Jae Keun Park

Background Computed tomography (CT) is widely used in the initial evaluation of suspected common bile duct (CBD) stones, but limited sensitivity often necessitates additional endoscopic procedures. We developed and validated an artificial neural network (ANN) to enhance CT interpretation and assessed its potential to support endoscopic decision-making. Methods We An ANN model integrating UNETR for segmentation and ResNet50 for classification was trained to detect CBD stones on CT. Patients who underwent abdominal CT for suspected CBD stones between March 2018 and June 2023 at Hallym University Kangnam Sacred Heart Hospital were included. A retrospective derivation cohort (n = 830) was used for model training, and a prospective validation cohort (n = 225) for testing, with endoscopic retrograde cholangiopancreato- graphy (ERCP) serving as the reference standard. ANN performance was compared with that of expert radiologists and trainee radiologists with ANN assistance. Multivariate analysis evaluated clinical factors influencing diagnostic accuracy, and heatmap visualization assessed interpretability relevant to endoscopic decision-making. Results The ANN achieved diagnostic accuracy comparable to expert radiologists (93.3% vs. 93.8%). When assisting trainees, accuracy improved from 82.2% (AUC 0.82) to 91.1% (AUC 0.91), approaching expert performance (93.8%; AUC 0.94). Stone type and bile duct diameter > 10 mm significantly increased ANN detection rates. Heatmap visualization confirmed the plausibility of ANN predictions in both clearly identifiable lesions and indeterminate CT findings, improving interpretability for endoscopic decision-making. Conclusions The ANN achieved expert-level diagnostic accuracy for detecting CBD stones. By enhancing CT interpretation, it may optimize ERCP indications, reduce unnecessary invasive procedures, and improve training for less-experienced clinicians. Prospective validation and integration into multimodal endoscopic workflows are warranted.

Background/Objectives: Aortic dissection (AD) and aortic intramural hematoma (IMH) are fatal diseases with similar clinical characteristics. Immediate computed tomography (CT) with a contrast medium is required to confirm the presence of AD or IMH. This retrospective study aimed to use CT images to differentiate AD and IMH from normal aorta (NA) using a deep learning algorithm. Methods: A 6-year retrospective study of non-contrast chest CT images was conducted at a university hospital in Seoul, Republic of Korea, from January 2016 to July 2021. The position of the aorta was analyzed in each CT image and categorized as NA, AD, or IMH. The images were divided into training, validation, and test sets in an 8:1:1 ratio. A deep learning model that can differentiate between AD and IMH from NA using non-contrast CT images alone, called YOLO (You Only Look Once) v4, was developed. The YOLOv4 model was used to analyze 8881 non-contrast CT images from 121 patients. Results: The YOLOv4 model can distinguish AD, IMH, and NA from each other simultaneously with a probability of over 92% using non-contrast CT images. Conclusions: This model can help distinguish AD and IMH from NA when applying a contrast agent is challenging.

Malignant hilar biliary obstruction (MHBO) is a malignant tumor that causes biliary stenosis and bile drainage problems, which can lead to elevated liver enzymes and jaundice. Endoscopic biliary drainage performed to treat MHBO is essential to improve treatment, survival rate, and patients’ quality of life. However, the biliary drainage in MHBO is technically challenging because of its anatomical complexity, and the successful palliation of cholestasis is less frequent in MHBO than in distal biliary obstruction. In this article, we report a case of proximal stent migration of fully covered self-expandable metal stent following side-by-side deployment in the patient with Klatskin tumor. It is expected that it will be of great help if various preventive methods to reduce stent migration are developed in the future.

Jae Hee Ahn, Hwehoon Chung, Jae Keun Park. Journal of Digestive Cancer Research 2024;12:31-7. https://doi.org/10.52927/jdcr.2024.12.1.31

Jae Keun Park, Ji Woong Hwang. Journal of Digestive Cancer Research 2024;12:23-30. https://doi.org/10.52927/jdcr.2024.12.1.23

This study aimed to investigate the efficacy and safety of human gut-derived multi-strain probiotics in patients with irritable bowel syndrome (IBS). This was an open-label, prospective, observational study. Patients with IBS were administered human gut-derived multi-strain probiotics for 4 weeks. The primary and secondary outcomes were based on the overall responder rate of the total IBS severity scoring system (IBS-SSS) score (>50-point decrease) and the IBS quality of life (IBS-QOL) score and IBS-SSS₁ subscore (>10-point decrease in both scores), respectively. The estimated response rate is 55%. Of 44 patients, the total IBS-SSS score responder rate was 18.2% and 63.6% of patients at 2 and 4 weeks, respectively (P = .018). Compared with baseline, a significant improvement in the IBS-QOL score was observed in 27.3% and 63.6% of patients at 2 and 4 weeks, respectively (P = .001). Overall improvement rates in the IBS-SSS₁ subscore were observed in 29.5% and 61.4% of patients at 2 and 4 weeks, respectively (P < .001). Primary and secondary outcomes were higher at 4 weeks (total IBS-SSS score, 63.6%; IBS-QOL score, 63.6%; IBS-SSS₁ subscore, 61.4%) than the estimated responder rate (55%). Human gut-derived multi-strain probiotics have the potential to become an effective and safe treatment option for IBS patients.

Hwehoon Chung, Jae Keun Park. Journal of Digestive Cancer Research 2023;11:93-8. https://doi.org/10.52927/jdcr.2023.11.2.93

Background: Malignant change from low-grade dysplasia to high-grade dysplasia and invasive carcinoma following an adenoma-carcinoma sequence is becoming more common in intraductal papillary mucinous neoplasm (IPMN) of the pancreas. The aim of this study is to analyze their main characteristics and recent research trends in IPMNs and consequently create better understandings of the current situation and trends. Methods: A comprehensive search was performed in The Science Citation Index Expanded of the Web of Science. All articles between 1990 and 2021 were searched. VOS viewer (Leiden University, Leiden, Netherlands) was used for a qualitative and quantitative analysis of keywords, constituting maps based on co-occurrence matrix. Results: A total of 1658 eligible articles were screened among the 3950 identified articles for this subject. Finally, 879 articles were included in this study. Many articles on IPMN have been published in Japan and South Korea. Tanaka published the highest number of articles (n = 26, citations = 11,143). The Pancreas published the highest number of articles. (n = 100, citations = 2533). These articles were grouped into 4 clusters including basic research, disease overview, management/prognosis and malignant IPMN by using bibliometric keywords network analysis. Overlay visualization demonstrates, a trend of the studies has been changed from basic research or disease to management or prognosis. Conclusions: In this study, we found and highlight the most cited and influential articles related to IPMN. Plus, this study analyzed global research trends in IPMN over the past 30 years and provides insight into the features and research hotspots of the articles in IPMN research.

Background: In patients with BD-IPMN, surgical indications have been focused on finding malignant lesions (HGD, high-grade dysplasia/IC, invasive carcinoma). The aim of this study was to compare the preoperative factors that distinguish HGD from LGD (low-grade dysplasia) and HGD from IC to find the optimal pathologic target for surgery according to individuals, considering surgical risks and outcomes. Methods: We retrospectively analyzed 232 patients with BD-IPMN diagnosed based on pathology after surgery and preoperative images. The primary outcome was identifying preoperative factors distinguishing HGD from LGD, and HGD from IC. Results: In patients with LGD/HGD, a solid component or an enhancing mural nodule ≥ 5 mm (OR = 9.29; 95% CI: 3.3–54.12; p < 0.000) and thickened/enhancing cyst walls (OR = 6.95; 95% CI: 1.68–33.13; p = 0.008) were associated with HGD. In patients with malignant lesions (HGD/IC), increased serum CA 19-9 (OR = 12.59; 95% CI: 1.81–87.44; p = 0.006) was associated with IC. Conclusions: The predictive factors for HGD were the presence of a solid component or an enhancing mural nodule ≥ 5 mm and thickened/enhancing cyst walls compared with LGD, and if accompanied by increased CA 19-9, it might be necessary to urgently evaluate the lesion due to the possibility of progression to IC. Based on this finding, we need to find HGD as the optimal pathologic target for surgery to improve survival in low-surgical-risk patients, and IC could be assumed to be the optimal pathologic target for surgery in high-surgical-risk patients because of high morbidity and mortality associated with surgery.