Background/Aims: The factors affecting the detection rate of lymphoplasmacytic sclerosing pancreatitis (LPSP) using endoscopic ultrasound-guided tissue acquisition (EUS-TA) in patients with type 1 autoimmune pancreatitis (AIP) have not been thoroughly studied. Therefore, we conducted a retrospective study to identify the predictive factors for histologically detecting level 1 or 2 LPSP using EUS-TA. Methods: Fifty patients with AIP were included in this study, and the primary outcome measures were the predictive factors for histologically detecting level 1 or 2 LPSP using EUS-TA. Results: Multivariate analysis identified the use of fine needle biopsy (FNB) needles as a significant predictive factor for LPSP detection (odds ratio, 15.1; 95% confidence interval, 1.62–141; ¬¬p=0.017). The rate of good-quality specimens (specimen adequacy score ≥4) was significantly higher for the FNB needle group than for the fine needle aspiration (FNA) needle group (97% vs. 56%; p<0.01), and the FNB needle group required significantly fewer needle passes than the FNA needle group (median, 2 vs. 3; p<0.01). Conclusions The use of FNB needles was the most important factor for the histological confirmation of LPSP using EUS-TA in patients with type 1 AIP.

Background/Aims: The factors affecting the detection rate of lymphoplasmacytic sclerosing pancreatitis (LPSP) using endoscopic ultrasound-guided tissue acquisition (EUS-TA) in patients with type 1 autoimmune pancreatitis (AIP) have not been thoroughly studied. Therefore, we conducted a retrospective study to identify the predictive factors for histologically detecting level 1 or 2 LPSP using EUS-TA. Methods: Fifty patients with AIP were included in this study, and the primary outcome measures were the predictive factors for histologically detecting level 1 or 2 LPSP using EUS-TA. Results: Multivariate analysis identified the use of fine needle biopsy (FNB) needles as a significant predictive factor for LPSP detection (odds ratio, 15.1; 95% confidence interval, 1.62–141; ¬¬p=0.017). The rate of good-quality specimens (specimen adequacy score ≥4) was significantly higher for the FNB needle group than for the fine needle aspiration (FNA) needle group (97% vs. 56%; p<0.01), and the FNB needle group required significantly fewer needle passes than the FNA needle group (median, 2 vs. 3; p<0.01). Conclusions The use of FNB needles was the most important factor for the histological confirmation of LPSP using EUS-TA in patients with type 1 AIP.

Background/Aims: The factors affecting the detection rate of lymphoplasmacytic sclerosing pancreatitis (LPSP) using endoscopic ultrasound-guided tissue acquisition (EUS-TA) in patients with type 1 autoimmune pancreatitis (AIP) have not been thoroughly studied. Therefore, we conducted a retrospective study to identify the predictive factors for histologically detecting level 1 or 2 LPSP using EUS-TA. Methods: Fifty patients with AIP were included in this study, and the primary outcome measures were the predictive factors for histologically detecting level 1 or 2 LPSP using EUS-TA. Results: Multivariate analysis identified the use of fine needle biopsy (FNB) needles as a significant predictive factor for LPSP detection (odds ratio, 15.1; 95% confidence interval, 1.62–141; ¬¬p=0.017). The rate of good-quality specimens (specimen adequacy score ≥4) was significantly higher for the FNB needle group than for the fine needle aspiration (FNA) needle group (97% vs. 56%; p<0.01), and the FNB needle group required significantly fewer needle passes than the FNA needle group (median, 2 vs. 3; p<0.01). Conclusions The use of FNB needles was the most important factor for the histological confirmation of LPSP using EUS-TA in patients with type 1 AIP.

Objective: Magnetic resonance imaging (MRI) is efficient for the diagnosis of preoperative uterine sarcoma; however, misdiagnoses may occur. In this study, we developed a new artificial intelligence (AI) system to overcome the limitations of requiring specialists to manually process datasets and a large amount of computer resources. Methods: The AI system comprises a tumor image filter, which extracts MRI slices containing tumors, and sarcoma evaluator, which diagnoses uterine sarcomas. We used 15 types of MRI patient sequences to train deep neural network (DNN) models used by tumor filter and sarcoma evaluator with 8 cross-validation sets. We implemented tumor filter and sarcoma evaluator using ensemble prediction technique with 9 DNN models. Ten tumor filters and sarcoma evaluator sets were developed to evaluate fluctuation accuracy. Finally, AutoDiag-AI was used to evaluate the new validation dataset, including 8 cases of sarcomas and 24 leiomyomas. Results: Tumor image filter and sarcoma evaluator accuracies were 92.68% and 90.50%, respectively. AutoDiag-AI with the original dataset accuracy was 89.32%, with 90.47% sensitivity and 88.95% specificity, whereas AutoDiag-AI with the new validation dataset accuracy was 92.44%, with 92.25% sensitivity and 92.50% specificity. Conclusion Our newly established AI system automatically extracts tumor sites from MRI images and diagnoses them as uterine sarcomas without human intervention. Its accuracy is comparable to that of a radiologist. With further validation, the system could be applied for diagnosis of other diseases. Further improvement of the system's accuracy may enable its clinical application in the future.

Objective: Magnetic resonance imaging (MRI) is efficient for the diagnosis of preoperative uterine sarcoma; however, misdiagnoses may occur. In this study, we developed a new artificial intelligence (AI) system to overcome the limitations of requiring specialists to manually process datasets and a large amount of computer resources. Methods: The AI system comprises a tumor image filter, which extracts MRI slices containing tumors, and sarcoma evaluator, which diagnoses uterine sarcomas. We used 15 types of MRI patient sequences to train deep neural network (DNN) models used by tumor filter and sarcoma evaluator with 8 cross-validation sets. We implemented tumor filter and sarcoma evaluator using ensemble prediction technique with 9 DNN models. Ten tumor filters and sarcoma evaluator sets were developed to evaluate fluctuation accuracy. Finally, AutoDiag-AI was used to evaluate the new validation dataset, including 8 cases of sarcomas and 24 leiomyomas. Results: Tumor image filter and sarcoma evaluator accuracies were 92.68% and 90.50%, respectively. AutoDiag-AI with the original dataset accuracy was 89.32%, with 90.47% sensitivity and 88.95% specificity, whereas AutoDiag-AI with the new validation dataset accuracy was 92.44%, with 92.25% sensitivity and 92.50% specificity. Conclusion Our newly established AI system automatically extracts tumor sites from MRI images and diagnoses them as uterine sarcomas without human intervention. Its accuracy is comparable to that of a radiologist. With further validation, the system could be applied for diagnosis of other diseases. Further improvement of the system's accuracy may enable its clinical application in the future.

Objective: Magnetic resonance imaging (MRI) is efficient for the diagnosis of preoperative uterine sarcoma; however, misdiagnoses may occur. In this study, we developed a new artificial intelligence (AI) system to overcome the limitations of requiring specialists to manually process datasets and a large amount of computer resources. Methods: The AI system comprises a tumor image filter, which extracts MRI slices containing tumors, and sarcoma evaluator, which diagnoses uterine sarcomas. We used 15 types of MRI patient sequences to train deep neural network (DNN) models used by tumor filter and sarcoma evaluator with 8 cross-validation sets. We implemented tumor filter and sarcoma evaluator using ensemble prediction technique with 9 DNN models. Ten tumor filters and sarcoma evaluator sets were developed to evaluate fluctuation accuracy. Finally, AutoDiag-AI was used to evaluate the new validation dataset, including 8 cases of sarcomas and 24 leiomyomas. Results: Tumor image filter and sarcoma evaluator accuracies were 92.68% and 90.50%, respectively. AutoDiag-AI with the original dataset accuracy was 89.32%, with 90.47% sensitivity and 88.95% specificity, whereas AutoDiag-AI with the new validation dataset accuracy was 92.44%, with 92.25% sensitivity and 92.50% specificity. Conclusion Our newly established AI system automatically extracts tumor sites from MRI images and diagnoses them as uterine sarcomas without human intervention. Its accuracy is comparable to that of a radiologist. With further validation, the system could be applied for diagnosis of other diseases. Further improvement of the system's accuracy may enable its clinical application in the future.

Background/Aims: We aimed to investigate (1) promising clinical findings for the recognition of focal type autoimmune pancreatitis (FAIP) and (2) the impact of endoscopic ultrasound (EUS)-guided tissue acquisition (EUS-TA) on the diagnosis of FAIP. Methods: Twenty-three patients with FAIP were involved in this study, and 44 patients with resected pancreatic ductal adenocarcinoma (PDAC) were included in the control group. Results: (1) Multivariate analysis revealed that homogeneous delayed enhancement on contrast-enhanced computed tomography was a significant factor indicative of FAIP compared to PDAC (90% vs. 7%, p=0.015). (2) For 13 of 17 FAIP patients (76.5%) who underwent EUS-TA, EUS-TA aided the diagnostic confirmation of AIPs, and only one patient (5.9%) was found to have AIP after surgery. On the other hand, of the six patients who did not undergo EUS-TA, three (50.0%) underwent surgery for pancreatic lesions. Conclusions Homogeneous delayed enhancement on contrast-enhanced computed tomography was the most useful clinical factor for discriminating FAIPs from PDACs. EUS-TA is mandatory for diagnostic confirmation of FAIP lesions and can contribute to a reduction in the rate of unnecessary surgery for patients with FAIP.

Background/Aims: This study aimed to clarify the efficacy and safety of pancreatic duct lavage cytology combined with a cell-block method (PLC-CB) for possible pancreatic ductal adenocarcinomas (PDACs). Methods: This study included 41 patients with suspected PDACs who underwent PLC-CB mainly because they were unfit for undergoing endoscopic ultrasonography-guided fine needle aspiration. A 6-Fr double lumen catheter was mainly used to perform PLC-CB. Final diagnoses were obtained from the findings of resected specimens or clinical outcomes during surveillance after PLC-CB. Results: Histocytological evaluations using PLC-CB were performed in 87.8% (36/41) of the patients. For 31 of the 36 patients, final diagnoses (invasive PDAC, 12; pancreatic carcinoma in situ, 5; benignancy, 14) were made, and the remaining five patients were excluded due to lack of surveillance periods after PLC-CB. For 31 patients, the sensitivity, specificity, and accuracy of PLC-CB for detecting malignancy were 94.1%, 100%, and 96.8%, respectively. In addition, they were 87.5%, 100%, and 94.1%, respectively, in 17 patients without pancreatic masses detectable using endoscopic ultrasonography. Four patients developed postprocedural pancreatitis, which improved with conservative therapy. Conclusions PLC-CB has an excellent ability to detect malignancies in patients with possible PDACs, including pancreatic carcinoma in situ.

Background/Aims: The use of endoscopic intervention (EI) for acute biliary pancreatitis (ABP) remains controversial because the severity of biliary obstruction/cholangitis/pancreatitis is not reflected in the indications for early EI (EEI). Methods: A total of 148 patients with ABP were included to investigate 1) the differences in the rate of worsening cholangitis/pancreatitis between the EEI group and the early conservative management (ECM) group, especially for each severity of cholangitis/pancreatitis, and 2) the diagnostic ability of imaging studies, including endoscopic ultrasound (EUS), to detect common bile duct stones (CBDSs) in the ECM group. Results: No differences were observed in the rate of worsening cholangitis between the EEI and ECM groups, regardless of the severity of cholangitis and/or the existence of impacted CBDSs. Among patients without impacted CBDSs and moderate/severe cholangitis, worsening pancreatitis was significantly more frequent in the EEI group (18% vs. 4%, p=0.048). In patients in the ECM group, the sensitivity and specificity for detecting CBDSs were 73% and 98%, respectively, for EUS, whereas the values were 13% and 92%, respectively, for magnetic resonance cholangiopancreatography. Conclusions EEI should be avoided in the absence of moderate/severe cholangitis and/or impacted CBDSs because of the high rate of worsening pancreatitis. EUS can contribute to the accurate detection of residual CBDSs, for the determination of the need for elective EI.

Background/Aims: The aim of this study was to evaluate outcomes of inside plastic stents (iPSs) versus those of metal stents (MSs) for treating unresectable perihilar malignant obstructions. Methods: For all patients who underwent endoscopic suprapapillary placement of iPS(s) or MS(s) as the first permanent biliary drainage for unresectable malignant perihilar obstructions between January 2014 and August 2019, clinical outcomes using iPSs (n=20) and MSs (n=85), including clinical efficacy, adverse events, and time to recurrence of biliary obstruction (RBO), were retrospectively evaluated. Results: There were no differences in clinical effectiveness (95% for the iPS group vs. 92% for the MS group, p=1.00). Procedure-related adverse events, including pancreatitis, acute cholangitis, acute cholecystitis, and death, were observed for 8% of the MS group, although no patient in the iPS group developed such adverse events. The median time to RBO was 561 days (95% confidence interval, 0–1,186 days) for iPSs and 209 days (127–291 days) for MSs, showing a significant difference (p=0.008). Conclusions Time to RBO after iPS placement was significantly longer than that after MS placement. IPSs, which are removable, unlike MSs, were an acceptable option.