Background And Aims: EUS-guided needle-based confocal laser endomicroscopy (nCLE) characteristics of common types of pancreatic cystic lesions (PCLs) have been identified; however, surgical histopathology was available in a minority of cases. We sought to assess the performance characteristics of EUS nCLE for differentiating mucinous from non-mucinous PCLs in a larger series of patients with a definitive diagnosis.
Methods: Six endosonographers (nCLE experience >30 cases each) blinded to all clinical data, reviewed nCLE images of PCLs from 29 patients with surgical (n = 23) or clinical (n = 6) correlation. After 2 weeks, the assessors reviewed the same images in a different sequence. A tutorial on available and novel nCLE image patterns was provided before each review. The performance characteristics of nCLE and the κ statistic for interobserver agreement (IOA, 95% confidence interval [CI]), and intraobserver reliability (IOR, mean ± standard deviation [SD]) for identification of nCLE image patterns were calculated. Landis and Koch interpretation of κ values was used.
Results: A total of 29 (16 mucinous PCLs, 13 non-mucinous PCLs) nCLE patient videos were reviewed. The overall sensitivity, specificity, and accuracy for the diagnosis of mucinous PCLs were 95%, 94%, and 95%, respectively. The IOA and IOR (mean ± SD) were κ = 0.81 (almost perfect); 95% CI, 0.71-0.90; and κ = 0.86 ± 0.11 (almost perfect), respectively. The overall specificity, sensitivity, and accuracy for the diagnosis of serous cystadenomas (SCAs) were 99%, 98%, and 98%, respectively. The IOA and IOR (mean ± SD) for recognizing the characteristic image pattern of SCA were κ = 0.83 (almost perfect); 95% CI, 0.73-0.92; and κ = 0.85 ± 0.11 (almost perfect), respectively.
Conclusions: EUS-guided nCLE can provide virtual histology of PCLs with a high degree of accuracy and inter- and intraobserver agreement in differentiating mucinous versus non-mucinous PCLs. These preliminary results support larger multicenter studies to evaluate EUS nCLE. (Clinical trial registration number: NCT02516488.).
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1016/j.gie.2017.03.002 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Thermal Liquid Biopsy: A Promising Tool for the Differential Diagnosis of Pancreatic Cystic Lesions and Malignancy Detection.
Cancers (Basel)
November 2024
Instituto de Investigación Sanitaria Aragón (IIS Aragón), 50009 Zaragoza, Spain.
Pancreatic cystic lesions (PCLs) are a heterogeneous group of lesions with increasing incidence, usually identified incidentally on imaging studies (multidetector computed tomography (MDCT), magnetic resonance imaging (MRI), or endoscopic ultrasound (EUS)) [...
View Article and Find Full Text PDFExploring a novel composite method using non-contrast EUS enhanced microvascular imaging and cyst fluid analysis to differentiate pancreatic cystic lesions.
Dig Liver Dis
November 2023
Humanitas Research Hospital -IRCCS-, Endoscopy Unit, Rozzano, Italy; Humanitas University, Department of Biomedical Sciences, Pieve Emanuele, Italy.
Background And Aims: Differentiating pancreatic cystic lesions (PCLs) remains a diagnostic challenge. The use of high-definition imaging modalities which detect tumor microvasculature have been described in solid lesions. We aim to evaluate the usefulness of cystic microvasculature when used in combination with cyst fluid biochemistry to differentiate PCLs.
View Article and Find Full Text PDFDeep-Learning-Enabled Computer-Aided Diagnosis in the Classification of Pancreatic Cystic Lesions on Confocal Laser Endomicroscopy.
Diagnostics (Basel)
March 2023
Department of Electrical Engineering, National Yunlin University of Science and Technology, Yunlin 64002, Taiwan.
Accurate classification of pancreatic cystic lesions (PCLs) is important to facilitate proper treatment and to improve patient outcomes. We utilized the convolutional neural network (CNN) of VGG19 to develop a computer-aided diagnosis (CAD) system in the classification of subtypes of PCLs in endoscopic ultrasound-guided needle-based confocal laser endomicroscopy (nCLE). From a retrospectively collected 22,424 nCLE video frames (50 videos) as the training/validation set and 11,047 nCLE video frames (18 videos) as the test set, we developed and compared the diagnostic performance of three CNNs with distinct methods of designating the region of interest.
View Article and Find Full Text PDFPancreatic cyst fluid interleukin-1 beta (IL-1β) level in predicting the risk of malignancy in pancreatic cysts.
Egypt J Immunol
October 2022
Department of Internal Medicine, Faculty of Medicine, Cairo University, Cairo, Egypt.
Pancreatic cystic lesions (PCLs) may be accidentally discovered in up to 13.5% of cases. These PCLs are of multiple types, including mucinous cysts (intra-ductal papillary mucinous neoplasms [IPMN] and mucinous cystic neoplasms [MCN]) that have a risk of malignant transformation.
View Article and Find Full Text PDFDeep Learning for Automatic Differentiation of Mucinous versus Non-Mucinous Pancreatic Cystic Lesions: A Pilot Study.
Diagnostics (Basel)
August 2022
Department of Gastroenterology, São João University Hospital, Alameda Professor Hernâni Monteiro, 4200-427 Porto, Portugal.
Endoscopic ultrasound (EUS) morphology can aid in the discrimination between mucinous and non-mucinous pancreatic cystic lesions (PCLs) but has several limitations that can be overcome by artificial intelligence. We developed a convolutional neural network (CNN) algorithm for the automatic diagnosis of mucinous PCLs. Images retrieved from videos of EUS examinations for PCL characterization were used for the development, training, and validation of a CNN for mucinous cyst diagnosis.
View Article and Find Full Text PDFWant AI Summaries of new PubMed Abstracts delivered to your In-box?
Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!