Feasibility of brain age predictions from clinical T1-weighted MRIs
An individual's brain predicted age minus chronological age (brain-PAD) obtained from MRIs could become a biomarker of disease in research studies. However, brain age reports from clinical MRIs are scant despite the rich clinical information hospitals provide. Since clinical MRI protocols are m...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2023-12-01
|
| Series: | Brain Research Bulletin |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S0361923023002368 |
| _version_ | 1797398719223562240 |
|---|---|
| author | Pedro A. Valdes-Hernandez Chavier Laffitte Nodarse James H. Cole Yenisel Cruz-Almeida |
| author_facet | Pedro A. Valdes-Hernandez Chavier Laffitte Nodarse James H. Cole Yenisel Cruz-Almeida |
| author_sort | Pedro A. Valdes-Hernandez |
| collection | DOAJ |
| description | An individual's brain predicted age minus chronological age (brain-PAD) obtained from MRIs could become a biomarker of disease in research studies. However, brain age reports from clinical MRIs are scant despite the rich clinical information hospitals provide. Since clinical MRI protocols are meant for specific clinical purposes, performance of brain age predictions on clinical data need to be tested. We explored the feasibility of using DeepBrainNet, a deep network previously trained on research-oriented MRIs, to predict the brain ages of 840 patients who visited 15 facilities of a health system in Florida. Anticipating a strong prediction bias in our clinical sample, we characterized it to propose a covariate model in group-level regressions of brain-PAD (recommended to avoid Type I, II errors), and tested its generalizability, a requirement for meaningful brain age predictions in new single clinical cases. The best bias-related covariate model was scanner-independent and linear in age, while the best method to estimate bias-free brain ages was the inverse of a scanner-independent and quadratic in brain age function. We demonstrated the feasibility to detect sex-related differences in brain-PAD using group-level regression accounting for the selected covariate model. These differences were preserved after bias correction. The Mean-Average Error (MAE) of the predictions in independent data was ∼8 years, 2–3 years greater than reports for research-oriented MRIs using DeepBrainNet, whereas an R2 (assuming no bias) was 0.33 and 0.76 for the uncorrected and corrected brain ages, respectively. DeepBrainNet on clinical populations seems feasible, but more accurate algorithms or transfer-learning retraining is needed. |
| first_indexed | 2024-03-09T01:29:32Z |
| format | Article |
| id | doaj.art-9a9d0fa8f7e04ba79517ed6e13eac0f5 |
| institution | Directory Open Access Journal |
| issn | 1873-2747 |
| language | English |
| last_indexed | 2024-03-09T01:29:32Z |
| publishDate | 2023-12-01 |
| publisher | Elsevier |
| record_format | Article |
| series | Brain Research Bulletin |
| spelling | doaj.art-9a9d0fa8f7e04ba79517ed6e13eac0f52023-12-10T06:13:45ZengElsevierBrain Research Bulletin1873-27472023-12-01205110811Feasibility of brain age predictions from clinical T1-weighted MRIsPedro A. Valdes-Hernandez0Chavier Laffitte Nodarse1James H. Cole2Yenisel Cruz-Almeida3Department of Community Dentistry and Behavioral Science, University of Florida, USA; Pain Research and Intervention Center of Excellence, University of Florida, USA; Center for Cognitive Aging and Memory, McKnight Brain Institute, University of Florida, USADepartment of Community Dentistry and Behavioral Science, University of Florida, USA; Pain Research and Intervention Center of Excellence, University of Florida, USA; Center for Cognitive Aging and Memory, McKnight Brain Institute, University of Florida, USACentre for Medical Image Computing, Department of Computer Science, University College London, UK; Dementia Research Centre, Queen Square Institute of Neurology, University College London, UKDepartment of Community Dentistry and Behavioral Science, University of Florida, USA; Pain Research and Intervention Center of Excellence, University of Florida, USA; Center for Cognitive Aging and Memory, McKnight Brain Institute, University of Florida, USA; Department of Neuroscience, College of Medicine, University of Florida, USA; Correspondence to: 1329 SW 16th Street, Ste. 5180 (zip 32608), Gainesville, FL 32610, USA.An individual's brain predicted age minus chronological age (brain-PAD) obtained from MRIs could become a biomarker of disease in research studies. However, brain age reports from clinical MRIs are scant despite the rich clinical information hospitals provide. Since clinical MRI protocols are meant for specific clinical purposes, performance of brain age predictions on clinical data need to be tested. We explored the feasibility of using DeepBrainNet, a deep network previously trained on research-oriented MRIs, to predict the brain ages of 840 patients who visited 15 facilities of a health system in Florida. Anticipating a strong prediction bias in our clinical sample, we characterized it to propose a covariate model in group-level regressions of brain-PAD (recommended to avoid Type I, II errors), and tested its generalizability, a requirement for meaningful brain age predictions in new single clinical cases. The best bias-related covariate model was scanner-independent and linear in age, while the best method to estimate bias-free brain ages was the inverse of a scanner-independent and quadratic in brain age function. We demonstrated the feasibility to detect sex-related differences in brain-PAD using group-level regression accounting for the selected covariate model. These differences were preserved after bias correction. The Mean-Average Error (MAE) of the predictions in independent data was ∼8 years, 2–3 years greater than reports for research-oriented MRIs using DeepBrainNet, whereas an R2 (assuming no bias) was 0.33 and 0.76 for the uncorrected and corrected brain ages, respectively. DeepBrainNet on clinical populations seems feasible, but more accurate algorithms or transfer-learning retraining is needed.http://www.sciencedirect.com/science/article/pii/S0361923023002368PatientsBrain-PADDeepBrainNetBrain age bias |
| spellingShingle | Pedro A. Valdes-Hernandez Chavier Laffitte Nodarse James H. Cole Yenisel Cruz-Almeida Feasibility of brain age predictions from clinical T1-weighted MRIs Brain Research Bulletin Patients Brain-PAD DeepBrainNet Brain age bias |
| title | Feasibility of brain age predictions from clinical T1-weighted MRIs |
| title_full | Feasibility of brain age predictions from clinical T1-weighted MRIs |
| title_fullStr | Feasibility of brain age predictions from clinical T1-weighted MRIs |
| title_full_unstemmed | Feasibility of brain age predictions from clinical T1-weighted MRIs |
| title_short | Feasibility of brain age predictions from clinical T1-weighted MRIs |
| title_sort | feasibility of brain age predictions from clinical t1 weighted mris |
| topic | Patients Brain-PAD DeepBrainNet Brain age bias |
| url | http://www.sciencedirect.com/science/article/pii/S0361923023002368 |
| work_keys_str_mv | AT pedroavaldeshernandez feasibilityofbrainagepredictionsfromclinicalt1weightedmris AT chavierlaffittenodarse feasibilityofbrainagepredictionsfromclinicalt1weightedmris AT jameshcole feasibilityofbrainagepredictionsfromclinicalt1weightedmris AT yeniselcruzalmeida feasibilityofbrainagepredictionsfromclinicalt1weightedmris |