Application of machine-learning model to optimize colonic adenoma detection in India.

AIMS: There is limited data on the prevalence and risk factors of colonic adenoma from the Indian sub-continent. We aimed at developing a machine-learning model to optimize colonic adenoma detection in a prospective cohort. METHODS: All consecutive adult patients undergoing diagnostic colonoscopy were enrolled between October 2020 and November 2022. Patients with a high risk of colonic adenoma were excluded. The predictive model was developed using the gradient-boosting machine (GBM)-learning method. The GBM model was optimized further by adjusting the learning rate and the number of trees and 10-fold cross-validation. RESULTS: Total 10,320 patients (mean age 45.18 ± 14.82 years; 69% men) were included in the study. In the overall population, 1152 (11.2%) patients had at least one adenoma. In patients with age > 50 years, hospital-based adenoma prevalence was 19.5% (808/4144). The area under the receiver operating curve (AUC) (SD) of the logistic regression model was 72.55% (4.91), while the AUCs for deep learning, decision tree, random forest and gradient-boosted tree model were 76.25% (4.22%), 65.95% (4.01%), 79.38% (4.91%) and 84.76% (2.86%), respectively. After model optimization and cross-validation, the AUC of the gradient-boosted tree model has increased to 92.2% (1.1%). CONCLUSIONS: Machine-learning models may predict colorectal adenoma more accurately than logistic regression. A machine-learning model may help optimize the use of colonoscopy to prevent colorectal cancers. TRIAL REGISTRATION: ClinicalTrials.gov (ID: NCT04512729).

Technical performance and diagnostic yield of motorised spiral enteroscopy compared with single-balloon enteroscopy in suspected Crohn's disease: a randomised controlled, open-label study (the MOTOR-CD trial).

OBJECTIVE: Recent studies have shown that motorised spiral enteroscopy (MSE) enables deeper and total small bowel evaluation compared with single-balloon enteroscopy (SBE) in suspected Crohn's disease (CD) when analysed per procedure. However, no randomised controlled study has compared bidirectional MSE with bidirectional SBE in suspected CD. DESIGN: Patients with suspected CD requiring small bowel enteroscopy were randomly assigned to either SBE or MSE between May 2022 and September 2022 in a high volume tertiary centre. Bidirectional enteroscopy was done if intended lesion could not be reached on unidirectional study. Comparison was made with regard to technical success (ability to reach lesion), diagnostic yield, depth of maximal insertion (DMI), procedure time and total enteroscopy rates. Depth:time ratio was calculated to avoid confounding for the location of lesion. RESULTS: Among 125 suspected patients with CD (28% female, 18-65 years, median 41 years), 62 and 63 underwent MSE and SBE, respectively. The overall technical success (98.4 %: MSE, 90.5 %: SBE; p=0.11), diagnostic yield (95.2%: MSE; 87.3%: SBE, p=0.2) and procedure time were not significantly different. However, MSE appeared to have higher technical success (96.8% vs 80.7%, p=0.08) in deeper small bowel (distal jejunum/proximal ileum) with higher DMI, higher depth:time ratio and total enteroscopy rates when attempted (77.8% vs 11.1%, p=0.0007). Both the modalities were safe although minor adverse events were more common with MSE. CONCLUSION: MSE and SBE have comparable technical success and diagnostic yield for small bowel evaluation in suspected CD. MSE scores over SBE with regard to deeper small bowel evaluation with complete small bowel coverage and higher depth of insertion in a shorter time. TRIAL REGISTRATION NUMBER: NCT05363930.

EUS versus MRCP to perform ERCP in patients with intermediate likelihood of choledocholithiasis: a randomised controlled trial.

OBJECTIVE: In patients with an intermediate likelihood of choledocholithiasis, European Society of Gastrointestinal Endoscopy (ESGE) guidelines recommend endoscopic ultrasound (EUS) or magnetic resonance cholangiopancreatography (MRCP) to diagnose choledocholithiasis to make the indication for endoscopic retrograde cholangiopancreatography (ERCP) treatment; there is no randomised control trial to compare both in this setting. DESIGN: Patients with suspected choledocholithiasis satisfying ESGE guideline's intermediate likelihood were screened for this single-centre randomised controlled trial between November 2019 and May 2020. The enrolled patients were randomised to either EUS or MRCP. ERCP was performed in stone positive cases or if clinical suspicion persisted during follow-up. Negative cases underwent a further 6-month clinical follow-up. Main outcome was accuracy (sensitivity/specificity) of both tests to diagnose choledocholithiasis, with ERCP or follow-up as a gold standard. RESULTS: Of 266 patients, 224 patients (mean age: 46.77±14.57 years; 50.9 % female) were enrolled; overall prevalence of choledocholithiasis was 49.6%, with a higher frequency in the MRCP group (63/112 vs 46/112 for EUS). Both sensitivity of EUS and MRCP were similarly high (92%-98%), without significant differences between the two groups. The negative predictive value and likelihood ratio + were significantly higher in EUS arm (p<0.05). The percentage of ERCPs either incorrectly halted back (false negatives: EUS: 2 vs MRCP: 5) or performed unnecessarily (false positives: EUS: 1 vs MRCP: 2) was low in both groups. CONCLUSION: The performance parameters of both EUS and MRCP are comparable for detecting choledocholithiasis in the intermediate-risk group of choledocholithiasis and the choice of a test should be based on local expertise, availability of resources and patient preference. TRIAL REGISTRATION NUMBER: NCT04173624.

Endoscopic full-thickness plication for the treatment of PPI-dependent GERD: results from a randomised, sham controlled trial.

BACKGROUND: The majority of endoscopic antireflux procedures for GERD are cumbersome to use and randomised long-term data are sparse. We conducted such a trial to determine the efficacy and safety of a novel, easy to use endoscopic full-thickness fundoplication (EFTP) device in patients with GERD. DESIGN: Patients with proton pump inhibitor (PPI)-dependent GERD were randomised to either EFTP or a sham procedure in 1:1 ratio. The primary endpoint was ≥50% improvement in the health-related quality of life (GERD-HRQL) score at 3 months. Secondary end points included improvement in GERD-HRQL, reflux symptom scores, PPI usage, oesophageal acid exposure and reflux episodes and endoscopic findings at 3, 6 and 12 months. RESULTS: Seventy patients were randomised; 35 in each group with a median (IQR) age of 36 (29-42) years, 71.4% males. 70% had non-erosive reflux disease on endoscopy with a mean DeMeester score of 18.9 (±19.93). The mean (±SD) duration of EFTP procedure was 17.4 (±4) min. The primary end point was more frequently achieved in the EFTP group (65.7% vs 2.9%; p<0.001). Median (IQR) % improvement in GERD-HRQL was significantly higher in the EFTP group at 6 (81.4 (60.9-100.0) versus 8.0 (2.2-21.6); p<0.001) and 12 (92.3 (84.4-100.0) versus 9.1 (4.8-36.0); p<0.001) months. In the EFTP group, 62.8% patients were off-PPI at 12 months compared with 11.4% in the sham group (p<0.001). pH-metry parameters partially improved at 3 months, (n=70; total reflux episodes in EFTP arm and non-acid reflux episodes for EFTP vs sham) but not at 12 months (n=27); endoscopic oesophagitis was seen in 0% in the treatment (n=18) and 5 (29.4%) in the control group (n=17) at 12 months. No major procedure-related adverse events were encountered in either group. CONCLUSION: EFTP using a novel device is safe and effective in improving quality of life in patients with PPI dependent mostly non-erosive reflux disease at short and long terms; objective parameters showed a limited response rate. TRIAL REGISTRATION NUMBER: NCT03322553.

Fatty acid ethyl ester (FAEE) associated acute pancreatitis: An ex-vivo study using human pancreatic acini.

BACKGROUND & AIM: Fatty acid ethyl esters (FAEEs), are produced by non-oxidative alcohol metabolism and can cause acinar cell damage and subsequent acute pancreatitis in rodent models. Even though experimental studies have elucidated the FAEE mediated early intra-acinar events, these mechanisms have not been well studied in humans. In the present study, we evaluate the early intra-acinar events and inflammatory response in human pancreatic acinar tissues and cells in an ex-vivo model. METHODS: Experiments were conducted using normal human pancreatic tissues exposed to FAEE. Subcellular fractionation was performed on tissue homogenates and trypsin and cathepsin B activities were estimated in these fractions. Acinar cell injury was evaluated by histology and immunohistochemistry. Cytokine release from exposed acinar cells was evaluated by performing Immuno-fluorescence. Serum was collected from patients with AP within the first 72 h of symptom onset for cytokine estimation using FACS. RESULTS: We observed significant trypsin activation and acinar cell injury in FAEE treated tissue. Cathepsin B was redistributed from lysosomal to zymogen compartment at 30 min of FAEE exposure. IHC results indicated the presence of apoptosis in pancreatic tissue at 1 & 2hrs of FAEE exposure. We also observed a time dependent increase in secretion of cytokines IL-6, IL-8, TNF-α from FAEE treated acinar tissue. There was also a significant elevation in plasma cytokines in patents with alcohol associated AP within 72 h of symptom onset. CONCLUSION: Our data suggest that alcohol metabolites can cause acute acinar cell damage and subsequent cytokine release which could eventually culminant in SIRS.