Machine Learning Models for Pancreatic Cancer Risk Prediction Using Electronic Health Record Data-A Systematic Review and Assessment.

INTRODUCTION: Accurate risk prediction can facilitate screening and early detection of pancreatic cancer (PC). We conducted a systematic review to critically evaluate effectiveness of machine learning (ML) and artificial intelligence (AI) techniques applied to electronic health records (EHR) for PC risk prediction. METHODS: Ovid MEDLINE(R), Ovid EMBASE, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, Scopus, and Web of Science were searched for articles that utilized ML/AI techniques to predict PC, published between January 1, 2012, and February 1, 2024. Study selection and data extraction were conducted by 2 independent reviewers. Critical appraisal and data extraction were performed using the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies checklist. Risk of bias and applicability were examined using prediction model risk of bias assessment tool. RESULTS: Thirty studies including 169,149 PC cases were identified. Logistic regression was the most frequent modeling method. Twenty studies utilized a curated set of known PC risk predictors or those identified by clinical experts. ML model discrimination performance (C-index) ranged from 0.57 to 1.0. Missing data were underreported, and most studies did not implement explainable-AI techniques or report exclusion time intervals. DISCUSSION: AI/ML models for PC risk prediction using known risk factors perform reasonably well and may have near-term applications in identifying cohorts for targeted PC screening if validated in real-world data sets. The combined use of structured and unstructured EHR data using emerging AI models while incorporating explainable-AI techniques has the potential to identify novel PC risk factors, and this approach merits further study.

An Unsuccessful Randomized Trial of Percutaneous vs Endoscopic Drainage of Suspected Malignant Hilar Obstruction.

Percutaneous transhepatic biliary drainage (PTBD) and endoscopic retrograde cholangiopancreatography (ERCP) are widely accepted but competing approaches for the management of malignant obstruction at the hilum of the liver. ERCP is favored in the United States on the basis of high success rates for non-hilar indications, the perceived safety and superior tissue sampling capability of ERCP relative to PTBD, and the avoidance of external drains that are undesirable to patients. A recent randomized controlled trial (RCT) comparing the 2 modalities in patients with resectable hilar cholangiocarcinoma was terminated prematurely because of higher mortality in the PTBD group.1 In contrast, most observational data suggest that PTBD is superior for achieving complete drainage.2-6 Because the preferred procedure remains uncertain, we aimed to compare PTBD and ERCP as the primary intervention in patients with cholestasis due to malignant hilar obstruction (MHO).

The Pancreatitis Activity Scoring System predicts clinical outcomes in acute pancreatitis: findings from a prospective cohort study.

OBJECTIVES: The Pancreatitis Activity Scoring System (PASS) has been derived by an international group of experts via a modified Delphi process. Our aim was to perform an external validation study to assess for concordance of the PASS score with high face validity clinical outcomes and determine specific meaningful thresholds to assist in application of this scoring system in a large prospectively ascertained cohort. METHODS: We analyzed data from a prospective cohort study of consecutive patients admitted to the Los Angeles County Hospital between March 2015 and March 2017. Patients were identified using an emergency department paging system and electronic alert system. Comprehensive characterization included substance use history, pancreatitis etiology, biochemical profile, and detailed clinical course. We calculated the PASS score at admission, discharge, and at 12 h increments during the hospitalization. We performed several analyses to assess the relationship between the PASS score and outcomes at various points during hospitalization as well as following discharge. Using multivariable logistic regression analysis, we assessed the relationship between admission PASS score and risk of severe pancreatitis. PASS score performance was compared to established systems used to predict severe pancreatitis. Additional inpatient outcomes assessed included local complications, length of stay, development of systemic inflammatory response syndrome (SIRS), and intensive care unit (ICU) admission. We also assessed whether the PASS score at discharge was associated with early readmission (re-hospitalization for pancreatitis symptoms and complications within 30 days of discharge). RESULTS: A total of 439 patients were enrolled, their mean age was 42 (±15) years, and 53% were male. Admission PASS score >140 was associated with moderately severe and severe pancreatitis (OR 3.5 [95% CI 2.0, 6.3]), ICU admission (OR 4.9 [2.5, 9.4]), local complications (3.0 [1.6, 5.7]), and development of SIRS (OR 2.9 [1.8, 4.5]) as well as prolongation of hospitalization by a mean of 1.5 (1.3-1.7) days. For the prediction of moderately severe/severe pancreatitis, the PASS score (AUC = 0.71) was comparable to the more established Ranson's (AUC = 0.63), Glasgow (AUC = 0.72), Panc3 (AUC = 0.57), and HAPS (AUC = 0.54) scoring systems. Discharge PASS score >60 was associated with early readmission (OR 5.0 [2.4, 10.7]). CONCLUSIONS: The PASS score is associated with important clinical outcomes in acute pancreatitis. The ability of the score to forecast important clinical events at different points in the disease course suggests that it is a valid measure of activity in patients with acute pancreatitis.

Percutaneous transhepatic vs. endoscopic retrograde biliary drainage for suspected malignant hilar obstruction: study protocol for a randomized controlled trial.

BACKGROUND: The optimal approach to the drainage of malignant obstruction at the liver hilum remains uncertain. We aim to compare percutaneous transhepatic biliary drainage (PTBD) to endoscopic retrograde cholangiography (ERC) as the first intervention in patients with cholestasis due to suspected malignant hilar obstruction (MHO). METHODS: The INTERCPT trial is a multi-center, comparative effectiveness, randomized, superiority trial of PTBD vs. ERC for decompression of suspected MHO. One hundred and eighty-four eligible patients across medical centers in the United States, who provide informed consent, will be randomly assigned in 1:1 fashion via a web-based electronic randomization system to either ERC or PTBD as the initial drainage and, if indicated, diagnostic procedure. All subsequent clinical interventions, including crossover to the alternative procedure, will be dictated by treating physicians per usual clinical care. Enrolled subjects will be assessed for successful biliary drainage (primary outcome measure), adequate tissue diagnosis, adverse events, the need for additional procedures, hospitalizations, and oncological outcomes over a 6-month follow-up period. Subjects, treating clinicians and outcome assessors will not be blinded. DISCUSSION: The INTERCPT trial is designed to determine whether PTBD or ERC is the better initial approach when managing a patient with suspected MHO, a common clinical dilemma that has never been investigated in a randomized trial. TRIAL REGISTRATION: ClinicalTrials.gov, Identifier: NCT03172832 . Registered on 1 June 2017.

Assessing Intra-arterial Complications of Planning and Treatment Angiograms for Y-90 Radioembolization.

PURPOSE: To report hepatic arterial-related complications encountered during planning and treatment angiograms for radioembolization and understand any potential-associated risk factors. MATERIALS AND METHODS: 518 mapping or treatment angiograms for 180 patients with primary or metastatic disease to the liver treated by Yttrium-90 radioembolization between 2/2010 and 12/2015 were retrospectively reviewed. Intra-procedural complications were recorded per SIR guidelines. Patient demographics, indication for treatment, prior exposure to chemotherapeutic agents, operator experience, and disease burden were reviewed. Technical variables including type of radioembolic (glass vs. resin microspheres), indication for angiography (mapping vs. treatment), variant anatomy, and attempts at coil embolization were also assessed. RESULTS: Thirteen (13/518, 2.5%) arterial-related complications occurred in 13 patients. All but two complications resulted during transcatheter coil embolization to prevent non-target embolization. Complications included coil migration (n = 6), arterial dissection (n = 2), focal vessel perforation (n = 2), arterial thrombus (n = 2), and vasospasm prohibiting further arterial sub-selection (n = 1). Transarterial coiling was identified as a significant risk factor of complications on both univariate and multivariate regression analysis (odds ratio 7.8, P = 0.004). Usage of resin microspheres was also a significant risk factor (odds ratio 9.5, P = 0.042). No other technical parameters or pre-procedural variables were significant after adjusting for confounding on multivariate analysis (P > 0.05). CONCLUSION: Intra-procedural hepatic arterial complications encountered during radioembolization were infrequent but occurred mainly during coil embolization to prevent non-target delivery to extra-hepatic arteries.