I Saizu, B Cotruta, R A Iacob, S Bunduc, R E Saizu, M Dumbrava, C Pietrareanu, G Becheanu, D Grigorie, C Gheorghe
{"title":"A MODEL TO PREDICT DIAGNOSIS OF PANCREATIC NEUROENDOCRINE TUMORS BASED ON EUS IMAGING FEATURES.","authors":"I Saizu, B Cotruta, R A Iacob, S Bunduc, R E Saizu, M Dumbrava, C Pietrareanu, G Becheanu, D Grigorie, C Gheorghe","doi":"10.4183/aeb.2023.407","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>This study aimed to determine predictive clinical and endoscopic ultrasound (EUS) features for pancreatic neuroendocrine tumor (PNET) diagnosis, utilizing EUS-guided tissue acquisition.</p><p><strong>Methods: </strong>A prospective study from 2018-2022 included patients with pancreatic masses undergoing EUS with elastography. Univariate binomial logistic regression followed by multiple logistic regression with significant predictors was employed. A forward selection algorithm identified optimal models based on predictor numbers. Variables encompassed EUS tumor characteristics (e.g., location, size, margins, echogenicity, vascularity on Doppler, main pancreatic duct dilation, elastography appearance, vascular invasion, and hypoechoic rim), alongside demographic and risk factors (smoking, alcohol, diabetes).</p><p><strong>Results: </strong>We evaluated 165 patients (24 PNETs). EUS features significantly linked with PNET diagnosis were well-defined margins (79% <i>vs</i>. 26%, p < 0.001), blue elastography appearance (46% <i>vs</i>. 9.9%, p < 0.001), vascularization (67% <i>vs</i>. 25%, p < 0.001), hypoechoic rim (46% <i>vs</i>. 10%, p < 0.001). The top-performing model, with 89.1% accuracy, included two predictors: a homogeneous lesion (OR, 95% CI) and a hypoechoic rim (OR, 95% CI).</p><p><strong>Conclusions: </strong>EUS appearance can differentiate PNETs from non-PNETs, with the hypoechoic rim being an independent predictor of PNET diagnosis. The most effective predictive model for PNETs combined the homogeneous lesion and presence of the hypoechoic rim.</p>","PeriodicalId":50902,"journal":{"name":"Acta Endocrinologica-Bucharest","volume":"19 4","pages":"407-414"},"PeriodicalIF":0.7000,"publicationDate":"2023-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11197819/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Acta Endocrinologica-Bucharest","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.4183/aeb.2023.407","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/6/24 0:00:00","PubModel":"Epub","JCR":"Q4","JCRName":"ENDOCRINOLOGY & METABOLISM","Score":null,"Total":0}
引用次数: 0
Abstract
Background: This study aimed to determine predictive clinical and endoscopic ultrasound (EUS) features for pancreatic neuroendocrine tumor (PNET) diagnosis, utilizing EUS-guided tissue acquisition.
Methods: A prospective study from 2018-2022 included patients with pancreatic masses undergoing EUS with elastography. Univariate binomial logistic regression followed by multiple logistic regression with significant predictors was employed. A forward selection algorithm identified optimal models based on predictor numbers. Variables encompassed EUS tumor characteristics (e.g., location, size, margins, echogenicity, vascularity on Doppler, main pancreatic duct dilation, elastography appearance, vascular invasion, and hypoechoic rim), alongside demographic and risk factors (smoking, alcohol, diabetes).
Results: We evaluated 165 patients (24 PNETs). EUS features significantly linked with PNET diagnosis were well-defined margins (79% vs. 26%, p < 0.001), blue elastography appearance (46% vs. 9.9%, p < 0.001), vascularization (67% vs. 25%, p < 0.001), hypoechoic rim (46% vs. 10%, p < 0.001). The top-performing model, with 89.1% accuracy, included two predictors: a homogeneous lesion (OR, 95% CI) and a hypoechoic rim (OR, 95% CI).
Conclusions: EUS appearance can differentiate PNETs from non-PNETs, with the hypoechoic rim being an independent predictor of PNET diagnosis. The most effective predictive model for PNETs combined the homogeneous lesion and presence of the hypoechoic rim.
期刊介绍:
Acta Endocrinologica (Buc) is an international journal covering the fields of basic and clinical Endocrinology, Neuroendocrinology, Reproductive Medicine, Chronobiology, Human Ethology published quarterly
Acta Endocrinologica (Buc) is the official international journal of the Romanian Society for Endocrinology. It continues the former Romanian Journal of Endocrinology