RESUMO
BACKGROUND AND AIMS: Accurately diagnosing malignant biliary strictures (MBSs) as benign or malignant remains challenging. It has been suggested that direct visualization and interpretation of cholangioscopy images provide greater accuracy for stricture classification than current sampling techniques (ie, brush cytology and forceps biopsy sampling) using ERCP. We aimed to develop a convolutional neural network (CNN) model capable of accurate stricture classification and real-time evaluation based solely on cholangioscopy image analysis. METHODS: Consecutive patients with cholangioscopy examinations from 2012 to 2021 were reviewed. A CNN was developed and tested using cholangioscopy images with direct expert annotations. The CNN was then applied to a multicenter, reserved test set of cholangioscopy videos. CNN performance was then directly compared with that of ERCP sampling techniques. Occlusion block heatmap analyses were used to evaluate and rank cholangioscopy features associated with MBSs. RESULTS: One hundred fifty-four patients with available cholangioscopy examinations were included in the study. The final image database comprised 2,388,439 still images. The CNN demonstrated good performance when tasked with mimicking expert annotations of high-quality malignant images (area under the receiver-operating characteristic curve, .941). Overall accuracy of CNN-based video analysis (.906) was significantly greater than that of brush cytology (.625, P = .04) or forceps biopsy sampling (.609, P = .03). Occlusion block heatmap analysis demonstrated that the most frequent image feature for an MBS was the presence of frond-like mucosa/papillary projections. CONCLUSIONS: This study demonstrates that a CNN developed using cholangioscopy data alone has greater accuracy for biliary stricture classification than traditional ERCP-based sampling techniques.
Assuntos
Colestase , Aprendizado Profundo , Humanos , Constrição Patológica/diagnóstico , Inteligência Artificial , Estudos Prospectivos , Colestase/diagnóstico por imagem , Colestase/etiologiaRESUMO
OBJECTIVE: The diagnosis of autoimmune pancreatitis (AIP) is challenging. Sonographic and cross-sectional imaging findings of AIP closely mimic pancreatic ductal adenocarcinoma (PDAC) and techniques for tissue sampling of AIP are suboptimal. These limitations often result in delayed or failed diagnosis, which negatively impact patient management and outcomes. This study aimed to create an endoscopic ultrasound (EUS)-based convolutional neural network (CNN) model trained to differentiate AIP from PDAC, chronic pancreatitis (CP) and normal pancreas (NP), with sufficient performance to analyse EUS video in real time. DESIGN: A database of still image and video data obtained from EUS examinations of cases of AIP, PDAC, CP and NP was used to develop a CNN. Occlusion heatmap analysis was used to identify sonographic features the CNN valued when differentiating AIP from PDAC. RESULTS: From 583 patients (146 AIP, 292 PDAC, 72 CP and 73 NP), a total of 1 174 461 unique EUS images were extracted. For video data, the CNN processed 955 EUS frames per second and was: 99% sensitive, 98% specific for distinguishing AIP from NP; 94% sensitive, 71% specific for distinguishing AIP from CP; 90% sensitive, 93% specific for distinguishing AIP from PDAC; and 90% sensitive, 85% specific for distinguishing AIP from all studied conditions (ie, PDAC, CP and NP). CONCLUSION: The developed EUS-CNN model accurately differentiated AIP from PDAC and benign pancreatic conditions, thereby offering the capability of earlier and more accurate diagnosis. Use of this model offers the potential for more timely and appropriate patient care and improved outcome.
Assuntos
Pancreatite Autoimune/diagnóstico por imagem , Carcinoma Ductal Pancreático/diagnóstico por imagem , Endossonografia , Interpretação de Imagem Assistida por Computador/métodos , Redes Neurais de Computação , Neoplasias Pancreáticas/diagnóstico por imagem , Área Sob a Curva , Diagnóstico Diferencial , Humanos , Aprendizado de Máquina , Variações Dependentes do Observador , Pâncreas/diagnóstico por imagem , Curva ROCRESUMO
BACKGROUND AND AIMS: Detection and characterization of focal liver lesions (FLLs) is key for optimizing treatment for patients who may have a primary hepatic cancer or metastatic disease to the liver. This is the first study to develop an EUS-based convolutional neural network (CNN) model for the purpose of identifying and classifying FLLs. METHODS: A prospective EUS database comprising cases of FLLs visualized and sampled via EUS was reviewed. Relevant still images and videos of liver parenchyma and FLLs were extracted. Patient data were then randomly distributed for the purpose of CNN model training and testing. Once a final model was created, occlusion heatmap analysis was performed to assess the ability of the EUS-CNN model to autonomously identify FLLs. The performance of the EUS-CNN for differentiating benign and malignant FLLs was also analyzed. RESULTS: A total of 210,685 unique EUS images from 256 patients were used to train, validate, and test the CNN model. Occlusion heatmap analyses demonstrated that the EUS-CNN model was successful in autonomously locating FLLs in 92.0% of EUS video assets. When evaluating any random still image extracted from videos or physician-captured images, the CNN model was 90% sensitive and 71% specific (area under the receiver operating characteristic [AUROC], 0.861) for classifying malignant FLLs. When evaluating full-length video assets, the EUS-CNN model was 100% sensitive and 80% specific (AUROC, 0.904) for classifying malignant FLLs. CONCLUSIONS: This study demonstrated the capability of an EUS-CNN model to autonomously identify FLLs and to accurately classify them as either malignant or benign lesions.
Assuntos
Inteligência Artificial , Neoplasias Hepáticas , Humanos , Neoplasias Hepáticas/diagnóstico por imagem , Redes Neurais de Computação , Estudos Prospectivos , Sensibilidade e EspecificidadeRESUMO
BACKGROUNDS: Endoscopic ultrasound (EUS)-guided placement of lumen-apposing metal stents (LAMSs) has gained popularity for the treatment of pancreatic walled-off necrosis (WON). We compared the 20-mm and 15-mm LAMSs for the treatment of symptomatic WON in terms of clinical success and adverse events. METHODS: We conducted a retrospective, case-matched study of 306 adults at 22 tertiary centers from 04/2014 to 10/2018. A total of 102 patients with symptomatic WON who underwent drainage with 20-mm LAMS (cases) and 204 patients who underwent drainage with 15-mm LAMS (controls) were matched by age, sex, and drainage approach. Conditional logistic regression analysis was performed to compare clinical success (resolution of WON on follow-up imaging without reintervention) and adverse events (according to American Society for Gastrointestinal Endoscopy criteria). RESULTS: Clinical success was achieved in 92.2â% of patients with 20-mm LAMS and 91.7â% of patients with 15-mm LAMS (odds ratio 0.92; Pâ=â0.91). Patients with 20-mm LAMS underwent fewer direct endoscopic necrosectomy (DEN) sessions (mean 1.3 vs. 2.1; Pâ<â0.001), despite having larger WON collections (transverse axis 118.2 vs. 101.9âmm, Pâ=â0.003; anteroposterior axis 95.9 vs. 80.1âmm, Pâ=â0.01). There was no difference in overall adverse events (21.6â% vs. 15.2â%; Pâ=â0.72) and bleeding events (4.9â% vs. 3.4â%; Pâ=â0.54) between the 20-mm and 15-mm LAMS groups, respectively. CONCLUSIONS: The 20-mm LAMS showed comparable clinical success and safety profile to the 15-mm LAMS, with the need for fewer DEN sessions for WON resolution.
Assuntos
Drenagem , Stents , Adulto , Humanos , Necrose/etiologia , Estudos Retrospectivos , Resultado do Tratamento , Ultrassonografia de IntervençãoRESUMO
Obstetrical management of patients with genetic disorders bears its own challenges in the overall care and outcomes of the patient. Hereditary hemorrhagic telangiectasia (HHT), or Osler-Weber-Rendu Syndrome, is an inherited genetic disorder that, due to the nature of the pathophysiology of arteriovenous malformation and dysplasia, has been highly associated with an increased tendency of bleeding. This case presents a patient diagnosed with HHT prior to pregnancy who developed severe postpartum hemorrhage (PPH) after cesarean section delivery. Clinical considerations were made proactively with the knowledge and understanding of the genetic disorder, but due to unforeseen changes in the manifestation of the PPH, clinical decisions were dynamically modified in order to save the patient from exsanguination, ultimately resulting in an emergency hysterectomy and a total of 15 units of transfused packed red blood cells (pRBC). As a result, this case report serves as a preliminary recount for future research and clinical management of similar cases.
RESUMO
Sheehan syndrome is a well-documented endocrinological disorder that appears to be closely associated as a secondary sequela to postpartum hemorrhage. Due to pregnancy-related physiological adaptations, namely the increase in blood volume but lack of hypertrophic or hyperplastic growth within the pituitary, pregnancy increases the likelihood of infarction of the pituitary. This, coupled with other complications, such as postpartum hemorrhage, can lead to ischemia and permanent damage to the pituitary, and thus, all the downstream endocrinological pathways regulated by the pituitary. Namely, this can include, but is not limited to, adrenal crisis from improper stimulation of steroid secretion. Individuals who have been diagnosed with Sheehan syndrome require lifelong steroid supplementation for appropriate regulation of multiple systems, specifically circulatory. Without appropriate steroid supplementation exogenously, patients can rapidly decline with adverse hypotension, altered mental status, and loss of vascular tone. This case presents a case of a patient who, after extensive chart review and history taking, was found to have had a complicated pregnancy many years ago with multiple transfusions needed to stabilize her and was placed on exogenous steroid management, presenting for adrenal crisis, hypotension, and altered mental status after not taking her home steroid medication.
RESUMO
Background: Global initiatives to mitigate COVID-19 transmission have shifted health system priorities to management of patients with prolonged long COVID symptoms. To better meet the needs of patients, clinicians, and systems, a learning health system approach can use rapid-cycle methods to integrate data and real-world experience to iteratively evaluate and adapt models of long COVID care. Observations: Employees in the Veterans Health Administration formed a multidisciplinary workgroup. We sought to develop processes to learn more about this novel long COVID syndrome and innovative long COVID care models that can be applied within and outside of our health care system. We describe our workgroup processes and goals to create a mechanism for cross-facility communication, identify gaps in care and research, and cocreate knowledge on best practices for long COVID care delivery. Conclusions: The learning health system approach will be critical in reimagining health care service delivery after the COVID-19 pandemic.
RESUMO
Artificial intelligence (AI) research for medical applications has expanded quickly. Advancements in computer processing now allow for the development of complex neural network architectures (eg, convolutional neural networks) that are capable of extracting and learning complex features from massive data sets, including large image databases. Gastroenterology and endoscopy are well suited for AI research. Video capsule endoscopy is an ideal platform for AI model research given the large amount of data produced by each capsule examination and the annotated databases that are already available. Studies have demonstrated high performance for applications of capsule-based AI models developed for various pathologic conditions.
Assuntos
Endoscopia por Cápsula , Gastroenterologia , Inteligência Artificial , Humanos , PesquisaRESUMO
Nova Scotia's healthcare policy direction seems well-defined and well-established for the foreseeable future. This is the case, despite the recent electoral transition from a Progressive Conservative to a New Democratic Party government for the first time in the province's history; and despite the threat of the province's net direct debt increasing through 2012, after eight years of declining net direct debt as a percentage of the province's gross domestic product. As well, little public consideration is being given to disrupting the current regional healthcare organizational structure by further consolidating the province's nine district health authorities (DHA), as occurred last year in Alberta and New Brunswick. Moreover, the Health Information Technology Services Program of Nova Scotia (HITS-NS), the province's shared IT services or provincial service delivery organization, is steadily expanding the inventory of clinical, financial, and administrative software applications hosted for eight DHAs on a common Meditech Client-Server platform, as well as some applications for Capital Health DHA 9 (CDHA) and IWK Health Centre (IWK), the province's consolidated women's and children's hospital located in Halifax.
Assuntos
Difusão de Inovações , Reforma dos Serviços de Saúde , Sistemas Computadorizados de Registros Médicos , Nova EscóciaRESUMO
In spring 2007, HIMSS Analytics began developing its first Canada Information and Communications Technology (ICT) Study. Less than one year later, 38 RHAs, DHAs and HAs are already on board, with some 20 more scheduled to participate by year's end. Why are so many Canadian provincial healthcare delivery organizations now participating in HIMSS Analytics' Canada ICT Study? The answer is tied to the character of the HIMSS study, the value offered to all participants and specific Canadian healthcare issues that are addressed by the study.
Assuntos
Difusão de Inovações , Informática Médica/tendências , Sistemas Computadorizados de Registros Médicos/estatística & dados numéricos , Sociedades , Benchmarking , CanadáRESUMO
BACKGROUND AND OBJECTIVES: Endoscopic drainage/debridement of symptomatic walled off necrosis (WON) using lumen-apposing metal stents (LAMS) is both safe and effective. While endoscopic management of WON is the standard approach to treatment, the ideal concomitant medical therapy remains unclear. The purpose of this study was to further elucidate the effect of proton pump inhibitor (PPIs) therapy on the technical and clinical success of endoscopic treatment of WON. METHODS: Two hundred and seventy-two patients in 8 centers with WON managed by endoscopic drainage using LAMS were evaluated. Patients were followed for at least 6 months following treatment. The patients were divided into two groups: Those that used PPIs continuously during the therapy and those not on PPIs continuously during the interval of therapy. Outcomes included but were not limited to technical success, clinical success, number of procedures performed, and adverse events. RESULTS: From 2013 to 2016, 272 patients underwent WON drainage with successful transmural LAMS placement. The two groups were split evenly into PPI users and non-PPI users, and matched in regards to demographics, etiology of pancreatitis, WON size, and location. There was no difference in the technical success between the two groups (100% vs. 98.8%, P = 1), or in clinical success rates (78.7% vs. 77.9%). There was a significant difference in the required number of direct endoscopic necrosectomies to achieve clinical success in the PPI vs. non-PPI group (3.2 vs. 4.6 respectively, P < 0.01). There were significantly more cases of stent occlusion in the non-PPI group vs. PPI group (9.5% vs. 20.1% P = 0.012), but all other documented adverse events were not significantly different. CONCLUSION: Discontinuing PPIs during endoscopic drainage and necrosectomy of symptomatic WON appears to reduce the number of endoscopic procedures required to achieve resolution. Continuous PPI results in higher rates of early stent occlusion.
RESUMO
BACKGROUND: Imaging of gas distribution in the lungs of patients with asthma has been restricted because of the lack of a suitable gaseous contrast agent. Hyperpolarized helium-3 (HHe3) provides a new technique for magnetic resonance imaging of lung diseases. OBJECTIVE: We sought to investigate the use of HHe3 gas to image the lungs of patients with moderate or severe asthma and to assess changes in gas distribution after methacholine and exercise challenge. METHODS: Magnetic resonance imaging was performed in asthmatic patients immediately after inhalation of HHe3 gas. In addition, images were obtained before and after methacholine challenge and a standard exercise test. RESULTS: Areas of the lung with no signal or sharply reduced HHe3 signal (ventilation defects) are common in patients with asthma, and the number of defects was inversely related to the percent predicted FEV(1) (r = 0.71, P <.002). After methacholine challenge (n = 3), the number of defects increased. Similarly, imaging of the lungs after exercise (n = 6) showed increased ventilation defects in parallel with decreases in FEV(1). The increase in defects after challenge in these 9 asthmatic patients was significant both for the number (P <.02) and extent (P <.02) of the defects. The variability and speed of changes in ventilation and the complete lack of signal in many areas is in keeping with a model in which the defects result from airway closure. CONCLUSION: HHe3 magnetic resonance provides a new technique for imaging the distribution of inhaled air in the lungs. The technique is suitable for following responses to treatment of asthma and changes after methacholine or exercise challenge.