Accepted Manuscript Agreement Between Rectosigmoidoscopy and Colonoscopy Analyses of Disease Activity and Healing in Patients With Ulcerative Colitis Jean-Frédéric Colombel, Ingrid Ordás, Thomas Ullman, Paul Rutgeerts, Akiko Chai, Sharon O’Byrne, Timothy T. Lu, Julián Panés PII: DOI: Reference:
S0016-5085(15)01515-2 10.1053/j.gastro.2015.10.016 YGAST 60090
To appear in: Gastroenterology Accepted Date: 19 October 2015 Please cite this article as: Colombel J-F, Ordás I, Ullman T, Rutgeerts P, Chai A, O’Byrne S, Lu TT, Panés J, Agreement Between Rectosigmoidoscopy and Colonoscopy Analyses of Disease Activity and Healing in Patients With Ulcerative Colitis, Gastroenterology (2015), doi: 10.1053/j.gastro.2015.10.016. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Page 1 of 20 Agreement Between Rectosigmoidoscopy and Colonoscopy Analyses of Disease Activity and Healing in Patients With Ulcerative Colitis
RI PT
Short title: Rectosigmoidoscopy vs colonoscopy in UC Jean-Frédéric Colombel,1 Ingrid Ordás,2 Thomas Ullman,1 Paul Rutgeerts,3 Akiko Chai,4 Sharon O’Byrne,4 Timothy T. Lu,4 Julián Panés2
SC
1
Division of Gastroenterology, Icahn School of Medicine at Mount Sinai New York, New
York
M AN U
2
Hospital Clinic de Barcelona, IDIBAPS, CIBERehd, Barcelona, Spain
3
Department of Gastroenterology, University Hospital Leuven, Leuven, Belgium
4
Genentech Research and Early Development, South San Francisco, California
TE D
Correspondence:
Jean-Frédéric Colombel, MD, Division of Gastroenterology, Icahn School of Medicine at Mount Sinai, 1 Gustave L. Levy Place, New York, New York 10029.
EP
Email: [email protected] Phone: (212) 241-0990
AC C
Fax: (646) 537-8647
Total word count (limit 6000): 4581 Abstract word count (limit 260): 294 Figures and Tables (limit 7): 7 Supplementary Figures and Tables: 4 References: 22
ACCEPTED MANUSCRIPT Page 2 of 20 Grant support: Research support was provided by Genentech, Inc. Abbreviations:
RI PT
CI, confidence interval; Etro, etrolizumab; IBD, inflammatory bowel disease; κ, kappa
coefficient; LD, loading dose; MCSe, Mayo Clinic endoscopic subscore; mITT, modified intentto-treat, NPV, negative predictive value; Pbo, placebo; PPV, positive predictive value; R/S,
UCEIS, Ulcerative Colitis Endoscopic Index of Severity.
M AN U
Disclosures:
SC
rectosigmoidoscopy; SC, subcutaneous; TNF, tumor necrosis factor; UC, ulcerative colitis;
Jean-Frédéric Colombel has served as a consultant or advisory board member for AbbVie, ABScience, Amgen, Bristol-Myers Squibb, Celltrion, Danone, Enterome, Ferring Pharmaceuticals, Genentech, Giuliani SPA, Given Imaging, Janssen, Immune Pharmaceuticals, Medimmune, Merck & Co., Millenium Pharmaceuticals Inc., Neovacs, Nutrition Science Partners
TE D
Ltd., Pfizer Inc., Prometheus Laboratories, Protagonist Therapeutics, Receptos, Sanofi, Schering-Plough Corporation, Second Genome, Shire, Takeda, Teva Pharmaceuticals, Tigenix, UCB Pharma, Vertex, and Dr. August Wolff GmbH & Co. He has also been a speaker for
Takeda.
EP
Abbvie, Ferring Pharmaceuticals, Janssen, Merck & Co., Nutrition Science Partners Ltd., and
AC C
Ingrid Ordás has been a speaker for Abbvie and MSD Pharmaceuticals. Thomas Ullman has served as a consultant to Takeda. Paul Rutgeerts has served as a consultant, advisory board member, or speaker for Abbvie, Bristol-Myers Squibb, Genentech, Janssen, Merck & Co., Millenium Pharmaceuticals Inc., Pfizer Inc., Prometheus Laboratories, Takeda, UCB Pharma, Dr. Falk Pharma, and Robarts Research Institute. Akiko Chai, Sharon O’Byrne, and Timothy T. Lu are employees of Genentech, Inc.
ACCEPTED MANUSCRIPT Page 3 of 20 Julián Panés has served as a consultant or advisory board member for AbbVie, Bristol-Myers Squibb, Celltrion, Genentech, Nutrition Science Partners Ltd., Pfizer Inc., Merck Sharp and Dohme, Takeda, and Tigenix. He has also been a speaker for Abbvie, Ferring Pharmaceuticals,
RI PT
Janssen, and Merck Sharp and Dohme. Writing Assistance:
Editing and writing support was provided by Deborah Solymar (Genentech, Inc., South San
SC
Francisco, CA, USA) and was funded by Genentech, Inc. Author Contributions:
M AN U
Jean-Frédéric Colombel contributed to the concept and study design, interpretation of data, and drafting and critical revision of the manuscript.
Ingrid Ordás contributed to the interpretation of data, and drafting and critical revision of the manuscript.
Thomas Ullman contributed to the study design, interpretation of data, and drafting and revision
TE D
of the manuscript.
Paul Rutgeerts was an endoscopy reviewer, and contributed to study concept and design, analysis of data, and the drafting and revision of the manuscript.
EP
Akiko Chai contributed to the study design, statistical analysis and interpretation of data, and drafting and revision of the manuscript.
AC C
Sharon O’Byrne contributed to the study design, interpretation of data, and drafting and revision of the manuscript.
Timothy T. Lu contributed to the study design, interpretation of data, and the drafting and revision of the manuscript.
Julián Panés contributed to the study design, interpretation of data, and drafting and critical revision of the manuscript.
ACCEPTED MANUSCRIPT Page 4 of 20 ABSTRACT Background & Aims: Endoscopy limited to the rectosigmoid colon is the standard technique used to measure endoscopic healing in ulcerative colitis (UC) clinical trials. We evaluated
RI PT
whether rectosigmoidoscopy adequately measures UC activity in the more proximal colon. Methods: We analyzed data from a phase 2 placebo-controlled study that evaluated the
efficacy and safety of etrolizumab in patients with moderate to severely active UC who had not
SC
responded to standard therapy. Central readers determined Mayo Clinic endoscopic subscores (MCSe) and ulcerative colitis endoscopic index of severity (UCEIS) scores from the
M AN U
rectosigmoid and proximal colon in videos of 331 examinations performed at baseline, week 6, and week 10. Rates of endoscopic healing (MCSe≤1, MCSe=0) and scores from rectosigmoidoscopy and colonoscopy analyses were compared among 239 examinations with endoscopic assessment proximal to the rectosigmoid colon.
Results: There was a high degree of correlation between findings from rectosigmoidoscopy vs
TE D
colonoscopy in assessment of disease activity based on MCSe≥2 (r=0.84) or MCSe≥1 (r=0.96), or the UCEIS (r=0.92). In 230/239 videos, findings from rectosigmoidoscopy agreed with those from colonoscopy in detection of active disease (MCSe≥2; n=205) or healing (MCSe≤1; n=25).
EP
In 9 videos (2 taken at baseline, 7 after treatment), colonoscopy found proximal disease activity not detected by rectosigmoidoscopy. Post-treatment discordance was more frequent in the
AC C
placebo group, affecting assessment of efficacy at week 10. When endoscopic healing was defined as MCSe=0, there were discordant findings from only 1 video. Conclusions: There is a high degree of correlation in assessments of UC activity made by rectosigmoidoscopy vs colonoscopy. For detection of endoscopic healing (MCSe≤1), colonoscopy found persistent proximal lesions in the placebo group, which affected efficacy analyses. When endoscopic healing was defined as MCSe=0, the concordance between rectosigmoidoscopy and colonoscopy was nearly perfect. Keywords: anti-integrin; EUCALYPTUS; Inflammatory Bowel Disease; IBD
ACCEPTED MANUSCRIPT Page 5 of 20 INTRODUCTION Ulcerative colitis (UC) is a chronic relapsing inflammatory bowel disease (IBD) characterized by superficial mucosal ulceration, rectal bleeding, diarrhea, and abdominal pain.1 Until the late
RI PT
1990s, the therapeutic goal for UC was clinical remission as defined by symptom management. However, a growing body of evidence indicates that healing of the mucosa is an important
therapeutic endpoint in clinical trials and in clinical practice.2 This altered therapeutic paradigm
SC
has further increased the role of endoscopy in UC.
Colonoscopy with intubation of the terminal ileum is the standard of care for diagnosis of
M AN U
patients with suspected UC.3-5 Once the diagnosis is established, and when UC patients are included in clinical trials, endoscopy limited to the rectum and sigmoid is the standard to assess disease activity and endoscopic healing because it is believed that the most severe activity of UC is located in the distal colon. However, little evidence supports this assertion, and some UC patients may harbor more severe endoscopic inflammation proximal to the sigmoid colon.
TE D
Etrolizumab is a humanized monoclonal anti-β7 antibody with a dual mechanism of action, inhibiting both α4β7:MAdCAM-1-mediated lymphocyte trafficking to the gut mucosa and αEβ7:E-cadherin-mediated lymphocyte retention in the intraepithelial compartment.6-12 In a
EP
double-blind, placebo-controlled, randomized, phase 2 study, etrolizumab led to significantly greater rates of clinical remission at Week 10 over placebo in patients with moderately-to-
AC C
severely active UC.13 During this study, 72% of endoscopic examinations assessed disease activity beyond the rectosigmoid. This circumstance allowed us to evaluate whether endoscopic assessment limited to the rectosigmoid adequately represents endoscopic activity of the more proximal colon, and how potential discrepancies might affect assessment of efficacy in the context of a randomized controlled clinical trial.
ACCEPTED MANUSCRIPT Page 6 of 20 METHODS Patient Cohort The EUCALYPTUS trial evaluated the efficacy and safety of etrolizumab (Genentech, South
RI PT
San Francisco, CA) in patients with moderately-to-severely active UC, who had not responded to conventional therapy, including immunosuppressants and/or at least one anti-tumor necrosis factor (anti-TNF) agent. For this randomized, double-blind, placebo-controlled, multi-center
SC
Phase 2 study, patients were recruited from 40 centers in 11 countries. Of 124 patients, 41 were assigned to 100 mg etrolizumab administered subcutaneously (SC) at Weeks 0, 4, and 8, 40
M AN U
were assigned to 300 mg etrolizumab (Weeks 2, 4, and 8) plus a 420 mg loading dose (LD) administered between Weeks 0–2, and 43 were assigned to monthly SC injections of placebo for 3 months. Five patients were excluded from the modified intent-to-treat (mITT) population as they had a centrally read screening endoscopic subscore of < 2, leaving 41 analyzed from the placebo group, 39 from the 100 mg etrolizumab group, and 39 from the 300 mg etrolizumab
TE D
plus LD group. Details of patient demographics, methods, and eligibility criteria have been described previously.13 Briefly, patients included were adults diagnosed with UC for a minimum of 12 weeks, who had a Mayo Clinic Score14 (MCS) ≥ 5, with an MCS endoscopic subscore
EP
(MCSe) ≥ 2, a rectal bleeding symptom score ≥ 1, and disease extending ≥ 25 cm from the anal verge. The primary efficacy endpoint was clinical remission at Week 10, defined as the
AC C
proportion of patients with an MCS ≤ 2, with no individual subscore > 1. The post hoc analyses described here received institutional review board approval under the EUCALYPTUS protocol (ClinicalTrials.gov, NCT01336465). Analysis of Endoscopic Examinations The current study examined 331 videos of endoscopies captured during the EUCALYPTUS trial, after exclusion of videos without associated patient identification numbers and videos of patients excluded due to protocol violations (Supplementary Figures 1 and 2). Videos of patient
ACCEPTED MANUSCRIPT Page 7 of 20 endoscopies were recorded at the baseline screening visit (Week 0), and after randomization at Week 6 and Week 10. Out of the 331 videos, 239 (100 from baseline and 139 from Weeks 6 and 10) included assessment of colonic segments proximal to the rectosigmoid region, and
RI PT
were analyzed for disease activity and endoscopic healing in this study (Supplementary Figure 2). Endoscopic severity was assessed using the MCSe and the Ulcerative Colitis Endoscopic Index of Severity (UCEIS).15,16 For the MCSe, active disease was defined as a score > 1, and
SC
endoscopic healing as MCSe ≤ 1 or MCSe = 0. We included analyses with endoscopic healing defined as MCSe = 0 because of emerging evidence that a score of 0 may be linked to different
M AN U
patient outcomes compared with a score of 1.17 Currently, the UCEIS does not specify a cutoff score to distinguish active disease from endoscopic healing and, therefore, was not used to calculate positive predictive values (PPV) and negative predictive values (NPV). Central readers performed a preliminary evaluation of videos to determine parameters for scoring the MCSe and the UCEIS. Of note, the location of segments explored was not
TE D
annotated on the videos during recording and not all patients received full colonoscopies, with examination of all 5 colonic segments (rectum, sigmoid, descending colon, transverse colon, and ascending colon). Four central readers (2 per video) from 2 sites estimated the number of
EP
colonic segments examined, guided by the length of the segments and identification of anatomical characteristics such as the hepatic, splenic, and rectosigmoid flexures. Readers also
AC C
determined the extent of disease.
For determining endoscopic healing based on the MCSe, the worst score of the rectosigmoid alone was compared with the worst score of the whole colon (including the rectosigmoid). A separate, independent, central reader adjudicated the final score when discrepancies emerged between the two readers of the same video. For the analysis of individual endoscopic severity scores (Figure 1), all 331 videos were scored with both the MCSe and the UCEIS, with 2 reads per video (one from each reader, without adjudication; 662 reads). Reads with ≤ 2 segments were excluded (251), resulting in a
ACCEPTED MANUSCRIPT Page 8 of 20 total of 411 reads (Supplementary Figure 1). The rectosigmoid and the proximal colon (descending, transverse, and ascending colon) were scored separately, according to the worstaffected segment. Each component of the UCEIS (vascular pattern, bleeding, and erosions and
RI PT
ulcers) was scored separately. Total UCEIS scores were generated from each reader from the sum of their respective subscores. Statistical Analysis
SC
To determine the level of correlation between rectosigmoidoscopy and colonoscopy for detecting disease activity and endoscopic healing, contingency tables were generated that
M AN U
tabulated the number of adjudicated MCSe scores above and below the indicated MCSe cutoffs. The degree of inter-method agreement and correlation was determined by the calculation of the kappa coefficient (κ) and the Spearman correlation coefficient (r) for each contingency table, respectively. The interpretation of the κ and the correlation coefficients was based on the proposed cutoffs by Landis and Koch,18 who recommended the following
TE D
standards for strength of agreement for the κ coefficient: 0, poor; 0.01–0.20, slight; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and 0.81–1, nearly perfect. NPVs and PPVs were also calculated. P-values were based on the two-sided Fisher exact test. Because this
EP
analysis was retrospective and not a prespecified goal of the EUCALYPTUS protocol, formal
AC C
sample size calculations were not performed.
ACCEPTED MANUSCRIPT Page 9 of 20 RESULTS Number of bowel segments evaluated and description of extent of disease Comparable numbers of videos were examined at each visit for each treatment group (Table
RI PT
1). Across the 3 treatment groups, the average total segments examined was 3.2, and the number of segments examined across different visits (Weeks 0, 6, and 10) and between treatment groups did not differ significantly.
SC
Across all treatment groups, 25% of patients had UC limited to the rectosigmoid region and 75% of patients had disease in the proximal colon, defined as any region beyond the
M AN U
rectosigmoid (Table 2). Overall evaluation
Among all of the videos, regardless of the number of segments read, 67 videos (67/354 = 19%) were sent for adjudication. Among videos with > 2 segments, 55 videos (55/239 = 23%) were sent for adjudication. The MCSe (mean ± SD) in videos where only two segments or fewer
TE D
were examined (2.66 ± 0.63) did not differ significantly from videos with > 2 segments examined (2.54 ± 0.75; Supplementary Table 1).
Across all visits and examinations, 239 videos examined colonic segments proximal to the
EP
rectosigmoid. For 230 videos, rectosigmoidoscopy and colonoscopy analyses agreed, determining that active disease (MCSe ≥ 2) was present in 205, but absent in 25 videos (Table
AC C
3; left). There was a strong correlation between the scores based on the two techniques (κ = 0.83; r = 0.84). However, for 9 of the 239 videos, the scores were discordant in assessing active disease, i.e. rectosigmoidoscopy found evidence of endoscopic healing, but examination of the more proximal colon with colonoscopy found mucosal lesions (MCSe ≥ 2). These 9 videos came from 7 patients: 3 with left-sided colitis, 2 with pancolitis, 1 with disease previously limited to the rectosigmoid, and 1 where the previous extent of disease was not specified. If endoscopic healing was defined as MCSe = 0 and endoscopic activity was defined as MCSe ≥ 1, the concordance between rectosigmoidoscopy and colonoscopy increased (κ = 0.95; r = 0.96), as
ACCEPTED MANUSCRIPT Page 10 of 20 only 1 video demonstrated endoscopic healing in the rectosigmoid while significant mucosal lesions were found in the proximal colon (Table 3; right). We then analyzed whether the severity of disease differed in the proximal colon compared
RI PT
with the rectosigmoid, using both the MCSe and the UCEIS (Figure 1; Supplementary Figure 3). For this analysis, each region was scored separately and there was no adjudication between readers; each score given by the readers was considered an individual piece of data. In general,
SC
we observed more severe disease in the rectum and sigmoid according to the MCSe, but 19 of the 411 (4.6%) individual reads with more than 2 segments showed greater disease severity in
M AN U
the proximal colon than in the rectosigmoid region (Figure 1A; gray cells). Out of those 19 scores, 14 (rectosigmoid ≤ 1, proximal colon ≥ 2) affected the assessment of endoscopic healing in the colon. In addition, assessment of endoscopic disease using the UCEIS confirmed the presence of more severe disease in the rectosigmoid compared with the proximal colon (3.98 ± 1.82 vs. 2.38 ± 1.98; P < .0001), but in a small proportion of the videos (31/411; 7.5%),
TE D
more severe disease was found in the proximal colon. Scoring with the UCEIS, using both the total UCEIS and individual subscores, also demonstrated a high correlation between rectosigmoidoscopy and colonoscopy (r = 0.92; Figure 1B; Supplementary Figure 3).
EP
Evaluation of endoscopic activity at baseline When we assessed endoscopic activity (MCSe ≥ 2) at baseline, we observed a moderate
AC C
agreement/correlation (κ = 0.49, r = 0.57) between the scores based on rectosigmoidoscopy and colonoscopy (Table 4). This level of correlation was related to the highly skewed distribution of MCSe values towards scores of 2 and 3, given the study inclusion criteria. In 2 videos out of 100 performed at baseline, rectosigmoidoscopy failed to find active disease while colonoscopy revealed active disease above the sigmoid colon. Evaluation of endoscopic healing After treatment with etrolizumab, patients were then assessed for endoscopic healing (MCSe ≤ 1) at Weeks 6 and 10 using a total of 139 available videos (Table 5; left). While 31
ACCEPTED MANUSCRIPT Page 11 of 20 videos showed endoscopic healing in the rectum and sigmoid colon, only 24 also had healing in the proximal colon, a difference of approximately 23% (PPV for endoscopic healing by rectosigmoidoscopy = 0.77). Colonoscopy and rectosigmoidoscopy demonstrated a high
RI PT
correlation for the detection of endoscopic healing (κ = 0.84; r = 0.85). With endoscopic healing defined as MCSe = 0 at Weeks 6 and 10 (Table 5; right), 12 videos showed healing in the rectum and sigmoid colon. Only 1 video did not show endoscopic healing
SC
in the proximal colon, a difference of approximately 9%. Therefore, when endoscopic healing was defined as an MCSe = 0, the correlation between the two techniques was very strong (κ =
M AN U
0.95; r = 0.95), and rectosigmoidoscopy was better able to predict endoscopic healing (PPV = 0.92) in the proximal colon.
When we investigated the efficacy of etrolizumab using rectosigmoidoscopy, we found that there was a trend towards higher rates of endoscopic healing, but there was no significant treatment effect for either dose group over placebo by Weeks 6 and 10 (Figure 2; white bars).
TE D
Colonoscopy also showed no significant treatment effect at Week 6 (Figure 2A). However, at Week 10, colonoscopy determined that the proportion of patients with endoscopic healing in the placebo group was 8%, rather than 14%, the proportion measured by rectosigmoidoscopy
EP
(Figure 2B). While this difference did not affect the efficacy assessment for the 300 mg dose group, this lower percentage of patients with healing increased the effect size to a significant
AC C
level between the placebo and the 100 mg etrolizumab treatment groups (P = .035).
ACCEPTED MANUSCRIPT Page 12 of 20 DISCUSSION In general, when we compared the ability of rectosigmoidoscopy with colonoscopy for determining either the presence of active disease or endoscopic healing in patients with
RI PT
moderately-to-severely active UC, we found strong correlations between the two examinations. Rectosigmoidoscopy was reliably able to detect active disease at the inclusion of patients in the study. We also found that the most severe disease was located more frequently in the
SC
rectosigmoid colon, a relevant finding for clinical practice because therapeutic decisions are based not only on the presence of active disease but also on severity.
M AN U
The ability of rectosigmoidoscopy to detect disease activity compared with colonoscopy differed between baseline and post treatment at Week 6 and Week 10. At baseline, rectosigmoidoscopy accurately reflected disease activity in the whole colon. In contrast, at Weeks 6 and 10, when endoscopic healing was defined as MCSe ≤ 1, rectosigmoidoscopy indicated healing of the distal mucosa, although there was evidence of mucosal ulceration more
TE D
proximally on colonoscopy. This resulted in a placebo remission rate that was almost two-fold higher based on rectosigmoidoscopy compared with colonoscopy (14% vs. 8%, respectively; Figure 2B), and a decreased treatment effect size. However, when endoscopic healing was
EP
defined as MCSe = 0, concordance improved between rectosigmoidoscopy and colonoscopy, as only 1 out of 11 videos had discordant scores between the two techniques (Table 5; right).
AC C
According to guidelines and consensus panels, once the diagnosis is established, rectosigmoidoscopy alone is usually sufficient to assess endoscopic activity/severity in the management of UC.3-5 Few studies, however, and none in the context of a therapeutic trial, have formally compared endoscopic severity in the rectosigmoid versus the whole colon. In a retrospective analysis of medical charts and endoscopic images from 545 UC patients, Kato et. al19 determined the colonic locations of the maximum inflammatory activity. While the majority of the patients analyzed (73%) had maximum inflammation in the rectum and sigmoid, 27% had maximum activity proximal to the sigmoid, in the descending colon, or in segments proximal to
ACCEPTED MANUSCRIPT Page 13 of 20 the splenic flexure. In addition, they found that 40% of patients had inflamed mucosa in the descending colon or in the more proximal portion of the colon, but showed no inflammation in the rectum and sigmoid. They concluded that total colonoscopy was warranted in patients who
RI PT
have a discrepancy between their symptoms and rectosigmoidoscopy results. A caveat of this study was that the use of rectal therapy with mesalazine or steroids was allowed. In the EUCALYPTUS trial, rectal therapy was withdrawn 2 weeks prior to randomization and
concomitant use of rectal therapy was contraindicated during the treatment phase of the trial,
SC
due to the possibility of rectal sparing.20
M AN U
In this endoscopy study, we used the UCEIS in addition to the MCSe to compare disease severity in the rectosigmoid and the proximal colon. We were unable to use the UCEIS to assess the endpoints of disease activity or endoscopic healing because a threshold score for mucosal healing has not yet been established.16 In general, the UCEIS confirmed the MCSe findings of more severe disease in the rectosigmoid for all three subscores (Supplementary
TE D
Figure 3) and for the overall score (Figure 1B). The UCEIS, as with the MCSe, also detected a small, but slightly higher proportion of videos where proximal disease was more severe than in the rectosigmoid (7.5% vs. 4.6% for the MCSe; Figure 1) and showed a very high correlation
EP
between rectosigmoidoscopy and colonoscopy.
One limitation of our study was that only 239 out of 331 examinations progressed beyond
AC C
the sigmoid colon. There may be some concern that the reasons for the unavailability of colonoscopy in those patients may have introduced bias into the study. Patients with severe disease above the sigmoid may have been underestimated because of concerns by the endoscopist in proceeding with the examination beyond the sigmoid colon. However, the number of segments analyzed between treatment groups was similar (Table 1) and disease severity (mean MCSe) was also similar between videos with fewer than 2 and those with more than 2 segments (Supplementary Table 1), which rules out a bias arising from performing rectosigmoidoscopy or colonoscopy based on the endoscopic severity of the disease. Another
ACCEPTED MANUSCRIPT Page 14 of 20 potential limitation was that because the explored segments were not annotated in the videos, the readers identified segments based on anatomical characteristics and the length of the segment; in particular, the boundary between the descending colon and the sigmoid could be
RI PT
difficult to distinguish, even for skilled endoscopists. It is, however, unlikely that this would have significantly affected evaluation of the rectosigmoid compared with the whole colon. We also acknowledge that, in this trial, the number of MCSe scores where the proximal colon had more
SC
severe disease than the rectosigmoid (9/239) was low. However, the number of patients that typically achieve remission in induction studies is also low, which means that factors that
M AN U
influence placebo rates can have a direct effect on statistical significance and influence decisions about the further development of experimental therapies. Finally, as this was an induction study, we cannot extrapolate these findings to maintenance studies. Longer treatment duration with therapies that are effective may improve the concordance between rectosigmoidoscopy and colonoscopy, and consequently demonstrate that rectosigmoidoscopy
TE D
is sufficient for measuring clinical effectiveness in maintenance trials. Despite these limitations, this study has important implications for the use of rectosigmoidoscopy and/or colonoscopy in both clinical practice and during clinical trials.
EP
Because of its strong ability to detect active disease, rectosigmoidoscopy should be sufficient in clinical practice for evaluating previously diagnosed patients with new symptoms. Two notable
AC C
exceptions may be pediatric patients and patients with primary sclerosing cholangitis where relative or complete rectal sparing has been observed.21,22 However, in cases where rectosigmoidoscopy has detected endoscopic healing in response to induction therapy, but symptoms persist, performance of a more extensive examination with colonoscopy is justified. In the clinical trial setting, this study suggests that if endoscopic healing is defined as an MCSe = 0, rectosigmoidoscopy is sufficient for efficacy analyses. However, if endoscopic healing is defined as an MCSe ≤ 1, these preliminary data suggest that colonoscopy is better than rectosigmoidoscopy for assessing the full extent of endoscopic healing for efficacy analyses of
ACCEPTED MANUSCRIPT Page 15 of 20 experimental therapies. Colonoscopy is more expensive than rectosigmoidoscopy because it frequently requires sedation and bowel preparation, which in turn, may impose additional burdens on the patient, such as missing work and requiring a caregiver to transport the patient
RI PT
to and from the endoscopy suite. The balance between cost and performance is an important consideration and is most relevant in the development of experimental therapies. It may be preferable to use full colonoscopy in relatively small Phase 2 induction studies. However, these
AC C
EP
TE D
M AN U
recommended for standard use in clinical trials.
SC
results warrant further evaluation prospectively in a larger study before colonoscopy can be
ACCEPTED MANUSCRIPT Page 16 of 20 Figure 1. Endoscopic severity in the rectosigmoid colon vs. the proximal colon (all visits; unadjudicated scores). Two readers assessed each video and scored the severity of disease for the rectosigmoid colon and the proximal colon, using the MCSe (A) or the UCEIS (B). Numbers
RI PT
in each cell represent the total number of scores of the individual site readers (two scores per video) without adjudication (pooled across readers). Data was pooled across the sites. Any reading with ≤ 2 segments was not included (Supplementary Figure 1). Gray cells in (A) and (B)
SC
indicate where the proximal colon score was higher than the rectosigmoid colon score. Scores in the light-gray cells in (A) affected the assessment of endoscopic healing when the proximal colon was included in the evaluation; the dark-gray cell in (A) indicates the MCSe scores that
M AN U
did not affect the overall endoscopic healing evaluation. κ and correlation coefficients are provided for the comparison of rectosigmoidoscopy and colonoscopy.
Figure 2. Comparison of endoscopic healing using rectosigmoidoscopy and colonoscopy.
TE D
The percentage of patients (mITT population) with endoscopic healing (MCSe ≤ 1) in placebo and etrolizumab-treated groups was calculated at Week 6 (A) and Week 10 (B) from videos based on rectosigmoidoscopy only (white bars) and colonoscopy (black bars). *Scoring was
AC C
EP
based on adjudicated results; ‡P < .05; Pbo, placebo; Etro, etrolizumab; LD, loading dose.
ACCEPTED MANUSCRIPT Page 17 of 20 REFERENCES 1. Kaser A, Zeissig S, Blumberg RS. Inflammatory bowel disease. Annu Rev Immunol 2010;28:573-621.
RI PT
2. Neurath MF, Travis SPL. Mucosal healing in inflammatory bowel diseases: a systematic review. Gut 2012;61:1619-1635.
3. Kornbluth A, Sachar DB, The Practice Parameters Committee of the American College of
SC
Gastroenterology. Ulcerative colitis practice guidelines in adults: American College of
Gastroenterology, Practice Parameters Committee. Am J Gastroenterol 2010;105:501-523.
M AN U
4. Dignass A, Eliakim R, Magro F, et al. Second European evidence-based consensus on the diagnosis and management of ulcerative colitis Part 1: definitions and diagnosis. J Crohns Colitis 2012;6:965-90.
5. Annese V, Daperno M, Rutter MD, et al. European evidence based consensus for endoscopy in inflammatory bowel disease. J Crohns Colitis 2013;7:982-1018.
TE D
6. Karecla PI, Bowden SJ, Green SJ, et al. Recognition of E-cadherin on epithelial cells by the mucosal T cell integrin alpha M290 beta 7 (alpha E beta 7). Eur J Immunol 1995;25:852-6. 7. Feagan BG, Greenberg GR, Wild G, et al. Treatment of ulcerative colitis with a humanized
EP
antibody to the α4β7 integrin. N Engl J Med 2005;352:2499-507. 8. Cepek KL, Parker CM, Madara JL, et al. Integrin alpha E beta 7 mediates adhesion of T
AC C
lymphocytes to epithelial cells. J Immunol 1993;150:3459-70. 9. Hesterberg PE, Winsor-Hines D, Briskin MJ, et al. Rapid resolution of chronic colitis in the cotton-top tamarin with an antibody to a gut-homing integrin alpha 4 beta 7. Gastroenterology 1996;111:1373-80. 10. Ludviksson BR, Strober W, Nishikomori R, et al. Administration of mAb against αEβ7 prevents and ameliorates immunization-induced colitis in IL-2-/- mice. J Immunol 1999;162:4975-82. 11. Sandborn WJ, Colombel JF, Enns R, et al. Natalizumab induction and maintenance therapy
ACCEPTED MANUSCRIPT Page 18 of 20 for Crohn's disease. N Engl J Med 2005;353:1912-25. 12. Elewaut D, Van Damme N, De Keyser F, et al. Altered expression of alpha E beta 7 integrin on intra-epithelial and lamina propria lymphocytes in patients with Crohn’s disease.
RI PT
Acta Gastroenterol Belg 1998;61:288-94. 13. Vermeire S, O'Byrne S, Keir M, et al. Etrolizumab as induction therapy for ulcerative colitis: a randomised, controlled, phase 2 trial. Lancet 2014;384:309-318.
SC
14. Schroeder KW, Tremaine WJ, Ilstrup DM. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis. A randomized study. N Engl J Med
M AN U
1987;317:1625-9.
15. Travis SP, Schnell D, Krzeski P, et al. Developing an instrument to assess the endoscopic severity of ulcerative colitis: the Ulcerative Colitis Endoscopic Index of Severity (UCEIS). Gut 2012;61:535-42.
16. Travis SP, Schnell D, Krzeski P, et al. Reliability and initial validation of the ulcerative colitis
TE D
endoscopic index of severity. Gastroenterology. 2013;145:987-95. 17. Manginot C, Baumann C, Peyrin-Biroulet L. An endoscopic Mayo score of 0 is associated with a lower risk of colectomy than a score of 1 in ulcerative colitis. Gut 2014;0:1-2.
EP
18. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-174.
AC C
19. Kato J, Kuriyama M, Hiraoka S, et al. Is sigmoidoscopy sufficient for evaluating inflammatory status of ulcerative colitis patients? J Gastroen Hepatol 2011;26:683-687. 20. Odze R, Antonioli D, Peppercorn M, et al. Effect of topical 5-aminosalicylic acid (5-ASA) therapy on rectal mucosal biopsy morphology in chronic ulcerative colitis. Am J Surg Pathol. 1993;17:869-75. 21. Rajwal SR, Puntis JW, McClean P, et al. Endoscopic rectal sparing in children with untreated ulcerative colitis. J Pediatr Gastroenterol Nutr 2004;38:66-9. 22. Loftus EV Jr, Harewood GC, Loftus CG, et al. PSC-IBD: a unique form of inflammatory
ACCEPTED MANUSCRIPT Page 19 of 20 bowel disease associated with primary sclerosing cholangitis. Gut 2005;54:91-6.
AC C
EP
TE D
M AN U
SC
RI PT
Author names in bold designate shared co-first authorship.
ACCEPTED MANUSCRIPT
Table 1. Mean number of colon segments examined by treatment group and visit Etrolizumab
100 mg
300 mg + LD
Placebo POOLED n
Mean (SD)
Segments
n
40
3.9 (1.2)
38
Week 6
35
2.7 (0.9)
34
Week 10
37
2.9 (0.9)
36
Total
112
3.2 (1.2)
108
AC C
EP
TE D
n, number of videos evaluated; LD, loading dose.
n
Mean (SD)
4.0 (1.1) 3.1 (1.0)
38
Mean (SD)
3.6 (1.3)
37
3.0 (1.0)
2.9 (1.0)
36
2.6 (0.9)
3.4 (1.1)
111
3.1 (1.1)
M AN U
Week 0/baseline
Segments
SC
Segments
RI PT
Etrolizumab
ACCEPTED MANUSCRIPT
Table 2. Extent of disease Etrolizumab
Etrolizumab
100 mg
300 mg
(n = 39)
(n = 39)
Placebo Extent of disease, n (%)
Total (n = 119)
RI PT
(n = 41)
Extensive/pancolitis
12 (29%)
14 (36%)
Left-sided colitis
17 (42%)
13 (33%)
Rectosigmoid
12 (29%)
10 (26%)
8 (21%)
30 (25%)
Non-specified
0 (0%)
2 (5%)
0 (0%)
2 (2%)
44 (37%)
13 (33%)
43 (36%)
SC
M AN U
AC C
EP
TE D
n, number of patients.
18 (46%)
ACCEPTED MANUSCRIPT
Table 3. Identification of endoscopically active disease as assessed by rectosigmoidoscopy vs. colonoscopy (all visits)
Colonoscopy
(n = 239)
Y
MCSe ≥ 1
Colonoscopy
(n = 239) Y
N
205
0
Y
227
N
1
R/S N
9
25
N 0
SC
R/S
Y
RI PT
MCSe ≥ 2
11
κ = 0.95; r = 0.96; P < .0001;
NPV = 0.74 (95% CI: 0.59, 0.88)
NPV = 0.92 (95% CI: 0.76, 1.07)
M AN U
κ = 0.83; r = 0.84; P < .0001;
This analysis only included assessments with > 2 segments and all visits were pooled. Gray
TE D
cells represent the number of discordant scores for endoscopic activity that were affected when the proximal colon was included in the evaluation. Presence of endoscopically active disease (MCSe ≥ 2, left; MCSe ≥ 1, right): Y = yes, N = no; CI, confidence interval; n, number of videos
AC C
EP
evaluated; NPV, negative predictive value; R/S, rectosigmoidoscopy.
ACCEPTED MANUSCRIPT
Table 4. Identification of endoscopically active disease (MCSe ≥ 2) as assessed by
Colonoscopy
(n = 100)
Y
N
Y
97
0
N
2
1
κ = 0.49; r = 0.57; P = .03; NPV = 0.33 (95% CI: -0.02, 0.87)
M AN U
R/S
SC
MCSe ≥ 2
RI PT
rectosigmoidoscopy vs. colonoscopy at Week 0/baseline
This analysis only included assessments with > 2 segments. The gray cell represents the
TE D
number of scores for endoscopic activity that were affected when the proximal colon was included in the evaluation. Presence of endoscopically active disease (MCSe ≥ 2): Y = yes, N = no; CI, confidence interval; n, number of videos evaluated; NPV, negative predictive value; R/S,
AC C
EP
rectosigmoidoscopy.
ACCEPTED MANUSCRIPT
Table 5. Evaluation of endoscopic healing as assessed by rectosigmoidoscopy vs. colonoscopy at Weeks 6 and 10 (pooled)
Colonoscopy
(n = 139)
N
Y
N
108
0
Y
7
24
(n = 139)
R/S
Y
N
127
0
Y
1
11
κ = 0.95; r = 0.95; P < .0001;
M AN U
κ = 0.84; r = 0.85; P < .0001; PPV = 0.77 (95% CI: 0.63, 0.92)
N
SC
R/S
Colonoscopy
MCSe = 0
RI PT
MCSe ≤ 1
PPV = 0.92 (95% CI: 0.76, 1.07)
This analysis only included assessments with > 2 segments by visit (excluding the screening visit). Gray cells represent the number of scores for endoscopic healing that were affected when
TE D
the proximal colon was included in the evaluation. Presence of endoscopic healing (MCS ≤ 1, left; MCSe = 0, right): Y = yes, N = no; CI, confidence interval; n, number of videos evaluated;
AC C
EP
PPV, positive predictive value; R/S, rectosigmoidoscopy.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Segments ≤ 2 (n = 92)
Segments > 2 (n = 239)
2.66 (0.63)
2.47 (0.80)
3 (1, 3)
3 (0, 3)
Mean (SD)
N/A
2.54 (0.75)
Median (range)
N/A
Mean (SD) Median (range)
SC
MCSe (rectosigmoid)
M AN U
MCSe (colon)
Extent of disease, n (%)
3 (0, 3)
35 (38%)
88 (37%)
Left-sided colitis
31 (34%)
89 (37%)
25 (27%)
59 (25%)
1 (1%)
3 (1%)
EP
Non-specified
TE D
Extensive/pancolitis
Rectosigmoid
n, number of videos evaluated; N/A, not applicable.
AC C
RI PT
Supplementary Table 1. MCSe of videos with ≤ 2 and > 2 segments
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S0016-5085(15)01515-2 10.1053/j.gastro.2015.10.016 YGAST 60090
To appear in: Gastroenterology Accepted Date: 19 October 2015 Please cite this article as: Colombel J-F, Ordás I, Ullman T, Rutgeerts P, Chai A, O’Byrne S, Lu TT, Panés J, Agreement Between Rectosigmoidoscopy and Colonoscopy Analyses of Disease Activity and Healing in Patients With Ulcerative Colitis, Gastroenterology (2015), doi: 10.1053/j.gastro.2015.10.016. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Page 1 of 20 Agreement Between Rectosigmoidoscopy and Colonoscopy Analyses of Disease Activity and Healing in Patients With Ulcerative Colitis
RI PT
Short title: Rectosigmoidoscopy vs colonoscopy in UC Jean-Frédéric Colombel,1 Ingrid Ordás,2 Thomas Ullman,1 Paul Rutgeerts,3 Akiko Chai,4 Sharon O’Byrne,4 Timothy T. Lu,4 Julián Panés2
SC
1
Division of Gastroenterology, Icahn School of Medicine at Mount Sinai New York, New
York
M AN U
2
Hospital Clinic de Barcelona, IDIBAPS, CIBERehd, Barcelona, Spain
3
Department of Gastroenterology, University Hospital Leuven, Leuven, Belgium
4
Genentech Research and Early Development, South San Francisco, California
TE D
Correspondence:
Jean-Frédéric Colombel, MD, Division of Gastroenterology, Icahn School of Medicine at Mount Sinai, 1 Gustave L. Levy Place, New York, New York 10029.
EP
Email: [email protected] Phone: (212) 241-0990
AC C
Fax: (646) 537-8647
Total word count (limit 6000): 4581 Abstract word count (limit 260): 294 Figures and Tables (limit 7): 7 Supplementary Figures and Tables: 4 References: 22
ACCEPTED MANUSCRIPT Page 2 of 20 Grant support: Research support was provided by Genentech, Inc. Abbreviations:
RI PT
CI, confidence interval; Etro, etrolizumab; IBD, inflammatory bowel disease; κ, kappa
coefficient; LD, loading dose; MCSe, Mayo Clinic endoscopic subscore; mITT, modified intentto-treat, NPV, negative predictive value; Pbo, placebo; PPV, positive predictive value; R/S,
UCEIS, Ulcerative Colitis Endoscopic Index of Severity.
M AN U
Disclosures:
SC
rectosigmoidoscopy; SC, subcutaneous; TNF, tumor necrosis factor; UC, ulcerative colitis;
Jean-Frédéric Colombel has served as a consultant or advisory board member for AbbVie, ABScience, Amgen, Bristol-Myers Squibb, Celltrion, Danone, Enterome, Ferring Pharmaceuticals, Genentech, Giuliani SPA, Given Imaging, Janssen, Immune Pharmaceuticals, Medimmune, Merck & Co., Millenium Pharmaceuticals Inc., Neovacs, Nutrition Science Partners
TE D
Ltd., Pfizer Inc., Prometheus Laboratories, Protagonist Therapeutics, Receptos, Sanofi, Schering-Plough Corporation, Second Genome, Shire, Takeda, Teva Pharmaceuticals, Tigenix, UCB Pharma, Vertex, and Dr. August Wolff GmbH & Co. He has also been a speaker for
Takeda.
EP
Abbvie, Ferring Pharmaceuticals, Janssen, Merck & Co., Nutrition Science Partners Ltd., and
AC C
Ingrid Ordás has been a speaker for Abbvie and MSD Pharmaceuticals. Thomas Ullman has served as a consultant to Takeda. Paul Rutgeerts has served as a consultant, advisory board member, or speaker for Abbvie, Bristol-Myers Squibb, Genentech, Janssen, Merck & Co., Millenium Pharmaceuticals Inc., Pfizer Inc., Prometheus Laboratories, Takeda, UCB Pharma, Dr. Falk Pharma, and Robarts Research Institute. Akiko Chai, Sharon O’Byrne, and Timothy T. Lu are employees of Genentech, Inc.
ACCEPTED MANUSCRIPT Page 3 of 20 Julián Panés has served as a consultant or advisory board member for AbbVie, Bristol-Myers Squibb, Celltrion, Genentech, Nutrition Science Partners Ltd., Pfizer Inc., Merck Sharp and Dohme, Takeda, and Tigenix. He has also been a speaker for Abbvie, Ferring Pharmaceuticals,
RI PT
Janssen, and Merck Sharp and Dohme. Writing Assistance:
Editing and writing support was provided by Deborah Solymar (Genentech, Inc., South San
SC
Francisco, CA, USA) and was funded by Genentech, Inc. Author Contributions:
M AN U
Jean-Frédéric Colombel contributed to the concept and study design, interpretation of data, and drafting and critical revision of the manuscript.
Ingrid Ordás contributed to the interpretation of data, and drafting and critical revision of the manuscript.
Thomas Ullman contributed to the study design, interpretation of data, and drafting and revision
TE D
of the manuscript.
Paul Rutgeerts was an endoscopy reviewer, and contributed to study concept and design, analysis of data, and the drafting and revision of the manuscript.
EP
Akiko Chai contributed to the study design, statistical analysis and interpretation of data, and drafting and revision of the manuscript.
AC C
Sharon O’Byrne contributed to the study design, interpretation of data, and drafting and revision of the manuscript.
Timothy T. Lu contributed to the study design, interpretation of data, and the drafting and revision of the manuscript.
Julián Panés contributed to the study design, interpretation of data, and drafting and critical revision of the manuscript.
ACCEPTED MANUSCRIPT Page 4 of 20 ABSTRACT Background & Aims: Endoscopy limited to the rectosigmoid colon is the standard technique used to measure endoscopic healing in ulcerative colitis (UC) clinical trials. We evaluated
RI PT
whether rectosigmoidoscopy adequately measures UC activity in the more proximal colon. Methods: We analyzed data from a phase 2 placebo-controlled study that evaluated the
efficacy and safety of etrolizumab in patients with moderate to severely active UC who had not
SC
responded to standard therapy. Central readers determined Mayo Clinic endoscopic subscores (MCSe) and ulcerative colitis endoscopic index of severity (UCEIS) scores from the
M AN U
rectosigmoid and proximal colon in videos of 331 examinations performed at baseline, week 6, and week 10. Rates of endoscopic healing (MCSe≤1, MCSe=0) and scores from rectosigmoidoscopy and colonoscopy analyses were compared among 239 examinations with endoscopic assessment proximal to the rectosigmoid colon.
Results: There was a high degree of correlation between findings from rectosigmoidoscopy vs
TE D
colonoscopy in assessment of disease activity based on MCSe≥2 (r=0.84) or MCSe≥1 (r=0.96), or the UCEIS (r=0.92). In 230/239 videos, findings from rectosigmoidoscopy agreed with those from colonoscopy in detection of active disease (MCSe≥2; n=205) or healing (MCSe≤1; n=25).
EP
In 9 videos (2 taken at baseline, 7 after treatment), colonoscopy found proximal disease activity not detected by rectosigmoidoscopy. Post-treatment discordance was more frequent in the
AC C
placebo group, affecting assessment of efficacy at week 10. When endoscopic healing was defined as MCSe=0, there were discordant findings from only 1 video. Conclusions: There is a high degree of correlation in assessments of UC activity made by rectosigmoidoscopy vs colonoscopy. For detection of endoscopic healing (MCSe≤1), colonoscopy found persistent proximal lesions in the placebo group, which affected efficacy analyses. When endoscopic healing was defined as MCSe=0, the concordance between rectosigmoidoscopy and colonoscopy was nearly perfect. Keywords: anti-integrin; EUCALYPTUS; Inflammatory Bowel Disease; IBD
ACCEPTED MANUSCRIPT Page 5 of 20 INTRODUCTION Ulcerative colitis (UC) is a chronic relapsing inflammatory bowel disease (IBD) characterized by superficial mucosal ulceration, rectal bleeding, diarrhea, and abdominal pain.1 Until the late
RI PT
1990s, the therapeutic goal for UC was clinical remission as defined by symptom management. However, a growing body of evidence indicates that healing of the mucosa is an important
therapeutic endpoint in clinical trials and in clinical practice.2 This altered therapeutic paradigm
SC
has further increased the role of endoscopy in UC.
Colonoscopy with intubation of the terminal ileum is the standard of care for diagnosis of
M AN U
patients with suspected UC.3-5 Once the diagnosis is established, and when UC patients are included in clinical trials, endoscopy limited to the rectum and sigmoid is the standard to assess disease activity and endoscopic healing because it is believed that the most severe activity of UC is located in the distal colon. However, little evidence supports this assertion, and some UC patients may harbor more severe endoscopic inflammation proximal to the sigmoid colon.
TE D
Etrolizumab is a humanized monoclonal anti-β7 antibody with a dual mechanism of action, inhibiting both α4β7:MAdCAM-1-mediated lymphocyte trafficking to the gut mucosa and αEβ7:E-cadherin-mediated lymphocyte retention in the intraepithelial compartment.6-12 In a
EP
double-blind, placebo-controlled, randomized, phase 2 study, etrolizumab led to significantly greater rates of clinical remission at Week 10 over placebo in patients with moderately-to-
AC C
severely active UC.13 During this study, 72% of endoscopic examinations assessed disease activity beyond the rectosigmoid. This circumstance allowed us to evaluate whether endoscopic assessment limited to the rectosigmoid adequately represents endoscopic activity of the more proximal colon, and how potential discrepancies might affect assessment of efficacy in the context of a randomized controlled clinical trial.
ACCEPTED MANUSCRIPT Page 6 of 20 METHODS Patient Cohort The EUCALYPTUS trial evaluated the efficacy and safety of etrolizumab (Genentech, South
RI PT
San Francisco, CA) in patients with moderately-to-severely active UC, who had not responded to conventional therapy, including immunosuppressants and/or at least one anti-tumor necrosis factor (anti-TNF) agent. For this randomized, double-blind, placebo-controlled, multi-center
SC
Phase 2 study, patients were recruited from 40 centers in 11 countries. Of 124 patients, 41 were assigned to 100 mg etrolizumab administered subcutaneously (SC) at Weeks 0, 4, and 8, 40
M AN U
were assigned to 300 mg etrolizumab (Weeks 2, 4, and 8) plus a 420 mg loading dose (LD) administered between Weeks 0–2, and 43 were assigned to monthly SC injections of placebo for 3 months. Five patients were excluded from the modified intent-to-treat (mITT) population as they had a centrally read screening endoscopic subscore of < 2, leaving 41 analyzed from the placebo group, 39 from the 100 mg etrolizumab group, and 39 from the 300 mg etrolizumab
TE D
plus LD group. Details of patient demographics, methods, and eligibility criteria have been described previously.13 Briefly, patients included were adults diagnosed with UC for a minimum of 12 weeks, who had a Mayo Clinic Score14 (MCS) ≥ 5, with an MCS endoscopic subscore
EP
(MCSe) ≥ 2, a rectal bleeding symptom score ≥ 1, and disease extending ≥ 25 cm from the anal verge. The primary efficacy endpoint was clinical remission at Week 10, defined as the
AC C
proportion of patients with an MCS ≤ 2, with no individual subscore > 1. The post hoc analyses described here received institutional review board approval under the EUCALYPTUS protocol (ClinicalTrials.gov, NCT01336465). Analysis of Endoscopic Examinations The current study examined 331 videos of endoscopies captured during the EUCALYPTUS trial, after exclusion of videos without associated patient identification numbers and videos of patients excluded due to protocol violations (Supplementary Figures 1 and 2). Videos of patient
ACCEPTED MANUSCRIPT Page 7 of 20 endoscopies were recorded at the baseline screening visit (Week 0), and after randomization at Week 6 and Week 10. Out of the 331 videos, 239 (100 from baseline and 139 from Weeks 6 and 10) included assessment of colonic segments proximal to the rectosigmoid region, and
RI PT
were analyzed for disease activity and endoscopic healing in this study (Supplementary Figure 2). Endoscopic severity was assessed using the MCSe and the Ulcerative Colitis Endoscopic Index of Severity (UCEIS).15,16 For the MCSe, active disease was defined as a score > 1, and
SC
endoscopic healing as MCSe ≤ 1 or MCSe = 0. We included analyses with endoscopic healing defined as MCSe = 0 because of emerging evidence that a score of 0 may be linked to different
M AN U
patient outcomes compared with a score of 1.17 Currently, the UCEIS does not specify a cutoff score to distinguish active disease from endoscopic healing and, therefore, was not used to calculate positive predictive values (PPV) and negative predictive values (NPV). Central readers performed a preliminary evaluation of videos to determine parameters for scoring the MCSe and the UCEIS. Of note, the location of segments explored was not
TE D
annotated on the videos during recording and not all patients received full colonoscopies, with examination of all 5 colonic segments (rectum, sigmoid, descending colon, transverse colon, and ascending colon). Four central readers (2 per video) from 2 sites estimated the number of
EP
colonic segments examined, guided by the length of the segments and identification of anatomical characteristics such as the hepatic, splenic, and rectosigmoid flexures. Readers also
AC C
determined the extent of disease.
For determining endoscopic healing based on the MCSe, the worst score of the rectosigmoid alone was compared with the worst score of the whole colon (including the rectosigmoid). A separate, independent, central reader adjudicated the final score when discrepancies emerged between the two readers of the same video. For the analysis of individual endoscopic severity scores (Figure 1), all 331 videos were scored with both the MCSe and the UCEIS, with 2 reads per video (one from each reader, without adjudication; 662 reads). Reads with ≤ 2 segments were excluded (251), resulting in a
ACCEPTED MANUSCRIPT Page 8 of 20 total of 411 reads (Supplementary Figure 1). The rectosigmoid and the proximal colon (descending, transverse, and ascending colon) were scored separately, according to the worstaffected segment. Each component of the UCEIS (vascular pattern, bleeding, and erosions and
RI PT
ulcers) was scored separately. Total UCEIS scores were generated from each reader from the sum of their respective subscores. Statistical Analysis
SC
To determine the level of correlation between rectosigmoidoscopy and colonoscopy for detecting disease activity and endoscopic healing, contingency tables were generated that
M AN U
tabulated the number of adjudicated MCSe scores above and below the indicated MCSe cutoffs. The degree of inter-method agreement and correlation was determined by the calculation of the kappa coefficient (κ) and the Spearman correlation coefficient (r) for each contingency table, respectively. The interpretation of the κ and the correlation coefficients was based on the proposed cutoffs by Landis and Koch,18 who recommended the following
TE D
standards for strength of agreement for the κ coefficient: 0, poor; 0.01–0.20, slight; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, substantial; and 0.81–1, nearly perfect. NPVs and PPVs were also calculated. P-values were based on the two-sided Fisher exact test. Because this
EP
analysis was retrospective and not a prespecified goal of the EUCALYPTUS protocol, formal
AC C
sample size calculations were not performed.
ACCEPTED MANUSCRIPT Page 9 of 20 RESULTS Number of bowel segments evaluated and description of extent of disease Comparable numbers of videos were examined at each visit for each treatment group (Table
RI PT
1). Across the 3 treatment groups, the average total segments examined was 3.2, and the number of segments examined across different visits (Weeks 0, 6, and 10) and between treatment groups did not differ significantly.
SC
Across all treatment groups, 25% of patients had UC limited to the rectosigmoid region and 75% of patients had disease in the proximal colon, defined as any region beyond the
M AN U
rectosigmoid (Table 2). Overall evaluation
Among all of the videos, regardless of the number of segments read, 67 videos (67/354 = 19%) were sent for adjudication. Among videos with > 2 segments, 55 videos (55/239 = 23%) were sent for adjudication. The MCSe (mean ± SD) in videos where only two segments or fewer
TE D
were examined (2.66 ± 0.63) did not differ significantly from videos with > 2 segments examined (2.54 ± 0.75; Supplementary Table 1).
Across all visits and examinations, 239 videos examined colonic segments proximal to the
EP
rectosigmoid. For 230 videos, rectosigmoidoscopy and colonoscopy analyses agreed, determining that active disease (MCSe ≥ 2) was present in 205, but absent in 25 videos (Table
AC C
3; left). There was a strong correlation between the scores based on the two techniques (κ = 0.83; r = 0.84). However, for 9 of the 239 videos, the scores were discordant in assessing active disease, i.e. rectosigmoidoscopy found evidence of endoscopic healing, but examination of the more proximal colon with colonoscopy found mucosal lesions (MCSe ≥ 2). These 9 videos came from 7 patients: 3 with left-sided colitis, 2 with pancolitis, 1 with disease previously limited to the rectosigmoid, and 1 where the previous extent of disease was not specified. If endoscopic healing was defined as MCSe = 0 and endoscopic activity was defined as MCSe ≥ 1, the concordance between rectosigmoidoscopy and colonoscopy increased (κ = 0.95; r = 0.96), as
ACCEPTED MANUSCRIPT Page 10 of 20 only 1 video demonstrated endoscopic healing in the rectosigmoid while significant mucosal lesions were found in the proximal colon (Table 3; right). We then analyzed whether the severity of disease differed in the proximal colon compared
RI PT
with the rectosigmoid, using both the MCSe and the UCEIS (Figure 1; Supplementary Figure 3). For this analysis, each region was scored separately and there was no adjudication between readers; each score given by the readers was considered an individual piece of data. In general,
SC
we observed more severe disease in the rectum and sigmoid according to the MCSe, but 19 of the 411 (4.6%) individual reads with more than 2 segments showed greater disease severity in
M AN U
the proximal colon than in the rectosigmoid region (Figure 1A; gray cells). Out of those 19 scores, 14 (rectosigmoid ≤ 1, proximal colon ≥ 2) affected the assessment of endoscopic healing in the colon. In addition, assessment of endoscopic disease using the UCEIS confirmed the presence of more severe disease in the rectosigmoid compared with the proximal colon (3.98 ± 1.82 vs. 2.38 ± 1.98; P < .0001), but in a small proportion of the videos (31/411; 7.5%),
TE D
more severe disease was found in the proximal colon. Scoring with the UCEIS, using both the total UCEIS and individual subscores, also demonstrated a high correlation between rectosigmoidoscopy and colonoscopy (r = 0.92; Figure 1B; Supplementary Figure 3).
EP
Evaluation of endoscopic activity at baseline When we assessed endoscopic activity (MCSe ≥ 2) at baseline, we observed a moderate
AC C
agreement/correlation (κ = 0.49, r = 0.57) between the scores based on rectosigmoidoscopy and colonoscopy (Table 4). This level of correlation was related to the highly skewed distribution of MCSe values towards scores of 2 and 3, given the study inclusion criteria. In 2 videos out of 100 performed at baseline, rectosigmoidoscopy failed to find active disease while colonoscopy revealed active disease above the sigmoid colon. Evaluation of endoscopic healing After treatment with etrolizumab, patients were then assessed for endoscopic healing (MCSe ≤ 1) at Weeks 6 and 10 using a total of 139 available videos (Table 5; left). While 31
ACCEPTED MANUSCRIPT Page 11 of 20 videos showed endoscopic healing in the rectum and sigmoid colon, only 24 also had healing in the proximal colon, a difference of approximately 23% (PPV for endoscopic healing by rectosigmoidoscopy = 0.77). Colonoscopy and rectosigmoidoscopy demonstrated a high
RI PT
correlation for the detection of endoscopic healing (κ = 0.84; r = 0.85). With endoscopic healing defined as MCSe = 0 at Weeks 6 and 10 (Table 5; right), 12 videos showed healing in the rectum and sigmoid colon. Only 1 video did not show endoscopic healing
SC
in the proximal colon, a difference of approximately 9%. Therefore, when endoscopic healing was defined as an MCSe = 0, the correlation between the two techniques was very strong (κ =
M AN U
0.95; r = 0.95), and rectosigmoidoscopy was better able to predict endoscopic healing (PPV = 0.92) in the proximal colon.
When we investigated the efficacy of etrolizumab using rectosigmoidoscopy, we found that there was a trend towards higher rates of endoscopic healing, but there was no significant treatment effect for either dose group over placebo by Weeks 6 and 10 (Figure 2; white bars).
TE D
Colonoscopy also showed no significant treatment effect at Week 6 (Figure 2A). However, at Week 10, colonoscopy determined that the proportion of patients with endoscopic healing in the placebo group was 8%, rather than 14%, the proportion measured by rectosigmoidoscopy
EP
(Figure 2B). While this difference did not affect the efficacy assessment for the 300 mg dose group, this lower percentage of patients with healing increased the effect size to a significant
AC C
level between the placebo and the 100 mg etrolizumab treatment groups (P = .035).
ACCEPTED MANUSCRIPT Page 12 of 20 DISCUSSION In general, when we compared the ability of rectosigmoidoscopy with colonoscopy for determining either the presence of active disease or endoscopic healing in patients with
RI PT
moderately-to-severely active UC, we found strong correlations between the two examinations. Rectosigmoidoscopy was reliably able to detect active disease at the inclusion of patients in the study. We also found that the most severe disease was located more frequently in the
SC
rectosigmoid colon, a relevant finding for clinical practice because therapeutic decisions are based not only on the presence of active disease but also on severity.
M AN U
The ability of rectosigmoidoscopy to detect disease activity compared with colonoscopy differed between baseline and post treatment at Week 6 and Week 10. At baseline, rectosigmoidoscopy accurately reflected disease activity in the whole colon. In contrast, at Weeks 6 and 10, when endoscopic healing was defined as MCSe ≤ 1, rectosigmoidoscopy indicated healing of the distal mucosa, although there was evidence of mucosal ulceration more
TE D
proximally on colonoscopy. This resulted in a placebo remission rate that was almost two-fold higher based on rectosigmoidoscopy compared with colonoscopy (14% vs. 8%, respectively; Figure 2B), and a decreased treatment effect size. However, when endoscopic healing was
EP
defined as MCSe = 0, concordance improved between rectosigmoidoscopy and colonoscopy, as only 1 out of 11 videos had discordant scores between the two techniques (Table 5; right).
AC C
According to guidelines and consensus panels, once the diagnosis is established, rectosigmoidoscopy alone is usually sufficient to assess endoscopic activity/severity in the management of UC.3-5 Few studies, however, and none in the context of a therapeutic trial, have formally compared endoscopic severity in the rectosigmoid versus the whole colon. In a retrospective analysis of medical charts and endoscopic images from 545 UC patients, Kato et. al19 determined the colonic locations of the maximum inflammatory activity. While the majority of the patients analyzed (73%) had maximum inflammation in the rectum and sigmoid, 27% had maximum activity proximal to the sigmoid, in the descending colon, or in segments proximal to
ACCEPTED MANUSCRIPT Page 13 of 20 the splenic flexure. In addition, they found that 40% of patients had inflamed mucosa in the descending colon or in the more proximal portion of the colon, but showed no inflammation in the rectum and sigmoid. They concluded that total colonoscopy was warranted in patients who
RI PT
have a discrepancy between their symptoms and rectosigmoidoscopy results. A caveat of this study was that the use of rectal therapy with mesalazine or steroids was allowed. In the EUCALYPTUS trial, rectal therapy was withdrawn 2 weeks prior to randomization and
concomitant use of rectal therapy was contraindicated during the treatment phase of the trial,
SC
due to the possibility of rectal sparing.20
M AN U
In this endoscopy study, we used the UCEIS in addition to the MCSe to compare disease severity in the rectosigmoid and the proximal colon. We were unable to use the UCEIS to assess the endpoints of disease activity or endoscopic healing because a threshold score for mucosal healing has not yet been established.16 In general, the UCEIS confirmed the MCSe findings of more severe disease in the rectosigmoid for all three subscores (Supplementary
TE D
Figure 3) and for the overall score (Figure 1B). The UCEIS, as with the MCSe, also detected a small, but slightly higher proportion of videos where proximal disease was more severe than in the rectosigmoid (7.5% vs. 4.6% for the MCSe; Figure 1) and showed a very high correlation
EP
between rectosigmoidoscopy and colonoscopy.
One limitation of our study was that only 239 out of 331 examinations progressed beyond
AC C
the sigmoid colon. There may be some concern that the reasons for the unavailability of colonoscopy in those patients may have introduced bias into the study. Patients with severe disease above the sigmoid may have been underestimated because of concerns by the endoscopist in proceeding with the examination beyond the sigmoid colon. However, the number of segments analyzed between treatment groups was similar (Table 1) and disease severity (mean MCSe) was also similar between videos with fewer than 2 and those with more than 2 segments (Supplementary Table 1), which rules out a bias arising from performing rectosigmoidoscopy or colonoscopy based on the endoscopic severity of the disease. Another
ACCEPTED MANUSCRIPT Page 14 of 20 potential limitation was that because the explored segments were not annotated in the videos, the readers identified segments based on anatomical characteristics and the length of the segment; in particular, the boundary between the descending colon and the sigmoid could be
RI PT
difficult to distinguish, even for skilled endoscopists. It is, however, unlikely that this would have significantly affected evaluation of the rectosigmoid compared with the whole colon. We also acknowledge that, in this trial, the number of MCSe scores where the proximal colon had more
SC
severe disease than the rectosigmoid (9/239) was low. However, the number of patients that typically achieve remission in induction studies is also low, which means that factors that
M AN U
influence placebo rates can have a direct effect on statistical significance and influence decisions about the further development of experimental therapies. Finally, as this was an induction study, we cannot extrapolate these findings to maintenance studies. Longer treatment duration with therapies that are effective may improve the concordance between rectosigmoidoscopy and colonoscopy, and consequently demonstrate that rectosigmoidoscopy
TE D
is sufficient for measuring clinical effectiveness in maintenance trials. Despite these limitations, this study has important implications for the use of rectosigmoidoscopy and/or colonoscopy in both clinical practice and during clinical trials.
EP
Because of its strong ability to detect active disease, rectosigmoidoscopy should be sufficient in clinical practice for evaluating previously diagnosed patients with new symptoms. Two notable
AC C
exceptions may be pediatric patients and patients with primary sclerosing cholangitis where relative or complete rectal sparing has been observed.21,22 However, in cases where rectosigmoidoscopy has detected endoscopic healing in response to induction therapy, but symptoms persist, performance of a more extensive examination with colonoscopy is justified. In the clinical trial setting, this study suggests that if endoscopic healing is defined as an MCSe = 0, rectosigmoidoscopy is sufficient for efficacy analyses. However, if endoscopic healing is defined as an MCSe ≤ 1, these preliminary data suggest that colonoscopy is better than rectosigmoidoscopy for assessing the full extent of endoscopic healing for efficacy analyses of
ACCEPTED MANUSCRIPT Page 15 of 20 experimental therapies. Colonoscopy is more expensive than rectosigmoidoscopy because it frequently requires sedation and bowel preparation, which in turn, may impose additional burdens on the patient, such as missing work and requiring a caregiver to transport the patient
RI PT
to and from the endoscopy suite. The balance between cost and performance is an important consideration and is most relevant in the development of experimental therapies. It may be preferable to use full colonoscopy in relatively small Phase 2 induction studies. However, these
AC C
EP
TE D
M AN U
recommended for standard use in clinical trials.
SC
results warrant further evaluation prospectively in a larger study before colonoscopy can be
ACCEPTED MANUSCRIPT Page 16 of 20 Figure 1. Endoscopic severity in the rectosigmoid colon vs. the proximal colon (all visits; unadjudicated scores). Two readers assessed each video and scored the severity of disease for the rectosigmoid colon and the proximal colon, using the MCSe (A) or the UCEIS (B). Numbers
RI PT
in each cell represent the total number of scores of the individual site readers (two scores per video) without adjudication (pooled across readers). Data was pooled across the sites. Any reading with ≤ 2 segments was not included (Supplementary Figure 1). Gray cells in (A) and (B)
SC
indicate where the proximal colon score was higher than the rectosigmoid colon score. Scores in the light-gray cells in (A) affected the assessment of endoscopic healing when the proximal colon was included in the evaluation; the dark-gray cell in (A) indicates the MCSe scores that
M AN U
did not affect the overall endoscopic healing evaluation. κ and correlation coefficients are provided for the comparison of rectosigmoidoscopy and colonoscopy.
Figure 2. Comparison of endoscopic healing using rectosigmoidoscopy and colonoscopy.
TE D
The percentage of patients (mITT population) with endoscopic healing (MCSe ≤ 1) in placebo and etrolizumab-treated groups was calculated at Week 6 (A) and Week 10 (B) from videos based on rectosigmoidoscopy only (white bars) and colonoscopy (black bars). *Scoring was
AC C
EP
based on adjudicated results; ‡P < .05; Pbo, placebo; Etro, etrolizumab; LD, loading dose.
ACCEPTED MANUSCRIPT Page 17 of 20 REFERENCES 1. Kaser A, Zeissig S, Blumberg RS. Inflammatory bowel disease. Annu Rev Immunol 2010;28:573-621.
RI PT
2. Neurath MF, Travis SPL. Mucosal healing in inflammatory bowel diseases: a systematic review. Gut 2012;61:1619-1635.
3. Kornbluth A, Sachar DB, The Practice Parameters Committee of the American College of
SC
Gastroenterology. Ulcerative colitis practice guidelines in adults: American College of
Gastroenterology, Practice Parameters Committee. Am J Gastroenterol 2010;105:501-523.
M AN U
4. Dignass A, Eliakim R, Magro F, et al. Second European evidence-based consensus on the diagnosis and management of ulcerative colitis Part 1: definitions and diagnosis. J Crohns Colitis 2012;6:965-90.
5. Annese V, Daperno M, Rutter MD, et al. European evidence based consensus for endoscopy in inflammatory bowel disease. J Crohns Colitis 2013;7:982-1018.
TE D
6. Karecla PI, Bowden SJ, Green SJ, et al. Recognition of E-cadherin on epithelial cells by the mucosal T cell integrin alpha M290 beta 7 (alpha E beta 7). Eur J Immunol 1995;25:852-6. 7. Feagan BG, Greenberg GR, Wild G, et al. Treatment of ulcerative colitis with a humanized
EP
antibody to the α4β7 integrin. N Engl J Med 2005;352:2499-507. 8. Cepek KL, Parker CM, Madara JL, et al. Integrin alpha E beta 7 mediates adhesion of T
AC C
lymphocytes to epithelial cells. J Immunol 1993;150:3459-70. 9. Hesterberg PE, Winsor-Hines D, Briskin MJ, et al. Rapid resolution of chronic colitis in the cotton-top tamarin with an antibody to a gut-homing integrin alpha 4 beta 7. Gastroenterology 1996;111:1373-80. 10. Ludviksson BR, Strober W, Nishikomori R, et al. Administration of mAb against αEβ7 prevents and ameliorates immunization-induced colitis in IL-2-/- mice. J Immunol 1999;162:4975-82. 11. Sandborn WJ, Colombel JF, Enns R, et al. Natalizumab induction and maintenance therapy
ACCEPTED MANUSCRIPT Page 18 of 20 for Crohn's disease. N Engl J Med 2005;353:1912-25. 12. Elewaut D, Van Damme N, De Keyser F, et al. Altered expression of alpha E beta 7 integrin on intra-epithelial and lamina propria lymphocytes in patients with Crohn’s disease.
RI PT
Acta Gastroenterol Belg 1998;61:288-94. 13. Vermeire S, O'Byrne S, Keir M, et al. Etrolizumab as induction therapy for ulcerative colitis: a randomised, controlled, phase 2 trial. Lancet 2014;384:309-318.
SC
14. Schroeder KW, Tremaine WJ, Ilstrup DM. Coated oral 5-aminosalicylic acid therapy for mildly to moderately active ulcerative colitis. A randomized study. N Engl J Med
M AN U
1987;317:1625-9.
15. Travis SP, Schnell D, Krzeski P, et al. Developing an instrument to assess the endoscopic severity of ulcerative colitis: the Ulcerative Colitis Endoscopic Index of Severity (UCEIS). Gut 2012;61:535-42.
16. Travis SP, Schnell D, Krzeski P, et al. Reliability and initial validation of the ulcerative colitis
TE D
endoscopic index of severity. Gastroenterology. 2013;145:987-95. 17. Manginot C, Baumann C, Peyrin-Biroulet L. An endoscopic Mayo score of 0 is associated with a lower risk of colectomy than a score of 1 in ulcerative colitis. Gut 2014;0:1-2.
EP
18. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-174.
AC C
19. Kato J, Kuriyama M, Hiraoka S, et al. Is sigmoidoscopy sufficient for evaluating inflammatory status of ulcerative colitis patients? J Gastroen Hepatol 2011;26:683-687. 20. Odze R, Antonioli D, Peppercorn M, et al. Effect of topical 5-aminosalicylic acid (5-ASA) therapy on rectal mucosal biopsy morphology in chronic ulcerative colitis. Am J Surg Pathol. 1993;17:869-75. 21. Rajwal SR, Puntis JW, McClean P, et al. Endoscopic rectal sparing in children with untreated ulcerative colitis. J Pediatr Gastroenterol Nutr 2004;38:66-9. 22. Loftus EV Jr, Harewood GC, Loftus CG, et al. PSC-IBD: a unique form of inflammatory
ACCEPTED MANUSCRIPT Page 19 of 20 bowel disease associated with primary sclerosing cholangitis. Gut 2005;54:91-6.
AC C
EP
TE D
M AN U
SC
RI PT
Author names in bold designate shared co-first authorship.
ACCEPTED MANUSCRIPT
Table 1. Mean number of colon segments examined by treatment group and visit Etrolizumab
100 mg
300 mg + LD
Placebo POOLED n
Mean (SD)
Segments
n
40
3.9 (1.2)
38
Week 6
35
2.7 (0.9)
34
Week 10
37
2.9 (0.9)
36
Total
112
3.2 (1.2)
108
AC C
EP
TE D
n, number of videos evaluated; LD, loading dose.
n
Mean (SD)
4.0 (1.1) 3.1 (1.0)
38
Mean (SD)
3.6 (1.3)
37
3.0 (1.0)
2.9 (1.0)
36
2.6 (0.9)
3.4 (1.1)
111
3.1 (1.1)
M AN U
Week 0/baseline
Segments
SC
Segments
RI PT
Etrolizumab
ACCEPTED MANUSCRIPT
Table 2. Extent of disease Etrolizumab
Etrolizumab
100 mg
300 mg
(n = 39)
(n = 39)
Placebo Extent of disease, n (%)
Total (n = 119)
RI PT
(n = 41)
Extensive/pancolitis
12 (29%)
14 (36%)
Left-sided colitis
17 (42%)
13 (33%)
Rectosigmoid
12 (29%)
10 (26%)
8 (21%)
30 (25%)
Non-specified
0 (0%)
2 (5%)
0 (0%)
2 (2%)
44 (37%)
13 (33%)
43 (36%)
SC
M AN U
AC C
EP
TE D
n, number of patients.
18 (46%)
ACCEPTED MANUSCRIPT
Table 3. Identification of endoscopically active disease as assessed by rectosigmoidoscopy vs. colonoscopy (all visits)
Colonoscopy
(n = 239)
Y
MCSe ≥ 1
Colonoscopy
(n = 239) Y
N
205
0
Y
227
N
1
R/S N
9
25
N 0
SC
R/S
Y
RI PT
MCSe ≥ 2
11
κ = 0.95; r = 0.96; P < .0001;
NPV = 0.74 (95% CI: 0.59, 0.88)
NPV = 0.92 (95% CI: 0.76, 1.07)
M AN U
κ = 0.83; r = 0.84; P < .0001;
This analysis only included assessments with > 2 segments and all visits were pooled. Gray
TE D
cells represent the number of discordant scores for endoscopic activity that were affected when the proximal colon was included in the evaluation. Presence of endoscopically active disease (MCSe ≥ 2, left; MCSe ≥ 1, right): Y = yes, N = no; CI, confidence interval; n, number of videos
AC C
EP
evaluated; NPV, negative predictive value; R/S, rectosigmoidoscopy.
ACCEPTED MANUSCRIPT
Table 4. Identification of endoscopically active disease (MCSe ≥ 2) as assessed by
Colonoscopy
(n = 100)
Y
N
Y
97
0
N
2
1
κ = 0.49; r = 0.57; P = .03; NPV = 0.33 (95% CI: -0.02, 0.87)
M AN U
R/S
SC
MCSe ≥ 2
RI PT
rectosigmoidoscopy vs. colonoscopy at Week 0/baseline
This analysis only included assessments with > 2 segments. The gray cell represents the
TE D
number of scores for endoscopic activity that were affected when the proximal colon was included in the evaluation. Presence of endoscopically active disease (MCSe ≥ 2): Y = yes, N = no; CI, confidence interval; n, number of videos evaluated; NPV, negative predictive value; R/S,
AC C
EP
rectosigmoidoscopy.
ACCEPTED MANUSCRIPT
Table 5. Evaluation of endoscopic healing as assessed by rectosigmoidoscopy vs. colonoscopy at Weeks 6 and 10 (pooled)
Colonoscopy
(n = 139)
N
Y
N
108
0
Y
7
24
(n = 139)
R/S
Y
N
127
0
Y
1
11
κ = 0.95; r = 0.95; P < .0001;
M AN U
κ = 0.84; r = 0.85; P < .0001; PPV = 0.77 (95% CI: 0.63, 0.92)
N
SC
R/S
Colonoscopy
MCSe = 0
RI PT
MCSe ≤ 1
PPV = 0.92 (95% CI: 0.76, 1.07)
This analysis only included assessments with > 2 segments by visit (excluding the screening visit). Gray cells represent the number of scores for endoscopic healing that were affected when
TE D
the proximal colon was included in the evaluation. Presence of endoscopic healing (MCS ≤ 1, left; MCSe = 0, right): Y = yes, N = no; CI, confidence interval; n, number of videos evaluated;
AC C
EP
PPV, positive predictive value; R/S, rectosigmoidoscopy.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Segments ≤ 2 (n = 92)
Segments > 2 (n = 239)
2.66 (0.63)
2.47 (0.80)
3 (1, 3)
3 (0, 3)
Mean (SD)
N/A
2.54 (0.75)
Median (range)
N/A
Mean (SD) Median (range)
SC
MCSe (rectosigmoid)
M AN U
MCSe (colon)
Extent of disease, n (%)
3 (0, 3)
35 (38%)
88 (37%)
Left-sided colitis
31 (34%)
89 (37%)
25 (27%)
59 (25%)
1 (1%)
3 (1%)
EP
Non-specified
TE D
Extensive/pancolitis
Rectosigmoid
n, number of videos evaluated; N/A, not applicable.
AC C
RI PT
Supplementary Table 1. MCSe of videos with ≤ 2 and > 2 segments
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
