Matches in SemOpenAlex for { <https://semopenalex.org/work/W2896593743> ?p ?o ?g. }
Showing items 1 to 84 of
84
with 100 items per page.
- W2896593743 abstract "Beta-amyloid (Aβ) is widely viewed as a major hallmark of Alzheimer's disease (AD) pathology. Standard uptake value ratio (SUVr) between a pre-specified target mean-cortical region and the cerebellum (or another) reference region is the most commonly used measurement to quantify Aβ burden for amyloid positron emission tomography (PET) technique. By utilizing a 3D convolutional neural network (3D-CNN), we examined the feasibility of estimating SUVr and determining Aβ positivity from florbetapir PET images directly without defining the target and reference regions and without extracting data from them. Florbetapir PET data from 1072 subjects (178 AD patients, 525 MCI patients and 369 cognitively unimpaired [CU] individuals) from Alzheimer's Disease Neuroimage Initiative (ADNI) were used for this analysis. Florbetapir PET data were first spatially normalized to MNI template space. We designed the Deep Learning 3D-CNN, consisting of two convolutional layers (each followed by a max pooling layer), and two fully connected layers with the mean-square-error (MSE) loss on top. To overcome data limitations (small size and noisy quality), we employ (1) a denoising auto-encoder to pre-train the 3D-CNN layer-wise; and (2) batch normalization to stabilize training. The estimated SUVr values from the 3D-CNN were subsequently used to define amyloid positivity using a previously defined threshold (Fleisher, et al., 2011). For florbetapir PET-based SUVr estimation, we achieve a mean absolute difference of 0.035, the correlation between original SUVr vs 3D-CNN estimated SUVr is 0.97 (p=2.2e-16). Using the traditional method as the current standard of truth, the deep learning method distinguished between positive and negative amyloid scans (i.e., florbetapir SUVr greater than or equal to 1.18) with 92% sensitivity, 98% specificity and 95% accuracy. Although preliminary, our deep-learning 3D-CNN based results point out the potential to produce highly accurately SUVr from entire PET images directly without extracting data from mean-cortical and the cerebellar reference region. Additional studies are being undertaken to improve SUVr estimates and explore the feasibility of directly using native space imaging data." @default.
- W2896593743 created "2018-10-26" @default.
- W2896593743 creator A5004794322 @default.
- W2896593743 creator A5026950221 @default.
- W2896593743 creator A5047768726 @default.
- W2896593743 creator A5048522863 @default.
- W2896593743 creator A5055454914 @default.
- W2896593743 creator A5082583182 @default.
- W2896593743 creator A5083352578 @default.
- W2896593743 date "2018-07-01" @default.
- W2896593743 modified "2023-10-16" @default.
- W2896593743 title "IC‐P‐026: QUANTIFICATION OF AMYLOID BURDEN FROM FLORBETAPIR PET IMAGES WITHOUT USING TARGET AND REFERENCE REGIONS: PRELIMINARY FINDINGS BASED ON THE DEEP LEARNING 3D CONVOLUTIONAL NEURAL NETWORK APPROACH" @default.
- W2896593743 doi "https://doi.org/10.1016/j.jalz.2018.06.2090" @default.
- W2896593743 hasPublicationYear "2018" @default.
- W2896593743 type Work @default.
- W2896593743 sameAs 2896593743 @default.
- W2896593743 citedByCount "0" @default.
- W2896593743 crossrefType "journal-article" @default.
- W2896593743 hasAuthorship W2896593743A5004794322 @default.
- W2896593743 hasAuthorship W2896593743A5026950221 @default.
- W2896593743 hasAuthorship W2896593743A5047768726 @default.
- W2896593743 hasAuthorship W2896593743A5048522863 @default.
- W2896593743 hasAuthorship W2896593743A5055454914 @default.
- W2896593743 hasAuthorship W2896593743A5082583182 @default.
- W2896593743 hasAuthorship W2896593743A5083352578 @default.
- W2896593743 hasConcept C105795698 @default.
- W2896593743 hasConcept C108583219 @default.
- W2896593743 hasConcept C116580362 @default.
- W2896593743 hasConcept C117220453 @default.
- W2896593743 hasConcept C136886441 @default.
- W2896593743 hasConcept C139945424 @default.
- W2896593743 hasConcept C144024400 @default.
- W2896593743 hasConcept C153180895 @default.
- W2896593743 hasConcept C154945302 @default.
- W2896593743 hasConcept C19165224 @default.
- W2896593743 hasConcept C199374082 @default.
- W2896593743 hasConcept C2524010 @default.
- W2896593743 hasConcept C2775842073 @default.
- W2896593743 hasConcept C2989005 @default.
- W2896593743 hasConcept C33923547 @default.
- W2896593743 hasConcept C41008148 @default.
- W2896593743 hasConcept C54170458 @default.
- W2896593743 hasConcept C70437156 @default.
- W2896593743 hasConcept C71924100 @default.
- W2896593743 hasConcept C81363708 @default.
- W2896593743 hasConceptScore W2896593743C105795698 @default.
- W2896593743 hasConceptScore W2896593743C108583219 @default.
- W2896593743 hasConceptScore W2896593743C116580362 @default.
- W2896593743 hasConceptScore W2896593743C117220453 @default.
- W2896593743 hasConceptScore W2896593743C136886441 @default.
- W2896593743 hasConceptScore W2896593743C139945424 @default.
- W2896593743 hasConceptScore W2896593743C144024400 @default.
- W2896593743 hasConceptScore W2896593743C153180895 @default.
- W2896593743 hasConceptScore W2896593743C154945302 @default.
- W2896593743 hasConceptScore W2896593743C19165224 @default.
- W2896593743 hasConceptScore W2896593743C199374082 @default.
- W2896593743 hasConceptScore W2896593743C2524010 @default.
- W2896593743 hasConceptScore W2896593743C2775842073 @default.
- W2896593743 hasConceptScore W2896593743C2989005 @default.
- W2896593743 hasConceptScore W2896593743C33923547 @default.
- W2896593743 hasConceptScore W2896593743C41008148 @default.
- W2896593743 hasConceptScore W2896593743C54170458 @default.
- W2896593743 hasConceptScore W2896593743C70437156 @default.
- W2896593743 hasConceptScore W2896593743C71924100 @default.
- W2896593743 hasConceptScore W2896593743C81363708 @default.
- W2896593743 hasIssue "7S_Part_1" @default.
- W2896593743 hasLocation W28965937431 @default.
- W2896593743 hasOpenAccess W2896593743 @default.
- W2896593743 hasPrimaryLocation W28965937431 @default.
- W2896593743 hasRelatedWork W2517027266 @default.
- W2896593743 hasRelatedWork W2731899572 @default.
- W2896593743 hasRelatedWork W2921836287 @default.
- W2896593743 hasRelatedWork W2999805992 @default.
- W2896593743 hasRelatedWork W3094279615 @default.
- W2896593743 hasRelatedWork W3116150086 @default.
- W2896593743 hasRelatedWork W3133861977 @default.
- W2896593743 hasRelatedWork W4200173597 @default.
- W2896593743 hasRelatedWork W4312417841 @default.
- W2896593743 hasRelatedWork W4321369474 @default.
- W2896593743 hasVolume "14" @default.
- W2896593743 isParatext "false" @default.
- W2896593743 isRetracted "false" @default.
- W2896593743 magId "2896593743" @default.
- W2896593743 workType "article" @default.