IMPROVEMENTS IN RISK STRATIFICATION AND COLONOSCOPIC HEMOSTASIS TO PREVENT DELAYED POST-POLYPECTOMY INDUCED ULCER HEMORRHAGE

Background: Biological treatment in inflammatory bowel disease (IBD) patients is currently hampered by high non-response rates. To enhance personalized medicine and predict response to biological therapeutics such as vedolizumab, the working mechanism should be elucidated. We aimed to visualize macroscopic and microscopic vedolizumab distribution and detect drug target cells during quantified fluorescence molecular endoscopy (QFME) to improve understanding of the mechanism of action.

Methods: Vedolizumab-800CW was developed and GMP produced. Forty-three QFME procedures were performed in thirty-seven IBD patients +/- three days after intravenous administration of vedolizumab-800CW. Each QFME procedure consisted out of endoscopic assessment of the inflammation status per colonic segment by high-definition white light endoscopy, followed by real-time in vivo assessment of the macro-distribution of fluorescent vedolizumab-800CW and quantification by spectroscopy of selected segments. Dose escalation was performed using 0.0 mg, 4.5 mg and 15 mg. Subsequently, two patient cohorts were added that received 75 mg or a therapeutic (300 mg) dose of unlabelled vedolizumab prior to vedolizumab-800CW to assess target saturation. Tissue biopsies were obtained for histopathological assessment, for further ex vivo analysis of the fluorescent signal and for visualization of the microscopic distribution and identification of vedolizumab-800CW target cells by fluorescence microscopy (fig 1).

Results: Macroscopically and microscopically a significant difference between inflamed and non-inflamed tissue was visualized (fig 2A) and quantified (0.0227 and 0.0470 Q*μfa,x [mm-1], p<0.0001) for the 15 mg cohort. In addition, ex vivo analysis showed a clear dose-dependent increase (p<0.0001) of the fluorescent drug signal, whereas a decrease could be established after adding an unlabelled dose (p<0.0001) (fig 2B). Fluorescence microscopy revealed clear membrane binding of vedolizumab-800CW to inflammatory cells and migration into the inflamed mucosa. Additional analyses to identify specific target cells are ongoing and immune compositions of regions with high and low vedolizumab signal will be unravelled.

Conclusion: QFME using vedolizumab-800CW elucidated novel detailed macroscopic and microscopic vedolizumab distribution in the inflamed target organ. In addition, its shows the potential of QFME to better understand local drug distribution, target cell identification and target engagement, which could improve understanding of targeted drugs over standard pharmacokinetic and pharmacodynamic analysis.

Background and study aim:
Dysplasia is frequent in Inflammatory Bowel Disease (IBD) patients and can be managed endoscopically providing organ sparing, in contrast to surgery. Endoscopic submucosal dissection (ESD) allows en-bloc resection and very low recurrence rates. ESD is a validated endoscopic resection technique for sporadic colorectal superficial neoplasms. However, data is still missing to confirm its short and long term efficacy for visible dysplasia in IBD patients.

Patients and methods:
We conducted a retrospective multicentric study including all consecutive ESD in IBD patients with visible dysplasia in 20 French centers.
The primary outcome was the rate of R0 resection.

Results:
89 lesions in 83 patients including 20 Crohn’s disease (CD) were resected. Mean follow-up was 26 months (+/- 25 SD). Median IBD duration was 21.3 years (+/- 11.9 SD). PSC was associated with 9 lesions (10.8%). 50% of ulcerative colitis (UC) patients were classified E3. Among CD patients, 13 were L2 (68.4%) and 5 were L3 (26.3%).
En-bloc resection, R0 resection and curative resection were achieved in 80 (91%), 71 (80%) and 69 (77.5%) lesions, respectively. 1 (1.2%) patient required surgery for complication, 3 (3.6%) for ESD failure and 6 (7.2%) for bad prognostic histological features. Traction strategy was used in 53 cases (59.55%) providing higher en-bloc resection rates (50 [96.2%] versus 30 [83.3%], p=0.040) of larger (48.84 mm +/-20.86 SD versus 34 mm +/-15.75 SD, p=0.001) and more severe fibrotic (34 [65.4%] versus 12 [38.7%], p=0.023) lesions than without traction. High-volume centers (> 100 colorectal ESD per year) performed larger (50.56 mm +/-20.37 SD versus 35.96 +/-17.63 SD, p=0,001) and faster (23.34 mm²/min +/-19.11 SD versus 10.36 mm²/min +/- 5.45 SD, p<0.0001) resections with less recurrences (0 [0%] versus 8 [16.7%], p=0.006) than low-volume centers. Recurrence was found to be more frequent in CD patients (n = 6, 26.1%) than in UC patients (n = 2, 3.1%) (p=0.004).

Discussion:
Concerning en-bloc and R0 resection, the results are close to those obtained from ESD for sporadic neoplasms. Perforation rate was higher than those from studies in sporadic lesions (14%) confirming the more complexity of these lesions that should be reserved for high volume experts centers.

Conclusion:
This study is the world’s largest reported so far concerning ESD for associated IBD visible dysplasia, and with the largest CD cohort. This study confirms the efficacy of ESD regarding en-bloc, R0 and curative resections, even in the CD population. ESD in IBD patients should be performed in high-volume centers with traction strategy to improve oncological outcomes and decrease complications. However, a long term follow-up is necessary to assess the impact of ESD on long-term surgery rate in the era of “treat to target”.

Background and aims
Patients with inflammatory bowel diseases (IBD) with an increased risk of developing colorectal carcinoma should undergo periodic surveillance colonoscopies. There is no quality metric for dysplasia detection rate (DDR) in IBD surveillance. We evaluated DDR in a dedicated surveillance program at a tertiary IBD referral center.

Methods
This cross-sectional study assessed DDR among consecutive patients with quiescent colitis enrolled in our surveillance program. Patients underwent high-definition colonoscopy with dye chromoendoscopy (DCE). A single specialized operator performed the procedures. Advanced dysplasia (AD) was defined as low-grade dysplasia ≥ 10mm, high-grade dysplasia, or colorectal cancer. We evaluated risk factors for dysplasia detection.

Results
Overall, 119 patients [female: 39.5%; median age: 54 years (IQR 43-66); median disease duration: 20 years (IQR 14-29)] underwent 151 procedures that revealed 206 lesions, of which 40 were dysplastic, and seven were considered AD. Per-procedure and per-lesion DDR were: 20.5%, and 19.4%, respectively. Per-procedure AD detection rate (ADDR) was 4.6%. Per-procedure dysplasia detection was associated with increased age at diagnosis at the index colonoscopy and past dysplasia or indefinite dysplasia, however, on multivariable analysis, only past dysplasia or indefinite dysplasia maintained a significant association (AdjOR 4.84, 95% CI 1.52-15.45, p=0.008). A Kudo pit pattern of II-V had a sensitivity of 92.5% for per-lesion dysplasia detection but a false positive rate of 64.8% (p<0.001).

Conclusions
DDR in a dedicated surveillance program in a real-world setting reached 20%. DDR should be considered as a quality measure while surveying high-risk patients with IBD.

Background and aims: The prognosis of patients with gastrointestinal bleeding (GIB) has improved with the technological advances of endoscopic hemostasis. Empirically, the clinical severity including bleeding-related death and bleeding source in patients with GIB is believed to differ between in-patient onset (I-O) and out-patient onset (O-O), but little is known about these clinical questions. The aim of this study was to evaluate the differences in prognosis including clinical characteristics of acute lower GIB (ALGIB) between I-O and O-O onset. Methods: From January 2011 to December 2019, patients who had undergone emergency endoscopy for GI bleeding were retrospectively reviewed. Of all 1,952 patients who had undergone emergency endoscopy, 723 patients with ALGIB (37.0%) were enrolled. They were divided into 2 cohorts; I-O cohort (172 patients, 23.8%) and O-O cohort (551 patients, 76.2%). The clinical differences between 2 cohorts were evaluated. Furthermore, subgroup analysis was performed. The severity of ALGIB was evaluated using NOBLADS scores. Patients in whom obvious ALGIB relapsed and/or iron deficiency anemia persisted after emergency endoscopy were considered to exhibit good clinical course. Results: The I-O cohort had a significantly higher proportion of males and older age than the O-O cohort (both P < 0.001). The proportions of patients with cardiovascular diseases, cerebrovascular diseases, chronic kidney disease under hemodialysis, and malignant tumors were significantly higher in the I-O cohort than the O-O cohort (all P<0.001). In a comparison of the bleeding sources, the I-O cohort had significantly higher proportions of patients with acute hemorrhagic rectal ulcer and nonspecific ulcer than the O-O cohort (all P<0.05). The proportions of patients with colonic diverticular bleeding, ischemic colitis, and hemorrhoids were significantly higher in the O-O cohort than the I-O cohort (all P<0.05) (Table 1). NOBLADS score were significantly higher than in the I-O cohort compare to O-O cohort (Figure 1A). Comparisons of patients with a poor clinical course and patients who underwent endoscopic hemostasis in the I-O and O-O cohorts are presented in Figure 1B and 1C (all P<0.001). Four patients in the I-O cohort died from ALGIB-related causes (2.3%). However, there were no bleeding-related deaths in the O-O cohort. Multivariate analysis showed that a platelet count >25×10⁴/μL and albumin concentration <3g/dL were significantly associated with a poor clinical course in the I-O cohort. Conclusions: The clinical course was significantly worse in patients with I-O ALGIB compared with that in patients with OO ALGIB. The bleeding source, clinical characteristics, and clinical course differed between the I-O and O-O cohorts. The clinical course in the I-O cohort may depend on the patient’s condition at ALGIB onset.

Background and Study Aims
Delayed post-polypectomy induced ulcer hemorrhage (DPPIUH) is a common complication of colonoscopic polypectomy. Empiric hemoclip closure (HC) of PPIU’s has mixed results, is expensive, and does not target invisible, submucosal arteries which are buried without blood flow obliteration by single layer closure of the mucosa by HC’s. Doppler endoscopic probe (DEP) interrogation of PPIUs to detect arterial blood flow just after polypectomy is easy to learn and has been reported to significantly reduce DPPIUH. We assessed relationships between arteries found on histology in resected polyps, DEP blood flow detection, and risk factors causing DPPIUH after empiric HC closure or DEP guided treatment.

Methods
Colon polyp vasculature on histopathology, DEP results, and DPPIUH were studied in high-risk patients on anti-thrombotic drugs and in polyp’s with PPIU’s > 10 mm. Resected polyps were coded and reviewed by expert GI pathologists for presence and size of submucosal vessels. Results were correlated with prospectively collected clinical outcomes, DEP results, and development of DPPIUH with DEP guided risk stratification and treatment, or empiric HC closure of PPIU’s.

Results
Resected polyp specimens from 98 patients were examined. 59 patients had both DEP interrogation and histopathology assessment. Of patients whose PPIU’s had positive DEP signals, 80% (36/45) had medium or small arteries at the cut margin of the polyp. No patient (0/31) with positive DEP and focal treatment with HC or bipolar coagulation in the PPIU had DPPIUH. For the negative Doppler group, 13 of 14 specimens (92.9%) had no arteries or only small arteries and no patient developed DPPIUH. For 39 other patients with severe DPPIUH, medium sized arteries were seen in 82% and 85% of the PPIU’s had arterial blood flow detected by DEP. In the patients with empiric HC and severe delayed hemorrhage, all (15/15) had open PPIU’s, and exposed submucosa, 86.7% (13/15) had a stigma of recent hemorrhage in the PPIU and positive arterial flow by Doppler detection, and 80% had a medium or large size artery in the previously resected polyp.

Conclusions
There were direct relationships between a positive arterial DEP signal in PPIUs, the size of arteries in resected colon polyps, and development of severe DPPIUH. When DEP was negative, no or small arteries were found on histology in 92.9% and no patient had DPPIUH. When DEP was positive, medium sized arteries were seen in the resected polyp, and with focal treatment in the PPIU, no patient developed DPPIUH. Based upon histopathology and Doppler studies, empiric hemoclip closure of PPIU’s did not reliably obliterate submucosal arteries nor prevent DPPIUH. These results support the anatomic rationale of DEP monitoring and focal treatment to obliterate the submucosal artery and prevent severe DPPIUH.