Introduction

Computer Aided Detection (CADe) systems have the potential to enhance the detection of early neoplasia in Barrett's Esophagus (BE) patients during endoscopic procedures. In this study, we aimed to evaluate a recently developed CADe system during live endoscopic procedures.

Aims & Methods

The CADe system was developed using a large and diverse training data set from 15 international endoscopy centers, comprising 6,237 neoplastic images from 1,304 patients and 7,595 non-dysplastic images from 1,103 patients. Multiple extensive ex-vivo benchmarking studies were performed for validation. The CADe system displayed consistent and reliable performance, leading to a significant improvement in the detection rate of neoplasia by endoscopists.
During this pilot study, the CADe system was evaluated in live endoscopic procedures of BE patients with neoplastic lesions or non-dysplastic Barrett's esophagus (NDBE) in two tertiary hospitals. The protocol involved a series of white light endoscopy videos obtained by an expert endoscopist with real-time evaluation and feedback provided by the CADe system. First, the Barrett's segment was comprehensively visualized using a standardized "pullback video", spanning from the gastric folds to the maximum extent of the Barrett's segment. Subsequently, at 2 centimeter intervals, a series of 10-second "level videos" was recorded, starting from the retrograde position. Ground truth, referring to the presence or absence of visible abnormalities requiring targeted biopsy, was determined by the endoscopist prior to starting the protocol. This was followed by post-hoc histopathological confirmation through targeted biopsies, endoscopic resection, or acquisition of random biopsies. The primary outcome measure focused on the standalone performance of the CADe system, which was evaluated in terms of per-patient sensitivity and specificity.

Results

A total of 15 neoplastic patients and 15 NDBE patients were included in the study (mean age of 67 years; mean Barrett length of C7M9; 25 males; no significant differences between groups). The CADe system exhibited a 100% sensitivity on a per patient basis, correctly identifying all neoplastic lesions. 14 out of 15 visible lesions were correctly diagnosed in the pullback videos. The initially missed lesion was detected in the subsequent level video. The system incorrectly predicted neoplasia in 8 NDBE patients. If applied in a clinical setting, this would result in a maximum of one additional targeted biopsy per patient, compared to the 16 random biopsies dictated by the mean Barrett length in this study. Histopathological examination confirmed neoplasia in 13 cases, with 11 cases being adenocarcinoma and 2 cases being high-grade dysplasia. Two cases did not contain any dysplasia. Additionally, in 4 cases from the non-dysplastic group, low-grade dysplasia was found through the random biopsy protocol.

Conclusion

This study represents one of the first evaluations of a CADe system for real-time detection of neoplasia in Barrett's esophagus in the endoscopy suite. The system accurately detected all neoplastic lesions, with an acceptable number of false positive detections.

Disclosure

This study received financial and logistical support from Olympus Tokyo.

Introduction

This report presents the results of a multicenter, randomized, double-blind, placebo-controlled phase 3 study (VP-VLY-686-3301) assessing the efficacy of tradipitant in relieving symptoms of gastroparesis. Tradipitant, a novel NK-1 receptor antagonist, previously demonstrated efficacy in diabetic and idiopathic gastroparesis in reducing gastroparesis symptoms in a phase 2 study (VP-VLY-686-2301).

Aims & Methods

N=201 idiopathic and diabetic gastroparesis patients with delayed gastric emptying, moderate to severe nausea, and at least 1 vomiting episode were included in the ITT (Intent-to-Treat) population. Subjects were randomized to receive 85 mg tradipitant twice a day (n=102) or placebo (n=99) for 12 weeks. Nausea was assessed daily using the 5-point Gastroparesis Core Symptom Daily Diary (GCSDD). Overall gastroparesis symptom improvement was evaluated using the Patient Global Impression of Change (PGI-C) and the Overall Patient Benefit (OPB) scales. Sensitivity analyses were also performed to control for confounders that may have masked the true treatment effect size.

Results

At the primary endpoint of change in nausea from baseline as measured by the GCSDD at week 12, both tradipitant and placebo showed significant and similar reductions from baseline at 1.55 and 1.49 respectively. Comparison between treatment arms did not reach statistical significance in this measure. However, in the PGI-C scale, more tradipitant treated patients demonstrated response as compared to placebo both at week 2 (74% v. 58%, p=0.019) and at week 12 (78% v. 66%, p=0.065). Similarly, in the OPB scale, more tradipitant treated patients demonstrated response as compared to placebo both at week 2 (81% v. 62%, p=0.0003) and week 12 (86% v. 71%, p=0.011).
Sensitivity analysis adjusting for rescue medication use further confirmed the ITT findings in both the PGI-C and OPB analysis at both week 2 and week 12.

Conclusion

Significant but similar improvements from baseline for tradipitant and placebo may have masked the true treatment effect size at the primary endpoint of the study of change in nausea severity as measured by daily diary at week 12 leading to no statistically significant difference between treatments. Despite this large placebo effect, tradipitant was shown to be significantly better than placebo in global measures of patient improvement including the Patient Global Impression of Change (PGI-C) and the Overall Patient Benefit (OPB) scales.

Disclosure

Study was sponsored by Vanda Pharmaceuticals, Inc. Authors are employees of Vanda Pharmaceuticals, Inc.

Introduction

The Janus kinase inhibitor (JAKi) upadacitinib (UPA) is approved for the treatment of moderate-to-severe ulcerative colitis (UC). However, safety warnings from the US Food and Drug Administration and the European Medicines Agency recommend caution when using JAKis in certain at-risk patients. These drug-class warnings are based on data from the ORAL Surveillance (OS) study, a safety outcomes trial comparing risk of cardiovascular (CV) events with the JAKi tofacitinib plus methotrexate vs a tumor necrosis factor inhibitor (TNFi) plus methotrexate in patients with rheumatoid arthritis (RA) who were ≥50 years of age with ≥1 CV risk factor.¹ An OS study post hoc analysis observed higher major adverse CV event risk with tofacitinib vs TNFi in patients with RA plus a history of atherosclerotic CV disease (ASCVD).² CV risk profiles may be different for patients with UC vs RA.

Aims & Methods

This analysis aimed to describe the CV risk profile for patients with UC relative to the CV risk profile for patients with RA included in the OS study. Patients with UC enrolled in the placebo-controlled UPA U-ACHIEVE and U-ACCOMPLISH induction trials³ were evaluated for the OS study inclusion criteria (age ≥50 years, ≥1 CV risk factor [CV event history, hypertension, diabetes mellitus, smoking/tobacco use, and high-density lipoprotein cholesterol <1.034 mmol/L]). In addition, history of coronary artery disease or ASCVD and 10-year CV disease risk for patients with UC with no prior CV disease were estimated using the American College of Cardiology ASCVD risk score calculation.

Results

Overall, 268/1097 (24.4%) patients with UC met the OS criteria and 624 (56.9%) had ≥1 CV risk factor, regardless of age. Few patients with UC (30; 2.7%) had a history of ASCVD, in contrast to the RA population (640/4362; 14.7%). The distribution of 10-year ASCVD risk scores for patients with UC was similar between the placebo and UPA treatment arms, but markedly different from patients with RA (Table). Most patients with UC were categorized as low risk (938; 85.5%), with another 52 (4.7%) categorized as borderline risk, 51 (4.6%) as intermediate risk, and 26 (2.4%) as high risk.

Conclusion

The majority of patients in the UC trials had low estimated CV risk, while the majority of patients with RA included in the OS study were at intermediate or high CV risk. Caution should be used when extrapolating adverse events data from patients with RA to risk discussions for patients with UC.

References

1. Ytterberg SR, Bhatt DL, Mikuls TR, et al. Cardiovascular and cancer risk with tofacitinib in rheumatoid arthritis. N Engl J Med 2022;386:316-326.
2. Charles-Schoeman C, Buch MH, Dougados M, et al. Risk of major adverse cardiovascular events with tofacitinib versus tumour necrosis factor inhibitors in patients with rheumatoid arthritis with or without a history of atherosclerotic cardiovascular disease: a post hoc analysis from ORAL Surveillance. Ann Rheum Dis 2023;82:119-129.
3. Danese S, Vermeire S, Zhou W, et al. Upadacitinib as induction and maintenance therapy for moderately to severely active ulcerative colitis: results from three phase 3, multicentre, double-blind, randomised trials. Lancet 2022;399:2113-2128.

Disclosure

L Peyrin-Biroulet has received personal fees from Galapagos, AbbVie, Janssen, Genentech, Alimentiv, Ferring Pharmaceuticals, Tillotts Pharma, Celltrion, Takeda, Pfizer, InDex Pharmaceuticals, Sandoz, Celgene, Biogen, Samsung Bioepis, Inotrem, Allergan, MSD, Roche, Arena Pharmaceuticals, Gilead, Amgen, Bristol Myers Squibb, CSL Vifor, Norgine, Mylan, Lilly, Fresenius Kabi, OSE Immunotherapeutics, Enthera, Theravance Biopharma, Pandion Therapeutics, Gossamer Bio, Viatris, and Thermo Fisher Scientific; grants from AbbVie, MSD, Takeda, and Fresenius Kabi; and stock options from CTMA.
RK Cross is an advisory board member and consultant for AbbVie, Adiso, BMS, Fresenius Kabi, Fzata, Janssen, Magellan Health, Pfizer, Samsung Bioepis, Sebela, and Takeda; and a member of the Executive Committee for the IBD Education Group and Scientific Co-Director of CorEvitas.
KM De Felice is an advisory board member and honorary speaker for AbbVie and BMS.
D Ilo, B Duncan, J Liu, and CTJ Holweg are employees of AbbVie Inc. and may hold AbbVie stock and/or stock options.
R Chowdhury has no conflicts of interest to declare.
This study was funded by AbbVie. Medical writing and editorial assistance were provided by Jeffrey Kohn, PhD, and Lauren Smith, BA (Hons), from Complete HealthVizion, an IPG Health Company, funded by AbbVie.

Introduction

Therapeutic monitoring of infliximab (INF) correlates with improved therapeutic outcomes in inflammatory bowel diseases (IBD).

Aims & Methods

The aim of this cross-sectional study was to assess the relationship between serum INF trough levels (TLs) and positive therapeutic outcomes in patients with ulcerative colitis (UC) and Crohn’s disease (CD) in a Greek population. The relationship between adverse events (AEs) and the existence of anti-infliximab antibodies (ADAs) was also investigated. Consecutive adult patients with IBD treated with intravenous INF at maintenance phase in a Greek tertiary centre were included. Investigated therapeutic outcomes included sustained clinical remission (CR), steroid-free clinical remission (SFCR), biomarker remission (CRP <5mg/L, BR) and combined (steroid-free clinical and biochemical) remission (SFCBR). Infliximab TLs and ADAs were assessed using ELISA-based assays.

Results

A total of 77 patients were enrolled (62.3% male, mean age 44 years), 48 (62.3%) with CD, 26 (33.8%) with UC and 3 (3.9%) with unclassified colitis. The mean disease duration was 11.7±8.4 years, while the mean duration of INF infusion was 5.1±4.6 years. Forty-seven (61%) patients had undergone treatment escalation, with a further dose escalation in 7 (9.1%) patients. Concomitant combined therapy was given in 13 (16.8%) patients, 4 (5.2%) of them were receiving azathioprine and 9 (11.7%) corticosteroids. Mean INF-TLs were 7.2±4.9 μg/mL and only correlated with treatment escalation (9.7 vs 3.6 μg/mL, p<0.001). Sixty-eight (88.3%) patients had CR, 62 (80.5%) had SFCR, 48 (62.3%) had BR and 43 (55.8%) had SFCBR. Compared to non-responders, mean INF-TLs were numerically higher, although not significantly different, in CR (7.5 vs 5.5 μg/mL, p=0.254), SFCR (7.3 vs 6.9 μg/mL, p=0.761), BR (7.4 vs 6.9 μg/mL, p=0.707) or SFCBR (7.7 vs 6.6 μg/mL, p=0.349).
In a subgroup analysis, among patients without treatment escalation, 27 (90%) had CR, 25 (83.3%) had SFCR, 22 (73.3%) had BR and 19 (63.3%) had SFCBR. Mean INF-TLs were significantly increased among patients with BR (4.2 vs 0.8 μg/mL, p=0.020) and SFCBR (4.3 vs 1.5 μg/mL, p=0.035). In ROC analysis, TLs could significantly predict (p=0.016) with a good accuracy (AUC 0.768, 95% Cl: 0.584=0.952) patients with SFCBR, with an optimal TL cut-off at 3.4 μg/mL. Among patients with treatment escalation, 41 (87.2%) had CR, 37 (78.7%) had SFCR, 26 (55.3%) had BR and 24 (51.1%) had SFCBR. In ROC analysis, TLs could significantly predict (p=0.018) with a fair accuracy (AUC 0.653, 95% CI: 0.527=0.755) patients with SFCBR, with an optimal TL cut-off at 11 μg/mL.
Twenty patients had TLs <1 μg/mL, of whom 10 had positive ADAs. Among them, 9 showed combined remission, 4 of whom had positive ADAs. Among 21 (27.3%) patients with reported AEs, 8 (10.4%) had an allergic reaction, which was attributed to INF. Allergic reactions had a tendency of correlation with ADA existence, as 4/5 patients with allergic reaction and TLs <1 μg/mL had positive ADAs, while 4/10 patients with positive ADAs reported allergic reaction, compared to 1/10 patients with negative antibodies (p=0.152).

Conclusion

Infliximab TLs correlate with treatment escalation. Higher TLs appear to result in favourable therapeutic outcomes. It is suggested that patients undergoing treatment escalation could benefit from higher INF-TLs in the achievement of favourable outcomes.