Artificial intelligence utilising corneal confocal microscopy for the diagnosis of peripheral neuropathy in diabetes mellitus and prediabetes.

Artificial intelligence utilising corneal confocal microscopy for the diagnosis of peripheral neuropathy in diabetes mellitus and prediabetes.
复制标题

人工智能利用角膜共聚焦显微镜诊断糖尿病和前驱糖尿病周围神经病变。

DOI:
10.1007/s00125-021-05617-x
复制
发表时间:
2022-03
期刊:
影响因子:
8.2
通讯作者:
Alam U
Alam U
中科院分区:
医学1区
文献类型:
--
作者:
Preston FG;Meng Y;Burgess J;Ferdousi M;Azmi S;Petropoulos IN;Kaye S;Malik RA;Zheng Y;Alam U

文献摘要

参考文献

被引文献

相似文献

我们的目标是开发一种基于人工智能(AI)的深度学习算法(DLA),将不需要图像分割的归因方法应用于角膜共焦显微镜图像,并准确分类周围神经病变(或缺乏)。基于AI的DLA利用具有数据增强的卷积神经网络来提高算法的通用性。该算法使用高端图形处理器在329张角膜神经图像上训练了300个时期,并在40张图像上进行了测试(1张图像/参与者)。参与者包括健康志愿者(HV)参与者(n = 90)和1型糖尿病(n = 88),2型糖尿病(n = 141)和前驱糖尿病(n = 50)参与者(定义为空腹血糖受损、糖耐量受损或两者的组合),并被分类为HV,无神经病变(PN-)(n = 149)和神经病变(PN+)(n = 130)。对于基于AI的DLA,开发了一种名为ResNet-50的改进型残差神经网络,用于从图像中提取特征并进行分类。该算法在40名参与者(15名HV,13名PN-,12名PN+)中进行了测试。归因方法梯度加权类激活标测(Grad-CAM)、引导Grad-CAM和遮挡敏感性显示了图像中对算法决策影响最大的区域。结果如下:HV:召回1.0(95% CI 1.0,1.0),精密度为0.83(95% CI 0.65,1.0),F1评分为0.91(95% CI 0.79,1.0); PN−:回忆0.85(95% CI 0.62,1.0),精密度0.92(95% CI 0.73,1.0),F1评分为0.88(95% CI 0.71,1.0); PN+:召回率为0.83(95% CI 0.58,1.0),精确度为1.0(95% CI 1.0,1.0),F1评分为0.91(95% CI 0.74,1.0)。归因方法显示的特征表明HV中有更多的角膜神经,PN-图像中角膜神经减少,PN+图像中没有角膜神经。我们证明了有前途的结果,在周围神经病变的快速分类使用一个单一的角膜图像。需要进行大规模的多中心验证研究,以评估基于AI的DLA在糖尿病神经病变筛查和诊断计划中的效用。本文的在线版本(10.1007/s 00125 -021-05617-x)包含同行评审但未经编辑的补充材料。
We aimed to develop an artificial intelligence (AI)-based deep learning algorithm (DLA) applying attribution methods without image segmentation to corneal confocal microscopy images and to accurately classify peripheral neuropathy (or lack of). The AI-based DLA utilised convolutional neural networks with data augmentation to increase the algorithm’s generalisability. The algorithm was trained using a high-end graphics processor for 300 epochs on 329 corneal nerve images and tested on 40 images (1 image/participant). Participants consisted of healthy volunteer (HV) participants (n = 90) and participants with type 1 diabetes (n = 88), type 2 diabetes (n = 141) and prediabetes (n = 50) (defined as impaired fasting glucose, impaired glucose tolerance or a combination of both), and were classified into HV, those without neuropathy (PN−) (n = 149) and those with neuropathy (PN+) (n = 130). For the AI-based DLA, a modified residual neural network called ResNet-50 was developed and used to extract features from images and perform classification. The algorithm was tested on 40 participants (15 HV, 13 PN−, 12 PN+). Attribution methods gradient-weighted class activation mapping (Grad-CAM), Guided Grad-CAM and occlusion sensitivity displayed the areas within the image that had the greatest impact on the decision of the algorithm. The results were as follows: HV: recall of 1.0 (95% CI 1.0, 1.0), precision of 0.83 (95% CI 0.65, 1.0), F1-score of 0.91 (95% CI 0.79, 1.0); PN−: recall of 0.85 (95% CI 0.62, 1.0), precision of 0.92 (95% CI 0.73, 1.0), F1-score of 0.88 (95% CI 0.71, 1.0); PN+: recall of 0.83 (95% CI 0.58, 1.0), precision of 1.0 (95% CI 1.0, 1.0), F1-score of 0.91 (95% CI 0.74, 1.0). The features displayed by the attribution methods demonstrated more corneal nerves in HV, a reduction in corneal nerves for PN− and an absence of corneal nerves for PN+ images. We demonstrate promising results in the rapid classification of peripheral neuropathy using a single corneal image. A large-scale multicentre validation study is required to assess the utility of AI-based DLA in screening and diagnostic programmes for diabetic neuropathy. The online version of this article (10.1007/s00125-021-05617-x) contains peer-reviewed but unedited supplementary material.
DOI: 10.1109/tbme.2016.2573642
发表时间: 2017-04
期刊: IEEE transactions on bio-medical engineering
影响因子: --
作者:
Chen X;Graham J;Dabbah MA;Petropoulos IN;Tavakoli M;Malik RA
通讯作者: Malik RA
DOI: 10.1016/j.diabres.2018.02.023
发表时间: 2018-04-01
影响因子: 5.1
作者:
Cho, N. H.;Shaw, J. E.;Malanda, B.
通讯作者: Malanda, B.
DOI: 10.1371/journal.pone.0183040
发表时间: 2017-08-17
期刊: PLOS ONE
影响因子: 3.7
作者:
Kalteniece, Alise;Ferdousi, Maryam;Malik, Rayaz A.
通讯作者: Malik, Rayaz A.
DOI: 10.1186/s40662-020-00192-5
发表时间: 2020-05-08
期刊: EYE AND VISION
影响因子: 4.2
作者:
Oakley, Jonathan D.;Russakoff, Daniel B.;Mankowski, Joseph L.
通讯作者: Mankowski, Joseph L.
DOI: 10.1136/bmjdrc-2020-002040
发表时间: 2021-05
影响因子: 4.1
作者:
Kirthi V;Perumbalath A;Brown E;Nevitt S;Petropoulos IN;Burgess J;Roylance R;Cuthbertson DJ;Jackson TL;Malik RA;Alam U
通讯作者: Alam U