Fast and Accurate Amyloid Brain PET Quantification Without MRI Using Deep Neural Networks

Kang, Seung Kwan; Kim, Daewoon; Shin, Seong A.; Kim, Yu Kyeong; Choi, Hongyoon; Lee, Jae Sung

doi:10.2967/jnumed.122.264414

Cited by 13 publications

(10 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The SUV at each voxel is standardized to the SUV of cerebellar gray matter to obtain the SUV ratio (SUVR). We used BTXBrain software (Brightonix Imaging, Republic of Korea) to measure SUVR in accordance with a previously described procedure …”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Association of Hearing Loss With Anatomical and Functional Connectivity in Patients With Mild Cognitive Impairment

Lee

et al. 2023

JAMA Otolaryngol Head Neck Surg

View full text Add to dashboard Cite

ImportanceHearing loss is the most important modifiable risk factor for cognitive impairment; however, the association of hearing loss with anatomical and functional connectivity is not fully understood. This association may be elucidated by evaluating the findings of newer imaging technologies.ObjectivesTo evaluate the association of hearing loss with anatomical and functional connectivity in patients with mild cognitive impairment (MCI) by using multimodal imaging technology.Design, Setting, and ParticipantsThis was a prospective cross-sectional study of patients with MCI under the care of a neurology clinic at the Soonchunhyang University Bucheon Hospital (Republic of Korea) from April to September 2021. Data were analyzed from April 1 to June 30, 2022.Main Outcomes and MeasuresPure tone averages (PTA) and word recognition scores were used to measure hearing acuity. Magnetic resonance imaging (MRI) and positron emission tomography scans of the brain were used to assess functional and anatomical connectivity. Results of diffusion MRI, voxel- and surface-based morphometric imaging, and global brain amyloid standardized uptake ratio were analyzed. Neuroimaging parameters of patients with MCI plus hearing loss were compared with those of patients with MCI and no hearing loss. Correlation analyses among neuroimaging parameters, PTA, and word recognition scores were performed.ResultsOf 48 patients with MCI, 30 (62.5%) had hearing loss (PTA &gt;25 dB) and 18 (37.5%) did not (PTA ≤25 dB). Median (IQR) age was 73.5 (69.0-78.0) years in the group with hearing loss and 75.0 (65.0-78.0) years in the group with normal hearing; there were 20 (66.7%) and 14 (77.8%) women in each group, respectively. The group with MCI plus hearing loss demonstrated decreased functional connectivity between the bilateral insular and anterior divisions of the cingulate cortex, and decreased fractional anisotropy in the bilateral fornix, corpus callosum forceps major and tapetum, left parahippocampal cingulum, and left superior thalamic radiation. Fractional anisotropy in the corpus callosum forceps major and bilateral parahippocampal cingulum negatively correlated with the severity of hearing loss shown by PTA testing. The 2 groups were not significantly different in global β-amyloid uptake, gray matter volume, and cortical thickness.Conclusion and RelevanceThe findings of this prospective cross-sectional study suggest that alterations in the salience network may contribute to the neural basis of cognitive impairment associated with hearing loss in patients who are on the Alzheimer disease continuum.

show abstract

Section: Methodsmentioning

confidence: 99%

“…We used BTXBrain software (Brightonix Imaging, Republic of Korea) to measure SUVR in accordance with a previously described procedure. 37…”

Section: Brain β-Amyloid Analysismentioning

confidence: 99%

Association of Hearing Loss With Anatomical and Functional Connectivity in Patients With Mild Cognitive Impairment

Lee

et al. 2023

JAMA Otolaryngol Head Neck Surg

View full text Add to dashboard Cite

show abstract

“…Therefore, the quantification results derived from the normalized image using predefined atlases remains accurate. While Kang et al have demonstrated the robustness of the BTXBrain through internal and external datasets [ 15 ], further investigations are warranted to validate its accuracy across diverse image sets and cohorts.…”

Section: Discussionmentioning

confidence: 99%

“…Moreover, MRI is not always available for all patients. Therefore, several approaches using artificial intelligence (AI) techniques have been proposed to achieve spatial normalization of brain PET images without the assistance of a matching MRI [ 5 , 14 , 15 ]. BTXBrain (Brightonix Imaging Inc., Seoul, Korea) software is a brain PET and single-photon emission computed tomography (SPECT) image quantification platform based on an AI-based fast and reliable spatial normalization algorithm.…”

Section: Introductionmentioning

confidence: 99%

Clinical Performance Evaluation of an Artificial Intelligence-Powered Amyloid Brain PET Quantification Method

Kang,

Heo,

Chung

et al. 2024

Nucl Med Mol Imaging

Self Cite

View full text Add to dashboard Cite

Purpose This study assesses the clinical performance of BTXBrain-Amyloid, an artificial intelligence-powered software for quantifying amyloid uptake in brain PET images. Methods 150 amyloid brain PET images were visually assessed by experts and categorized as negative and positive. Standardized uptake value ratio (SUVR) was calculated with cerebellum grey matter as the reference region, and receiver operating characteristic (ROC) and precision-recall (PR) analysis for BTXBrain-Amyloid were conducted. For comparison, same image processing and analysis was performed using Statistical Parametric Mapping (SPM) program. In addition, to evaluate the spatial normalization (SN) performance, mutual information (MI) between MRI template and spatially normalized PET images was calculated and SPM group analysis was conducted. Results Both BTXBrain and SPM methods discriminated between negative and positive groups. However, BTXBrain exhibited lower SUVR standard deviation (0.06 and 0.21 for negative and positive, respectively) than SPM method (0.11 and 0.25). In ROC analysis, BTXBrain had an AUC of 0.979, compared to 0.959 for SPM, while PR curves showed an AUC of 0.983 for BTXBrain and 0.949 for SPM. At the optimal cut-off, the sensitivity and specificity were 0.983 and 0.921 for BTXBrain and 0.917 and 0.921 for SPM12, respectively. MI evaluation also favored BTXBrain (0.848 vs. 0.823), indicating improved SN. In SPM group analysis, BTXBrain exhibited higher sensitivity in detecting basal ganglia differences between negative and positive groups. Conclusion BTXBrain-Amyloid outperformed SPM in clinical performance evaluation, also demonstrating superior SN and improved detection of deep brain differences. These results suggest the potential of BTXBrain-Amyloid as a valuable tool for clinical amyloid PET image evaluation.

show abstract

“…In addition, deep learning has been applied to predict quantitative values from amyloid PET images. Deep learning-based anatomical standardization method for 18 F-florbetaben or 18 F-flutemetamol PET without MRI has been proposed [12]. A deep learning model was developed to quantify SUVR from 18 F-florbetapir or 18 F-florbetaben PET images in a native space [13].…”

Section: Introductionmentioning

confidence: 99%

Deep Learning-Driven Estimation of Centiloid Scales from Amyloid PET Images with 11C-PiB and 18F-Labeled Tracers in Alzheimer’s Disease

Yamao,

Miwa,

Kaneko

et al. 2024

Brain Sciences

View full text Add to dashboard Cite

Background: Standard methods for deriving Centiloid scales from amyloid PET images are time-consuming and require considerable expert knowledge. We aimed to develop a deep learning method of automating Centiloid scale calculations from amyloid PET images with 11C-Pittsburgh Compound-B (PiB) tracer and assess its applicability to 18F-labeled tracers without retraining. Methods: We trained models on 231 11C-PiB amyloid PET images using a 50-layer 3D ResNet architecture. The models predicted the Centiloid scale, and accuracy was assessed using mean absolute error (MAE), linear regression analysis, and Bland–Altman plots. Results: The MAEs for Alzheimer’s disease (AD) and young controls (YC) were 8.54 and 2.61, respectively, using 11C-PiB, and 8.66 and 3.56, respectively, using 18F-NAV4694. The MAEs for AD and YC were higher with 18F-florbetaben (39.8 and 7.13, respectively) and 18F-florbetapir (40.5 and 12.4, respectively), and the error rate was moderate for 18F-flutemetamol (21.3 and 4.03, respectively). Linear regression yielded a slope of 1.00, intercept of 1.26, and R2 of 0.956, with a mean bias of −1.31 in the Centiloid scale prediction. Conclusions: We propose a deep learning means of directly predicting the Centiloid scale from amyloid PET images in a native space. Transferring the model trained on 11C-PiB directly to 18F-NAV4694 without retraining was feasible.

show abstract

Fast and Accurate Amyloid Brain PET Quantification Without MRI Using Deep Neural Networks

Cited by 13 publications

References 45 publications

Association of Hearing Loss With Anatomical and Functional Connectivity in Patients With Mild Cognitive Impairment

Association of Hearing Loss With Anatomical and Functional Connectivity in Patients With Mild Cognitive Impairment

Clinical Performance Evaluation of an Artificial Intelligence-Powered Amyloid Brain PET Quantification Method

Deep Learning-Driven Estimation of Centiloid Scales from Amyloid PET Images with 11C-PiB and 18F-Labeled Tracers in Alzheimer’s Disease

Contact Info

Product

Resources

About