Introduction

Tofacitinib is an oral small molecule Janus kinase inhibitor for the treatment of ulcerative colitis (UC). Histological outcomes and the immune landscape were assessed in an UC patient cohort at baseline and after 8 weeks tofacitinib.

Aims & Methods

In this prospective study, 40 patients with moderate to severely active UC received tofacitinib 10 mg twice daily for 8 weeks. Clinical, endoscopic and histological outcomes were assessed at baseline and week 8 and colon biopsies were stained for JAK1-3, Tyrosine kinase 2 (TYK2) and Signal transducer and activator of transcription (STAT) 1-6 total protein expression using immunohistochemistry (IHC). Histological remission was defined as a Robarts Histopathology Index (RHI) ≤3 points without mucosal neutrophils and histological response as 50% decrease in RHI compared to baseline. Response was defined as histological remission and endoscopic Mayo score (EMS) ≤1 (histo-endoscopic mucosal improvement). Biopsies were scored by a blinded pathologist. Whole slide images with multiple biopsies per patients were stained using IHC, analyzed using automated cell detection (QuPath) and quantified in percentages by dividing the number of positive stained cells by the total number of cells on the slide. Outcomes were compared between responders and non-responders (NR).

Results

Out of 40 patients, 26 (65%) had severe endoscopic disease (EMS 3) with a median RHI of 14. Thirty-one patients (78%) had failed prior anti-Tumor Necrosis Factor (anti-TNF) agents and 18 (55%) failed vedolizumab. At week 8, 23 patients (58%) achieved histological remission, 34 (85%) histological response and 15 (38%) histo-endoscopic mucosal improvement. A significant decline in RHI compared to baseline was observed after 8 weeks both in responders (14 vs 0, p<0.001) and NR (15 vs 6, p=0.002). In the entire cohort, total STAT1 (p<0.001), STAT3 (p=0.002) and STAT5 (p=0.026) expression significantly decreased after 8 weeks treatment compared to baseline, which was not different between responders and NR (Table). However, responders to tofacitinib expressed significantly lower total STAT1 levels compared to NR at 8 weeks (p<0.001). Likewise, JAK2 (p=0.008) and total STAT2 expression (p=0.004) were significantly lower in responders compared to NR at 8 weeks, although no significant decrease compared to baseline was observed. A trend towards lower baseline JAK1, JAK2 and total STAT4 expression and higher total STAT6 expression was seen in responders compared to NR.

	Responders Baseline	NR Baseline	p-value Baseline Responders vs NR	Responders Week 8	NR Week 8	p-value Week 8 Responders vs NR
JAK1, median % [IQR]	6.2 [1.1-12.1]	13.7 [6.9-17.6]	0.113	8.4 [5.2-13.1]	9.7 [5.6-13.6]	0.811
JAK2, median % [IQR]	0.4 [0.1-2.1]	2.7 [0.1-7.7]	0.125	0.2 [0.1-1.7]	1.8 [0.6-8.6]	0.008
TYK2, median % [IQR]	13.3 [1.3-27.6]	10.0 [5.5-25.4]	0.761	11.9 [7.4-15.4]	17.5 [10.3-26.6]	0.156
STAT1, median % [IQR]	13.0 [5.2-24.9]	13.5 [7.5-22.0]	0.709	0.2 [0.1-1.8]*	4.3 [1.2-11.9**	<0.001
STAT2, median % [IQR]	0.3 [0.1-1.7]	0.7 [0.2-1.8]	0.393	0.3 [0.1-0.6]	0.8 [0.4-2.6]	0.004
STAT3, median % [IQR]	0.3 [0.0-6.7]	0.8 [0.1-6.6]	0.361	0.0 [0.0-0.2]	0.1 [0.0-1.0]**	0.106
STAT4, median % [IQR]	0.9 [0.3-2.4]	1.9 [0.4-3.3]	0.315	1.2 [0.5-3.2]	2.1 [1.4-3.9]	0.175
STAT5, median % [IQR]	0.1 [0.0-1.1]	0.2 [0.2-1.3]	0.235	0.0 [0.0-0.3]	0.2 [0.0-0.9	0.159
STAT6, median % [IQR]	23.6 [4.2-28.8]	17.4 [7.6-32.1]	0.455	14.1 [5.3-45.7]	22.1 [10.9-32.5]	0.472

Table legend: IQR, interquartile range; * statistically significant decrease after 8 weeks compared to baseline in responders; ** statistically significant decrease after 8 weeks compared to baseline in NR.

Conclusion

Tofacitinib resulted in histological response in the majority of UC patients and led to a substantial decline of total STAT1, STAT3 and STAT5 expression both in responders and NR. Yet, after 8 weeks significantly lower total STAT1 levels were observed in responders compared to NR. Responders tended to express lower JAK1, JAK2 and total STAT4 and higher total STAT6 levels at baseline.

Disclosure

This study was funded and supported by Pfizer.

Introduction

Although CADe/x systems increase colorectal polyp detection, their ability to predict histological diagnosis is limited.

Aims & Methods

The aim of this study is to determine which factors related to morphology, histology and the image characteristics are associated with histology prediction by PolyDeep, a CADe/x system.
For this analysis, 5546 still images of colorectal polyps collected in PIBAdb database (https://www.iisgaliciasur.es/home/biobanco/colorectal-polyp-image-cohort-pibadb/?lang=en) were used. The images were obtained from videos recorded by endoscopist. Each image was associated with the endoscopy description (morphology, size), the final histology diagnosis and the characteristics (size, area) related with the Red, Green and Blue model (RGB model) and the Hue, Saturation and Value model (HSV model). Finally, the Convolutional Neural Network (CNN) determined the probability a polyp was classified as neoplastic lession (adenomatous, Traditional Serrated Adenoma,TSA, Sessile Serrated Lession, SSL). We determined Polydeep classified lesions as neoplastic if the probability was ≥0.5. We performed three analysis: (1) linear multiple regression in order to determine which variables are associated with the probability of classification as neoplastic lesions. (2). A multivariate logistic regresion analysis to define the variables associated with the classification as a neoplastic lesions. (3) A multivariate logistic regression to establish which variables are associated with missclassification of polyps. We use the R software for statistical computing.

Results

PolyDeep classified 84,51% of the lesions as neoplastic lession and managed to correctly classify 85.02% of the evaluated polyps (92.8% adenomas, 82.5% TSA, 84.2% SSL and 51.6% hyperplastic). The R² in the multiple linear regression is the 0.3655. The model associated to the neoplastic diagnosis has an Akaike Information Criterion (AIC) of 3363.3 and the model associated with missclassification, has an AIC of 3547.5. The variables associated to each regression model showed in the attached table.

Predictor variables		Neoplastic lesion classification - Predictive model		Missclassification - Predictive model
		Multiple linear regression OR / (CI 95%) R2 = 0.3655	Multivariate logistic regression OR / (CI 95%) AIC = 3363.3	Multivariate logistic regression OR / (CI 95%) AIC = 3547.5
Histology	TSA SSL Hyperplastic	-0.09 / (-0.11 - 0.07) -0.05 / (-0.07 - 0.02) -0.37 / (-0.39 - 0.35)	0.44 / (0.32 - 0.61) 0.65 / (0.47 - 0.92) 0.17 / (0.13 - 0.22)	1.88 / (1.34 - 2.60) 2.52 / (1.77 - 3.53) 15.67 / (11.98 - 20.63)
NICE classification	Type 2 Type 3	0.10 / (0.09 - 0.12) 0.03 / (-0.11 - 0.17)	1.55 / (1.23 - 1.96) 8.29 / (0.00 - NA)	1.23 / (0.97 - 1.57) 0.00 / (NA- 5,36x10+87)
Paris classification	Slightly elevated Sessile Pedunculated	0.06 / (0.02 - 0.10) 0.04 / (-0.01 - 0.07) 0.04 / (-0.01 - 0.07)		1.31 / (0.80 - 2.19) 0.88 / (0.54 - 1.46) 0.42 / (0.23 - 0.79)
Polyp size (mm)		0.23 / (0.23 - 0.24)	1.25 / (1.20 - 1.30)	1.06 / (1.02 - 1.09)
Width (píxels)		-0.31 / (-0.31 - 0.31)	0.99 / (0.98 - 0.99)
Area (píxels2)		0.68 / (0.68 - 0.68)
Mean model RGB	Mean box red (0º-255º) Mean box blue (0º-255º) Mean box color (0º-255º) Mean image green (0º-255º) Mean image blue (0º-255º) Mean outside box red (0-255º) Mean outside box green (0º-255º) Mean outside box blue (0º-255º)	0.02 / (0.02 - 0.02) -0.52 / (-0.53 - 0.52) 0.36 / (0.35 - 0.36) 0.60 / (0.59 - 0.60) 0.82 / (0.81 - 0.82)	0.97 / (0.96 - 0.97) 0.73 / (0.67 - 0.79) 1.50 / (1.35 - 1.68)	3.38 / (2.30 -5.02) 1.82 / (1.54 - 2.16) 0.20 / (0.12 - 0.32) 2.91 / (2.12 - 4.01) 0.53 / (0.36 - 0.78) 0.76 / (0.66 - 0.89) 0.70 / (0.65 - 0.76)
Mean Model HSV	Mean box Hue (0º-179º) Mean box saturation (0º-179º) Mean image hue (0º-179º) Mean image saturation (0º-179º) Mean image value (0º-179º) Mean outside box hue (0º-179º) Mean outside box saturation (0º-179º) Mean outside box value (0º-179º)	0.02 / (0.02 - 0.02) 0.19 / (0.19 - 0.20) -0.21 / (-0.22 - 0.21) 0.17 / (0.17 - 0.17) -0.84 / (-0.85 - 0.83)	1.09 / (1.07 - 1.12) 1.07 / (1.05 - 1.08) 1.20 / (1.14 - 1.26) 0.74 / (0.68 - 0.79) 1.25 / (1.16 - 1.34)	0.97 / (0.94 - 0.99) 1.26 / (1.10 - 1.43) 0.55 / (0.47 - 0.64) 0.88 / (0.85 - 0.92) 0.84 / (0.75 - 0.94) 1.71 / (1.49 - 1.95)

Table 1: The table show the results of the multiple linear regression and both multivariate logistic regressions with the variables associated to correct classifications or missclassification of colorectal lesions. Traditional Serrated Adenoma (TSA); Sessile Serrated Lesion (SSL); Odd Ratio (OR); Confindence Interval (CI)

Conclusion

The results obtained in this study allow us to establish strategies to improve Polydeep predicted histological diagnosis.

Introduction

Inflammatory bowel disease (IBD) is characterized by chronic inflammation due to an uncontrolled immune response orchestrated through, i.e. neutrophils. The basement membrane acts as a front-line defense promoting tissue integrity, where the α-3 chain of type IV collagen is solely expressed on the outermost layer of the mucosa. Detection of subclinical ulceration is crucial since early-stage IBD presents a unique window of opportunity for intervention, as the disease might become a self-sustaining process once deep inflammation becomes established. We therefore, wanted to develop a biomarker reflecting early tissue injury, aiding in the implementation of prompt treatment.

Aims & Methods

A competitive enzyme-linked immunosorbent assay quantifying a human neutrophil elastase (HNE) degraded fragment of type IV collagen α-3 chain was developed (C4A3-HNE). Dilution- and spiking recovery, inter- and intra-variability, and interference were evaluated. Acceptance criteria for technical validation were a recovery of 100% ±25 and an inter-intra variability of <15%. Commercial serum samples were purchased from ProteoGenex. Biological relevance was validated in an observational study from Odense University Hospital (OUH) and healthy controls. Groups were compared using Kruskal Wallis (Dunn‘s corrected).

Results

The technical validation of the C4A3-HNE assay demonstrated acceptable results for recovery, variation, and interference. Biological relevance was assessed in commercial IBD samples and healthy donors. C4A3-HNE was elevated in both CD and UC compared to healthy (p=0.001 and <0.0001)(AUC [95% CI]: 0.820 [0.669-0.971]; 0.875 [0.775-1.000]). The results were validated in an independent cohort of IBD patients (both p<0.0001). For UC, analysis showed that compared to patients in endoscopic remission, C4A3-HNE is elevated in patients with mild endoscopic activity, decreasing dose-dependently for moderate and severe disease (T1). Fecal calprotectin and CRP levels were also lower in IBD patients with high levels of C4A3-HNE (T1). With a cut-off value of 54.1 ng/ml, C4A3-HNE could significantly discriminate between mild and severe endoscopic disease (T1).

	Median conc. (ng/ml)	[IQR]	Patients (n)
Remission Mild Moderate Severe	61.16 71.77 56.07 40.52	49.74, 65.52 53.96, 79.17 39.21, 68.71 25.65, 69.01	10 18 19 10
	Low C4A3-HNE	High C4A3-HNE	Δ %
CD: Fecal Calprotectin (µg/g) UC: Fecal Calprotectin (µg/g) CD: CRP (mg/L) UC: CRP (mg/L)	721.8 840.0 8.4 7.4	365.7 491.6 6.5 5.0	67.7 52.3 26.5 38.5
	AUC [95% CI]	Cut-off value	Sens./Spec. (%)	OR
Remission vs. mild Mild vs. moderate Mild vs. severe	0.68 [0.48–0.85] 0.68 [0.51–0.83] 0.74 (*) [0.54–0.89]	68.8 59.9 54.1	61.1/90.0 72.2/73.3 77.8/70.0	14.1 (*) [1.7–166.6] 6.9 (*) [1.8–31.1] 6.4 (*) [1.2–33.5]

Conclusion

Our study demonstrates the biological and possible clinical relevance of the novel C4A3-HNE biomarker. The biomarker is elevated in patients with IBD compared with healthy, and high levels may reflect early mucosal damage in patients with UC. Furthermore, inflammatory markers were lower in IBD patients with high levels of C4A3-HNE, pointing towards high levels being indicative of early disease. C4A3-HNE may provide early detection of mucosal damage or prediction of relapse; however, further studies are required to assess the clinical validity of this biomarker.

Disclosure

MA, MP, MK, and JM are employees of Nordic Bioscience A/S. MK owns stocks in Nordic Bioscience A/S.

Introduction

Colorectal cancer (CRC) is the 3rd most common cancer worldwide. Approximately 85% of CRC develop on a background of Wnt/β-catenin signalling impairment. Familial adenomatous polyposis (FAP), caused by inherited APC gene mutations (part of the β-catenin destruction complex), is characterized by the development of hundreds to thousands adenomatous polyps. FAP patients have up to 100% risk of developing CRC during their lifetime. The Wnt/β-catenin together with the PI3K/mTOR signalling control the self-renewal and homeostasis of intestinal stem cells and are frequently deregulated in CRC therefore representing attractive targets for chemoprevention and treatment [1], [2]. It has been shown that each pathway may be deregulated by an altered composition of the gut microbiota [3], [4], which plays an important role in the onset, development and progression of CRC thus representing potential target for prevention and potentially biomarkers for the early detection of deranged Wnt/β-catenin-driven lesions [5].

Aims & Methods

The aim of the study was to identify potential targets for the prevention and early detection of CRC in the setting of Wnt/β-catenin disruption.
Transcriptional and translational levels of Wnt/β-catenin and PI3K/mTOR pathways were evaluated by RT-qPCR and immunohistochemical analysis, respectively, in FAP adenomas (n=17 patients) and in sporadic CRC tissues (tumor stages I-III; n=15 patients). In FAP adenomas germline and somatic APC mutations were confirmed by Sanger and Targeted Next Generation Sequencing. The composition of oral, fecal and mucosal-associated microbiota was sequenced by the bacterial 16S rRNA gene. Non-diseased Faecal Immunochemical Testing positive subjects (FIT+; n=17) were included as control group.

Results

We found increased expressions of AXIN2 and cMYC, downstream targets of Wnt/β-catenin, in FAP adenomas. The activation status of PI3K/mTOR was confirmed by an increase of RPS6 gene and pS6R protein expression, indicating that germline and somatic APC mutations of FAP adenomas induce an activation of both pathways. On the other hand, compared to FAP adenomas, a higher expression of AXIN2, cMYC and cytosolic and nuclear β-catenin was found in CRC, while no changes were detected in pS6R, suggesting that the hyperactivation of Wnt/β-catenin in sporadic CRC contributes to the impairment of mTOR signalling. Importantly, we analyzed the microbiome composition in FAP and CRC settings. The oral microbiota of FAP was significantly enriched of Proteobacteria (p<0.01), which decreased in FAP adenomatous mucosal samples compared to FIT+ and CRC. In contrast, Firmicutes showed an opposite trend in FAP patients, with a decrease in the oral composition (p<0.01) and an increase in the adenomatous mucosa (p<0.05). Interestingly, a significant decrease of the CRC-protective phylum Actinobacteria was found in oral and adenomas microbiota of FAP patients compared to FIT+. In contrast, CRC fecal microbiota was enriched of Proteobacteria and a higher abundance of the well-known cancer associated families Fusobacteriaceae, Clostridiaceae and Gemellaceae was found in CRC mucosal samples compared to FIT+.

Conclusion

These findings reveal new insights into the role of Wnt/β-catenin and PI3K/mTOR pathways in influencing the homeostasis of precancerous CRC lesions. Moreover, by defining different microbial signatures in FAP patients, these results will contribute to the enhancement of non-invasive screening tools for the early detection of WNT signalling-driven colorectal cancers.

References

[1] X. Cheng, X. Xu, D. Chen, F. Zhao, and W. Wang, “Therapeutic potential of targeting the Wnt/β-catenin signaling pathway in colorectal cancer,” Biomed. Pharmacother., vol. 110, pp. 473–481, 2019, doi: https://doi.org/10.1016/j.biopha.2018.11.082.
[2] J. Zhang, T. M. Roberts, and R. A. Shivdasani, “Targeting PI3K Signaling as a Therapeutic Approach for Colorectal Cancer,” Gastroenterology, vol. 141, no. 1, pp. 50–61, 2011, doi: https://doi.org/10.1053/j.gastro.2011.05.010.
[3] M. H. Noureldein and A. A. Eid, “Gut microbiota and mTOR signaling: Insight on a new pathophysiological interaction,” Microb. Pathog., vol. 118, no. March 2018, pp. 98–104, 2018, doi: 10.1016/j.micpath.2018.03.021.
[4] O. Silva-García, J. J. Valdez-Alarcón, and V. M. Baizabal-Aguirre, “Wnt/β-catenin signaling as a molecular target by pathogenic bacteria,” Front. Immunol., vol. 10, no. SEP, pp. 1–14, 2019, doi: 10.3389/fimmu.2019.02135.
[5] J. H. Kim, Y. J. Kim, G. M. Oh, W. Jung, and S. J. Park, “How is gut microbiome of patients with familial adenomatous polyposis different from healthy people?,” Medicine (Baltimore)., vol. 101, no. 49, p. e32194, Dec. 2022, doi: 10.1097/MD.0000000000032194.

Introduction

Acute severe colitis (ASC) is an important cause of morbidity and mortality in Ulcerative Colitis (UC), requiring hospitalisation and, frequently, colectomy. Intravenous corticosteroids (CS) remain the first-line medical treatment. Approximately 40% of episodes of ASC do not respond to this treatment and require rescue therapy with ciclosporin (CIC) or infliximab (IFX)¹. In the past decades, these treatments and protocol-driven management algorithms led to reduced mortality and colectomy rates. A simple clinical scoring system (the Admission Model for the Intensification of Therapy in Acute Severe Colitis (ADMIT-ASC))³, based on CRP, albumin, and the Ulcerative Colitis Endoscopic Index of Severity (UCEIS) score, was recently developed to predict the steroid response on admission.

Aims & Methods

We aimed to retrospectively assess the steroid non-response during admission with ASC and the usefulness of the new prognostic clinical index ADMIT-ASC at admission to predict steroid response in a Belgian tertiary center. All patients admitted with ASC between 1^st January 2015 and 1^st April 2023 were identified from our inflammatory bowel disease database at Erasme University Hospital. Inclusion criteria were ASC patients admitted at the hospital, standard treatment with at least 3 days of intravenous CS and endoscopy evaluation upon admission. ASC was defined according to modified Truelove and Witts’ criteria as ≥6 bloody stools/day with one or more markers of systemic disturbance (tachycardia >90 beats/min, temperature >37.8°C, haemoglobin <105 g/L or CRP >30 mg/L). Steroid non-response was defined as the need for medical rescue therapy (CIC of IFX) or colectomy during the admission. ADMIT-ASC index consists of four points: CRP ≥100 mg/L (1 point), albumin ≤25 g/L (1 point) and UCEIS score ≥4 (1 point) or ≥7 (2 points).

Results

Table I Admission Demographics data (n= 84)	Steroid response (n=49)	Steroid non-response (n=35)	P value
Median disease duration (years) (IQR)	3.6 (0.3-0.9)	2.1 (0.8-5.5)	0.5
Montreal Classification	E1 : 5/49 (10%) E2 : 20/49 (41%) E3 : 24/49 (49%)	E1 : 0/35 E2 : 12/35 (34%) E3 : 23/35 (66%)	0.084
5-ASA, n (%)	22 (45%)	10 (29%)	0.13
Methotrexate or thiopurine, n (%)	6 (12%)	9 (26%)	0.11
Infliximab, n (%)	4 (8.2%)	8 (23%)	0.058
Vedolizumab, n (%)	7 (14%)	5 (14%)	>0.9
Tofacitinib, n (%)	2 (4.1%)	2 (5.7%)	>0.9
Readmission, n (%)	12 (24%)	10 (29%)	0.7
5-ASA, 5-aminosalicylate P values from Wilcoxon rank-sum test or Fisher’s exact test.

Out of 127 patients admitted with UC and treated with intravenous CS during three days, 84 (45% male, median age of 31 years old) were included in the study. Admission demographics and clinical characteristics are described in Table 1.
For 17 patients (20%), the admission was their index presentation of UC. Response to CS was seen in 49 patients (58%). Thirty-three patients (39%) received medical rescue therapy (17/33 IFX and 16/33 CIC). Colectomy rate during admission was 13% after a median duration of 18 days (IQR: 16-54). Admission C reactive protein (CRP) (p=0.036), albumin (p=0.026) and UCEIS scores (p=0.022) were significantly associated with steroid non-response. Patients scoring 0, 1, 2, 3 and 4 had steroid response rates of 90%,75%, 55%, 19% and 0%, respectively. Scoring of ≥3 was 75% (95% CI, 0.43 to 0.95) predictive of steroid failure (OR 5.08, 95%, CI 1.38 to 24.5).

Conclusion

In this tertiary referral cohort study, we confirmed that the ADMIT-ASC index is a simple clinical tool that can be used on the day of admission to predict steroid non-response in ASC with a threshold of ≥3 to identify patients who are unlikely to respond to intravenous CS and enabling timely treatment escalation or surgery.

References

1. Travis S, Satsangi J, Lémann M. Predicting the need for colectomy in severe ulcerativecolitis: a critical appraisal of clinical parameters and currently available biomarkers. Gut 2011;60:3–9.
2. Travis SP, Farrant JM, Ricketts C, et al. Predicting outcome in severe ulcerative colitis.Gut 1996;38:905–10.
3. Adams A, Gupta V, Mohsen W, et al. Gut 2023;72:433–442.

Disclosure

VB has not disclosure.
CDT has not disclosure.
CM has not disclosure.
AC has received lecture fees from Galapagos, Abbvie, Janssen, celltrion, Pfizer, Takeda.
DF has received lecture fees from Abbvie, Astrazeneca, Dr.Falk Pharma, Ferring, Janssen, Mundipharma, MSD, Pfizer, Takeda, Amgen, and advisory board fees from Abbvie, Janssen,Takeda, Celltrion, Amgen.
CL has received consultancy fees from Janssen, Abbvie, Takeda and Galapago and speaker fees from Sandoz, Janssen, Abbvie