Bayesian Posterior Distribution Estimation of Kinetic Parameters in Dynamic Brain PET Using Generative Deep Learning Models

Positron Emission Tomography (PET) is a valuable imaging method for studying molecular-level processes in the body, such as hyperphosphorylated tau (ptau) protein aggregates, a hallmark of several neurodegenerative diseases including Alzheimer's disease. P-tau density and cerebral perfusion can...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on medical imaging Vol. PP; p. 1
Main Authors: Djebra, Yanis, Liu, Xiaofeng, Marin, Thibault, Tiss, Amal, Dhaynaut, Maeva, Guehl, Nicolas, Johnson, Keith, Fakhri, Georges El, Ma, Chao, Ouyang, Jinsong
Format: Journal Article
Language:English
Published: United States IEEE 15.07.2025
Subjects:
ISSN:0278-0062, 1558-254X, 1558-254X
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract Positron Emission Tomography (PET) is a valuable imaging method for studying molecular-level processes in the body, such as hyperphosphorylated tau (ptau) protein aggregates, a hallmark of several neurodegenerative diseases including Alzheimer's disease. P-tau density and cerebral perfusion can be quantified from dynamic PET images using tracer kinetic modeling techniques. However, noise in PET images leads to uncertainty in the estimated kinetic parameters, which can be quantified by estimating the posterior distribution of kinetic parameters using Bayesian inference (BI). Markov Chain Monte Carlo (MCMC) techniques are commonly used for posterior estimation but with significant computational needs. This work proposes an Improved Denoising Diffusion Probabilistic Model (iDDPM)-based method to estimate the posterior distribution of kinetic parameters in dynamic PET, leveraging the high computational efficiency of deep learning. The performance of the proposed method was evaluated on a [18F]MK6240 study and compared to a Conditional Variational Autoencoder with dual decoder (CVAE-DD)-based method and a Wasserstein GAN with gradient penalty (WGAN-GP)-based method. Posterior distributions inferred from Metropolis-Hasting MCMC were used as reference. Our approach consistently outperformed the CVAE-DD and WGAN-GP methods and offered significant reduction in computation time than the MCMC method (over 230 times faster), inferring accurate (< 0.67% mean error) and precise (< 7.23% standard deviation error) posterior distributions.
AbstractList Positron Emission Tomography (PET) is a valuable imaging method for studying molecular-level processes in the body, such as hyperphosphorylated tau (ptau) protein aggregates, a hallmark of several neurodegenerative diseases including Alzheimer's disease. P-tau density and cerebral perfusion can be quantified from dynamic PET images using tracer kinetic modeling techniques. However, noise in PET images leads to uncertainty in the estimated kinetic parameters, which can be quantified by estimating the posterior distribution of kinetic parameters using Bayesian inference (BI). Markov Chain Monte Carlo (MCMC) techniques are commonly used for posterior estimation but with significant computational needs. This work proposes an Improved Denoising Diffusion Probabilistic Model (iDDPM)-based method to estimate the posterior distribution of kinetic parameters in dynamic PET, leveraging the high computational efficiency of deep learning. The performance of the proposed method was evaluated on a [18F]MK6240 study and compared to a Conditional Variational Autoencoder with dual decoder (CVAE-DD)-based method and a Wasserstein GAN with gradient penalty (WGAN-GP)-based method. Posterior distributions inferred from Metropolis-Hasting MCMC were used as reference. Our approach consistently outperformed the CVAE-DD and WGAN-GP methods and offered significant reduction in computation time than the MCMC method (over 230 times faster), inferring accurate (< 0.67% mean error) and precise (< 7.23% standard deviation error) posterior distributions.
Positron Emission Tomography (PET) is a valuable imaging method for studying molecular-level processes in the body, such as hyperphosphorylated tau (p-tau) protein aggregates, a hallmark of several neurodegenerative diseases including Alzheimer’s disease. P-tau density and cerebral perfusion can be quantified from dynamic PET images using tracer kinetic modeling techniques. However, noise in PET images leads to uncertainty in the estimated kinetic parameters, which can be quantified by estimating the posterior distribution of kinetic parameters using Bayesian inference (BI). Markov Chain Monte Carlo (MCMC) techniques are commonly used for posterior estimation but with significant computational needs. This work proposes an Improved Denoising Diffusion Probabilistic Model (iDDPM)-based method to estimate the posterior distribution of kinetic parameters in dynamic PET, leveraging the high computational efficiency of deep learning. The performance of the proposed method was evaluated on a [18F]MK6240 study and compared to a Conditional Variational Autoencoder with dual decoder (CVAE-DD)-based method and a Wasserstein GAN with gradient penalty (WGAN-GP)-based method. Posterior distributions inferred from Metropolis-Hasting MCMC were used as reference. Our approach consistently outperformed the CVAE-DD and WGAN-GP methods and offered significant reduction in computation time than the MCMC method (over 230 times faster), inferring accurate (< 0.67% mean error) and precise (< 7.23% standard deviation error) posterior distributions.
Positron Emission Tomography (PET) is a valuable imaging method for studying molecular-level processes in the body, such as hyperphosphorylated tau (p-tau) protein aggregates, a hallmark of several neurodegenerative diseases including Alzheimer's disease. P-tau density and cerebral perfusion can be quantified from dynamic PET images using tracer kinetic modeling techniques. However, noise in PET images leads to uncertainty in the estimated kinetic parameters, which can be quantified by estimating the posterior distribution of kinetic parameters using Bayesian inference (BI). Markov Chain Monte Carlo (MCMC) techniques are commonly used for posterior estimation but with significant computational needs. This work proposes an Improved Denoising Diffusion Probabilistic Model (iDDPM)-based method to estimate the posterior distribution of kinetic parameters in dynamic PET, leveraging the high computational efficiency of deep learning. The performance of the proposed method was evaluated on a [18F]MK6240 study and compared to a Conditional Variational Autoencoder with dual decoder (CVAE-DD)-based method and a Wasserstein GAN with gradient penalty (WGAN-GP)-based method. Posterior distributions inferred from Metropolis-Hasting MCMC were used as reference. Our approach consistently outperformed the CVAE-DD and WGAN-GP methods and offered significant reduction in computation time than the MCMC method (over 230 times faster), inferring accurate (< 0.67% mean error) and precise (< 7.23% standard deviation error) posterior distributions.Positron Emission Tomography (PET) is a valuable imaging method for studying molecular-level processes in the body, such as hyperphosphorylated tau (p-tau) protein aggregates, a hallmark of several neurodegenerative diseases including Alzheimer's disease. P-tau density and cerebral perfusion can be quantified from dynamic PET images using tracer kinetic modeling techniques. However, noise in PET images leads to uncertainty in the estimated kinetic parameters, which can be quantified by estimating the posterior distribution of kinetic parameters using Bayesian inference (BI). Markov Chain Monte Carlo (MCMC) techniques are commonly used for posterior estimation but with significant computational needs. This work proposes an Improved Denoising Diffusion Probabilistic Model (iDDPM)-based method to estimate the posterior distribution of kinetic parameters in dynamic PET, leveraging the high computational efficiency of deep learning. The performance of the proposed method was evaluated on a [18F]MK6240 study and compared to a Conditional Variational Autoencoder with dual decoder (CVAE-DD)-based method and a Wasserstein GAN with gradient penalty (WGAN-GP)-based method. Posterior distributions inferred from Metropolis-Hasting MCMC were used as reference. Our approach consistently outperformed the CVAE-DD and WGAN-GP methods and offered significant reduction in computation time than the MCMC method (over 230 times faster), inferring accurate (< 0.67% mean error) and precise (< 7.23% standard deviation error) posterior distributions.
Author Ouyang, Jinsong
Dhaynaut, Maeva
Ma, Chao
Tiss, Amal
Djebra, Yanis
Marin, Thibault
Guehl, Nicolas
Johnson, Keith
Fakhri, Georges El
Liu, Xiaofeng
Author_xml – sequence: 1
  givenname: Yanis
  orcidid: 0000-0003-0756-571X
  surname: Djebra
  fullname: Djebra, Yanis
  organization: Department of Radiology & Biomedical Imaging, Yale School of Medicine, Yale Biomedical Imaging Institute, Yale University, New Haven, CT, USA
– sequence: 2
  givenname: Xiaofeng
  orcidid: 0000-0002-4514-2016
  surname: Liu
  fullname: Liu, Xiaofeng
  organization: Department of Radiology & Biomedical Imaging, Yale School of Medicine, Yale Biomedical Imaging Institute, Yale University, New Haven, CT, USA
– sequence: 3
  givenname: Thibault
  orcidid: 0000-0002-8669-1003
  surname: Marin
  fullname: Marin, Thibault
  organization: Department of Radiology & Biomedical Imaging, Yale School of Medicine, Yale Biomedical Imaging Institute, Yale University, New Haven, CT, USA
– sequence: 4
  givenname: Amal
  surname: Tiss
  fullname: Tiss, Amal
  organization: Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
– sequence: 5
  givenname: Maeva
  surname: Dhaynaut
  fullname: Dhaynaut, Maeva
  organization: Department of Radiology & Biomedical Imaging, Yale School of Medicine, Yale Biomedical Imaging Institute, Yale University, New Haven, CT, USA
– sequence: 6
  givenname: Nicolas
  orcidid: 0000-0001-6876-8292
  surname: Guehl
  fullname: Guehl, Nicolas
  organization: Department of Radiology & Biomedical Imaging, Yale School of Medicine, Yale Biomedical Imaging Institute, Yale University, New Haven, CT, USA
– sequence: 7
  givenname: Keith
  surname: Johnson
  fullname: Johnson, Keith
  organization: Department of Radiology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, USA
– sequence: 8
  givenname: Georges El
  surname: Fakhri
  fullname: Fakhri, Georges El
  organization: Department of Radiology & Biomedical Imaging, Yale School of Medicine, Yale Biomedical Imaging Institute, Yale University, New Haven, CT, USA
– sequence: 9
  givenname: Chao
  orcidid: 0000-0002-2016-8084
  surname: Ma
  fullname: Ma, Chao
  email: chao.ma.cm2943@yale.edu
  organization: Department of Radiology & Biomedical Imaging, Yale School of Medicine, Yale Biomedical Imaging Institute, Yale University, New Haven, CT, USA
– sequence: 10
  givenname: Jinsong
  orcidid: 0000-0003-4245-170X
  surname: Ouyang
  fullname: Ouyang, Jinsong
  email: jinsong.ouyang@yale.edu
  organization: Department of Radiology & Biomedical Imaging, Yale School of Medicine, Yale Biomedical Imaging Institute, Yale University, New Haven, CT, USA
BackLink https://www.ncbi.nlm.nih.gov/pubmed/40663684$$D View this record in MEDLINE/PubMed
BookMark eNpVkUFvEzEQhS1URNPCnQNCPnLZMPauvfYJ0Sa0FanIIZW4Wd7d2WK0sVN7Uyniz-M0ocDJ9rxv3mj8zsiJDx4Jectgyhjoj6vbmykHLqalUEoJ_YJMmBCq4KL6fkImwGtVAEh-Ss5S-gnAKgH6FTmtQMpSqmpCfl3YHSZnPV2GNGJ0IdKZS2N0zXZ0wdN5Gt3aPl1DT786j6Nr6dJGu8bMJ-o8ne28XefqRbT5tZyv6F1y_p5eoceYex-RzhA3dIE2-r1wGzoc0mvysrdDwjfH85zcfZmvLq-Lxberm8vPi6LlSumCt51om6YW1uq-qmVp83rQ1NBp1TagK1VJ1VXcyrrWWmLZ52IW2rqUqOuuPCefDr6bbbPGrkU_RjuYTcyLxZ0J1pn_Fe9-mPvwaBgvmaoYyw4fjg4xPGwxjWbtUovDYD2GbTIlL0FwoYBn9P2_w56n_PnyDMABaGNIKWL_jDAw-1RNTtXsUzXHVHPLu0OLQ8S_OAMFoGT5G5z_nzk
CODEN ITMID4
ContentType Journal Article
DBID 97E
RIA
RIE
AAYXX
CITATION
NPM
7X8
5PM
DOI 10.1109/TMI.2025.3588859
DatabaseName IEEE All-Society Periodicals Package (ASPP) 2005–Present
IEEE All-Society Periodicals Package (ASPP) 1998–Present
IEEE Electronic Library (IEL)
CrossRef
PubMed
MEDLINE - Academic
PubMed Central (Full Participant titles)
DatabaseTitle CrossRef
PubMed
MEDLINE - Academic
DatabaseTitleList

MEDLINE - Academic
PubMed
Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: RIE
  name: IEEE Electronic Library (IEL)
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
– sequence: 3
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod fulltext_linktorsrc
Discipline Medicine
Engineering
EISSN 1558-254X
EndPage 1
ExternalDocumentID PMC12318411
40663684
10_1109_TMI_2025_3588859
11080086
Genre orig-research
Journal Article
GrantInformation_xml – fundername: Common Fund
  grantid: P01AG036694; P41EB022544; R01AG076153; R01AG085561; R01EB033582; R01EB035093; R21AG070714; R21EB034911
  funderid: 10.13039/100015326
– fundername: NIA NIH HHS
  grantid: P01 AG036694
– fundername: NIA NIH HHS
  grantid: R01 AG085561
– fundername: NIBIB NIH HHS
  grantid: R01 EB035093
– fundername: NIA NIH HHS
  grantid: R21 AG070714
– fundername: NIBIB NIH HHS
  grantid: P41 EB022544
– fundername: NIBIB NIH HHS
  grantid: R21 EB034911
– fundername: NIA NIH HHS
  grantid: R01 AG076153
– fundername: NIBIB NIH HHS
  grantid: R01 EB033582
GroupedDBID ---
-DZ
-~X
0R~
29I
4.4
5GY
5RE
6IK
97E
AAJGR
AARMG
AASAJ
AAWTH
ABAZT
ABQJQ
ABVLG
ACGFO
ACGFS
ACIWK
ACNCT
ACPRK
AENEX
AFRAH
AGQYO
AHBIQ
AKJIK
AKQYR
ALMA_UNASSIGNED_HOLDINGS
ASUFR
ATWAV
BEFXN
BFFAM
BGNUA
BKEBE
BPEOZ
CS3
DU5
EBS
F5P
HZ~
IFIPE
IPLJI
JAVBF
LAI
M43
MS~
O9-
OCL
P2P
PQQKQ
RIA
RIE
RNS
RXW
TAE
TN5
.GJ
53G
5VS
AAYXX
AETIX
AGSQL
AI.
AIBXA
ALLEH
CITATION
EJD
H~9
IBMZZ
ICLAB
IFJZH
VH1
NPM
7X8
5PM
ID FETCH-LOGICAL-c2889-2cd5cbb75aa9f4763a2540b70d98cb0948468d42a677996e3fb09cb0c736e97d3
IEDL.DBID RIE
ISSN 0278-0062
1558-254X
IngestDate Tue Sep 30 17:02:43 EDT 2025
Thu Oct 02 22:23:28 EDT 2025
Wed Aug 06 16:36:32 EDT 2025
Sat Nov 29 07:41:56 EST 2025
Wed Aug 06 18:00:03 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed false
IsScholarly true
Language English
License https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
https://doi.org/10.15223/policy-029
https://doi.org/10.15223/policy-037
Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, including reprinting/republishing this material for advertising or promotional purposes, collecting new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The final version of record is available at: https://doi.org/10.1109/TMI.2025.3588859
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c2889-2cd5cbb75aa9f4763a2540b70d98cb0948468d42a677996e3fb09cb0c736e97d3
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ORCID 0000-0002-4514-2016
0000-0002-8669-1003
0000-0002-2016-8084
0000-0003-4245-170X
0000-0001-6876-8292
0000-0003-0756-571X
OpenAccessLink https://www.ncbi.nlm.nih.gov/pmc/articles/12318411
PMID 40663684
PQID 3230525802
PQPubID 23479
PageCount 1
ParticipantIDs ieee_primary_11080086
pubmed_primary_40663684
crossref_primary_10_1109_TMI_2025_3588859
pubmedcentral_primary_oai_pubmedcentral_nih_gov_12318411
proquest_miscellaneous_3230525802
PublicationCentury 2000
PublicationDate 20250715
PublicationDateYYYYMMDD 2025-07-15
PublicationDate_xml – month: 7
  year: 2025
  text: 20250715
  day: 15
PublicationDecade 2020
PublicationPlace United States
PublicationPlace_xml – name: United States
PublicationTitle IEEE transactions on medical imaging
PublicationTitleAbbrev TMI
PublicationTitleAlternate IEEE Trans Med Imaging
PublicationYear 2025
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0014509
Score 2.484378
Snippet Positron Emission Tomography (PET) is a valuable imaging method for studying molecular-level processes in the body, such as hyperphosphorylated tau (ptau)...
Positron Emission Tomography (PET) is a valuable imaging method for studying molecular-level processes in the body, such as hyperphosphorylated tau (p-tau)...
SourceID pubmedcentral
proquest
pubmed
crossref
ieee
SourceType Open Access Repository
Aggregation Database
Index Database
Publisher
StartPage 1
SubjectTerms Conditional Variational Autoencoder
Data models
Deep learning
Diffusion models
Diffusion processes
Dynamic PET imaging
Estimation
Imaging
Kinetic modeling
Kinetic theory
Noise
Positron emission tomography
Posterior distribution
Radiotracer
Training
Title Bayesian Posterior Distribution Estimation of Kinetic Parameters in Dynamic Brain PET Using Generative Deep Learning Models
URI https://ieeexplore.ieee.org/document/11080086
https://www.ncbi.nlm.nih.gov/pubmed/40663684
https://www.proquest.com/docview/3230525802
https://pubmed.ncbi.nlm.nih.gov/PMC12318411
Volume PP
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
journalDatabaseRights – providerCode: PRVIEE
  databaseName: IEEE Electronic Library (IEL)
  customDbUrl:
  eissn: 1558-254X
  dateEnd: 99991231
  omitProxy: false
  ssIdentifier: ssj0014509
  issn: 0278-0062
  databaseCode: RIE
  dateStart: 19820101
  isFulltext: true
  titleUrlDefault: https://ieeexplore.ieee.org/
  providerName: IEEE
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1Rb9MwED6xCSF4GDA2yBiTkXjhIZvjxLH9uK2dQLCpDwX1LXIcBypNydSuSNP-_O6ctKwTe-Atiu3I8jm--3x33wF84hXqoTxTce51FWdo08fGahN7VMWJU2WahQDZn9_VxYWeTMyoT1YPuTDe-xB85g_pMfjyq9Yt6KrsiELWyQbfgA2lVJestXIZZLKL5xBEGctzsfRJcnM0Pv-KSFDIw1Qi4JPEFJqRqs11tqaOQn2Vf5maDyMm76mgs5f_OflXsNXbmuy42xyv4YlvtuHFPQbCbXh23vvW38Dtib3xlFLJqIAvtrczNiBa3b4iFhviadAlOrK2Zt9wEH6XjSyFdxFHJ5s2bNAVuGcnVHmCjYZjFmISWMduTUcrG3h_xXpa11-MarFdznfgx9lwfPol7kszxE6EuChXSVeWSlpr6gzPKItAk5eKV0a7EiEjmjW6yoTNlUJE5dMaX2KDU2nujarSXdhs2sa_A8Zra3hdpklqEZrhiSCEr2XtlMjLXHAVweelhIqrjoGjCMiFmwIFW5Bgi16wEezQwv_t1695BB-XMi3w9yGfiG18u5gXKUIwKaTmIoK3nYxXo5d7JAK9Jv1VB6LmXm9ppr8DRTfaAwidk2TvkQm9h-c0b7okTuQ-bF7PFv4DPHV_rqfz2QFu74k-CNv7DuM_9aI
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3db9MwED_BQAwe-NgGhE8j8cJDNsex4_iR0U6b1lZ9KGhvkeM4UGlKpnZFQvzz3DlpWRE88BbFdpT4HN_9fHe_A3jPK9RDmdRx5vMqlmjTx8bmJvaoihOny1SGANkvIz2Z5BcXZtonq4dcGO99CD7zh3QZfPlV61Z0VHZEIetkg9-GO0pKkXTpWhungVRdRIcg0lieibVXkpuj2fgMsaBQh6lCyKeIK1SSss1yuaWQQoWVvxmbf8ZM3lBCJ4_-8_Ufw8Pe2mQfu-XxBG75Zg8e3OAg3IN74967vg8_j-0PT0mVjEr4Ynu7YAMi1u1rYrEh7gddqiNra3aOg_C5bGopwItYOtm8YYOuxD07ptoTbDqcsRCVwDp-a9pc2cD7K9YTu35lVI3tcnkAn0-Gs0-ncV-cIXYiREa5Srmy1MpaU0vcpSxCTV5qXpnclQga0bDJKylspjViKp_WeBMbnE4zb3SVPoWdpm38c2C8tobXZZqkFsEZ7glC-FrVTouszATXEXxYS6i46jg4ioBduClQsAUJtugFG8EBTfzvfv2cR_BuLdMCfyDyitjGt6tlkSIIU0LlXETwrJPxZvR6jUSQb0l_04HIubdbmvm3QNKNFgGC5yR58Y8Xegu7p7PxqBidTc5fwn36BjoyTtQr2LlerPxruOu-X8-Xizdhkf8C_7T4AQ
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Bayesian+Posterior+Distribution+Estimation+of+Kinetic+Parameters+in+Dynamic+Brain+PET+Using+Generative+Deep+Learning+Models&rft.jtitle=IEEE+transactions+on+medical+imaging&rft.au=Djebra%2C+Yanis&rft.au=Liu%2C+Xiaofeng&rft.au=Marin%2C+Thibault&rft.au=Tiss%2C+Amal&rft.date=2025-07-15&rft.eissn=1558-254X&rft.volume=PP&rft_id=info:doi/10.1109%2FTMI.2025.3588859&rft_id=info%3Apmid%2F40663684&rft.externalDocID=40663684
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0278-0062&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0278-0062&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0278-0062&client=summon