Multicenter study of the efficacy and safety of electrocautery-enhanced lumen-apposing metal stents for the internal drainage of pancreatic fluid collections.

Pancreatic fluid collections (PFCs) including pancreatic pseudocyst (PP) and walled-off necrosis (WON) are complications after acute pancreatitis. We aimed to evaluate the efficacy and safety of endoscopic ultrasound (EUS)-guided lumen-apposing metal stent (LAMS) placement to manage PFCs. Between June 2019 and May 2023, patients with symptomatic PFCs who underwent EUS-guided electrocautery-enhanced LAMS drainage were enrolled retrospectively from eight tertiary centers in Taiwan. In total, 33 [14 (42.42%) PP and 19 (57.58%) WON] patients were enrolled. Gallstones (27.27%) and abdominal pain (72.73%) were the most common etiology and indication for drainage. The technical and clinical success rates were 100% and 96.97%, respectively, and the mean procedure time was 30.55 (± 16.17) min. Complications included one (3.03%) case of self-limited bleeding; there were no cases of mortality. Seven (21.21%) patients had recurrence. Patients with disconnected pancreatic duct syndrome (DPDS) had a higher recurrence rate than those without (71.43% vs. 38.46%, p = 0.05). After replacing LAMSs with transmural double-pigtail plastic stents (DPSs) in the DPDS patients, the DPS migration rate was higher in the patients with recurrence (100% vs. 33.33%, p = 0.04). In conclusion, drainage of symptomatic PFCs with EUS-guided electrocautery-enhanced LAMS appears to be efficient and safe. Replacing LAMSs with DPSs in DPDS patients was associated with a lower recurrence rate.

Diagnostic Accuracy for Acute Rib Fractures: A Cross-sectional Study Utilizing Automatic Rib Unfolding and 3D Volume-Rendered Reformation.

RATIONALE AND OBJECTIVES: The aim of this study was to compare the use of computed tomography (CT) with automatic rib unfolding and three-dimensional (3D) volume-rendered imaging in the detection and characterization of rib fractures and flail chest. MATERIALS AND METHODS: A total of 130 patients with blunt chest trauma underwent whole-body CT, and five independent readers assessed the presence and characterization of rib fractures using traditional CT images, automatic rib unfolding, and 3D volume-rendered images in separate readout sessions at least 2 weeks apart. A gold standard was established by consensus among the readers based on the combined analysis of conventional and reformatted images. RESULTS: Automatic rib unfolding significantly reduced mean reading time by 47.5%-74.9% (P < 0.0001) while maintaining a comparable diagnostic performance for rib fractures (positive predictive value [PPV] of 82.1%-93.5%, negative predictive value [NPV] of 96.8%-98.2%, and 69.4%-94.2% and 96.9%-99.1% for conventional axial images and 70.4%-85.1% and 95.2%-96.6% for 3D images) and better interobserver agreement (kappa of 0.74-0.87). For flail chest, automatic rib unfolding showed a PPV of 85.7%-100%, NPV of 90.4%-99.0%, and 80.0%-100% and 89.7%-100% for conventional axial images and 76.9%-100% and 89.0%-92.1% for 3D images. CONCLUSION: Automatic rib unfolding demonstrated equivalent diagnostic performance to conventional images in detecting acute rib fractures and flail chest, with good interobserver agreement and time-saving benefits.

Precise mapping of hilar cholangiocarcinoma with a skip lesion by SpyGlass cholangioscopy: A case report.

BACKGROUND: Cholangiocarcinoma (CC) is a very aggressive cancer with a poor prognosis. As surgery is the only curative therapy, preoperative evaluation of the tumor extent is essential for surgical planning. Although high-quality image modalities such as computed tomography and magnetic resonance imaging have been used extensively in preoperative evaluation, the accuracy is low. To obtain precise localization of tumor spread arising from the hilar region preoperatively, the development of an acceptable imaging modality is still an unmet need. CASE SUMMARY: A 52-year-old female presented to our emergency department with jaundice, abdominal pain, and fever. Initially, she was treated for cholangitis. Endoscopic retrograde cholangiopancreatography with the cholangiogram showed long segment filling defect in the common hepatic duct with dilatation of bilateral intrahepatic ducts. Transpapillary biopsy was performed, and the pathology suggested intraductal papillary neoplasm with high-grade dysplasia. After treatment of cholangitis, contrasted-enhanced computed tomography revealed a hilar lesion with undetermined Bismuth-Corlette classification. SpyGlass cholangioscopy showed that the lesion involved the confluence of the common hepatic duct with one skip lesion in the posterior branch of the right intrahepatic duct, which was not detected by previous image modalities. The surgical plan was modified from extended left hepatectomy to extended right hepatectomy. The final diagnosis was hilar CC, pT2aN0M0. The patient has remained disease-free for more than 3 years. CONCLUSION: SpyGlass cholangioscopy may have a role in precision localization of hilar CC to provide surgeons with more information before the operation.

The Role of Neutrophil-to-Lymphocyte Ratio in Advanced EGFR-Mutant NSCLC Patients Treated with First-Line Osimertinib.

Purpose: Although serum neutrophil-to-lymphocyte ratio (NLR) is correlated with the outcome of various cancer types, its role in treatment-naïve, advanced, epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC) patients treated with osimertinib remains uncertain. We have the intention to use this biomarker to evaluate the outcomes in NSCLC. Patients and Methods: Advanced EGFR-mutant NSCLC patients receiving osimertinib as the first-line treatment were included. We evaluated the prognostic role of baseline NLR and explored its association with patients' characteristics. A high NLR was defined as pretreatment serum NLR ≥ 5. Results: A total of 112 eligible patients were included. The objective response rate was 83.7%. The median progression-free survival (PFS) and overall survival (OS) were 20.5 months (95% CI 14.5-26.5) and 47.3 months (95% CI 36.7-58.2), respectively. A high NLR predicted an inferior PFS (HR 1.90 [95% CI 1.02-3.51], P = 0.042) and OS (HR 3.85 [95% CI 1.39-10.66], P = 0.009). Patients with stage IVB disease were more likely to have a high baseline NLR than those with stage IIIB-IVA (33.9% vs 15.1%, P = 0.029). Other patients' characteristics did not correlate with the baseline NLR significantly. Patients with a high NLR had significantly more metastatic organs than those with a low NLR (2.5 ± 1.3 vs 1.8 ± 0.9, P = 0.012), particularly brain, liver, and bone metastasis. There was no significant association between NLR and intrathoracic metastasis. Conclusion: Baseline serum NLR could act as an important prognostic marker for EGFR-mutant NSCLC patients receiving first-line osimertinib. A high NLR was associated with higher metastatic burden, more extrathoracic metastases, and therefore, a worse outcome.

Endoscopic subserosal dissection for a GI stromal tumor.

Video 1Endoscopic subserosal dissection was carried out to remove the subepithelial tumors.

Training a Deep Contextualized Language Model for International Classification of Diseases, 10th Revision Classification via Federated Learning: Model Development and Validation Study.

BACKGROUND: The automatic coding of clinical text documents by using the International Classification of Diseases, 10th Revision (ICD-10) can be performed for statistical analyses and reimbursements. With the development of natural language processing models, new transformer architectures with attention mechanisms have outperformed previous models. Although multicenter training may increase a model's performance and external validity, the privacy of clinical documents should be protected. We used federated learning to train a model with multicenter data, without sharing data per se. OBJECTIVE: This study aims to train a classification model via federated learning for ICD-10 multilabel classification. METHODS: Text data from discharge notes in electronic medical records were collected from the following three medical centers: Far Eastern Memorial Hospital, National Taiwan University Hospital, and Taipei Veterans General Hospital. After comparing the performance of different variants of bidirectional encoder representations from transformers (BERT), PubMedBERT was chosen for the word embeddings. With regard to preprocessing, the nonalphanumeric characters were retained because the model's performance decreased after the removal of these characters. To explain the outputs of our model, we added a label attention mechanism to the model architecture. The model was trained with data from each of the three hospitals separately and via federated learning. The models trained via federated learning and the models trained with local data were compared on a testing set that was composed of data from the three hospitals. The micro F1 score was used to evaluate model performance across all 3 centers. RESULTS: The F1 scores of PubMedBERT, RoBERTa (Robustly Optimized BERT Pretraining Approach), ClinicalBERT, and BioBERT (BERT for Biomedical Text Mining) were 0.735, 0.692, 0.711, and 0.721, respectively. The F1 score of the model that retained nonalphanumeric characters was 0.8120, whereas the F1 score after removing these characters was 0.7875-a decrease of 0.0245 (3.11%). The F1 scores on the testing set were 0.6142, 0.4472, 0.5353, and 0.2522 for the federated learning, Far Eastern Memorial Hospital, National Taiwan University Hospital, and Taipei Veterans General Hospital models, respectively. The explainable predictions were displayed with highlighted input words via the label attention architecture. CONCLUSIONS: Federated learning was used to train the ICD-10 classification model on multicenter clinical text while protecting data privacy. The model's performance was better than that of models that were trained locally.

Pancreatic Heterotopia in the Small Bowel: A Case Report and Literature Review.

ABSTRACT: Diagnosing pancreatic heterotopia (PH) is challenging, because it is a rare congenital anomaly. We report the case of a 66-year-old man who presented with ileus and jejunal PH diagnosed using single-balloon enteroscopy. Reviewing 178 cases of small intestinal PH in the literature showed that (1) PH mostly occurred at the duodenum (61.8%), followed by the jejunum (22.5%) and ileum (14.6%); (2) patients with ileal PH were younger (mean, 40.7 vs 54.6 years; P < 0.001), while those with jejunal PH were older (mean, 59.5 vs 51.7 years; P = 0.006); (3) most patients with jejunal PH were females, while those with duodenal or ileal PH were predominantly males (M:F = 0.8:1 vs 2.5:1; P = 0.003); (4) asymptomatic cases were rare (15.2% vs 83.4% in small and nonsmall bowel, respectively); (5) most patients with ileal PH presented with bleeding (61.5%), while the remaining had abdominal pain (42.8%); (6) emergent operation was performed in 11 cases (6.2%), mainly at the jejunum (12.5%) and Meckel's diverticulum (25%); and (7) Heinrich's classification may be inconclusive. Enteroscopy was performed in only 6 cases (3.4%), and earlier diagnosis and further minimally invasive procedures could have been achieved with enteroscopy.

Automatic International Classification of Diseases Coding System: Deep Contextualized Language Model With Rule-Based Approaches.

BACKGROUND: The tenth revision of the International Classification of Diseases (ICD-10) is widely used for epidemiological research and health management. The clinical modification (CM) and procedure coding system (PCS) of ICD-10 were developed to describe more clinical details with increasing diagnosis and procedure codes and applied in disease-related groups for reimbursement. The expansion of codes made the coding time-consuming and less accurate. The state-of-the-art model using deep contextual word embeddings was used for automatic multilabel text classification of ICD-10. In addition to input discharge diagnoses (DD), the performance can be improved by appropriate preprocessing methods for the text from other document types, such as medical history, comorbidity and complication, surgical method, and special examination. OBJECTIVE: This study aims to establish a contextual language model with rule-based preprocessing methods to develop the model for ICD-10 multilabel classification. METHODS: We retrieved electronic health records from a medical center. We first compared different word embedding methods. Second, we compared the preprocessing methods using the best-performing embeddings. We compared biomedical bidirectional encoder representations from transformers (BioBERT), clinical generalized autoregressive pretraining for language understanding (Clinical XLNet), label tree-based attention-aware deep model for high-performance extreme multilabel text classification (AttentionXLM), and word-to-vector (Word2Vec) to predict ICD-10-CM. To compare different preprocessing methods for ICD-10-CM, we included DD, medical history, and comorbidity and complication as inputs. We compared the performance of ICD-10-CM prediction using different preprocesses, including definition training, external cause code removal, number conversion, and combination code filtering. For the ICD-10 PCS, the model was trained using different combinations of DD, surgical method, and key words of special examination. The micro F1 score and the micro area under the receiver operating characteristic curve were used to compare the model's performance with that of different preprocessing methods. RESULTS: BioBERT had an F1 score of 0.701 and outperformed other models such as Clinical XLNet, AttentionXLM, and Word2Vec. For the ICD-10-CM, the model had an F1 score that significantly increased from 0.749 (95% CI 0.744-0.753) to 0.769 (95% CI 0.764-0.773) with the ICD-10 definition training, external cause code removal, number conversion, and combination code filter. For the ICD-10-PCS, the model had an F1 score that significantly increased from 0.670 (95% CI 0.663-0.678) to 0.726 (95% CI 0.719-0.732) with a combination of discharge diagnoses, surgical methods, and key words of special examination. With our preprocessing methods, the model had the highest area under the receiver operating characteristic curve of 0.853 (95% CI 0.849-0.855) and 0.831 (95% CI 0.827-0.834) for ICD-10-CM and ICD-10-PCS, respectively. CONCLUSIONS: The performance of our model with the pretrained contextualized language model and rule-based preprocessing method is better than that of the state-of-the-art model for ICD-10-CM or ICD-10-PCS. This study highlights the importance of rule-based preprocessing methods based on coder coding rules.

Predicting Postoperative Mortality With Deep Neural Networks and Natural Language Processing: Model Development and Validation.

BACKGROUND: Machine learning (ML) achieves better predictions of postoperative mortality than previous prediction tools. Free-text descriptions of the preoperative diagnosis and the planned procedure are available preoperatively. Because reading these descriptions helps anesthesiologists evaluate the risk of the surgery, we hypothesized that deep learning (DL) models with unstructured text could improve postoperative mortality prediction. However, it is challenging to extract meaningful concept embeddings from this unstructured clinical text. OBJECTIVE: This study aims to develop a fusion DL model containing structured and unstructured features to predict the in-hospital 30-day postoperative mortality before surgery. ML models for predicting postoperative mortality using preoperative data with or without free clinical text were assessed. METHODS: We retrospectively collected preoperative anesthesia assessments, surgical information, and discharge summaries of patients undergoing general and neuraxial anesthesia from electronic health records (EHRs) from 2016 to 2020. We first compared the deep neural network (DNN) with other models using the same input features to demonstrate effectiveness. Then, we combined the DNN model with bidirectional encoder representations from transformers (BERT) to extract information from clinical texts. The effects of adding text information on the model performance were compared using the area under the receiver operating characteristic curve (AUROC) and the area under the precision-recall curve (AUPRC). Statistical significance was evaluated using P<.05. RESULTS: The final cohort contained 121,313 patients who underwent surgeries. A total of 1562 (1.29%) patients died within 30 days of surgery. Our BERT-DNN model achieved the highest AUROC (0.964, 95% CI 0.961-0.967) and AUPRC (0.336, 95% CI 0.276-0.402). The AUROC of the BERT-DNN was significantly higher compared to logistic regression (AUROC=0.952, 95% CI 0.949-0.955) and the American Society of Anesthesiologist Physical Status (ASAPS AUROC=0.892, 95% CI 0.887-0.896) but not significantly higher compared to the DNN (AUROC=0.959, 95% CI 0.956-0.962) and the random forest (AUROC=0.961, 95% CI 0.958-0.964). The AUPRC of the BERT-DNN was significantly higher compared to the DNN (AUPRC=0.319, 95% CI 0.260-0.384), the random forest (AUPRC=0.296, 95% CI 0.239-0.360), logistic regression (AUPRC=0.276, 95% CI 0.220-0.339), and the ASAPS (AUPRC=0.149, 95% CI 0.107-0.203). CONCLUSIONS: Our BERT-DNN model has an AUPRC significantly higher compared to previously proposed models using no text and an AUROC significantly higher compared to logistic regression and the ASAPS. This technique helps identify patients with higher risk from the surgical description text in EHRs.

Randomized controlled trial of early endoscopy for upper gastrointestinal bleeding in acute coronary syndrome patients.

Acute upper gastrointestinal bleeding (UGIB) in acute coronary syndrome (ACS) patients are not uncommon, particularly under dual antiplatelet therapy (DAPT). The efficiency and safety of early endoscopy (EE) for UGIB in these patients needs to be elucidated. This multicenter randomized controlled trial randomized recent ACS patients presenting acute UGIB to non-EE and EE groups. All eligible patients received intravenous proton pump inhibitor therapy. Those in EE group underwent therapeutic endoscopy within 24 h after bleeding. The data regarding efficacy and safety of EE were analyzed. It was early terminated because the UGIB rate was lower than expected and interim analysis was done. In total, 43 patients were randomized to non-EE (21 patients) and EE (22 patients) groups. The failure rate of control hemorrhage (intention-to-treat [ITT] 4.55% vs. 23.81%, p < 0.001; per-protocol [PP] 0% vs. 4.55%, p = 0.058) and 3-day rebleeding rate (ITT 4.55% vs. 28.57%, p = 0.033; PP 0% vs. 21.05%, p = 0.027) were lower in EE than non-EE group. The mortality, minor and major complication rates were not different between two groups. Male patients were at higher risk of minor and major complications after EE with OR (95% CI) of 3.50 (1.15-10.63) and 4.25 (1.43-12.63), respectively. In multivariate analysis, EE was associated with lower needs for blood transfusion (HR 0.13, 95% CI 0.02-0.98). Among patients who discontinued DAPT during acute UGIB, a higher risk (OR 5.25, 95% CI 1.21-22.74) of coronary artery stent re-thrombosis within 6 months was noticed. EE for acute UGIB in recent ACS patients has higher rate of bleeding control, lower 3-day rebleeding rate and lower needs for blood transfusion, but more complications in male patients. Further enrollment is mandatory to avoid bias from small sample size (ClinicalTrial.gov Number NCT02618980, registration date 02/12/2015).

Efficacy and safety of gastric peroral endoscopic myotomy (G-POEM) for refractory gastroparesis: 3-year follow up results.

BACKGROUND/PURPOSE: Gastroparesis is a common but easily overlooked disease. Gastric peroral endoscopic myotomy (G-POEM) is one of the third-space endoscopy techniques to treat gastroparesis. In this study, we aimed to evaluate the efficacy and safety of G-POEM for patients with refractory gastroparesis. METHODS: Between December 2017 and 2020, we consecutively enrolled patients with gastroparesis who failed after the administration of several kinds of medication and repeated admission for nutritional support. All patients underwent gastric emptying scintigraphy and answered a questionnaire on Gastroparesis Cardinal Symptom Index (GCSI). Demographic data, endoscopic procedure, and post procedural outcome were analyzed. RESULTS: A total of 11 (9 women and 2 men) patients with refractory gastroparesis (nine with diabetes mellitus, one systemic lupus erythematosus, and one idiopathic) were enrolled. The mean (±standard deviation (SD)) procedure time was 61.82 (±18.99) min with technical and clinical success rates of 100% and 81.82%, respectively. A statistically significant improvement was observed in the clinical severity (mean GCSI score 36.00 vs. 14.73, p < 0.0001) and gastric emptying time (mean T1/2 341.92 vs. 65.92 min, p = 0.016) after G-POEM. Hospital stay was 7.18 (±4.49) days without mortality. Complications included 4 (36.36%) patients with self-limited postprocedural abdominal pain and 3 (27.27%) patients with intra-procedural pneumoperitoneum. During the mean follow-up period of 554.36 days, one (9.09%) patient had relapsed clinical symptoms after 6 months. CONCLUSION: G-POEM is an efficient and safe pylorus-directed endoscopic therapy for refractory gastroparesis with promising results.

Case 296: Phlebosclerotic Colitis.

History A 55-year-old woman without systemic underlying disease, such as diabetes mellitus, inflammatory bowel disease, autoimmune disease, or chronic kidney disease, presented with generalized dull abdominal pain of 1-week duration. She had ingested herbal medicine for physical conditioning for several years. Laboratory findings, including biochemistry, electrolyte levels, and complete blood count, were all within normal limits, except for elevated serum C-reactive protein level (7.719 mg/dL; normal range, <1 mg/dL). The patient underwent initial evaluation with conventional abdominal radiography. She underwent subsequent evaluation with noncontrast CT of the abdomen and colonoscopy.

The effect of the head-up position on cardiopulmonary resuscitation: a systematic review and meta-analysis.

OBJECTIVE: Experimental studies of head-up positioning (HUP) during cardiopulmonary resuscitation (CPR) have had some degree of conflicting published results. The current study aim was to analyze and reconcile those discrepancies in order to better clarify the effects of HUP CPR compared to conventional supine (SUP) CPR. METHODS: Three databases (PubMed, EMBASE and Cochrane Library) were searched comprehensively (from each respective database's inception to May 2021) for articles addressing HUP CPR. The primary outcome to be observed was cerebral perfusion pressure (CerPP), and secondary outcomes were mean intracranial pressure (ICP), mean arterial pressure (MAP), coronary perfusion pressure (CoPP) and frequencies of return of spontaneous circulation (ROSC). RESULTS: Seven key studies involving 131 animals were included for analysis. Compared to SUP CPR, CerPP (MD 10.37; 95% CI 7.11-13.64; p < 0.01; I2 = 58%) and CoPP (MD 7.56; 95% CI 1.84-13.27, p = 0.01; I2 = 75%) increased significantly with HUP CPR, while ICP (MD - 13.66; 95% CI - 18.6 to -8.71; p < 0.01; I2 = 96%) decreased significantly. Combining all study methodologies, there were no significant differences detected in MAP (MD - 1.63; 95% CI - 10.77-7.52; p = 0.73; I2 = 93%) or frequency of ROSC (RR 0.9; 95% CI 0.31-2.60; p = 0.84; I2 = 65%). However, in contrast to worse outcomes in studies using immediate elevation of the head in a reverse Trendelenburg position, study outcomes were significantly improved when HUP (head and chest only) was introduced in a steady, graduated manner following a brief period of basic CPR augmented by active compression-decompression (ACD) and impedance threshold (ITD) devices. CONCLUSION: In experimental models, gradually elevating the head and chest following a brief interval of circulatory priming with ACD and ITD devices can enhance CoPP, lower ICP and improve CerPP significantly while maintaining MAP. This effect is immediate, remains sustained and is associated with improved outcomes.

Automatic ICD-10 Coding and Training System: Deep Neural Network Based on Supervised Learning.

BACKGROUND: The International Classification of Diseases (ICD) code is widely used as the reference in medical system and billing purposes. However, classifying diseases into ICD codes still mainly relies on humans reading a large amount of written material as the basis for coding. Coding is both laborious and time-consuming. Since the conversion of ICD-9 to ICD-10, the coding task became much more complicated, and deep learning- and natural language processing-related approaches have been studied to assist disease coders. OBJECTIVE: This paper aims at constructing a deep learning model for ICD-10 coding, where the model is meant to automatically determine the corresponding diagnosis and procedure codes based solely on free-text medical notes to improve accuracy and reduce human effort. METHODS: We used diagnosis records of the National Taiwan University Hospital as resources and apply natural language processing techniques, including global vectors, word to vectors, embeddings from language models, bidirectional encoder representations from transformers, and single head attention recurrent neural network, on the deep neural network architecture to implement ICD-10 auto-coding. Besides, we introduced the attention mechanism into the classification model to extract the keywords from diagnoses and visualize the coding reference for training freshmen in ICD-10. Sixty discharge notes were randomly selected to examine the change in the F1-score and the coding time by coders before and after using our model. RESULTS: In experiments on the medical data set of National Taiwan University Hospital, our prediction results revealed F1-scores of 0.715 and 0.618 for the ICD-10 Clinical Modification code and Procedure Coding System code, respectively, with a bidirectional encoder representations from transformers embedding approach in the Gated Recurrent Unit classification model. The well-trained models were applied on the ICD-10 web service for coding and training to ICD-10 users. With this service, coders can code with the F1-score significantly increased from a median of 0.832 to 0.922 (P<.05), but not in a reduced interval. CONCLUSIONS: The proposed model significantly improved the F1-score but did not decrease the time consumed in coding by disease coders.

Case 296.

History A 55-year-old woman without systemic underlying disease, such as diabetes mellitus, inflammatory bowel disease, autoimmune disease, or chronic kidney disease, presented with generalized dull abdominal pain of 1-week duration. She had ingested herbal medicine for physical conditioning for several years. Laboratory findings, including biochemistry, electrolyte levels, and complete blood count, were all within normal limits, except for elevated serum C-reactive protein level (7.719 mg/dL; normal range,<1 mg/dL). The patient underwent initial evaluation with conventional abdominal radiography (Fig 1). She underwent subsequent evaluation with noncontrast CT of the abdomen (Figs 2, 3) and colonoscopy (Fig 4).

Peroral endoscopic tumor resection (POET) with preserved mucosa technique for management of upper gastrointestinal tract subepithelial tumors.

BACKGROUND: Third space endoscopy technique facilitates therapeutic endoscopy in subepithelial space. This study aimed to investigate peroral endoscopic tumor resection (POET) with preserved mucosa technique for upper gastrointestinal tract subepithelial tumors (UGI-SETs) removal. METHODS: Between February 2011 and December 2019, consecutive patients with SETs of esophagus and stomach who underwent POET for enlarging size during follow-up, malignant endoscopic ultrasound features or by patient's request were enrolled. Demographic, endoscopic and pathological data were analyzed retrospectively. RESULTS: Totally 18 esophageal (mean ± SD age, 55.23 ± 4.15 year-old, 38.89% female) and 30 gastric (52.65 ± 2.43 year-old, 53.33% female) SETs in 47 patients (one with both esophageal and gastric lesions) were resected. The mean (± SD) endoscopic/pathological tumor size, procedure time, en-bloc/complete resection rate, and hospital stays of esophageal and gastric SET patients were 12.36 (± 7.89)/11.86 (± 5.67) and 12.57 (± 6.25)/12.35 (± 5.73) mm, 14.86 (± 6.15) and 38.21 (± 15.29) minutes, 88.89%/94.44% and 86.77%/93.30%, and 4.14 (± 0.21) and 4.17 (± 0.20) days, respectively. The overall complication rate was 18.75%, including 6 self-limited fever and 3 pneumoperitoneum relieved by needle puncture. There was no mortality or recurrence reported with mean follow-up period of 23.74 (± 4.12) months. CONCLUSIONS: POET is a safe and efficient third space endoscopic resection technique for removal of UGI-SETs less than 20 mm. Long term data are warranted to validate these results.

Endoscopic ultrasound ablation in a patient with multiple metastatic pancreatic tumors from adrenocorticotropic hormone-producing thymic neuroendocrine neoplasm.

Adrenocorticotropic hormone (ACTH)-producing neuroendocrine neoplasm (NEN) of the thymus is rare. Lymph nodes and bones are the most common metastatic sites. Most cases present with florid Cushing's syndrome (CS). Here, we reported a 58-year-old woman, who presented with intermittent flush and weight loss. Imaging studies revealed tumors in the mediastinum, pancreas, and bones. Contrast-enhanced harmonic endoscopic ultrasound (EUS) of the pancreatic tumors showed heterogeneous and hyperenhancing characteristics. EUS elastography revealed a heterogeneous stiff pattern. EUS-fine needle biopsy to the pancreatic lesion confirmed the NEN nature. Serum ACTH and cortisol levels were abnormally high. Immunohistochemical staining of the thymic and pancreatic specimens was positive for ACTH. However, the patient did not have obvious CS appearance. The patient underwent surgery, radiation, EUS-guided ethanol injection, and anti-cancer medications, but the disease still progressed. The patient died from infection 16 months after NEN was diagnosed. In conclusion, the pancreas can be a metastatic site for ACTH-producing thymic NEN. EUS-associated procedures can help in the diagnosis and treatment of pancreatic metastatic NEN.

Surveillance of Rejection After Intestinal Transplantation Using an Image Enhanced Endoscopy "VENCH" Scoring System.

BACKGROUND: Graft rejection after intestinal transplantation remains challenging. We aimed to use endoscopy for rejection prediction. MATERIALS AND METHODS: Patients ≥7 years old who underwent intestinal transplantation between November 2016 and September 2019 were prospectively enrolled. Magnifying endoscopy under narrow-band imaging was performed through ileostomy. Endoscopic findings were reported as five components (each graded from 0-2): "V" (villi appearance), "E" (erythema), "N" (capillary network), "C" (crypt widening), and "H" (heterogeneity). The correlation between histological severity and endoscopic score was analyzed. RESULTS: Ninety-nine endoscopic biopsies from three female and one male patient were analyzed. The mean ± SD age was of 41.25±13.77 (range 29-58) years. Three short bowel syndrome patients after multiple intestinal resections and one with chronic intestinal pseudo-obstruction were indicated for intestinal transplantation. Sensitivity, specificity, and accuracy of V, E, N, C, and H scores for predicting rejection were 97.4%, 45.9%, 65.7%; 94.7%, 70.5%, 79.8%; 97.4%, 52.5%, 69.7%; 94.7%, 54.1%, 69.7%; and 97.4%, 62.3%, 75.8%, respectively. Pearson's correlation coefficients between total and individual V, E, N, C, H scores and histological rejection were 0.79, 0.64, 0.70, 0.71, 0.73, and 0.66, respectively (P < .001). To predict mild and moderate/severe rejection, total scores more than 4 and 6 had the sensitivity/specificity of 87.50%/57.38% and 96.67%/85.25%, respectively (area under the ROC 0.791 and 0.987). CONCLUSION: Endoscopic VENCH scoring is promising for predicting rejection after IT. More studies are warranted to validate such results. (ClinicalTrials.gov number, NCT03616548.).