Ethical Implications of Artificial Intelligence in Gastroenterology: The Co-pilot or the Captain?

Though artificial intelligence (AI) is being widely implemented in gastroenterology (GI) and hepatology and has the potential to be paradigm shifting for clinical practice, its pitfalls must be considered along with its advantages. Currently, although the use of AI is limited in practice to supporting clinical judgment, medicine is rapidly heading toward a global environment where AI will be increasingly autonomous. Broader implementation of AI will require careful ethical considerations, specifically related to bias, privacy, and consent. Widespread use of AI raises concerns related to increasing rates of systematic errors, potentially due to bias introduced in training datasets. We propose that a central repository for collection and analysis for training and validation datasets is essential to overcoming potential biases. Since AI does not have built-in concepts of bias and equality, humans involved in AI development and implementation must ensure its ethical use and development. Moreover, ethical concerns regarding data ownership and health information privacy are likely to emerge, obviating traditional methods of obtaining patient consent that cover all possible uses of patient data. The question of liability in case of adverse events related to use of AI in GI must be addressed among the physician, the healthcare institution, and the AI developer. Though the future of AI in GI is very promising, herein we review the ethical considerations in need of additional guidance informed by community experience and collective expertise.

The incremental yield of adenoma detection with I-Scan versus high-definition white light colonoscopy-a systematic review and meta-analysis of randomized studies.

INTRODUCTION: The incremental yield of I-Scan virtual chromoendoscopy compared to high-definition white light endoscopy (HD-WLE) in detection of colorectal adenomas has not been thoroughly elucidated. METHODS: A systematic search from inception to April 2023 was conducted to identify randomized controlled trials (RCTs) comparing I-Scan to HD-WLE for detection of adenomas. A random effects model was used to compute risk difference (RD) with corresponding 95% confidence intervals in adenoma detection rate (ADR). Influence analysis was done to assess robustness of findings. The number needed to diagnose was computed. Heterogeneity was assessed using the I2 statistic and explored further by subgroup analyses defined a priori. Certainty in effect estimates was assessed using the GRADE approach. RESULTS: We identified four studies (I-Scan n = 730, HD-WLE n = 765). I-Scan increased adenoma detection by 9% (risk difference (RD), 0.09; 0.04, 0.14; I2 02%; certainty, low). Influence analysis revealed that the gain in yield remained statistically significant with exclusion of all but one study. The number needed to capture one additional adenomatous polyp with I-Scan use was 11.2. I-Scan 1 use was associated with a statistically significant gain in ADR, whereas no significant difference in ADR was noted with I-Scan use on subgroup analysis. DISCUSSION: In conclusion, I-Scan increases the yield of adenoma detection by 9% compared to HD-WLE, with low certainty in the estimate of this effect. Data on the gain in yield of detecting large polyps, sessile serrated lesions, and on the impact of formally training endoscopists and trainees in I-Scan use and similar technology on adenoma detection rate are needed.

Cost-effectiveness analysis of optimal diagnostic strategy for patients with symptomatic cholelithiasis with intermediate probability for choledocholithiasis.

BACKGROUND AND AIMS: EUS, MRCP, and intraoperative cholangiogram (IOC) are the recommended diagnostic modalities for patients with intermediate probability for choledocholithiasis (IPC). The relative cost-effectiveness of these modalities in patients with cholelithiasis and IPC is understudied. METHODS: We developed a decision tree for diagnosing IPC (base-case probability, 50%; range, 10%-70%); patients with a positive test were modeled to undergo therapeutic ERCP. The strategies tested were laparoscopic cholecystectomy with IOC (LC-IOC), MRCP, single-session EUS + ERCP, and separate-session EUS + ERCP. Costs and probabilities were extracted from the published literature. Effectiveness was assessed by assigning utility scores to health states, average proportion of true-positive diagnosis of IPC, and the mean length of stay (LOS) per strategy. Cost-effectiveness was assessed by extrapolating a net-monetary benefit (NMB) and average cost per true-positive diagnosis. RESULTS: LC-IOC was the most cost-effective strategy to diagnose IPC (base-case probability of 50%) among patients with cholelithiasis in health state-based effectiveness analysis (NMB of $34,612), diagnostic test accuracy-based effectiveness analysis (average cost of $13,260 per true-positive diagnosis), and LOS-based effectiveness analysis (mean LOS of 4.13) compared with strategies 2 (MRCP), 3 (single-session EUS + ERCP), and 4 (separate-session EUS + ERCP). These findings were robust on deterministic and probabilistic sensitivity analyses. CONCLUSIONS: For patients with cholelithiasis with IPC, LC-IOC is a cost-effective approach that should limit preoperative testing and may shorten hospital LOS. Our findings may be used to design institutional and organizational management protocols.

(#WomeninDDW) Persistent Speaker Gender Gap at the Premier Digestive Disease Event.

BACKGROUND: Over the last few decades, advances have been made regarding gender equality starting from medical students to trainees, to leadership in academics. The female representation in specialty academic conferences not only reflects the existing gender disparities in that specialty but also can influence young female trainees to join that field. Digestive Disease Week (DDW) is the premier digestive disease event. We aimed to calculate the proportion of female representation among speakers and moderators at the DDW meetings held from 2018 to 2020. METHODS: The data for DDW 2018-2020 were collected via the online web-based planner. The gender of speakers of presentations and moderators of sessions were identified by a google search. We further categorized the data by each participating society (AGA, ASGE, AASLD, and SSAT), by presentation track, by session track, and total overall representation in each year. RESULTS: Despite the subject of the gender gap being in focus, the proportion of female moderators and speakers was low in DDW in the last 3 years. The female speakers constituted 31.6% in 2018, 33.8% in 2019 and 34.6% in 2020. There was slightly improved female representation in sessions of Inflammatory Bowel Disease, Stomach, and Small Bowel Disorders, Microbiome in GI & Liver disease, and Basic Science over the last 3 years. CONCLUSION: Based on our study and those referenced in this article, we believe that strategies to promote the inclusivity of female moderators and speakers at DDW provide a huge opportunity to influence gender equity within GI.

Kras mutation rate precisely orchestrates ductal derived pancreatic intraepithelial neoplasia and pancreatic cancer.

Pancreatic ductal adenocarcinoma (PDAC) is the third leading cause of cancer-related death in the United States. Despite the high prevalence of Kras mutations in pancreatic cancer patients, murine models expressing the oncogenic mutant Kras (Krasmut) in mature pancreatic cells develop PDAC at a low frequency. Independent of cell of origin, a second genetic hit (loss of tumor suppressor TP53 or PTEN) is important for development of PDAC in mice. We hypothesized ectopic expression and elevated levels of oncogenic mutant Kras would promote PanIN arising in pancreatic ducts. To test our hypothesis, the significance of elevating levels of K-Ras and Ras activity has been explored by expression of a CAG driven LGSL-KrasG12V allele (cKras) in pancreatic ducts, which promotes ectopic Kras expression. We predicted expression of cKras in pancreatic ducts would generate neoplasia and PDAC. To test our hypothesis, we employed tamoxifen dependent CreERT2 mediated recombination. Hnf1b:CreERT2;KrasG12V (cKrasHnf1b/+) mice received 1 (Low), 5 (Mod) or 10 (High) mg per 20 g body weight to recombine cKras in low (cKrasLow), moderate (cKrasMod), and high (cKrasHigh) percentages of pancreatic ducts. Our histologic analysis revealed poorly differentiated aggressive tumors in cKrasHigh mice. cKrasMod mice had grades of Pancreatic Intraepithelial Neoplasia (PanIN), recapitulating early and advanced PanIN observed in human PDAC. Proteomics analysis revealed significant differences in PTEN/AKT and MAPK pathways between wild type, cKrasLow, cKrasMod, and cKrasHigh mice. In conclusion, in this study, we provide evidence that ectopic expression of oncogenic mutant K-Ras in pancreatic ducts generates early and late PanIN as well as PDAC. This Ras rheostat model provides evidence that AKT signaling is an important early driver of invasive ductal derived PDAC.

Bleeding Risk and Mortality Associated With Uninterrupted Antithrombotic Therapy During Percutaneous Endoscopic Gastrostomy Tube Placement.

INTRODUCTION: Antithrombotic therapy is often interrupted before the placement of a percutaneous endoscopic gastrostomy (PEG) tube because of potentially increased risk of hemorrhagic events. The aim of our study was to evaluate the risk of bleeding events and overall complication rates after PEG in patients on uninterrupted antiplatelet and anticoagulation therapy in a high-volume center. METHODS: Data regarding demographics, diagnoses, comorbidities, and clinical outcomes pertinent to PEG were collected from 2010 to 2016. Furthermore, data regarding antithrombotic therapy along with the rate of minor or major complications including bleeding associated with this procedure were analyzed. Significant bleeding was defined as postprocedure bleeding from PEG site requiring a blood transfusion and/or surgical/endoscopic intervention. RESULTS: We included 1,613 consecutive PEG procedures in this study, of which 1,540 patients (95.5%) received some form of uninterrupted antithrombotic therapy. Of those patients, 535 (34.7%) were on aspirin, 256 (16.6%) on clopidogrel, and 119 (7.7%) on both aspirin and clopidogrel. Subcutaneous heparin was uninterrupted in 980 (63.6%), intravenous heparin in 34 (2.1%), warfarin in 168 (10.9%), and direct-acting oral anticoagulation in 82 (5.3%) patients who overlapped on multiple drugs. We observed 6 significant bleeding events in the entire cohort (0.39%), and all were in subcutaneous heparin groups either alone or in combination with aspirin. No clinically significant bleeding was noted in patients on uninterrupted aspirin, warfarin, clopidogrel, or direct-acting oral anticoagulation groups. Only 5 patients (0.31%) had PEG-related mortality. DISCUSSION: The risk of significant bleeding associated with the PEG placement was minimal in patients on uninterrupted periprocedural antithrombotic therapy.

A potential signaling axis between RON kinase receptor and hypoxia-inducible factor-1 alpha in pancreatic cancer.

The Cancer Genome Atlas (TCGA) of a pancreatic cancer cohort identified high MST1R (RON tyrosine kinase receptor) expression correlated with poor prognosis in human pancreatic cancer. RON expression is null/minimal in normal pancreas but elevates from pan-in lesions through invasive carcinomas. We report using multiple approaches RON directly regulates HIF-1α, a critical driver of genes involved in cancer cell invasion and metastasis. RON and HIF-1α are highly co-expressed in the 101 human PDAC tumors analyzed and RON expression correlated with HIF-1α expression in a subset of PDAC cell lines. knockdown of RON expression in RON positive cells blocked HIF-1α expression, whereas ectopic RON expression in RON null cells induced HIF-1α expression suggesting the direct regulation of HIF-1α by RON kinase receptor. RON regulates HIF-1α through an unreported transcriptional mechanism involving PI3 kinase-mediated AKT phosphorylation and Sp1-dependent HIF-1α promoter activity leading to increased HIF-1α mRNA expression. RON/HIF-1α modulation altered the invasive behavior of PDAC cells. A small-molecule RON kinase inhibitor decreased RON ligand, MSP-induced HIF-1α expression, and invasion of PDAC cells. Immunohistochemical analysis on RON knockdown orthotopic PDAC tumor xenograft confirmed that RON inhibition significantly blocked HIF-1α expression. RON/HIF-1α co-expression also exists in triple-negative breast cancer cells, a tumor type that also lacks molecular therapeutic targets. This is the first report describing RON/HIF-1α axis in any tumor type and is a potential novel therapeutic target.

Comparing diagnostic accuracy of current practice guidelines in predicting choledocholithiasis: outcomes from a large healthcare system comprising both academic and community settings.

BACKGROUND AND AIMS: The American Society for Gastrointestinal Endoscopy (ASGE) 2010 guidelines for suspected choledocholithiasis were recently updated by proposing more specific criteria for selection of high-risk patients to undergo direct ERCP while advocating the use of additional imaging studies for intermediate- and low-risk individuals. We aim to compare the performance and diagnostic accuracy of 2019 versus 2010 ASGE criteria for suspected choledocholithiasis. METHODS: We performed a retrospective chart review of a prospectively maintained database (2013-2019) of over 10,000 ERCPs performed by 70 gastroenterologists in our 14-hospital system. We randomly selected 744 ERCPs in which the primary indication was suspected choledocholithiasis. Patients with a history of cholecystectomy or prior sphincterotomy were excluded. The same patient cohort was assigned as low, intermediate, or high risk according to the 2010 and 2019 guideline criteria. Overall accuracy of both guidelines was compared against the presence of stones and/or sludge on ERCP. RESULTS: Of 744 patients who underwent ERCP, 544 patients (73.1%) had definite stones during ERCP and 696 patients (93.5%) had stones and/or sludge during ERCP. When classified according to the 2019 guidelines, fewer patients were high risk (274/744, 36.8%) compared with 2010 guidelines (449/744, 60.4%; P < .001). Within the high-risk group per both guidelines, definitive stone was found during ERCP more frequently in the 2019 guideline cohort (226/274, 82.5%) compared with the 2010 guideline cohort (342/449, 76.2%; P < .001). In our patient cohort, overall specificity of the 2010 guideline was 46.5%, which improved to 76.0% as per 2019 guideline criteria (P < .001). However, no significant change was noted for either positive predictive value or negative predictive value between 2019 and 2010 guidelines. CONCLUSIONS: The 2019 ASGE guidelines are more specific for detection of choledocholithiasis during ERCP when compared with the 2010 guidelines. However, a large number of patients are categorized as intermediate risk per 2019 guidelines and will require an additional confirmatory imaging study.

Protein kinase D up-regulates transcription of VEGF receptor-2 in endothelial cells by suppressing nuclear localization of the transcription factor AP2ß.

Vascular endothelial growth factor A (VEGF) signals primarily through its cognate receptor VEGF receptor-2 (VEGFR-2) to control vasculogenesis and angiogenesis, key physiological processes in cardiovascular disease and cancer. In human umbilical vein endothelial cells (HUVECs), knockdown of protein kinase D-1 (PKD1) or PKD2 down-regulates VEGFR-2 expression and inhibits VEGF-induced cell proliferation and migration. However, how PKD regulates VEGF signaling is unclear. Previous bioinformatics analyses have identified binding sites for the transcription factor activating enhancer-binding protein 2 (AP2) in the VEGFR-2 promoter. Using ChIP analyses, here we found that PKD knockdown in HUVECs increases binding of AP2ß to the VEGFR-2 promoter. Luciferase reporter assays with serial deletions of AP2-binding sites within the VEGFR-2 promoter revealed that its transcriptional activity negatively correlates with the number of these sites. Next we demonstrated that AP2ß up-regulation decreases VEGFR-2 expression and that loss of AP2ß enhances VEGFR-2 expression in HUVECs. In vivo experiments confirmed increased VEGFR-2 immunostaining in the spinal cord of AP2ß knockout mouse embryos. Mechanistically, we observed that PKD phosphorylates AP2ß at Ser258 and Ser277 and suppresses its nuclear accumulation. Inhibition of PKD activity with a pan-PKD inhibitor increased AP2ß nuclear localization, and overexpression of both WT and constitutively active PKD1 or PKD2 reduced AP2ß nuclear localization through a Ser258- and Ser277-dependent mechanism. Furthermore, substitution of Ser277 in AP2ß increased its binding to the VEGFR-2 promoter. Our findings uncover evidence of a molecular pathway that regulates VEGFR-2 expression, insights that may shed light on the etiology of diseases associated with aberrant VEGF/VEGFR signaling.

A Combination of MUC5AC and CA19-9 Improves the Diagnosis of Pancreatic Cancer: A Multicenter Study.

OBJECTIVES: Pancreatic cancer (PC) is a lethal malignancy that lacks specific diagnostic markers. The present study explores the diagnostic potential of the most differentially overexpressed secretory mucin MUC5AC alone and in combination with CA19-9 using multi-center training and validation sets. METHODS: The expression of MUC5AC in benign pancreatic pathologies, PC precursor lesions, primary PC tissues and metastatic lesions was evaluated by immunohistochemistry. Circulating MUC5AC levels were measured using sandwich ELISA assay developed in-house, and CA19-9 was measured using radioimmunoassay. A combined training set (n=346) was used to evaluate the diagnostic (n=241) and predictive (n=105, total samples 201 from pre- and post-surgical and chemotherapy set) significance of MUC5AC. Results were further validated with a pre-defined cut-off value using independent sets from the Mayo Clinic (n=94) and the University of Pittsburgh Medical Center (n=321). RESULTS: Tissue expression analyses indicated the de novo expression of MUC5AC in pancreatic intraepithelial precursor lesions 1A (PanIN1A); the expression was maintained through all stages of progression to invasive adenocarcinoma. The median circulating MUC5AC levels in patients with resectable early-stage PC (EPC) (stage 1/2; 67.2 ng/ml, IQR: 23.9-382.1) and unresectable late-stage PC (LPC) (stage 3/4; 389.7 ng/ml, IQR: 87.7-948.6) were significantly higher compared with (P-value ≤0.0001) benign controls (BC) (7.2 ng/ml, IQR: 0.4-26.5) and (P-value ≤0.0001) chronic pancreatitis (CP) controls (8.4 ng/ml, IQR: 1.5-19.2). In the diagnostic training set (n=241), MUC5AC efficiently differentiated EPC from healthy controls (HC) (83%/80% sensitive (SN)/specific (SP)), BC (67%/87% SN/SP), and CP (83%/77% SN/SP). Independent validation sets from the Mayo Clinic and UPMC confirmed the diagnostic potential of MUC5AC to differentiate EPC from BC (68%/73%; 65%/83%) and CP (68%/79%; 65%/72%). Furthermore, MUC5AC and CA19-9 combination significantly improved (p-value < 0.001) the diagnostic accuracy for differentiating resectable cases from controls. CONCLUSIONS: MUC5AC is a valuable diagnostic biomarker, either alone or in combination with CA19-9, to differentiate PC from CP and benign controls.

Increased Incidence of Second Primary Pancreatic Cancer in Patients with Prior Colorectal Cancer: A Population-Based US Study.

BACKGROUND: Identifying high-risk groups for pancreatic cancer (PC) may lead to earlier detection. We determined the risk of subsequent PC among survivors of sporadic colorectal cancer (CRC). METHODS: We evaluated data from the US Surveillance, Epidemiology, and End Results registry to identify individuals with primary CRC between the years 1973-2006. Standardized incidence ratios (SIRs) and 95 % confidence interval (95 % CI) were calculated to compare the risk of subsequent primary PC in the study cohort to that of the standard population. Analysis was stratified by age at diagnosis of CRC and sex. CRC characteristics were compared among CRC survivors with and without PC. Multivariate sub-hazard ratios were calculated to identify factors associated with subsequent primary PC, using death from non-PC causes as a competing event. RESULTS: Of the 273,144 patients with first primary CRC, 657 (0.24 %) developed subsequent PC. CRC survivors were more likely to develop PC (SIR 1.22; 95 % CI 1.09-1.35). Mean latency period (time between CRC and PC diagnosis) was 1, 3, and 5 years from index age of CRC 20-49, 50-64, and >65 years, respectively. Multivariate analysis showed CRC survivors >50 years had 3.5-fold, those with right-sided CRC had 1.2-fold, and those with localized and regional CRCs had sixfold and fivefold increased risk of PC, respectively. CONCLUSION: This study suggests that CRC survivors have an increased risk of developing subsequent PC within 1-5 years. CRC survivors age >50 with localized/regional stage, and right-sided CRC have higher predisposition to PC.

A preoperative serum signature of CEA+/CA125+/CA19-9 ≥ 1000 U/mL indicates poor outcome to pancreatectomy for pancreatic cancer.

Pancreatectomy is associated with significant morbidity and unpredictable outcome, with few diagnostic tools to determine, which patients gain the most benefit from this treatment, especially before the operation. This study aimed to define a preoperative signature panel of serum markers to indicate response to pancreatectomy for pancreatic cancer. Over 1000 patients with pancreatic cancer treated at two independent high-volume institutions were included in this study and were divided into three groups, including resected, locally advanced and metastatic. Eight serum tumor markers most commonly used in gastrointestinal cancers were analyzed for patient outcome. Preoperative CA19-9 independently indicated surgical response in pancreatic cancer. Patients with CA19-9 ≥1000 U/mL generally had a poor surgical benefit. However, a subset of these patients still achieved a survival advantage when CA19-9 levels decreased postoperatively. CEA and CA125 in the presence of CA19-9 ≥1000 U/mL could independently predict the non-decrease of CA19-9 postoperatively. The combination of the three markers was useful for predicting a worse surgical outcome with a median survival of 5.1 months vs. 23.0 months (p < 0.001) for the training cohort and 7.0 months vs. 18.2 months (p < 0.001) for the validation cohort and also suggested a higher prevalence of early distant metastasis after surgery. Resected patients with this proposed signature showed no survival advantage over patients in the locally advanced group who did not receive pancreatectomy. Therefore, a preoperative serum signature of CEA(+)/CA125(+)/CA19-9 ≥1000 U/mL is associated with poor surgical outcome and can be used to select appropriate patients with pancreatic cancer for pancreatectomy.