1
|
Lin S, Xue M, Sun J, Xu C, Wang T, Lian J, Lv M, Yang P, Sheng C, Cheng Z, Wang W. MRI Radiomics Nomogram for Predicting Disease Transition Time and Risk Stratification in Preclinical Alzheimer's Disease. Acad Radiol 2024:S1076-6332(24)00646-9. [PMID: 39332990 DOI: 10.1016/j.acra.2024.08.059] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/11/2024] [Revised: 08/20/2024] [Accepted: 08/30/2024] [Indexed: 09/29/2024]
Abstract
RATIONALE AND OBJECTIVES Accurate prediction of the progression of preclinical Alzheimer's disease (AD) is crucial for improving clinical management and disease prognosis. The objective of this study was to develop and validate clinical-radimoics integrated model to predict the time to progression (TTP) and disease risk stratification of preclinical AD. MATERIALS AND METHODS A total of 244 cases (mean age: 73.8 ± 5.5 years, 120 women) from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database were randomly divided into the training cohort (n = 172) and validation cohort (n = 72) using a 7:3 ratio. Clinical factors were identified by univariate and multivariate COX regression. Radiomics features were extracted from GM, WM and CSF of T1WI images and selected by Spearman correlation analysis and least absolute shrinkage and selection operator (LASSO). Using selected clinical factors and radiomics features, the clinical, radimocis and clinical-radiomics nomogram models were developed for predicting the TTP. The performance of each model was assessed by C-index. The risk stratification ability and predicting efficacy of the clinical-radiomics model were utilizing the Kaplan-Meier curve and receiver operator characteristic (ROC) curve. RESULTS The C-index of clinical, radimocis and clinical-radiomics models were 0.852 (95% confidence interval[CI]:0.810-0.893), 0.863 (95%CI:0.816-0.910) and 0.903 (95%:0.870-0.936) in the training cohort and 0.725 (95%CI:0.630-0.820), 0.788 (95%CI:0.678-0.898), 0.813(95%CI:0.734-0.892) in the validation cohort. The AUCs of the multi-predictor nomogram at 1-, 3-, 5- and 7-year were 0.894, 0.908, 0.930, 0.979 in the training cohort and 0.671, 0.726, 0.839, 0.931 in the validation cohort. CONCLUSION In this study, we constructed a clinical-radimoics integrated model to predict the progression of preclinical AD and stratified the risk of disease progression in preclinical AD.
Collapse
Affiliation(s)
- Shuai Lin
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | - Ming Xue
- Department of Radiology, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | - Jiali Sun
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | - Chang Xu
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | - Tianqi Wang
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | | | - Min Lv
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | - Ping Yang
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | - Chenjun Sheng
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | - Zijian Cheng
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China
| | - Wei Wang
- Department of MRI, First Affiliated Hospital of Harbin Medical University, Harbin, China.
| |
Collapse
|
2
|
Krüger DM, Pena-Centeno T, Liu S, Park T, Kaurani L, Pradhan R, Huang YN, Risacher SL, Burkhardt S, Schütz AL, Wan Y, Shaw LM, Brodsky AS, DeStefano AL, Lin H, Schroeder R, Krunic A, Hempel N, Sananbenesi F, Blusztajn JK, Saykin AJ, Delalle I, Nho K, Fischer A. The plasma miRNAome in ADNI: Signatures to aid the detection of at-risk individuals. Alzheimers Dement 2024. [PMID: 39291752 DOI: 10.1002/alz.14157] [Citation(s) in RCA: 1] [Impact Index Per Article: 1.0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/06/2024] [Revised: 07/08/2024] [Accepted: 07/09/2024] [Indexed: 09/19/2024]
Abstract
INTRODUCTION MicroRNAs are short non-coding RNAs that control proteostasis at the systems level and are emerging as potential prognostic and diagnostic biomarkers for Alzheimer's disease (AD). METHODS We performed small RNA sequencing on plasma samples from 847 Alzheimer's Disease Neuroimaging Initiative (ADNI) participants. RESULTS We identified microRNA signatures that correlate with AD diagnoses and help predict the conversion from mild cognitive impairment (MCI) to AD. DISCUSSION Our data demonstrate that plasma microRNA signatures can be used to not only diagnose MCI, but also, critically, predict the conversion from MCI to AD. Moreover, combined with neuropsychological testing, plasma microRNAome evaluation helps predict MCI to AD conversion. These findings are of considerable public interest because they provide a path toward reducing indiscriminate utilization of costly and invasive testing by defining the at-risk segment of the aging population. HIGHLIGHTS We provide the first analysis of the plasma microRNAome for the ADNI study. The levels of several microRNAs can be used as biomarkers for the prediction of conversion from MCI to AD. Adding the evaluation of plasma microRNA levels to neuropsychological testing in a clinical setting increases the accuracy of MCI to AD conversion prediction.
Collapse
Affiliation(s)
- Dennis M Krüger
- Department for Epigenetics and Systems Medicine in Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
- Bioinformatics Unit, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
| | - Tonatiuh Pena-Centeno
- Department for Epigenetics and Systems Medicine in Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
- Bioinformatics Unit, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
| | - Shiwei Liu
- Center for Computational Biology and Bioinformatics, Indiana University School of Medicine, Indianapolis, Indiana, USA
| | - Tamina Park
- Center for Computational Biology and Bioinformatics, Indiana University School of Medicine, Indianapolis, Indiana, USA
| | - Lalit Kaurani
- Department for Epigenetics and Systems Medicine in Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
| | - Ranjit Pradhan
- Department for Epigenetics and Systems Medicine in Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
| | - Yen-Ning Huang
- Center for Computational Biology and Bioinformatics, Indiana University School of Medicine, Indianapolis, Indiana, USA
| | - Shannon L Risacher
- Center for Computational Biology and Bioinformatics, Indiana University School of Medicine, Indianapolis, Indiana, USA
| | - Susanne Burkhardt
- Department for Epigenetics and Systems Medicine in Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
| | - Anna-Lena Schütz
- Research Group for Genome Dynamics in Brain Diseases, German Center for Neurodegenerative Diseases, Göttingen, Germany
| | - Yang Wan
- Perelman School of Medicine, Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA
| | - Leslie M Shaw
- Perelman School of Medicine, Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, Pennsylvania, USA
| | - Alexander S Brodsky
- Department of Pathology and Laboratory Medicine, Rhode Island Hospital, Warren Alpert Medical School at Brown University, Providence, Rhode Island, USA
| | - Anita L DeStefano
- Department of Biostatistics, Boston University School of Public Health, Boston, Massachusetts, USA
| | - Honghuang Lin
- Department of Medicine, UMass Chan Medical School, Worcester, Massachusetts, USA
| | - Robert Schroeder
- Department for Epigenetics and Systems Medicine in Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
| | - Andre Krunic
- Department of Pathology & Laboratory Medicine, Boston University Chobanian & Avedisian School of Medicine, Boston, Massachusetts, USA
| | - Nina Hempel
- Department for Epigenetics and Systems Medicine in Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
| | - Farahnaz Sananbenesi
- Research Group for Genome Dynamics in Brain Diseases, German Center for Neurodegenerative Diseases, Göttingen, Germany
| | - Jan Krzysztof Blusztajn
- Department of Pathology & Laboratory Medicine, Boston University Chobanian & Avedisian School of Medicine, Boston, Massachusetts, USA
| | - Andrew J Saykin
- Department of Radiology and Imaging Sciences and the Indiana Alzheimer's Disease Research Center, Indiana University School of Medicine, Indianapolis, Indiana, USA
| | - Ivana Delalle
- Department of Pathology & Laboratory Medicine, Boston University Chobanian & Avedisian School of Medicine, Boston, Massachusetts, USA
| | - Kwangsik Nho
- Department of Radiology and Imaging Sciences and the Indiana Alzheimer's Disease Research Center, Indiana University School of Medicine, Indianapolis, Indiana, USA
| | - Andre Fischer
- Department for Epigenetics and Systems Medicine in Neurodegenerative Diseases, German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany
- Department for Psychiatry and Psychotherapy, University Medical Center of Göttingen, Georg-August University, Göttingen, Germany
- Cluster of Excellence "Multiscale Bioimaging: from Molecular Machines to Networks of Excitable Cells" (MBExC), University of Göttingen, Göttingen, Germany
- German Center for Cardiovascular Diseases (DZKH) Göttingen, Göttingen, Germany
| |
Collapse
|
3
|
Jack CR, Arani A, Borowski BJ, Cash DM, Crawford K, Das SR, DeCarli C, Fletcher E, Fox NC, Gunter JL, Ittyerah R, Harvey DJ, Jahanshad N, Maillard P, Malone IB, Nir TM, Reid RI, Reyes DA, Schwarz CG, Senjem ML, Thomas DL, Thompson PM, Tosun D, Yushkevich PA, Ward CP, Weiner MW. Overview of ADNI MRI. Alzheimers Dement 2024. [PMID: 39258539 DOI: 10.1002/alz.14166] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/06/2024] [Revised: 07/09/2024] [Accepted: 07/11/2024] [Indexed: 09/12/2024]
Abstract
The magnetic resonance imaging (MRI) Core has been operating since Alzheimer's Disease Neuroimaging Initiative's (ADNI) inception, providing 20 years of data including reliable, multi-platform standardized protocols, carefully curated image data, and quantitative measures provided by expert investigators. The overarching purposes of the MRI Core include: (1) optimizing and standardizing MRI acquisition methods, which have been adopted by many multicenter studies and trials worldwide and (2) providing curated images and numeric summary values from relevant MRI sequences/contrasts to the scientific community. Over time, ADNI MRI has become increasingly complex. To remain technically current, the ADNI MRI protocol has changed substantially over the past two decades. The ADNI 4 protocol contains nine different imaging types (e.g., three dimensional [3D] T1-weighted and fluid-attenuated inversion recovery [FLAIR]). Our view is that the ADNI MRI data are a greatly underutilized resource. The purpose of this paper is to educate the scientific community on ADNI MRI methods and content to promote greater awareness, accessibility, and use. HIGHLIGHTS: The MRI Core provides multi-platform standardized protocols, carefully curated image data, and quantitative analysis by expert groups. The ADNI MRI protocol has undergone major changes over the past two decades to remain technically current. As of April 25, 2024, the following numbers of image series are available: 17,141 3D T1w; 6877 FLAIR; 3140 T2/PD; 6623 GRE; 3237 dMRI; 2846 ASL; 2968 TF-fMRI; and 2861 HighResHippo (see Table 1 for abbreviations). As of April 25, 2024, the following numbers of quantitative analyses are available: FreeSurfer 10,997; BSI 6120; tensor based morphometry (TBM) and TBM-SYN 12,019; WMH 9944; dMRI 1913; ASL 925; TF-fMRI NFQ 2992; and medial temporal subregion volumes 2726 (see Table 4 for abbreviations). ADNI MRI is an underutilized resource that could be more useful to the research community.
Collapse
Affiliation(s)
- Clifford R Jack
- Department of Radiology, Mayo Clinic, Rochester, Minnesota, USA
| | - Arvin Arani
- Department of Radiology, Mayo Clinic, Rochester, Minnesota, USA
| | - Bret J Borowski
- Department of Radiology, Mayo Clinic, Rochester, Minnesota, USA
| | - Dave M Cash
- Dementia Research Centre, University College London Institute of Neurology, Queen Square, London, UK
| | - Karen Crawford
- Laboratory of Neuro Imaging (LONI), University of Southern California, Los Angeles, California, USA
| | - Sandhitsu R Das
- Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, USA
| | - Charles DeCarli
- Department of Neurology, University of California, Davis, California, USA
| | - Evan Fletcher
- Department of Neurology, University of California, Davis, California, USA
| | - Nick C Fox
- Dementia Research Centre, University College London Institute of Neurology, Queen Square, London, UK
| | | | - Ranjit Ittyerah
- Department of Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, USA
| | - Danielle J Harvey
- Department of Public Health Sciences, Division of Biostatistics, University of California, Davis, California, USA
| | - Neda Jahanshad
- Keck School of Medicine of USC, Los Angeles, California, USA
| | - Pauline Maillard
- Department of Neurology, University of California, Davis, California, USA
| | - Ian B Malone
- Dementia Research Centre, University College London Institute of Neurology, Queen Square, London, UK
| | - Talia M Nir
- Keck School of Medicine of USC, Los Angeles, California, USA
| | - Robert I Reid
- Department of Radiology, Mayo Clinic, Rochester, Minnesota, USA
| | - Denise A Reyes
- Department of Radiology, Mayo Clinic, Rochester, Minnesota, USA
| | | | - Matthew L Senjem
- Department of Information Technology, Mayo Clinic, Rochester, Minnesota, USA
| | - David L Thomas
- Department of Brain Repair and Rehabilitation, UCL Queen Square Institute of Neurology, London, UK
| | - Paul M Thompson
- Laboratory of Neuro Imaging (LONI), University of Southern California, Los Angeles, California, USA
| | - Duygu Tosun
- Department of Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, California, USA
| | - Paul A Yushkevich
- Department of Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, USA
| | - Chadwick P Ward
- Department of Radiology, Mayo Clinic, Rochester, Minnesota, USA
| | - Michael W Weiner
- Department of Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, California, USA
| |
Collapse
|
4
|
Hu J, Huang B, Chen K. The impact of physical exercise on neuroinflammation mechanism in Alzheimer's disease. Front Aging Neurosci 2024; 16:1444716. [PMID: 39233828 PMCID: PMC11371602 DOI: 10.3389/fnagi.2024.1444716] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/21/2024] [Accepted: 08/07/2024] [Indexed: 09/06/2024] Open
Abstract
Introduction Alzheimer's disease (AD), a major cause of dementia globally, imposes significant societal and personal costs. This review explores the efficacy of physical exercise as a non-pharmacological intervention to mitigate the impacts of AD. Methods This review draws on recent studies that investigate the effects of physical exercise on neuroinflammation and neuronal enhancement in individuals with AD. Results Consistent physical exercise alters neuroinflammatory pathways, enhances cognitive functions, and bolsters brain health among AD patients. It favorably influences the activation states of microglia and astrocytes, fortifies the integrity of the blood-brain barrier, and attenuates gut inflammation associated with AD. These changes are associated with substantial improvements in cognitive performance and brain health indicators. Discussion The findings underscore the potential of integrating physical exercise into comprehensive AD management strategies. Emphasizing the necessity for further research, this review advocates for the refinement of exercise regimens to maximize their enduring benefits in decelerating the progression of AD.
Collapse
Affiliation(s)
- Junhui Hu
- School of Physical Education, West Anhui University, Lu'an, China
| | - Baiqing Huang
- School of Physical Education, Yunnan Minzu University, Kunming, China
| | - Kang Chen
- Tianjin Key Laboratory of Exercise Physiology and Sports Medicine, Tianjin University of Sport, Tianjin, China
| |
Collapse
|
5
|
Ding X, Yin L, Zhang L, Zhang Y, Zha T, Zhang W, Gui B. Diabetes accelerates Alzheimer's disease progression in the first year post mild cognitive impairment diagnosis. Alzheimers Dement 2024; 20:4583-4593. [PMID: 38865281 PMCID: PMC11247667 DOI: 10.1002/alz.13882] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/28/2023] [Revised: 02/28/2024] [Accepted: 03/18/2024] [Indexed: 06/14/2024]
Abstract
BACKGROUND Mild cognitive impairment (MCI) heightens Alzheimer's disease (AD) risk, with diabetes mellitus (DM) potentially exacerbating this vulnerability. This study identifies the optimal intervention period and neurobiological targets in MCI to AD progression using the Alzheimer's Disease Neuroimaging Initiative dataset. METHODS Analysis of 980 MCI patients, categorized by DM status, used propensity score matching and inverse probability treatment weighting to assess rate of conversion from MCI to AD, neuroimaging, and cognitive changes. RESULTS DM significantly correlates with cognitive decline and an increased risk of progressing to AD, especially within the first year of MCI follow-up. It adversely affects specific brain structures, notably accelerating nucleus accumbens atrophy, decreasing gray matter volume and sulcal depth. DISCUSSION Findings suggest the first year after MCI diagnosis as the critical window for intervention. DM accelerates MCI-to-AD progression, targeting specific brain areas, underscoring the need for early therapeutic intervention. HIGHLIGHTS Diabetes mellitus (DM) accelerates mild cognitive impairment (MCI)-to-Alzheimer's disease (AD) progression within the first year after MCI diagnosis. DM leads to sharper cognitive decline within 12 months of follow-up. There is notable nucleus accumbens atrophy observed in MCI patients with DM. DM causes significant reductions in gray matter volume and sulcal depth. There are stronger correlations between cognitive decline and brain changes due to DM.
Collapse
Affiliation(s)
- Xiahao Ding
- Department of AnesthesiologyNanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityNanjingChina
- Department of Anesthesiology and Perioperative MedicineThe First Affiliated Hospital of Nanjing Medical UniversityNanjingChina
| | - Li Yin
- Department of Anesthesiology and Perioperative MedicineThe First Affiliated Hospital of Nanjing Medical UniversityNanjingChina
| | - Lin Zhang
- Department of Anesthesiology and Perioperative MedicineThe First Affiliated Hospital of Nanjing Medical UniversityNanjingChina
| | - Yang Zhang
- Department of Anesthesiology and Perioperative MedicineThe First Affiliated Hospital of Nanjing Medical UniversityNanjingChina
| | - Tianming Zha
- Department of Anesthesiology and Perioperative MedicineThe First Affiliated Hospital of Nanjing Medical UniversityNanjingChina
| | - Wen Zhang
- Department of RadiologyNanjing Drum Tower Hospital, Affiliated Hospital of Medical School, Nanjing UniversityNanjingChina
- Medical Imaging Centerthe Affiliated Drum Tower Hospital, Medical School of Nanjing UniversityNanjingChina
- Institute of Medical Imaging and Artificial IntelligenceNanjing UniversityNanjingChina
| | - Bo Gui
- Department of Anesthesiology and Perioperative MedicineThe First Affiliated Hospital of Nanjing Medical UniversityNanjingChina
| | | |
Collapse
|
6
|
Noda K, Lim Y, Goto R, Sengoku S, Kodama K. Cost-effectiveness comparison between blood biomarkers and conventional tests in Alzheimer's disease diagnosis. Drug Discov Today 2024; 29:103911. [PMID: 38311028 DOI: 10.1016/j.drudis.2024.103911] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/15/2023] [Revised: 01/26/2024] [Accepted: 01/30/2024] [Indexed: 02/06/2024]
Abstract
Dementia management has evolved with drugs such as lecanemab, shifting management from palliative care to early diagnosis and intervention. However, the administration of these drugs presents challenges owing to the invasiveness, high cost and limited availability of amyloid-PET and cerebrospinal fluid tests for guiding drug administration. Our manuscript explores the potential of less invasive blood biomarkers as a diagnostic method, with a cost-effectiveness analysis and a comparison with traditional tests. Our findings suggest that blood biomarkers are a cost-effective alternative, but with lower accuracy, indicating the need for multiple specific biomarkers for precision. This underscores the importance of future research on new blood biomarkers and their clinical efficacy.
Collapse
Affiliation(s)
- Kenta Noda
- Graduate School of Design and Architecture, Nagoya City University, Nagoya 464-0083, Japan
| | | | - Rei Goto
- Graduate School of Health Management, Keio University, Fujisawa 252-0883, Kanagawa, Japan; Graduate School of Business Administration, Keio University, Yokohama 223-8526, Japan
| | - Shintaro Sengoku
- School of Environment and Society, Tokyo Institute of Technology, Tokyo 108-0023, Japan
| | - Kota Kodama
- Graduate School of Design and Architecture, Nagoya City University, Nagoya 464-0083, Japan; Ritsumeikan University, Osaka 567-8570, Japan; Faculty of Data Science, Nagoya City University, Nagoya 467-8501, Japan; Center for Research and Education on Drug Discovery, The Graduate School of Pharmaceutical Sciences, Hokkaido University, Sapporo 060-0812, Japan.
| |
Collapse
|
7
|
Dhinagar NJ, Thomopoulos SI, Laltoo E, Thompson PM. Counterfactual MRI Generation with Denoising Diffusion Models for Interpretable Alzheimer's Disease Effect Detection. BIORXIV : THE PREPRINT SERVER FOR BIOLOGY 2024:2024.02.05.578983. [PMID: 38370616 PMCID: PMC10871297 DOI: 10.1101/2024.02.05.578983] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 02/20/2024]
Abstract
Generative AI models have recently achieved mainstream attention with the advent of powerful approaches such as stable diffusion, DALL-E and MidJourney. The underlying breakthrough generative mechanism of denoising diffusion modeling can generate high quality synthetic images and can learn the underlying distribution of complex, high-dimensional data. Recent research has begun to extend these models to medical and specifically neuroimaging data. Typical neuroimaging tasks such as diagnostic classification and predictive modeling often rely on deep learning approaches based on convolutional neural networks (CNNs) and vision transformers (ViTs), with additional steps to help in interpreting the results. In our paper, we train conditional latent diffusion models (LDM) and denoising diffusion probabilistic models (DDPM) to provide insight into Alzheimer's disease (AD) effects on the brain's anatomy at the individual level. We first created diffusion models that could generate synthetic MRIs, by training them on real 3D T1-weighted MRI scans, and conditioning the generative process on the clinical diagnosis as a context variable. We conducted experiments to overcome limitations in training dataset size, compute time and memory resources, testing different model sizes, effects of pretraining, training duration, and latent diffusion models. We tested the sampling quality of the disease-conditioned diffusion using metrics to assess realism and diversity of the generated synthetic MRIs. We also evaluated the ability of diffusion models to conditionally sample MRI brains using a 3D CNN-based disease classifier relative to real MRIs. In our experiments, the diffusion models generated synthetic data that helped to train an AD classifier (using only 500 real training scans) -and boosted its performance by over 3% when tested on real MRI scans. Further, we used implicit classifier-free guidance to alter the conditioning of an encoded individual scan to its counterfactual (representing a healthy subject of the same age and sex) while preserving subject-specific image details. From this counterfactual image (where the same person appears healthy), a personalized disease map was generated to identify possible disease effects on the brain. Our approach efficiently generates realistic and diverse synthetic data, and may create interpretable AI-based maps for neuroscience research and clinical diagnostic applications.
Collapse
|