Diagnosis of Alzheimer's Disease via Multi-modality 3D Convolutional Neural Network (1902.09904v1)

Published 26 Feb 2019 in cs.CV

Abstract: Alzheimer's Disease (AD) is one of the most concerned neurodegenerative diseases. In the last decade, studies on AD diagnosis attached great significance to AI-based diagnostic algorithms. Among the diverse modality imaging data, T1-weighted MRI and 18F-FDGPET are widely researched for this task. In this paper, we propose a novel convolutional neural network (CNN) to fuse the multi-modality information including T1-MRI and FDG-PDT images around the hippocampal area for the diagnosis of AD. Different from the traditional machine learning algorithms, this method does not require manually extracted features, and utilizes the stateof-art 3D image-processing CNNs to learn features for the diagnosis and prognosis of AD. To validate the performance of the proposed network, we trained the classifier with paired T1-MRI and FDG-PET images using the ADNI datasets, including 731 Normal (NL) subjects, 647 AD subjects, 441 stable MCI (sMCI) subjects and 326 progressive MCI (pMCI) subjects. We obtained the maximal accuracies of 90.10% for NL/AD task, 87.46% for NL/pMCI task, and 76.90% for sMCI/pMCI task. The proposed framework yields comparative results against state-of-the-art approaches. Moreover, the experimental results have demonstrated that (1) segmentation is not a prerequisite by using CNN, (2) the hippocampal area provides enough information to give a reference to AD diagnosis. Keywords: Alzheimer's Disease, Multi-modality, Image Classification, CNN, Deep Learning, Hippocampal

View on arXiv

Authors (6)

Yechong Huang (4 papers)
Jiahang Xu (14 papers)
Yuncheng Zhou (5 papers)
Tong Tong (26 papers)
Xiahai Zhuang (66 papers)
the Alzheimer's Disease Neuroimaging Initiative (13 papers)

Citations (192)

View on Semantic Scholar

Summary

Alzheimer's Disease Diagnosis Using Multi-Modality 3D CNNs

The paper "Diagnosis of Alzheimer's Disease via Multi-modality 3D Convolutional Neural Network" presents a significant advancement in the classification of Alzheimer's Disease (AD) using deep learning techniques. The research primarily focuses on leveraging convolutional neural networks (CNNs) to integrate information from T1-weighted magnetic resonance imaging (MRI) and 18F-FDG positron emission tomography (PET) scans, targeting the hippocampal region for accurate diagnostic outcomes.

Methodology and Experimental Setup

This work utilizes state-of-the-art 3D CNN architectures without the need for manual feature extraction or segmentation of images. The 3D VGG variant introduced offers a streamlined approach to multi-modality classification, emphasizing object detection, feature extraction, and image classification—all critical components in the diagnostic workflow for neurodegenerative diseases like AD.

The datasets employed originate from the Alzheimer's Disease Neuroimaging Initiative (ADNI), which includes both normal (NL) and clinically diagnosed AD subjects, as well as patients with stable and progressive mild cognitive impairment (sMCI and pMCI). The experimental framework adopts a robust validation approach, sharing results on three distinct classification tasks: NL vs. AD, NL vs. pMCI, and sMCI vs. pMCI, with commendable accuracy rates of 90.10%, 87.46%, and 76.90%, respectively.

Key Findings and Contributions

One of the pivotal conclusions drawn is that segmentation of substructures such as the hippocampi is not essential when employing CNN-based classification, a departure from traditional methodologies in this space. Contrariwise, this research underscores the adequacy of a region of interest (ROI) surrounding the hippocampal area, providing enough data to infer diagnostic details relevant to AD.

The paper also effectively illustrates the potential benefits of multi-modality fusion in medical imaging via the CNN framework, showing improved classification performance compared to single-modality analyses. Notably, the multi-modality classifier demonstrated superior outcomes compared to reference methods in diagnosing preclinical stages of AD, crucial for timely intervention strategies.

Implications and Future Directions

The implications of these findings are significant for clinical diagnostics. By enhancing non-invasive imaging techniques for early-stage AD detection, this method could lead to increased efficacy in monitoring disease progression and tailoring treatment paths. The preference for high-resolution images of key brain regions informs future research strategies, potentially reducing computational load while retaining diagnostic precision.

Future explorations may incorporate additional imaging modalities, such as T2-MRI or alternative PET tracers, to further refine data integration and improve classification algorithms. Another avenue is enhancing interpretability of the features detected by CNNs, which could be facilitated through attention mechanisms or other advanced visualization methods.

In summary, the research contributes to a more efficient and less resource-intensive process in the clinical application of deep learning for the early detection of Alzheimer's Disease. By focusing on hippocampal data, the approach maximizes both computational efficiency and diagnostic accuracy, setting a foundation for further development in AI-based neurodegenerative disease analytics.

PDF Markdown

Related Papers

Find Related Papers