IoT and Signal Processing Research Group, ICT Research Institute, Faculty of Intelligent Systems Engineering and Data Science, Persian Gulf University, Bushehr, Iran.
IoT and Signal Processing Research Group, ICT Research Institute, Faculty of Intelligent Systems Engineering and Data Science, Persian Gulf University, Bushehr, Iran.
Clin Imaging. 2024 Nov;115:110301. doi: 10.1016/j.clinimag.2024.110301. Epub 2024 Sep 16.
Alzheimer's disease (AD) is a common neurodegenerative disorder that primarily affects older individuals. Due to its high incidence, an accurate and efficient stratification system could greatly aid in the clinical diagnosis and prognosis of AD patients. Convolutional neural networks (CNN) approaches have demonstrated exceptional performance in the automated stratification of AD, mild cognitive impairment (MCI) and cognitively normal (CN) participants using MRI, owing to their high predictive accuracy and reliability. Therefore, we aimed to develop an algorithm based on CNN and radiomic features derived from ROIs of bilateral hippocampus and amygdala in brain MRI for stratification between AD, MCI and CN.
In this study, we proposed a CNN and radiomic features-based algorithm using the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. T1-weighted images were used. We utilized three datasets, including AD (199 cases, 602 images), MCI (200 cases, 948 images), and CN (200 cases, 853 images), to perform binary classification (AD vs. CN, AD vs. MCI, and MCI vs. CN). Finally, we obtained the accuracy (ACC) and the area under the curve of the receiver operating characteristic curve (AUC) to evaluate the performance of the algorithm.
Our proposed algorithm achieved acceptable overall discrimination accuracy. In the term of AD vs CN, radiomic-based algorithm alone obtained ACC of 82.6 % and AUC of 88.8, CNN-based algorithm obtained ACC of 80 % and AUC of 87.2 and their fusion showed ACC of 84.4 % and AUC of 90. In the term of MCI vs CN, radiomic-based algorithm alone obtained ACC of 71.6 % and AUC of 77.8, CNN-based algorithm obtained ACC of 69 % and AUC of 75 and their fusion showed ACC of 72.7 % and AUC of 80. In the term of AD vs MCI, radiomic-based algorithm alone obtained ACC of 57 % and AUC of 57.5, CNN-based algorithm obtained ACC of 56.6 % and AUC of 57.7 and their fusion showed ACC of 58 % and AUC of 59.5.
In conclusion, it has been determined that hippocampus and amygdala-based stratification using CNN features and radiomic features-based algorithm is a promising method for the classification of AD, MCI, and CN participants.
This study proposed an automated procedures based on MRI-derived radiomic features and CNN for classification between AD, MCI and CN.
阿尔茨海默病（AD）是一种常见的神经退行性疾病，主要影响老年人。由于其发病率高，一个准确、高效的分层系统可以极大地帮助 AD 患者的临床诊断和预后。卷积神经网络（CNN）方法在使用 MRI 对 AD、轻度认知障碍（MCI）和认知正常（CN）参与者进行自动分层方面表现出了优异的性能，因为它们具有较高的预测准确性和可靠性。因此，我们旨在开发一种基于 CNN 和从双侧海马体和杏仁核 ROI 提取的放射组学特征的算法，用于对 AD、MCI 和 CN 进行分层。
在这项研究中，我们使用阿尔茨海默病神经影像学倡议（ADNI）数据库提出了一种基于 CNN 和放射组学特征的算法。使用 T1 加权图像。我们利用三个数据集，包括 AD（199 例，602 张图像）、MCI（200 例，948 张图像）和 CN（200 例，853 张图像），进行二分类（AD 与 CN、AD 与 MCI、MCI 与 CN）。最后，我们获得了准确性（ACC）和受试者工作特征曲线（ROC）的曲线下面积（AUC），以评估算法的性能。
我们提出的算法在整体判别准确性方面表现良好。在 AD 与 CN 的情况下，基于放射组学的算法单独获得了 82.6%的 ACC 和 88.8%的 AUC，基于 CNN 的算法获得了 80%的 ACC 和 87.2%的 AUC，融合算法获得了 84.4%的 ACC 和 90%的 AUC。在 MCI 与 CN 的情况下，基于放射组学的算法单独获得了 71.6%的 ACC 和 77.8%的 AUC，基于 CNN 的算法获得了 69%的 ACC 和 75%的 AUC，融合算法获得了 72.7%的 ACC 和 80%的 AUC。在 AD 与 MCI 的情况下，基于放射组学的算法单独获得了 57%的 ACC 和 57.5%的 AUC，基于 CNN 的算法获得了 56.6%的 ACC 和 57.7%的 AUC，融合算法获得了 58%的 ACC 和 59.5%的 AUC。
总之，已经确定使用 CNN 特征和基于放射组学特征的算法对海马体和杏仁核进行分层是一种有前途的方法，可用于 AD、MCI 和 CN 参与者的分类。
本研究提出了一种基于 MRI 衍生的放射组学特征和 CNN 的自动处理程序，用于 AD、MCI 和 CN 之间的分类。
