiles with the establishment of cut-off values ([Table 5]; [Fig. 2]).[165] Later, the incremental number of ICCE criteria was found to be the predictive factor
of significant inflammatory activity (LS > 135) on VCE.[62] LS ≥ 135 was shown to have a positive predictive value of 73.9%, and a score <135
had a negative predictive value of 91.8%.[168] Similar findings were seen in another validation study. A strong agreement was seen
for global as well as for each tertile score in interobserver study.[166] Recent studies have evaluated the prognostic role of LS to predict CD-related emergency
hospitalization and risk of cumulative relapse.[171]
[172] Correlation of VCE with disease activity and small bowel transit time was weak in
adults, whereas correlation with inflammatory markers was moderate in both pediatric
age group and adults.[169] LS score correlates well with the MRE global score (r = 0.71) except the proximal LS score (r = 0.55).[170]
Fig. 2 Video capsule endoscopy in Crohn's disease. (A) Aphthous ulcers, (B) linear ulcer, (C) transverse hemicircumferential ulcer, (D) circumferential ulcerated stricture, (E) fibrotic stricture, and (F) mucosal edema.
Table 5
Summary of studies on video capsule endoscopy scoring systems
Author and year
|
Scoring system
|
Area of bowel
|
N
|
Study type
|
Study objective
|
Results
|
Granlek et al 2008
|
LS
|
Small bowel
|
44
|
Prospective
|
Development of scoring index
|
Based on villous edema, ulcer and stenosis in three tertiles, LS was developed, score
<135: clinically insignificant,135–790: mild, ≥790: moderate to severe.
|
Rosa et al 2012
|
LS
|
Small bowel
|
56
|
Retrospective
|
Usefulness of LS in suspected CD as per the ICCE criteria
|
Patients with suspected CD based on one or more ICCE criteria were more likely have
inflammatory activity (LS > 135) compared with those in whom none of the ICCE criteria
were present (incremental increase with increase in the number of ICCE criteria)
|
Cotter et al 2015
|
LS
|
Small bowel
|
70
|
Retrospective
|
Interobserver agreement
|
Strong interobserver agreement in each tertile and global score (k = 0.852–0.960; p < 0.0001)
|
Monteiro et al 2015
|
LS
|
Small bowel
|
95
|
Retrospective
|
Diagnostic accuracy of the LS in patients with suspected CD undergoing capsule endoscopy
|
LS > 135 had an overall diagnostic accuracy of 83.2% with a sensitivity, specificity,
positive predictive value, and negative predictive value of 89.5, 78.9, 73.9, and
91.8%, respectively for the diagnosis of CD.
|
De Castro et al 2015
|
LS
|
Small bowel
|
53
|
Retrospective
|
Assess prognostic value of the severity of inflammatory lesions quantified by the
LS
|
Increased need for steroid (RR: 5) and hospitalization (RR: 13.7) on multivariate
analysis
|
He et al 2017
|
LS
|
Small bowel
|
150
(30 pediatric)
|
Retrospective
|
Correlation of LS with disease activity, inflammatory markers and small bowel transit
time (SBTT)
|
Correlation with disease activity: moderate (pediatric), weak (adults)
Correlation with inflammatory markers: Moderate (both)
Correlation with SBTT: none (pediatric), weak (adults)
|
Nishikawa et al 2019
|
LS
|
Small bowel
|
125
|
Retrospective
|
Predicting emergency hospitalization and clinical relapse
|
An LS of 264 was an useful cutoff value that could predict CD-related emergency hospitalization
and cumulative risk of relapse (AUC: 0.92)
|
Nishikawa et al 2021
|
LS
|
Small bowel
|
102 (retrospective) + 66 (prospective)
|
Retrospective +prospective
|
Predicting emergency hospitalization and clinical relapse based on retrospective analysis
followed by prospective validation
|
LS ≥ 270 or prognostic nutrition index (PNI) < 45 had a significantly higher risk
of Crohn's disease-related emergency hospitalization. Additional treatment in these
patients
|
Klang et al 2018
|
LS for validation of MRE global score
|
Small bowel
|
50
|
Prospective
|
Validation of MRE global score using LS and fecal biomarkers
|
Significant correlation of LS with global MRE score (r = 0.71, p < 0.001), the correlation of MRE global score with Proximal LS score (r = 0.55). Correlation with fecal calprotectin was higher with MRE global score compared
with LS
|
Gal et al 2008
|
CECDAI
|
Small bowel
|
20
|
Prospective
|
Assessment and validation
|
Strong interobserver agreement (k = 0.87) Convenient, reliable and reproducible diagnostic and follow-up tool
|
Niv et al 2012
|
CECDAI
|
Small bowel
|
62
|
Prospective
|
Validation of CECDAI score
|
The correlation between endoscopists between different centers was good (r = 0.767)
|
Miyazu et al 2021
|
CECDAI
|
Small bowel
|
21
|
Prospective
|
To assess use of CECDAI to predict need of additional treatment for patients in clinical
remission
|
CECDAI was useful in assessing requirement of additional treatment for CD patients
in clinical remission (more in those with CECDAI ≥ 5.8)
|
Koulaouzidis et al 2012
|
LS and CECDAI
|
Small bowel
|
49
|
Retrospective
|
Comparison of correlation with fecal calprotectin with LS and CECDAI
|
In patients with fecal calprotectin <100 μg/g, correlation was better with LS compared
with CECDAI. In patients with elevated fecal calprotectin (>100 µg/g), neither LS
and CECDAI correlated with fecal calprotectin.
|
Omori et al 2019
|
CECDAI and LS
|
Small bowel
|
132
|
Retrospective
|
Compare the usefulness of CECDAI and LS
|
CECDAI better reflect the status and severity of intestinal inflammation than LS
Those with high LS but normal CECDAI may reflect strictures rather than active inflammation
|
Ponte et al 2017
|
CECDAI and LS
|
Small bowel
|
53
|
Retrospective
|
To identify cut off values of CECDAI as corresponding to LS cut offs
|
LS threshold values of 135–790 in LS corresponds to CECDAI cutoff values of 7.7–10.3,
both scores did not have any correlation to CRP or Harvey–Bradshaw index
|
Elaikim et al 2020
|
Elaikim score
|
Panenteric
|
41
|
Substudy of a RCT
|
Correlation with LS and reliability
|
Excellent interobserver agreement (k = 0.9) and strong correlation with calprotectin levels (r = 0.54) which was better than with LS (r = 0.32)
|
Niv et al 2016
|
CECDAIic
|
Small bowel and colon
|
10
|
Prospective
|
Extension of Niv score into colon to establish a new score for small bowel and colon
|
Concordance high (0.85 for small bowel and 0.77 for entire bowel) except for proximal
small bowel and distal colonic strictures
|
Arieira et al 2019
|
CECDAIic
|
Panenteric
|
22
|
Retrospective
|
Interobserver agreement and the correlation with inflammatory parameters.
|
Excellent interobserver agreement (k = 0.94) and strong correlation with calprotectin levels (r = 0.82) moderate correlation with CRP (r = 0.5)
|
Hosoe et al 2018
|
Capsule Scoring of Ulcerative Colitis (CSUC)
|
Large bowel
|
40
|
Prospective
|
Development of endoscopic score for UC with colon capsule endoscopy 2
|
Correlation of newly developed CSUC (score 0–14 based on with fecal calprotectin and
Lichtiger index)
|
Macedo Silva et al 2022
|
APEX score
|
Small bowel
|
47
|
Retrospective
|
Prediction of flare in small bowel CD
|
Age ≤ 30 y (+ 3 points), platelet count ≥ 280 × 103/L (+2 points) and extraintestinal manifestations (+2 points) to calculate APEX score
(low: 0–3, high: 4–7) to predict CD flare during the first year after achieving mucosal
healing
|
Abbreviations: CD, Crohn's disease; CECDAIic, Capsule Endoscopy Crohn's Disease Activity
Index; CRP, c-reactive protein; ICCE, International Conference on Capsule Endoscopy;
LS, Lewis score; MRE, magnetic resonance enterography.
3.7.2. Capsule Endoscopy Crohn's Disease Activity Index
Capsule Endoscopy Crohn's Disease Activity Index (CECDAI) or Niv score was also developed
at the same time as the LS (2008) which was simpler and based on severity of inflammation,
extent of disease, and narrowing in proximal and distal small bowel. Interobserver
agreement was strong (k = 0.87) in single center and good between different centers (k = 0.767).[173]
[174] The score has been validated by Ponte et al in 2018 which showed that the corresponding
cut-off value of CECDAI for LS between 135 and 790 was 7.7 to 10.3.[175] Another study showed the cut-off value to be 3.8 to 5.8 which also showed that LS
better correlates with fecal calprotectin (<100 μg/g) than CECDAI.[176]
[177] In comparison to LS, a retrospective study has shown that CECDAI may better predict
intestinal inflammation. Those with high LS and normal CECDAI may reflect strictures
rather than active inflammation.[58]
3.7.3. Panenteric Capsule Endoscopy Scores
As panenteric evaluation became feasible with VCE, panenteric scores were developed.
The first one was CECDAlic which was an extension of CECDAI score into colon. Inflammation,
extent of disease, and narrowing were evaluated in proximal small bowel, distal small
bowel, right colon, and left colon. The concordance was high for small bowel (Kendell's
coefficient: k = 0.85) and panenteric evaluation (k = 0.77) except for strictures in proximal small intestine and distal colon.[178] Later, it was validated and was shown to have excellent interobserver agreement
(k = 0.94).[179]
The second panenteric score was based on novel PillCam Crohn's (PCC) (Medtronic, Dublin,
Ireland) capsule. Panenteric scores were calculated from five areas: three tertiles
of small intestine, right, and left colon. Each subscore was calculated using most
common lesion (1), most severe lesion (2), extent of disease (3), and stricture (4)
(each parameter rated from 0 to 3). Each segmental score was ([A + B] × C) +D. This
score also named as the Elaikim score was shown to have an excellent correlation with
LS and had excellent interobserver agreement (k = 0.9).[180]
3.7.4. Capsule Scoring of Ulcerative Colitis
Similar to ulcerative colitis endoscopic activity index (UCEIS), Capsule Scoring of
Ulcerative Colitis (CSUC) is based on parameters like vascular pattern (0: none, patchy
obliteration[1] <30%, obliterated[2] >30%), bleeding (0: none, automated suspected blood indicator <10 = 1, >10 = 2),
and erosions/ulcers (0: none, 1: <5 mm erosion, 2: >5 mm superficial ulcer, 3: excavated
deep ulcer ± excavation/raised margins) in second-generation colon capsule endoscopy.
Each item was subdivided into proximal and distal parts with the reference point being
the splenic flexure. The total score was 0 to 14. Its correlation with fecal calprotectin,
CRP, and clinical Lichtiger index was similar to UCEIS.[181]
3.7.5 APEX Score
This is based on age (≤30 years) (+3), platelet count (≥280 × 103/L) (+2), and extraintestinal manifestations (+2) which were shown to predict risk of 1 year relapse
after achieving mucosal healing in small bowel CD based on a recent retrospective
study.[182]
3.8. Interobserver Agreement in Diagnosing Small Bowel Crohn's Disease with Video
Capsule Endoscopy
There is substantial interobserver agreement (IOA) for the detection of small bowel
CD with VCE (k = 0.68). IOA was moderate for localization (k = 0.44) and only fair for aphthous ulcers (k = 0.38). Although small bowel CD can be diagnosed confidently with VCE, diagnosis
can be observer dependent in those with few lesions. Differentiating ileal from cecal
lesions can be difficult in a minority of patients.[183]
3.9. Artificial Intelligence
Study Selection and Study Characteristics
We found nine original articles on the use of artificial intelligence in VCE related
to IBD.
Results
AI technology was used for Pillcam SB 3, panenteric capsule, and colon capsule. A
variable number of training images (469–483,444) were used to develop the various
AI technology followed by validation. The sensitivity and specificity of the AI models
were 80 to 97.1% and 89 to 98.1%, respectively[184]
[185]
[186]
[187]
[188]
[189]
[190]
[191]
[192] ([Supplementary Table S4], available in the online version). Hence, AI can significantly reduce the examination
time with excellent sensitivity and specificity.
3.10 Novel Techniques and Future Directions
Study Selection and Study Characteristics
We found eight original articles describing various technical advances for VCE in
IBD[5]
[193]
[194]
[195]
[196]
[197]
[198]
[199]([Table 6]).
Table 6
Summary of studies studying technological advances in video capsule endoscopy pertaining
to inflammatory bowel disease
Author and year
|
Indication
|
N
|
Capsule technology
|
Basic principle
|
Advantages
|
Drawbacks
|
Yung et al 2021
|
Small bowel inflammation and reassessment of known IBD
|
84
|
MiroCam MC2000
|
Double head capsule instead of conventional single head capsule
|
13.1% clinically significant different finding with new technology
|
Potential to overreport if the same lesion is visualized at different time points
by different camera heads
|
Yau et al 2021
|
IBD
|
Pre-clinical
|
Recoverable sampling system
|
Automatic sampling of gastrointestinal fluids and storage of analytes using preservatives
to stabilize DNA and proteins
|
May enable sampling of GI fluid without endoscopy
|
Intact ileo-cecal valve is mandatory for triggering tissue sampling
|
Tontini et al 2020
|
Suspected or known CD
|
41
|
PillCamTM Crohn's System, PCS; Medtronic, Dublin, Ireland
|
344 degree panoramic view
|
Higher diagnostic yield (56 vs. 39%)
Better clinical management (48.8 vs. 31.7%)
|
Overestimation of lesion
Higher reading time
Lower image quality
|
Nam et al 2020
|
Suspected or known CD
|
14
|
MiroCam MC4000
|
3D reconstruction using stereo camera-based technology
|
3D reconstruction
Size estimation for lesions
|
The value in altering clinical management not clear
Size estimation function needs validation
|
Koulaouzidis et al 2012
|
Suspected or known CD
|
81
|
QuickView (QV) mode RAPID capsule view software
|
Rapid capsule video review
|
Reduction in capsule reading time
|
Blue mode does not add any advantage over white light
Decreased overall diagnostic yield
|
Halling et al 2013
|
Suspected CD
|
40
|
QV mode RAPID capsule view software
|
Rapid capsule video review
|
Reduction in capsule reading time
Sensitivity 94%
Diagnostic accuracy 98%
|
False negative in terminal ileal lesions
Significant number of missed lesions
|
Freitas et al 2020
|
Suspected or known CD
|
115
|
TOP 100 software tool of the RAPID Reader
|
Automatic selection of 100 images that will most likely contain abnormalities
|
Prompt calculation of Lewis score and high agreement in moderate to severe inflammatory
activity
|
Needs further validation
Agreement less in mild inflammatory activity
|
Tontini et al 2014
|
CD
|
1
|
CapsoCam SV-1 (Capso- Vision, Inc. Saratoga, CA, United States)
|
Lateral panoramic 360 degree viewing
|
Improved diagnostic yield
|
Needs further validation
|
Abbreviations: CD, Crohn's disease; IBD, inflammatory bowel disease.
Results
To increase the visibility and diagnostic yield, double head capsule and 344 degree
panoramic view capsules have been developed.[198] For rapid review, the QuickView mode of RAPID capsule view software can reduce the
reading time with excellent diagnostic accuracy up to 98%.[194] 3D reconstruction can help in the estimation of size of lesions.[197]
A novel RSS capsule technology has been developed which may allow noninvasive sampling,
preservation, and storage of analytes found in gastrointestinal fluids which can correlate
with inflammation and gut permeability. The preservative contained in the novel capsule
stabilizes DNA and proteins for analysis after expulsion.[5]
3.11. Cleansing Regimen
A randomized controlled trial in pediatric patients has shown that 25 mL/kg of polyethylene
glycol (PEG) solution plus 20 mL (376 mg) of oral simethicone was superior to high
volume PEG (50 mL/kg), oral simethicone, and low-dose PEG alone (25 mL/kg) with regard
to better visualization but not diagnostic yield.[200] According to a prospective study, the addition of 15 mL of castor oil to 1 L of
Moviprep and 10 mg bisacodyl significantly improved colon capsule endoscopy completion
rates (87 vs. 73%) and polyp detection (82 vs. 44%)[201].