利用深度学习从多发性硬化症中分辨神经退行性变与衰老：脑预测的疾病持续时间差距。
Disentangling Neurodegeneration From Aging in Multiple Sclerosis Using Deep Learning: The Brain-Predicted Disease Duration Gap.
机构信息
From the Queen Square Multiple Sclerosis Centre (G.P., F.P., J.C., B.K., O.A.-M., S.A.-A., A. Bianchi, W.J.B., R. Christensen, E.C., S. Collorone, M.A.F., Y.H., A.H., S. Mohamud, R.N., A.T.T., J.W., C.Y., O.C., F.B.), Department of Neuroinflammation, UCL Queen Square Institute of Neurology, University College London, United Kingdom; MS Center Amsterdam (G.P., H.V., F.B.), Radiology and Nuclear Medicine, Vrije Universiteit Amsterdam, Amsterdam Neuroscience, Amsterdam UMC location VUmc, the Netherlands; Departments of Advanced Biomedical Sciences and Electrical Engineering and Information Technology (G.P., A. Brunetti, S. Cocozza), University of Naples "Federico II," Italy; Centre for Medical Image Computing (F.P., B.K., F.B.), Department of Medical Physics and Biomedical Engineering, University College London, United Kingdom; E-Health Center (F.P.), Universitat Oberta de Catalunya, Barcelona, Spain; Institute of Neuroradiology (B.B., C.L.), St. Josef Hospital, Ruhr-University Bochum, Germany; Department of Advanced Medical and Surgical Sciences (A. Bisecco, A.G.), University of Campania "Luigi Vanvitelli," Naples, Italy; Translational Imaging in Neurology (ThINK) Basel (A.C., C. Granziera), Department of Biomedical Engineering, Faculty of Medicine, University Hospital Basel, University of Basel; Neurologic Clinic and Policlinic (A.C., C. Granziera, J.K.), MS Center and Research Center for Clinical Neuroimmunology and Neuroscience Basel (RC2NB), University Hospital Basel, University of Basel, Switzerland; Department of Neurosciences, Biomedicine and Movement Sciences (M. Calabrese, M. Castellaro), University of Verona; Department of Information Engineering (M. Castellaro), University of Padova; Department of Medicine, Surgery and Neuroscience (R. Cortese, N.D.S.), University of Siena, Italy; Department of Neurology (C.E., D.P.), Medical University of Graz, Austria; Neuroimaging Research Unit (M.F., M.A.R., P.V.), Division of Neuroscience, IRCCS San Raffaele Scientific Institute, Neurology Unit, Neurorehabilitation Unit, Neurophysiology Service, IRCCS San Raffaele Scientific Institute; Vita-Salute San Raffaele University (M.F., M.A.R., P.V.), Milan; Department of Neurosciences (C. Gasperini, S.R.), San Camillo-Forlanini Hospital, Rome, Italy; Department of Neurology (G.G.-E., S.G.), Focus Program Translational Neuroscience (FTN) and Immunotherapy (FZI), Rhine Main Neuroscience Network (rmn2), University Medical Center of the Johannes Gutenberg University Mainz, Germany; Department of Neurology (H.F.F.H., E.A.H., G.O.N.), Oslo University Hospital; Institute of Clinical Medicine (H.F.F.H., E.A.H., G.O.N.), and Department of Psychology (E.A.H.), University of Oslo, Norway; Neuroimmunology and Multiple Sclerosis Unit Laboratory of Advanced Imaging in Neuroimmunological Diseases (ImaginEM) (S.L., E.M.-H.), Hospital Clinic Barcelona, Fundació de Recerca Clínic Barcelona-Institut d'Investigacions Biomèdiques August Pi i Su, Barcelona, Spain; Department of Neurology (C.L.), St. Josef Hospital, Ruhr-University Bochum, Germany; Nuffield Department of Clinical Neurosciences (S. Messina, J.P.), University of Oxford, United Kingdom; Department of Molecular Medicine and Medical Biotechnology (M.M.), and Department of Neurosciences and Reproductive and Odontostomatological Sciences (M.P.), University of Naples "Federico II"; Department of Human Neurosciences (M.P.), Sapienza University of Rome, Italy; Section of Neuroradiology (A.R.), Department of Radiology, and Centre d'Esclerosi Múltiple de Catalunya (Cemcat) (J.S.-G.), Department of Neurology/Neuroimmunology, Hospital Universitari Vall d'Hebron, Universitat Autònoma de Barcelona, Spain; MS Center Amsterdam (E.M.M.S.), Neurology, Vrije Universiteit Amsterdam, Amsterdam Neuroscience, Amsterdam UMC location VUmc, the Netherlands; Department of Neurology and Center of Clinical Neuroscience (T.U.), and Department of Radiology (M.V.), First Faculty of Medicine, Charles University and General University Hospital, Prague, Czech Republic; MS Center Amsterdam (M.M.S.), Anatomy and Neurosciences, Vrije Universiteit Amsterdam, Amsterdam Neuroscience, Amsterdam UMC location VUmc, the Netherlands; Centre for Medical Image Computing (J.H.C.), Department of Computer Science, and Dementia Research Centre (J.H.C., F.B.), UCL Queen Square Institute of Neurology, University College London, United Kingdom.
出版信息
Neurology. 2024 Nov 26;103(10):e209976. doi: 10.1212/WNL.0000000000209976. Epub 2024 Nov 4.
BACKGROUND AND OBJECTIVES
Disentangling brain aging from disease-related neurodegeneration in patients with multiple sclerosis (PwMS) is increasingly topical. The brain-age paradigm offers a window into this problem but may miss disease-specific effects. In this study, we investigated whether a disease-specific model might complement the brain-age gap (BAG) by capturing aspects unique to MS.
METHODS
In this retrospective study, we collected 3D T1-weighted brain MRI scans of PwMS to build (1) a cross-sectional multicentric cohort for age and disease duration (DD) modeling and (2) a longitudinal single-center cohort of patients with early MS as a clinical use case. We trained and evaluated a 3D DenseNet architecture to predict DD from minimally preprocessed images while age predictions were obtained with the DeepBrainNet model. The brain-predicted DD gap (the difference between predicted and actual duration) was proposed as a DD-adjusted global measure of MS-specific brain damage. Model predictions were scrutinized to assess the influence of lesions and brain volumes while the DD gap was biologically and clinically validated within a linear model framework assessing its relationship with BAG and physical disability measured with the Expanded Disability Status Scale (EDSS).
RESULTS
We gathered MRI scans of 4,392 PwMS (69.7% female, age: 42.8 ± 10.6 years, DD: 11.4 ± 9.3 years) from 15 centers while the early MS cohort included 749 sessions from 252 patients (64.7% female, age: 34.5 ± 8.3 years, DD: 0.7 ± 1.2 years). Our model predicted DD better than chance (mean absolute error = 5.63 years, = 0.34) and was nearly orthogonal to the brain-age model (correlation between DD and BAGs: = 0.06 [0.00-0.13], = 0.07). Predictions were influenced by distributed variations in brain volume and, unlike brain-predicted age, were sensitive to MS lesions (difference between unfilled and filled scans: 0.55 years [0.51-0.59], < 0.001). DD gap significantly explained EDSS changes ( = 0.060 [0.038-0.082], < 0.001), adding to BAG (Δ = 0.012, < 0.001). Longitudinally, increasing DD gap was associated with greater annualized EDSS change ( = 0.50 [0.39-0.60], < 0.001), with an incremental contribution in explaining disability worsening compared with changes in BAG alone (Δ = 0.064, < 0.001).
DISCUSSION
The brain-predicted DD gap is sensitive to MS-related lesions and brain atrophy, adds to the brain-age paradigm in explaining physical disability both cross-sectionally and longitudinally, and may be used as an MS-specific biomarker of disease severity and progression.
背景与目的
在多发性硬化症患者（PwMS）中，将大脑老化与与疾病相关的神经退行性变区分开来变得越来越重要。大脑年龄模型为此提供了一个窗口，但可能会错过特定于疾病的影响。在这项研究中，我们研究了一种特定于疾病的模型是否可以通过捕获 MS 特有的方面来补充大脑年龄差距（BAG）。
方法
在这项回顾性研究中，我们收集了 PwMS 的 3D T1 加权脑 MRI 扫描，以构建（1）用于年龄和疾病持续时间（DD）建模的横断面多中心队列，以及（2）作为临床用例的早期 MS 的纵向单中心队列。我们训练并评估了 3D DenseNet 架构，以从最小预处理的图像中预测 DD，同时使用 DeepBrainNet 模型获得大脑预测的 DD 差距（预测的 DD 与实际持续时间之间的差异），作为 DD 调整的 MS 特异性脑损伤的全局指标。我们仔细检查了模型预测，以评估病变和脑体积的影响，同时在线性模型框架内对 DD 差距进行了生物学和临床验证，评估其与 BAG 和用扩展残疾状况量表（EDSS）测量的身体残疾之间的关系。
结果
我们从 15 个中心收集了 4392 名 PwMS 的 MRI 扫描（69.7%为女性，年龄：42.8±10.6 岁，DD：11.4±9.3 年），而早期 MS 队列则包括 252 名患者的 749 次就诊（64.7%为女性，年龄：34.5±8.3 岁，DD：0.7±1.2 年）。我们的模型比偶然预测 DD 更好（平均绝对误差=5.63 年， =0.34），并且几乎与大脑年龄模型正交（DD 和 BAGs 之间的相关性： =0.06[0.00-0.13]， =0.07）。预测受到大脑体积分布变化的影响，与大脑预测年龄不同，它们对 MS 病变敏感（未填充和填充扫描之间的差异：0.55 年[0.51-0.59]， <0.001）。DD 差距显著解释了 EDSS 的变化（ =0.060[0.038-0.082]， <0.001），增加了 BAG（Δ=0.012， <0.001）。纵向研究显示，DD 差距的增加与 EDSS 年变化率的增加有关（ =0.50[0.39-0.60]， <0.001），与单独改变 BAG 相比，在解释残疾恶化方面具有更大的增量贡献（Δ=0.064， <0.001）。
讨论
大脑预测的 DD 差距对 MS 相关病变和脑萎缩敏感，在横断面上和纵向增加了大脑年龄模型在解释身体残疾方面的作用，并且可以用作 MS 特异性疾病严重程度和进展的生物标志物。
Zotero 插件