Introduction

Crohn’s disease (CD) is a chronic, auto-immune, inflammatory condition affecting the digestive tract. Accurate evaluation of CD activity is crucial to ongoing treatment monitoring. Magnetic resonance enterography (MRE) is used to characterise disease behaviour, but its interpretation remains subjective. There are validated MRE activity scores available, such as the simplified magnetic resonance enterography index (sMARIA), but they often fail to capture early treatment response. Much current evaluation is based on an isolated 2D evaluation of a subjectively-chosen area of diseased bowel rather than assessing the full disease volume; even in this limited context, reductions in bowel wall thickness are associated with treatment response. It is believed that segmentation of abnormal bowel volume may allow more objective, quantitative assessment for tracking disease activity and treatment response. This is not feasible without automation due to constraints on clinician's time. To address this issue, we developed a segmentation protocol and semi-automated pipeline for use on coronal T2-weighted MRI, routinely acquired during Crohn’s monitoring.​

Aims & Methods

Imaging data from patients with confirmed CD from the METRIC / METRIC EF trials [1,2] were collected from 6 different hospitals. Abnormal bowel segments, identified based on pre-existing clinical information, had centrelines drawn by expert radiologists on coronal T2-weighted noncontrast MRI. Volumetric segmentation was performed based on these by a technician or radiologist. All annotations were performed on Entrolytics® (Motilent Ltd, UK).
Patients were randomly split to train, validation and test at project start. The train set (60 segments, 49 patients) was used for explicit optimisation (parameter fitting). The validation set (20 segments, 18 patients) was used for interim performance quantification. The test set (37 segments, 34 patients) was held out during development. To examine inter-reader agreement, N=18 cases from the VIGOR++ [3] study were annotated with centrelines drawn through the Terminal Ileum, and segmentation performed independently by two readers in isolation, from the same centreline.
The algorithm uses k-means clustering to divide images into partitions, which are characterised using a custom feature set. The fraction of their voxels contributing to the segmentation is predicted by a trained Random Forest Regressor.
We quantify agreement for bowel wall segmentations based on the same centreline, using three standard metrics for segmentation quality: Dice score, symmetric Hausdorff distance and mean contour distance. We compare agreement between readers with agreement between algorithm and human readers using the Kolmogorov-Smirnov test.

Results

Between human readers (N=18), levels of agreement are moderate (mean Dice score = 0.603, Symmetric Hausdorff Distance = 13.1mm, Mean Contour Distance = 1.30mm). Between algorithm and human (N=37), levels of agreement are comparable (mean Dice score = 0.521, Symmetric Hausdorff Distance = 15.1mm, Mean Contour Distance = 2.28mm). Difference in Dice scores between these two distributions cannot be detected (K-S statistic = 0.285, p = 0.230). Difference in SHD cannot be detected (K-S statistic = 0.339, p = 0.099). Differences in MCD can be detected (K-S statistic = 0.697, p = 4.51×10^-6).

Conclusion

Our semi-automated algorithm agrees with humans at a similar level to inter-reader agreement. This algorithm will b​e embedded within a web-based platform in order to further aid researchers and clinicians in deriving quantitative imaging biomarkers.

References

1. Taylor SA, Mallett S, Bhatnagar G, Baldwin-Cleland R, Bloom S, Gupta A, Hamlin PJ, Hart AL, Higginson A, Jacobs I, McCartney S, Miles A, Murray CD, Plumb AA, Pollok RC, Punwani S, Quinn L, Rodriguez-Justo M, Shabir Z, Slater A, Tolan D, Travis S, Windsor A, Wylie P, Zealley I, Halligan S; METRIC study investigators. Diagnostic accuracy of magnetic resonance enterography and small bowel ultrasound for the extent and activity of newly diagnosed and relapsed Crohn's disease (METRIC): a multicentre trial. Lancet Gastroenterol Hepatol. 2018 Aug;3(8):548-558. doi: 10.1016/S2468-1253(18)30161-4. Epub 2018 Jun 18.

2. Kumar S, Plumb A, Mallett S, Bhatnagar G, Bloom S, Clarke CS, Hamlin J, Hart AL, Jacobs I, Travis S, Vega R, Halligan S, Taylor SA. METRIC-EF: magnetic resonance enterography to predict disabling disease in newly diagnosed Crohn's disease-protocol for a multicentre, non-randomised, single-arm, prospective study. BMJ Open. 2022 Oct 3;12(10):e067265. doi: 10.1136/bmjopen-2022-067265.

3. Puylaert CAJ et al., Semiautomatic Assessment of the Terminal Ileum and Colon in Patients with Crohn Disease Using MRI (the VIGOR++ Project). Acad Radiol. 2018 Aug;25(8):1038-1045. doi: 10.1016/j.acra.2017.12.024. Epub 2018 Feb 7.

Disclosure

Benjamin Barrow, Gauraang Bhatnagar, Alex Menys and Andrew Bard are ​employees of Motilent Ltd (UK).​

Introduction

FOLFIRINOX and Gemcitabine/Nab-paclitaxel are widely used 1st-line regimens in metastatic pancreatic cancer. We would like to present real-world data of 12 years of experience with both regimens including genomics insights

Aims & Methods

Patients with metastatic pancreatic cancer who received FOLFIRINOX or Gemcitabine/Nab-paclitaxel were retrospectively included from January 2010 to December 2022. Outcomes were overall survival (OS) and progression free survival (PFS). Median survival time was calculated using Kaplan-Meier estimates adjusted with 1:1 Propensity Score Matching (PSM). Additional tissue Next Generation Sequencing (NGS) panel data was analyzed.

Results

A total of 521 patients, 251 patients received FOLFIRINOX and 270 patients received Gemcitabine/Nab-paclitaxel as a first-line regimen. FOLFIRINOX group were younger (61.8 vs. 65.9, p <0.001). Before PSM, FOLFIRINOX group showed prolonged median OS (11.1 vs 7.8 months) and PFS (8.8 vs. 5.8 months) ( p = 0.002, both) compared to Gemcitabine/Nab-paclitaxel group. After PSM, FOLFIRINOX group still showed prolonged median OS (11.1 vs 9.0 months, p = 0.010) and PFS (9.2 vs 6.4 months, p = 0.001) than Gemcitabine/Nab-paclitaxel group. NGS was available from 31 and 22 patients in the FOFLIRINOX and Gemcitabine/Abraxane group, respectively. The most frequently mutated genes were KRAS (94.3%), TP53 (84.9%), CDKN2A (24.5%), SMAD4 (24.5%), ARID1A(13.2%) and RNF43(9.4%). On Cox Regression Analysis, RNF43 mutation was correlated with poor PFS with Gemcitabine/Nab-paclitaxel (HR, 4.849; 95% CI, 1.150-20.450, P= 0.032).

Conclusion

FOLFIRINOX showed prolonged OS and PFS compared to Gemcitabine/Nab-paclitaxel. FOLFIRINOX can be considered 1st-line regimen in metastatic disease. The NGS data suggest RNF43 could be a potential predictive marker for the response to Gemcitabine/Nab-Paclitaxel, however, larger size confirmation study is needed to validate this finding.

Introduction

Eosinophilic esophagitis (EoE) is a complex disease that often coexists with different conditions. Given the complexity of this pathology, a multidisciplinary approach, with gastroenterologist, allergist, dietitian, psychologist, and endoscopist may benefit to diagnosis and management of patients, reducing time to diagnosis as well.

Aims & Methods

To evaluate impact of multidisciplinary team in diagnosis and management of eosinophilic esophagitis disease. A retrospective study was conducted to include all patients referred to Fondazione Policlinico Universitario A. Gemelli between September 2016 and April 2023. Prevalence of allergic, nutritional, and psychological disturbs was evaluated. Referral for each patient was reviewed. Diagnostic delay (time interval between first occurrence of EoE symptoms and confirmed diagnosis with endoscopy) was analyzed in patients with suspected eosinophilic esophagitis and diagnosis made in our center, compared to data reported in literature. Publications from 2014 to 2023 were reviewed.

Results

A total of 98 patients, 76 males and 22 females, with a mean age at diagnosis of 25 years, were included in the study. For all patients, each visit was conducted in team with gastroenterologist, allergist, dietitian, and psychologist. 81 patients presented allergic diseases (82 %), 51 patients (52 %) needed nutritional support by a dietitian and 63 patients (64 %) presented psychological implication by the disease to necessitate clinical support. 42 patients (43%) were referred by team members. 13 patients (13%) presented to our center with suspect eosinophilic esophagitis. For these patients, ab initio evaluated with multidisciplinary approach, mean time to diagnosis was 2.25 months. Diagnostic delay reported in literature was 3 years.

Conclusion

Multidisciplinary team is needed to manage patients with eosinophilic esophagitis because its several associated diseases. In fact, most of our patients presented allergic, nutritional, and psychological disturbs. This approach assures more appropriate medical management through shared decision making. Almost half of patients was referred by team members. Thank to awareness of each physician of the team and his early referral, we assisted in a reduction of diagnostic delay in our center compared to data reported in literature.

References

Lenti MV, Savarino E, Mauro A, Penagini R, Racca F, Ghisa M, Laserra G, Merli S, Arsiè E, Longoni V, de Bortoli N, Sostilio A, Marabotto E, Ziola S, Vanoli A, Zingone F, Barberio B, Tolone S, Docimo L, Pellegatta G, Paoletti G, Ribolsi M, Repici A, Klersy C, Di Sabatino A. Diagnostic delay and misdiagnosis in eosinophilic oesophagitis. Dig Liver Dis. 2021 Dec;53(12):1632-1639. doi: 10.1016/j.dld.2021.05.017. Epub 2021 Jun 8. PMID: 34116974
Sauer BG, West A, McGowan EC. Multidisciplinary Eosinophilic Esophagitis Care: A Model for Comprehensive Patient-Centered Care Through Shared Decision Making Between Gastroenterology, Allergy, and Nutrition. Clin Gastroenterol Hepatol. 2021 Nov;19(11):2226-2229. doi: 10.1016/j.cgh.2021.07.025. Epub 2021 Jul 16. PMID: 34280551.
Murray FR, Kreienbuehl AS, Greuter T, Nennstiel S, Safroneeva E, Saner C, Schindler V, Schlag C, Schoepfer AM, Schreiner P, Straumann A, Biedermann L. Diagnostic Delay in Patients With Eosinophilic Esophagitis Has Not Changed Since the First Description 30 Years Ago: Diagnostic Delay in Eosinophilic Esophagitis. Am J Gastroenterol. 2022 Nov 1;117(11):1772-1779. doi: 10.14309/ajg.0000000000001950. Epub 2022 Aug 12. PMID: 35971224.

Introduction

Several factors affect the difficulty of endoscopic retrograde cholangiopancreatography (ERCP) (Facciorusso et al.). This study aimed to evaluate whether the intra- and peri-diverticular location of the papilla has an impact on the results of ERCP for common bile duct stones.

Aims & Methods

Retrospective evaluation of patients without previous sphincterotomy undergoing ERCP for common bile duct stones between August 2020 and February 2023. Data on the patient (age, sex, comorbidities), procedure (elective/urgent, location and type of papilla, techniques used, successful cannulation, diagnosis, immediate complications), and outcomes (complications at 3, 30, and 90 days, need for repeating ERCP) were collected.

Results

Of the 769 patients undergoing ERCP during this period, 224 met the inclusion criteria (median age 76 years, 58% female). The papilla had an intra-diverticular location in 20 and a peri-diverticular location in 45, while it was not dependent on the diverticula in the remaining 159. Regarding the endoscopic classification (SADE; Haraldsson et al.), most were type 1 (55%) or type 2 (20%). The cannulation rate was 93%, with precut being used in 18% (fistulotomy/papillotomy n=22; transpancreatic precut n=14). The cannulation rate was significantly lower in patients with intra-diverticular and peri-diverticular papilla (75% and 87%, respectively, versus 97%; p<0.01), being the precut use also higher (20% and 31%, p=0.038). The ability to completely remove stones in the index ERCP (efficacy) was lower in patients with intra- or peri-diverticular papilla (71% and 70% versus 85%, p=0.04). There were no differences in intraprocedural complications (n=8, p=0.42), at 3 (n=20, p>0.05), 30, and 90 days (n=1). However, the need for repeating ERCP was different between groups (38% peri-diverticular, 25% intra-diverticular, versus 15%; p<0.01).

Conclusion

The intra- or peri-diverticular location of the papilla limits the cannulation rate and efficacy of ERCP, requiring more frequent use of advanced cannulation techniques and repeat procedures. However, it should be noted that no more complications occurred in these locations.

References

Haraldsson, E., et al. Endoscopic classification of the papilla of Vater. Results of an inter-and intraobserver agreement study. United European Gastroenterology Journal 5.4 (2017): 504-510.
Facciorusso A, et al. Comparative efficacy of different methods for difficult biliary cannulation in ERCP: systematic review and network meta-analysis. Gastrointest Endosc. 2022;95(1):60-71.

Introduction

The endoscopic retrograde cholnagio-pancreatography (ERCP) with endoscopic sphincterotomy results in a complete bile duct stones (CBDSs) clearance in about 80-90%. However, in case of incomplete CBDSs clearance, an endoprosthesis could be placed. A plastic stent is commonly used, with a success rate after a previous ERCP, ranging from 44% to 96%. However, there are no comparative studies between plastic and metal stent. Moreover, data concerning the safety and efficacy of antimigration Fully covered Metal stents – flare type (aFCMS) are lacking.

Aims & Methods

In this single-center retrospective study we collected data concerning all patients who underwent ERCP with antimigration metal stent placement for incomplete removal of CBDSs between January 2021 to January 2023. All patients were contacted in order to program stent remotion within 3-6 months (considering their general clinical condition) or to evaluate their clinical status. We aim to estimate the incidence of migration rate, and adverse events, and complete biliary clearance after 3-6 months from the index ERCP.

Results

Out of 34 patients included patients, 20 underwent a second ERCP after a median of 164 (±78) days. The reasons for the lack of a second ERCP evaluation are reported in the table. In 11/20 (55%) a distal migration occurred. Only one proximal migration was reported. Among the 34 included patients only one presented a clinical recurrence linked to the proximal migration of aFCMS. 19/20 patients achieved a complete clearance after the second ERCP. No serious adverse events were reported.

Conclusion

The aFCMS represents an effective and safe therapeutic chance in case of incomplete CBDSc clearance in particular in case of difficult CBDSs. Moreover, the very low incidence of stent proximal migration makes this device particularly useful in the case of this clinical condition.

Disclosure

None

Introduction

The role of prophylactic biliary stenting after clearance of common bile duct stones (CBDS) in patients awaiting cholecystectomy for concomitant gallstones is controversial.

Aims & Methods

Aim: To study the effect of prophylactic biliary stenting after achieving biliary clearance in reducing CBDS recurrence and biliary complications in patients awaiting cholecystectomy.
Methods: Patients with concomitant gallstones and CBDS were included and those who had evidence of clearance of CBDS that was documented on occlusion cholangiogram during endoscopic retrograde cholangiography (ERC) were randomized to prophylactic biliary stenting (group A) or no stenting (group B). After clearance, the patients were advised to undergo cholecystectomy as early as possible. The patients were followed up at 1 week, 1, 2, and 3 months after clearance, and at 3 months or after cholecystectomy (whichever was early) they were assessed for recurrence of CBDS by either endoscopic ultrasound or MRCP or ultrasound abdomen and liver function test whichever was convenient for the patient. During each follow-up visit patients were assessed for symptoms and were requested to do a hemogram and liver function test. The primary outcome was to compare the CBDS recurrence rate, while the secondary outcomes were to compare the biliary complications and the need for repeat/ emergency ERC. Statistical analysis was performed using IBM SPSS version 22. Quantitative variables were analysed by mean and standard deviation (SD) while categorical variables were analysed by frequency and proportion. Median and interquartile range (IQR) were used for non-normally distributed quantitative variables. Comparison of categorical outcomes was done between study groups using Chi-square test/Fisher's Exact test. At different time periods, categorical variables were compared using the McNemar test. For quantitative parameters with normal distribution, the mean values were compared between study groups using an independent sample t-test (2 groups). Change in the quantitative parameters before and after the intervention was assessed by paired t-test. P value < 0.05 was considered statistically significant.

Results

Between September 2021 to July 2022, 70 patients were randomized into Group A-stenting (n-35) and Group B- no stenting (n-35). Sixty-six patients were included in the final analysis. The baseline characteristics were comparable between the 2 groups. Ten (15.15%) patients had a recurrence of CBDS and it was comparable between the 2 groups [7/34 (20.59%) and 3/32 (9.38%), p=0.306]. Two patients in the stent group developed biliary complication (cholecystitis) while none of the patients in the no stent group developed any biliary complication. Three patients in the stent group while none in the no-stent group developed post-ERC pancreatitis (p=0.085). None of the patients in both the groups needed emergency ERC during the follow-up.

Conclusion

This randomized trial shows no advantage of prophylactic stenting in preventing CBDS recurrence or biliary complications in patients waiting for cholecystectomy after biliary clearance. (CTRI registration number: CTRI/2021/09/036538)

Introduction

Post-ERCP pancreatitis (PEP) is the most important adverse event of ERCP. In order to make a decision to take adequate preventive measures, it is necessary to estimate the risk of PEP in individual cases. Several models have been reported to predict PEP, but these models do not have high generalizability.

Aims & Methods

The aim of this study is to develop a reliable prediction rule for PEP in biliary ERCP for patients with intact papilla.
We had conducted a multicenter prospective cohort study which investigated the incidence of adverse events of biliary ERCP in patients with intact papilla (1). This study is a sub-analysis of the cohort study. A total of 3739 patients were prospectively enrolled at 36 hospitals of the Biliopancreatic Study Group of West Japan between April 2017 and March 2018. Exclusion criteria were acute pancreatitis, altered gastrointestinal anatomy, and an ASA physical status greater than 3.
The primary outcomes were the incidence of PEP. The diagnosis and severity of PEP were defined by the lexicon for endoscopic adverse events (2). Patient-related factors, operator-related factors, procedure-related factors and preventative measures were accumulated.
Multivariable logistic regression analysis with stepwise methods were used to identify predictors for PEP and a scoring model was developed based on their β coefficients.
The performance of the model was evaluated based on area under the receiver-operating characteristic curve (AUC), calibration and clinical adequacy. In addition, internal validity was assessed using the bootstrap method by sampling with replacements for 2000 iterations.

Results

PEP developed in 258 cases (6.9%), 201 mild, 39 moderate, 17 severe and 1 fatal. Significant factors with adjusted odds ratios (OR) in multivariate analysis were under 50 years old (OR: 1.78 [95% confidence interval (CI), 1.10-2.79]), female (OR: 1.33 [95% CI, 1.02-1.73]), ASA 3 (OR: 0.62 [95% CI, 0.39-0.95]), cholangitis (OR: 0.72 [95% CI, 0.52-0.98]), hypoamylasemia (OR: 0.62 [95% CI, 0.37-0.97]), obstruction of the main pancreatic duct (MPD) at the pancreatic head (OR: 0.41 [95% CI, 0.22-0.71]), cannulation time > 20 min (OR: 1.43 [95% CI, 1.03-1.97]), guidewire insertion into pancreatic duct (OR: 2.88 [95% CI, 2.17-3.82]), intraductal ultrasonography (IDUS) (OR: 1.49 [95% CI, 1.00-2.17]), biopsy (OR: 1.51 [95% CI, 1.00-2.23]) and total procedure time > 60 min (OR:1.64 [95% CI, 1.17-2.28]).
A scoring system was constructed from the eleven clinical variables as below. Under 50 years old: 2 points, female: 1 point, ASA 3: -2 points, cholangitis: -1 point, hypoamylasemia: -1 point, obstruction of the MPD at the pancreatic head: -2 points, cannulation time > 20 min: 1 point, guidewire insertion into pancreatic duct: 3 points, IDUS: 1 point, biopsy: 1 point and total procedure time > 60 min: 2 points. The points were added up to a total score that predicts the risk of pancreatitis.
AUC of this scoring model was 0.721. A bootstrap analysis revealed a corrected AUC of 0.728 (95%CI: 0.697 - 0.759).
When all cases were divided into three groups; a low-risk group (scoring -1 points or fewer), an intermediate-risk group (between 0 and 3 points) and a high-risk group (more than 4 points), the incidence of PEP were 1.6% (17/1048), 6.0% (121/2012) and 17.7% (120/679), respectively.

Conclusion

This scoring model is useful for predicting and stratifying the risk of PEP in individual cases. Aggressive preventive measures are recommended for patients in the high-risk group with a score of 4 or higher.

References

1. Fujita K, Yazumi S, Matsumoto H, Asada M, Nebiki H, Matsumoto K, et al. Multicenter prospective cohort study of adverse events associated with biliary endoscopic retrograde cholangiopancreatography: Incidence of adverse events and preventive measures for post-endoscopic retrograde cholangiopancreatography pancreatitis. Dig Endosc. 2022;34(6):1198-204.
2. Cotton PB, Eisen GM, Aabakken L, Baron TH, Hutter MM, Jacobson BC, et al. A lexicon for endoscopic adverse events: report of an ASGE workshop. Gastrointest Endosc. 2010;71(3):446-54.