Dina Seyedi
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Dina Seyedi has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Shahriar Kolahi
Advanced Diagnostic and Interventional Radiology Research Center (ADIR), Tehran University of Medical Sciences, Tehran, Iran
2 shared publications - Ali Darzi
School of Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran
1 shared publication - Ali Mousavi
Student Research Committee, Tabriz University of Medical Sciences, Tabriz, Iran
1 shared publication Ali Sareminia
Shiraz University of Medical Sciences, Shiraz, Iran
1 shared publicationAli Shahriari
Department of Radiation Oncology, Dana Farber Cancer Institute, Brigham and Women’s Hospital, Harvard Medical School, Boston, USA
1 shared publication- Alireza Ghaedamini
Neurosurgery Department, Kerman University of Medical Sciences, Kerman, Iran
1 shared publication
Latest Publications
- Open Access
- 20 September 2026
- Published by Wiley in Human Brain Mapping
- Vol. 47 (14) , e70649
DTI ‐ ALPS as a Biomarker of Small Vessel Disease Progression and Amyloid‐β in Normal Aging: A 3‐Year Longitudinal Study
Cerebral small vessel disease (SVD) is a leading cause of stroke, vascular cognitive impairment, and functional decline in older adults. Emerging experimental and translational data implicate glymphatic dysfunction in SVD pathogenesis and in vascular amyloid accumulation. The analysis along the perivascular space (ALPS) index, derived from diffusion tensor imaging (DTI), noninvasively estimates water diffusivity along perivascular pathways and has been proposed as an indirect imaging marker related to glymphatic function. In this study, we aim to determine whether the baseline DTI‐ALPS index is associated with baseline SVD burden and subsequent longitudinal changes in SVD markers and amyloid‐β deposition in cognitively normal aging. We analyzed 204 cognitively normal participants from the Harvard Aging Brain Study with baseline and follow‐up visits separated by 3 years. Primary MRI SVD outcomes were intracranial‐volume–normalized white matter hyperintensities (WMH/ICV), and visual SVD ratings (Fazekas, ARWMC, perivascular space, and BOMBS [microbleed scale]). Secondary outcomes included cortical amyloid PET burden (PIB_FS_DVR_FLR). We tested the effects of time, baseline ALPS ( z ‐scored), and the time‐ALPS interaction using Bayesian mixed‐effects models with subject‐specific random intercepts adjusted for baseline age, sex, and education; amyloid models were additionally adjusted for APOE4 status. Continuous outcomes were modeled with Gaussian mixed‐effects models, binary outcomes with Bernoulli mixed‐effects models, and ordinal outcomes with Bayesian cumulative mixed‐effects models. In a sensitivity analysis, ALPS was recalculated after excluding voxels with λ 2 / λ 3 > 1.8, and the models were repeated using the crossing‐fiber‐adjusted index. Posterior means and 95% HDIs were the primary summaries. Multiplicity was controlled using the Benjamini–Hochberg false discovery rate procedure, and the resulting FDR‐adjusted two‐sided posterior tail‐area p ‐analogues are reported as p corrected . Over the 3‐year follow‐up, SVD and amyloid‐β burden increased: WMH/ICV ( β = 0.096, 95% HDI [0.069, 0.121], p corrected < 0.001), cortical amyloid‐β ( β = 0.130, 95% HDI [0.106, 0.152], p corrected < 0.001), Fazekas ( β = 0.320, 95% HDI [0.088, 0.541], p corrected = 0.010), ARWMC ( β = 0.404, 95% HDI [0.165, 0.649], p corrected = 0.004), basal ganglia ( β = 0.318, 95% HDI [0.079, 0.546], p corrected = 0.010), and centrum semiovale ( β = 0.244, 95% HDI [0.026, 0.450], p corrected = 0.025). PVS scores also increased over time. Higher baseline ALPS was associated with lower Fazekas severity ( β = −0.742, 95% HDI [−1.281, −0.201], p corrected = 0.026) and with attenuated amyloid accumulation over time ( β = −0.025, 95% HDI [−0.049, −0.001], p corrected = 0.039). In a sensitivity analysis excluding voxels with a high likelihood of substantial fiber crossing, the ALPS‐by‐time interaction for amyloid burden remained significant ( β = −0.028, 95% HDI [−0.052, −0.003], p corrected = 0.022), whereas the baseline association with Fazekas severity was attenuated ( β = −0.413, 95% HDI [−0.966, 0.181], p corrected = 0.751). Higher baseline WMH/ICV was associated with poorer baseline global cognition (PACC5) ( β = −0.092, 95% HDI [−0.177, −0.011], p corrected = 0.036), processing speed ( β = −0.116, 95% HDI [−0.223, −0.010], p corrected = 0.036), and executive function ( β = −0.144, 95% HDI [−0.251, −0.029], p corrected = 0.035). Higher baseline amyloid burden was associated with steeper PACC5 decline ( β = −0.065, 95% HDI [−0.096, −0.033], p corrected < 0.001). In cognitively normal older adults, higher conventional baseline ALPS was associated with lower Fazekas severity, although attenuation after crossing‐fiber adjustment indicates that this cross‐sectional relationship was sensitive to the microstructural composition of the ALPS regions. In contrast, higher baseline ALPS was associated with slower short‐term amyloid‐β accumulation, and this longitudinal association remained significant after crossing‐fiber adjustment. Greater baseline amyloid burden was associated with more rapid decline in global cognition, whereas higher baseline WMH volume was associated with poorer baseline global cognition, processing speed, and executive function. These findings support ALPS as a candidate noninvasive diffusion marker associated with longitudinal amyloid‐related change, while emphasizing the need for further validation of its relationship with SVD burden and its biological specificity.
- 20 September 2026
- Published by Wiley in Human Brain Mapping
- Vol. 47 (14) , e70649
Open AccessDTI ‐ ALPS as a Biomarker of Small Vessel Disease Progression and Amyloid‐β in Normal Aging: A 3‐Year Longitudinal Study
Cerebral small vessel disease (SVD) is a leading cause of stroke, vascular cognitive impairment, and functional decline in older adults. Emerging experimental and translational data implicate glymphatic dysfunction in SVD pathogenesis and in vascular amyloid accumulation. The analysis along the perivascular space (ALPS) index, derived from diffusion tensor imaging (DTI), noninvasively estimates water diffusivity along perivascular pathways and has been proposed as an indirect imaging marker related to glymphatic function. In this study, we aim to determine whether the baseline DTI‐ALPS index is associated with baseline SVD burden and subsequent longitudinal changes in SVD markers and amyloid‐β deposition in cognitively normal aging. We analyzed 204 cognitively normal participants from the Harvard Aging Brain Study with baseline and follow‐up visits separated by 3 years. Primary MRI SVD outcomes were intracranial‐volume–normalized white matter hyperintensities (WMH/ICV), and visual SVD ratings (Fazekas, ARWMC, perivascular space, and BOMBS [microbleed scale]). Secondary outcomes included cortical amyloid PET burden (PIB_FS_DVR_FLR). We tested the effects of time, baseline ALPS ( z ‐scored), and the time‐ALPS interaction using Bayesian mixed‐effects models with subject‐specific random intercepts adjusted for baseline age, sex, and education; amyloid models were additionally adjusted for APOE4 status. Continuous outcomes were modeled with Gaussian mixed‐effects models, binary outcomes with Bernoulli mixed‐effects models, and ordinal outcomes with Bayesian cumulative mixed‐effects models. In a sensitivity analysis, ALPS was recalculated after excluding voxels with λ 2 / λ 3 > 1.8, and the models were repeated using the crossing‐fiber‐adjusted index. Posterior means and 95% HDIs were the primary summaries. Multiplicity was controlled using the Benjamini–Hochberg false discovery rate procedure, and the resulting FDR‐adjusted two‐sided posterior tail‐area p ‐analogues are reported as p corrected . Over the 3‐year follow‐up, SVD and amyloid‐β burden increased: WMH/ICV ( β = 0.096, 95% HDI [0.069, 0.121], p corrected < 0.001), cortical amyloid‐β ( β = 0.130, 95% HDI [0.106, 0.152], p corrected < 0.001), Fazekas ( β = 0.320, 95% HDI [0.088, 0.541], p corrected = 0.010), ARWMC ( β = 0.404, 95% HDI [0.165, 0.649], p corrected = 0.004), basal ganglia ( β = 0.318, 95% HDI [0.079, 0.546], p corrected = 0.010), and centrum semiovale ( β = 0.244, 95% HDI [0.026, 0.450], p corrected = 0.025). PVS scores also increased over time. Higher baseline ALPS was associated with lower Fazekas severity ( β = −0.742, 95% HDI [−1.281, −0.201], p corrected = 0.026) and with attenuated amyloid accumulation over time ( β = −0.025, 95% HDI [−0.049, −0.001], p corrected = 0.039). In a sensitivity analysis excluding voxels with a high likelihood of substantial fiber crossing, the ALPS‐by‐time interaction for amyloid burden remained significant ( β = −0.028, 95% HDI [−0.052, −0.003], p corrected = 0.022), whereas the baseline association with Fazekas severity was attenuated ( β = −0.413, 95% HDI [−0.966, 0.181], p corrected = 0.751). Higher baseline WMH/ICV was associated with poorer baseline global cognition (PACC5) ( β = −0.092, 95% HDI [−0.177, −0.011], p corrected = 0.036), processing speed ( β = −0.116, 95% HDI [−0.223, −0.010], p corrected = 0.036), and executive function ( β = −0.144, 95% HDI [−0.251, −0.029], p corrected = 0.035). Higher baseline amyloid burden was associated with steeper PACC5 decline ( β = −0.065, 95% HDI [−0.096, −0.033], p corrected < 0.001). In cognitively normal older adults, higher conventional baseline ALPS was associated with lower Fazekas severity, although attenuation after crossing‐fiber adjustment indicates that this cross‐sectional relationship was sensitive to the microstructural composition of the ALPS regions. In contrast, higher baseline ALPS was associated with slower short‐term amyloid‐β accumulation, and this longitudinal association remained significant after crossing‐fiber adjustment. Greater baseline amyloid burden was associated with more rapid decline in global cognition, whereas higher baseline WMH volume was associated with poorer baseline global cognition, processing speed, and executive function. These findings support ALPS as a candidate noninvasive diffusion marker associated with longitudinal amyloid‐related change, while emphasizing the need for further validation of its relationship with SVD burden and its biological specificity.
- 11 December 2025
- Published by Elsevier in Seminars in Arthritis and Rheumatism
- Vol. 76, 152892
Diffuse idiopathic skeletal hyperostosis in the oncologic population: a cross-sectional analysis of 1053 patients
- 11 December 2025
- Published by Elsevier in Seminars in Arthritis and Rheumatism
- Vol. 76, 152892
Diffuse idiopathic skeletal hyperostosis in the oncologic population: a cross-sectional analysis of 1053 patients
- Open Access
- 26 August 2025
- Published by Springer Nature in Cancer Imaging
- Vol. 25 (1) , 1-14
Machine Learning–Driven radiomics on 18 F-FDG PET for glioma diagnosis: a systematic review and meta-analysis
Machine learning (ML) applied to radiomics has revolutionized neuro-oncological imaging, yet the diagnostic performance of ML models based specifically on ^18F-FDG PET features in glioma remains poorly characterized. To systematically evaluate and quantitatively synthesize the diagnostic accuracy of ML models trained on ^18F-FDG PET radiomics for glioma classification. We conducted a PRISMA-compliant systematic review and meta-analysis registered on OSF ( https://doi.org/10.17605/OSF.IO/XJG6P ). PubMed, Scopus, and Web of Science were searched up to January 2025. Studies were included if they applied ML algorithms to ^18F-FDG PET radiomic features for glioma classification and reported at least one performance metric. Data extraction included demographics, imaging protocols, feature types, ML models, and validation design. Meta-analysis was performed using random-effects models with pooled estimates of accuracy, sensitivity, specificity, AUC, F1 score, and precision. Heterogeneity was explored via meta-regression and Galbraith plots. Twelve studies comprising 2,321 patients were included. Pooled diagnostic metrics were: accuracy 92.6% (95% CI: 91.3–93.9%), AUC 0.95 (95% CI: 0.94–0.95), sensitivity 85.4%, specificity 89.7%, F1 score 0.78, and precision 0.90. Heterogeneity was high across all domains (I² >75%). Meta-regression identified ML model type and validation strategy as partial moderators. Models using CNNs or PET/MRI integration achieved superior performance. ML models based on ^18F-FDG PET radiomics demonstrate strong and balanced diagnostic performance for glioma classification. However, methodological heterogeneity underscores the need for standardized pipelines, external validation, and transparent reporting before clinical integration.
- 26 August 2025
- Published by Springer Nature in Cancer Imaging
- Vol. 25 (1) , 1-14
Open AccessMachine Learning–Driven radiomics on 18 F-FDG PET for glioma diagnosis: a systematic review and meta-analysis
Machine learning (ML) applied to radiomics has revolutionized neuro-oncological imaging, yet the diagnostic performance of ML models based specifically on ^18F-FDG PET features in glioma remains poorly characterized. To systematically evaluate and quantitatively synthesize the diagnostic accuracy of ML models trained on ^18F-FDG PET radiomics for glioma classification. We conducted a PRISMA-compliant systematic review and meta-analysis registered on OSF ( https://doi.org/10.17605/OSF.IO/XJG6P ). PubMed, Scopus, and Web of Science were searched up to January 2025. Studies were included if they applied ML algorithms to ^18F-FDG PET radiomic features for glioma classification and reported at least one performance metric. Data extraction included demographics, imaging protocols, feature types, ML models, and validation design. Meta-analysis was performed using random-effects models with pooled estimates of accuracy, sensitivity, specificity, AUC, F1 score, and precision. Heterogeneity was explored via meta-regression and Galbraith plots. Twelve studies comprising 2,321 patients were included. Pooled diagnostic metrics were: accuracy 92.6% (95% CI: 91.3–93.9%), AUC 0.95 (95% CI: 0.94–0.95), sensitivity 85.4%, specificity 89.7%, F1 score 0.78, and precision 0.90. Heterogeneity was high across all domains (I² >75%). Meta-regression identified ML model type and validation strategy as partial moderators. Models using CNNs or PET/MRI integration achieved superior performance. ML models based on ^18F-FDG PET radiomics demonstrate strong and balanced diagnostic performance for glioma classification. However, methodological heterogeneity underscores the need for standardized pipelines, external validation, and transparent reporting before clinical integration.