Matches in SemOpenAlex for { <https://semopenalex.org/work/W2806644002> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W2806644002 endingPage "AB434" @default.
- W2806644002 startingPage "AB434" @default.
- W2806644002 abstract "Deep Learning is an automatic machine learning using a cutting edge artificial intellectual possessing the structure and function of the brain called artificial neural networks, so called convolutional neural networks. Several research has demonstrated CAD using deep learning achieved favorable diagnostic performance in differentiation of benign from malignancy in ultrasound or cytology image. EUS-FNA has been common diagnostic technique in pancreatic cancer. Rapid onsite evaluation (ROSE) of EUS-FNA specimen improves diagnostic accuracy and reduces the number of EUS-FNA pass. High performance CAD using deep learning will potentially support ROSE with additional diagnostic value. We evaluated diagnostic performance of our CAD system using deep learning in EUS-FNA cytology of pancreatic ductal adenocarcinoma. We reviewed cytology specimen obtained by EUS-FNA performed in 2015 for histological evidence of pancreatic tumor. Cytology specimen was papanicolaou stained on 200 times magnification and every image was captured with fine quality of 1280 by 1024 pixel. Based on cyotopathology reports, 300 images with suggestive or suspicious of malignancy ( group A) and 150 images with suggestive of benign ( group B) were collected and grouped. 200/100 images in group A and 100/50 images in group B was respectively prepared for training set/test set. Computer spec was GPU NIVIDIA TITAN X and deep learning was conducted with single shot multibox detector. We performed sequential transfer training of 100 images in group A and 50 images in group B at the first learning stage and 200 images in group A and 100 images in group B at the second learning stage. We tested diagnostic performance of our deep learning system with the test set. Our deep learning system using single shot multibox detector showed diagnostic performance sensitivity 78%, specificity 60% and accuracy 69% at the first learning stage and sensitivity 80%, specificity 80% and accuracy 80% at the second learning stage. CAD performance using our deep learning system in EUS-FNA cytology of pancreatic cancer has demonstrated promising preliminary result with improving diagnostic performance by step-by-step learning. Higher training volume and more efficient system development is required to anticipate optimal CAD performance in ROSE of EUS-FNA cytology." @default.
- W2806644002 created "2018-06-13" @default.
- W2806644002 creator A5006683327 @default.
- W2806644002 creator A5020779902 @default.
- W2806644002 creator A5021407007 @default.
- W2806644002 creator A5022537134 @default.
- W2806644002 creator A5034737415 @default.
- W2806644002 creator A5035908413 @default.
- W2806644002 creator A5040580329 @default.
- W2806644002 creator A5062757390 @default.
- W2806644002 creator A5064282637 @default.
- W2806644002 creator A5072969214 @default.
- W2806644002 creator A5076202712 @default.
- W2806644002 creator A5078772421 @default.
- W2806644002 date "2018-06-01" @default.
- W2806644002 modified "2023-10-06" @default.
- W2806644002 title "Mo1296 RELIMINARY RESULT OF COMPUTER AIDED DIAGNOSIS (CAD) PERFORMANCE USING DEEP LEARNING IN EUS-FNA CYTOLOGY OF PANCREATIC CANCER" @default.
- W2806644002 doi "https://doi.org/10.1016/j.gie.2018.04.1946" @default.
- W2806644002 hasPublicationYear "2018" @default.
- W2806644002 type Work @default.
- W2806644002 sameAs 2806644002 @default.
- W2806644002 citedByCount "7" @default.
- W2806644002 countsByYear W28066440022019 @default.
- W2806644002 countsByYear W28066440022020 @default.
- W2806644002 countsByYear W28066440022021 @default.
- W2806644002 countsByYear W28066440022023 @default.
- W2806644002 crossrefType "journal-article" @default.
- W2806644002 hasAuthorship W2806644002A5006683327 @default.
- W2806644002 hasAuthorship W2806644002A5020779902 @default.
- W2806644002 hasAuthorship W2806644002A5021407007 @default.
- W2806644002 hasAuthorship W2806644002A5022537134 @default.
- W2806644002 hasAuthorship W2806644002A5034737415 @default.
- W2806644002 hasAuthorship W2806644002A5035908413 @default.
- W2806644002 hasAuthorship W2806644002A5040580329 @default.
- W2806644002 hasAuthorship W2806644002A5062757390 @default.
- W2806644002 hasAuthorship W2806644002A5064282637 @default.
- W2806644002 hasAuthorship W2806644002A5072969214 @default.
- W2806644002 hasAuthorship W2806644002A5076202712 @default.
- W2806644002 hasAuthorship W2806644002A5078772421 @default.
- W2806644002 hasBestOaLocation W28066440021 @default.
- W2806644002 hasConcept C108583219 @default.
- W2806644002 hasConcept C121608353 @default.
- W2806644002 hasConcept C126322002 @default.
- W2806644002 hasConcept C126838900 @default.
- W2806644002 hasConcept C142724271 @default.
- W2806644002 hasConcept C154945302 @default.
- W2806644002 hasConcept C18823058 @default.
- W2806644002 hasConcept C2775934546 @default.
- W2806644002 hasConcept C2776170712 @default.
- W2806644002 hasConcept C2778220009 @default.
- W2806644002 hasConcept C2778654104 @default.
- W2806644002 hasConcept C2779399171 @default.
- W2806644002 hasConcept C2780210213 @default.
- W2806644002 hasConcept C2780390042 @default.
- W2806644002 hasConcept C2781182431 @default.
- W2806644002 hasConcept C41008148 @default.
- W2806644002 hasConcept C65051434 @default.
- W2806644002 hasConcept C71924100 @default.
- W2806644002 hasConcept C81363708 @default.
- W2806644002 hasConceptScore W2806644002C108583219 @default.
- W2806644002 hasConceptScore W2806644002C121608353 @default.
- W2806644002 hasConceptScore W2806644002C126322002 @default.
- W2806644002 hasConceptScore W2806644002C126838900 @default.
- W2806644002 hasConceptScore W2806644002C142724271 @default.
- W2806644002 hasConceptScore W2806644002C154945302 @default.
- W2806644002 hasConceptScore W2806644002C18823058 @default.
- W2806644002 hasConceptScore W2806644002C2775934546 @default.
- W2806644002 hasConceptScore W2806644002C2776170712 @default.
- W2806644002 hasConceptScore W2806644002C2778220009 @default.
- W2806644002 hasConceptScore W2806644002C2778654104 @default.
- W2806644002 hasConceptScore W2806644002C2779399171 @default.
- W2806644002 hasConceptScore W2806644002C2780210213 @default.
- W2806644002 hasConceptScore W2806644002C2780390042 @default.
- W2806644002 hasConceptScore W2806644002C2781182431 @default.
- W2806644002 hasConceptScore W2806644002C41008148 @default.
- W2806644002 hasConceptScore W2806644002C65051434 @default.
- W2806644002 hasConceptScore W2806644002C71924100 @default.
- W2806644002 hasConceptScore W2806644002C81363708 @default.
- W2806644002 hasIssue "6" @default.
- W2806644002 hasLocation W28066440021 @default.
- W2806644002 hasOpenAccess W2806644002 @default.
- W2806644002 hasPrimaryLocation W28066440021 @default.
- W2806644002 hasRelatedWork W1706872257 @default.
- W2806644002 hasRelatedWork W2057092851 @default.
- W2806644002 hasRelatedWork W2071108161 @default.
- W2806644002 hasRelatedWork W2885508485 @default.
- W2806644002 hasRelatedWork W3033709430 @default.
- W2806644002 hasRelatedWork W3035783210 @default.
- W2806644002 hasRelatedWork W3212722508 @default.
- W2806644002 hasRelatedWork W3215838009 @default.
- W2806644002 hasRelatedWork W4200057142 @default.
- W2806644002 hasRelatedWork W4386418371 @default.
- W2806644002 hasVolume "87" @default.
- W2806644002 isParatext "false" @default.
- W2806644002 isRetracted "false" @default.
- W2806644002 magId "2806644002" @default.
- W2806644002 workType "article" @default.