FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2897071176> ?p ?o ?g. }

• W2897071176 abstract "Amyloid imaging has become a clinically important diagnostic tool of Alzheimer's disease and related disorders. In many recent studies using amyloid PET, the images were classified as normal or abnormal based on their standard uptake value ratios(SUVRs). However, the image processing pipelines to obtain SUVR differ among research centers and are not readily accessible in clinical practice. Therefore, reading amyloid PET images in the clinical setting largely depends on visual assessment. In this preliminary study, we evaluated whether a supervised machine learning algorithm can replicate the classifications made by human raters blinded to the clinical diagnosis. Visual assessment of 18F-florbetaben scans was implemented by two blinded raters based on the instructions provided by Piramal. In brief, regional cortical tracer uptake(RCTU) scores were rated for eight anatomical regions of interest(ROIs) and brain amyloid plaque load(BAPL) scores were derived from the rated RCTU scores. Cases which both raters assigned BAPL score 1 (N= 135) or 3 (N= 95) in agreement were included in this study. With these cases, we trained our support vector machine(SVM) classifier to classify each case into amyloid positive or negative group and 10-fold cross validation scheme was used. Accuracies of classification using each single ROI or a composite ROI including all 8 ROIs were obtained. Accuracy of classification using the composite ROI was 99.1% with 100% specificity. Only one amyloid positive scan was misclassified. Accuracies obtained using each ROI ranged from 87.7% to 99.1%, with highest accuracy from right temporal and left precuneus regions, and relatively low accuracies from both lateral parietal cortices (87.7% in the left, 91.2% in the right parietal region). Our results demonstrated high classification performance in distinguishing amyloid positive scans from negative scans. While classifications using frontal, lateral temporal, and precuneus regions each showed high accuracies, using lateral temporal regions for classification resulted in relatively low performances. In current study, we did not include cases with BAPL score 2. In our further study, we are planning to develop our algorithm so that it can eventually classify amyloid scans with more equivocal tracer uptakes ." @default.
• W2897071176 created "2018-10-26" @default.
• W2897071176 creator A5002819999 @default.
• W2897071176 creator A5003132353 @default.
• W2897071176 creator A5007768734 @default.
• W2897071176 creator A5013118725 @default.
• W2897071176 creator A5014452353 @default.
• W2897071176 creator A5022350068 @default.
• W2897071176 creator A5023107984 @default.
• W2897071176 creator A5064204310 @default.
• W2897071176 creator A5081007211 @default.
• W2897071176 creator A5083862622 @default.
• W2897071176 date "2018-07-01" @default.
• W2897071176 modified "2023-10-16" @default.
• W2897071176 title "P4‐096: REPLICATING VISUAL ASSESSMENTS OF ¹⁸ F‐FLORBETABEN PET USING MACHINE LEARNING TECHNIQUE" @default.
• W2897071176 doi "https://doi.org/10.1016/j.jalz.2018.06.2499" @default.
• W2897071176 hasPublicationYear "2018" @default.
• W2897071176 type Work @default.
• W2897071176 sameAs 2897071176 @default.
• W2897071176 citedByCount "0" @default.
• W2897071176 crossrefType "journal-article" @default.
• W2897071176 hasAuthorship W2897071176A5002819999 @default.
• W2897071176 hasAuthorship W2897071176A5003132353 @default.
• W2897071176 hasAuthorship W2897071176A5007768734 @default.
• W2897071176 hasAuthorship W2897071176A5013118725 @default.
• W2897071176 hasAuthorship W2897071176A5014452353 @default.
• W2897071176 hasAuthorship W2897071176A5022350068 @default.
• W2897071176 hasAuthorship W2897071176A5023107984 @default.
• W2897071176 hasAuthorship W2897071176A5064204310 @default.
• W2897071176 hasAuthorship W2897071176A5081007211 @default.
• W2897071176 hasAuthorship W2897071176A5083862622 @default.
• W2897071176 hasConcept C12267149 @default.
• W2897071176 hasConcept C153180895 @default.
• W2897071176 hasConcept C154945302 @default.
• W2897071176 hasConcept C168820333 @default.
• W2897071176 hasConcept C19609008 @default.
• W2897071176 hasConcept C199374082 @default.
• W2897071176 hasConcept C2775842073 @default.
• W2897071176 hasConcept C2989005 @default.
• W2897071176 hasConcept C41008148 @default.
• W2897071176 hasConcept C71924100 @default.
• W2897071176 hasConceptScore W2897071176C12267149 @default.
• W2897071176 hasConceptScore W2897071176C153180895 @default.
• W2897071176 hasConceptScore W2897071176C154945302 @default.
• W2897071176 hasConceptScore W2897071176C168820333 @default.
• W2897071176 hasConceptScore W2897071176C19609008 @default.
• W2897071176 hasConceptScore W2897071176C199374082 @default.
• W2897071176 hasConceptScore W2897071176C2775842073 @default.
• W2897071176 hasConceptScore W2897071176C2989005 @default.
• W2897071176 hasConceptScore W2897071176C41008148 @default.
• W2897071176 hasConceptScore W2897071176C71924100 @default.
• W2897071176 hasIssue "7S_Part_28" @default.
• W2897071176 hasLocation W28970711761 @default.
• W2897071176 hasOpenAccess W2897071176 @default.
• W2897071176 hasPrimaryLocation W28970711761 @default.
• W2897071176 hasRelatedWork W2041399278 @default.
• W2897071176 hasRelatedWork W2056016498 @default.
• W2897071176 hasRelatedWork W2136184105 @default.
• W2897071176 hasRelatedWork W2336974148 @default.
• W2897071176 hasRelatedWork W2389470892 @default.
• W2897071176 hasRelatedWork W3013515612 @default.
• W2897071176 hasRelatedWork W4205671340 @default.
• W2897071176 hasRelatedWork W4293087713 @default.
• W2897071176 hasRelatedWork W2187500075 @default.
• W2897071176 hasRelatedWork W2345184372 @default.
• W2897071176 hasVolume "14" @default.
• W2897071176 isParatext "false" @default.
• W2897071176 isRetracted "false" @default.
• W2897071176 magId "2897071176" @default.
• W2897071176 workType "article" @default.