Estimation of brain amyloid accumulation using deep learning in clinical [11C]PiB PET imaging

Research output: Contribution to journal › Journal article › Research › peer-review

Standard

Estimation of brain amyloid accumulation using deep learning in clinical [11C]PiB PET imaging. / for the Alzheimer’s Disease Neuroimaging Initiative.

In: EJNMMI Physics, Vol. 10, No. 1, 44, 2023.

Research output: Contribution to journal › Journal article › Research › peer-review

Harvard

for the Alzheimer’s Disease Neuroimaging Initiative 2023, 'Estimation of brain amyloid accumulation using deep learning in clinical [11C]PiB PET imaging', EJNMMI Physics, vol. 10, no. 1, 44. https://doi.org/10.1186/s40658-023-00562-7

APA

for the Alzheimer’s Disease Neuroimaging Initiative (2023). Estimation of brain amyloid accumulation using deep learning in clinical [11C]PiB PET imaging. EJNMMI Physics, 10(1), [44]. https://doi.org/10.1186/s40658-023-00562-7

Vancouver

for the Alzheimer’s Disease Neuroimaging Initiative. Estimation of brain amyloid accumulation using deep learning in clinical [11C]PiB PET imaging. EJNMMI Physics. 2023;10(1). 44. https://doi.org/10.1186/s40658-023-00562-7

Author

for the Alzheimer’s Disease Neuroimaging Initiative. / Estimation of brain amyloid accumulation using deep learning in clinical [11C]PiB PET imaging. In: EJNMMI Physics. 2023 ; Vol. 10, No. 1.

Bibtex

@article{5ae77df72f534b99b82067bd46374fba,
title = "Estimation of brain amyloid accumulation using deep learning in clinical [11C]PiB PET imaging",
abstract = "Introduction: Estimation of brain amyloid accumulation is valuable for evaluation of patients with cognitive impairment in both research and clinical routine. The development of high throughput and accurate strategies for the determination of amyloid status could be an important tool in patient selection for clinical trials and amyloid directed treatment. Here, we propose the use of deep learning to quantify amyloid accumulation using standardized uptake value ratio (SUVR) and classify amyloid status based on their PET images. Methods: A total of 1309 patients with cognitive impairment scanned with [11C]PIB PET/CT or PET/MRI were included. Two convolutional neural networks (CNNs) for reading-based amyloid status and SUVR prediction were trained using 75% of the PET/CT data. The remaining PET/CT (n = 300) and all PET/MRI (n = 100) data was used for evaluation. Results: The prevalence of amyloid positive patients was 61%. The amyloid status classification model reproduced the expert reader{\textquoteright}s classification with 99% accuracy. There was a high correlation between reference and predicted SUVR (R 2 = 0.96). Both reference and predicted SUVR had an accuracy of 97% compared to expert classification when applying a predetermined SUVR threshold of 1.35 for binary classification of amyloid status. Conclusion: The proposed CNN models reproduced both the expert classification and quantitative measure of amyloid accumulation in a large local dataset. This method has the potential to replace or simplify existing clinical routines and can facilitate fast and accurate classification well-suited for a high throughput pipeline.",
keywords = "AI, Alzheimer{\textquoteright}s disease, Amyloid, Automatic diagnosis, Convolutional neural network, Decision support, Deep learning, Dementia, PET, Stratification",
author = "Ladefoged, {Claes N{\o}hr} and Lasse Anderberg and Karine Madsen and Henriksen, {Otto M{\o}lby} and Hasselbalch, {Steen Gregers} and Andersen, {Flemming Littrup} and Liselotte H{\o}jgaard and Ian Law and {for the Alzheimer{\textquoteright}s Disease Neuroimaging Initiative}",
note = "Publisher Copyright: {\textcopyright} 2023, The Author(s).",
year = "2023",
doi = "10.1186/s40658-023-00562-7",
language = "English",
volume = "10",
journal = "E J N M M I Physics",
issn = "2197-7364",
publisher = "SpringerOpen",
number = "1",

}

RIS

TY - JOUR

T1 - Estimation of brain amyloid accumulation using deep learning in clinical [11C]PiB PET imaging

AU - Ladefoged, Claes Nøhr

AU - Anderberg, Lasse

AU - Madsen, Karine

AU - Henriksen, Otto Mølby

AU - Hasselbalch, Steen Gregers

AU - Andersen, Flemming Littrup

AU - Højgaard, Liselotte

AU - Law, Ian

AU - for the Alzheimer’s Disease Neuroimaging Initiative

N1 - Publisher Copyright: © 2023, The Author(s).

PY - 2023

Y1 - 2023

N2 - Introduction: Estimation of brain amyloid accumulation is valuable for evaluation of patients with cognitive impairment in both research and clinical routine. The development of high throughput and accurate strategies for the determination of amyloid status could be an important tool in patient selection for clinical trials and amyloid directed treatment. Here, we propose the use of deep learning to quantify amyloid accumulation using standardized uptake value ratio (SUVR) and classify amyloid status based on their PET images. Methods: A total of 1309 patients with cognitive impairment scanned with [11C]PIB PET/CT or PET/MRI were included. Two convolutional neural networks (CNNs) for reading-based amyloid status and SUVR prediction were trained using 75% of the PET/CT data. The remaining PET/CT (n = 300) and all PET/MRI (n = 100) data was used for evaluation. Results: The prevalence of amyloid positive patients was 61%. The amyloid status classification model reproduced the expert reader’s classification with 99% accuracy. There was a high correlation between reference and predicted SUVR (R 2 = 0.96). Both reference and predicted SUVR had an accuracy of 97% compared to expert classification when applying a predetermined SUVR threshold of 1.35 for binary classification of amyloid status. Conclusion: The proposed CNN models reproduced both the expert classification and quantitative measure of amyloid accumulation in a large local dataset. This method has the potential to replace or simplify existing clinical routines and can facilitate fast and accurate classification well-suited for a high throughput pipeline.

AB - Introduction: Estimation of brain amyloid accumulation is valuable for evaluation of patients with cognitive impairment in both research and clinical routine. The development of high throughput and accurate strategies for the determination of amyloid status could be an important tool in patient selection for clinical trials and amyloid directed treatment. Here, we propose the use of deep learning to quantify amyloid accumulation using standardized uptake value ratio (SUVR) and classify amyloid status based on their PET images. Methods: A total of 1309 patients with cognitive impairment scanned with [11C]PIB PET/CT or PET/MRI were included. Two convolutional neural networks (CNNs) for reading-based amyloid status and SUVR prediction were trained using 75% of the PET/CT data. The remaining PET/CT (n = 300) and all PET/MRI (n = 100) data was used for evaluation. Results: The prevalence of amyloid positive patients was 61%. The amyloid status classification model reproduced the expert reader’s classification with 99% accuracy. There was a high correlation between reference and predicted SUVR (R 2 = 0.96). Both reference and predicted SUVR had an accuracy of 97% compared to expert classification when applying a predetermined SUVR threshold of 1.35 for binary classification of amyloid status. Conclusion: The proposed CNN models reproduced both the expert classification and quantitative measure of amyloid accumulation in a large local dataset. This method has the potential to replace or simplify existing clinical routines and can facilitate fast and accurate classification well-suited for a high throughput pipeline.

KW - AI

KW - Alzheimer’s disease

KW - Amyloid

KW - Automatic diagnosis

KW - Convolutional neural network

KW - Decision support

KW - Deep learning

KW - Dementia

KW - PET

KW - Stratification

U2 - 10.1186/s40658-023-00562-7

DO - 10.1186/s40658-023-00562-7

M3 - Journal article

C2 - 37450069

AN - SCOPUS:85165290294

VL - 10

JO - E J N M M I Physics

JF - E J N M M I Physics

SN - 2197-7364

IS - 1

M1 - 44

ER -

ID: 366005008