XLIM Research Institute, URM CNRS 7252, University of Poitiers, France; I3M, Common Laboratory CNRS-Siemens-Healthineers, University and Hospital of Poitiers, France.
XLIM Research Institute, URM CNRS 7252, University of Poitiers, France; I3M, Common Laboratory CNRS-Siemens-Healthineers, University and Hospital of Poitiers, France.
Comput Med Imaging Graph. 2022 Jul;99:102074. doi: 10.1016/j.compmedimag.2022.102074. Epub 2022 May 27.
Imaging bio-markers have been widely used for Computer-Aided Diagnosis (CAD) of Alzheimer's Disease (AD) with Deep Learning (DL). However, the structural brain atrophy is not detectable at an early stage of the disease (namely for Mild Cognitive Impairment (MCI) and Mild Alzheimer's Disease (MAD)). Indeed, potential biological bio-markers have been proved their ability to early detect brain abnormalities related to AD before brain structural damage and clinical manifestation. Proton Magnetic Resonance Spectroscopy (H-MRS) provides a promising solution for biological brain changes detection in a no invasive manner. In this paper, we propose an attention-guided supervised DL framework for early AD detection using H-MRS data. In the early stages of AD, features may be closely related and often complex to delineate between subjects. Hence, we develop a 1D attention mechanism that explicitly guides the classifier to focus on diagnostically relevant metabolites for classes discrimination. Synthetic data are used to tackle the lack of data problem and to help in learning the feature space. Data used in this paper are collected in the University Hospital of Poitiers, which contained 111 H-MRS samples extracted from the Posterior Cingulate Cortex (PCC) brain region. The data contain 33 Normal Control (NC), 49 MCI due to AD, and 29 MAD subjects. The proposed model achieves an average classification accuracy of 95.23%. Our framework outperforms state of the art imaging-based approaches, proving the robustness of learning metabolites features against traditional imaging bio-markers for early AD detection.
成像生物标志物已广泛应用于基于深度学习的阿尔茨海默病（AD）计算机辅助诊断（CAD）。然而，在疾病的早期阶段（即轻度认知障碍（MCI）和轻度 AD（MAD）），结构脑萎缩无法检测到。事实上，已经证明潜在的生物生物标志物具有在脑结构损伤和临床表现之前早期检测与 AD 相关的脑异常的能力。质子磁共振波谱（H-MRS）提供了一种有前途的非侵入性方法来检测生物脑变化。在本文中，我们提出了一种使用 H-MRS 数据进行早期 AD 检测的注意力引导监督深度学习框架。在 AD 的早期阶段，特征可能密切相关，并且通常难以区分主体之间的特征。因此，我们开发了一种 1D 注意力机制，该机制明确指导分类器关注用于分类的诊断相关代谢物。合成数据用于解决数据缺乏的问题，并帮助学习特征空间。本文使用的数据是在普瓦提埃大学医院收集的，包含从后扣带回皮质（PCC）脑区提取的 111 个 H-MRS 样本。数据包含 33 个正常对照组（NC）、49 个 AD 所致 MCI 和 29 个 MAD 患者。所提出的模型实现了平均分类准确率为 95.23%。我们的框架优于最先进的基于成像的方法，证明了学习代谢物特征对早期 AD 检测的抗传统成像生物标志物的稳健性。
