Validation of deep learning techniques for quality augmentation in diffusion MRI for clinical studies.
Validation of deep learning techniques for quality augmentation in diffusion MRI for clinical studies.
复制标题
验证深度学习技术在临床研究中用于弥散MRI质量增强。
DOI：
10.1016/j.nicl.2023.103483
复制
发表时间：
2023
影响因子：
4.2
通讯作者：
Pieciak, Tomasz
中科院分区：
文献类型：
--
作者：
Aja-Fernandez, Santiago;Martin-Martin, Carmen;Planchuelo-Gomez, Alvaro;Faiyaz, Abrar;Uddin, Md Nasir;Schifitto, Giovanni;Tiwari, Abhishek;Shigwan, Saurabh J.;Singh, Rajeev Kumar;Zheng, Tianshu;Cao, Zuozhen;Wu, Dan;Blumberg, Stefano B.;Sen, Snigdha;Goodwin-Allcock, Tobias;Slator, Paddy J.;Avci, Mehmet Yigit;Li, Zihan;Bilgic, Berkin;Tian, Qiyuan;Wang, Xinyi;Tang, Zihao;Cabezas, Mariano;Rauland, Amelie;Merhof, Dorit;Maria, Renata Manzano;Campos, Vinicius Paraniba;Santini, Tales;Vieira, Marcelo Andrade da Costa;Hashemizadehkolowri, Seyyedkazem;Dibella, Edward;Peng, Chenxu;Shen, Zhimin;Chen, Zan;Ullah, Irfan;Mani, Merry;Abdolmotalleby, Hesam;Eckstrom, Samuel;Baete, Steven H.;Filipiak, Patryk;Dong, Tanxin;Fan, Qiuyun;de Luis-Garcia, Rodrigo;Tristan-Vega, Antonio;Pieciak, Tomasz
关键词：
分析了14种不同的AI方法：4种未显示任何改善; 6种方法改善超过25%。大多数方法显示假阳性的恒定增长率与新的真阳性成比例。基于视觉质量或误差的评价不足以评估AI重建数据。在MRI临床研究中使用DL进行质量增强时，建议谨慎。本研究的目的是评估深度学习（DL）技术在改善临床应用中弥散MRI（dMRI）数据质量方面的有效性。该研究旨在确定在医学图像中使用人工智能（AI）方法是否会导致关键临床信息的丢失和/或虚假信息的出现。为了评估这一点，重点是dMRI的角分辨率，并对偏头痛进行了临床试验，特别是在发作性和慢性偏头痛患者之间。梯度方向的数量对白色物质分析结果有影响，当使用21个梯度方向而不是原始的61个梯度方向时，组间的统计学显著差异大幅降低。来自不同机构的14个团队的任务是使用DL增强从21个梯度方向和1000 s/mm 2的b值获得的数据计算的三个扩散指标（FA，AD和MD）。目标是产生与从61个梯度方向计算的结果相当的结果。使用标准图像质量指标和基于轨道的空间统计（TBSS）对结果进行评估，以比较发作性和慢性偏头痛患者。研究结果表明，虽然大多数DL技术提高了检测组间统计差异的能力，但它们也导致了假阳性的增加。结果表明，假阳性的增长率与新的真阳性成线性比例，这突出了在评估不同的临床队列和使用单个组的数据进行训练时，基于AI的任务泛化的风险。这些方法在复制数据的原始分布时也表现出不同的性能，有些表现出显着的偏差。总之，在处理临床研究中的异质数据时，在临床研究中使用AI方法进行协调或合成时应极其谨慎，因为重要信息可能会改变，即使结构相似性或峰值信噪比等全局指标似乎表明并非如此。
14 different AI methods were analyzed: 4 did not show any improvement; 6 methods improved over 25%. Most methods showed a constant growth rate of false positives proportional to the new true positives. Metrics based on visual quality or errors are inadequate for evaluating AI reconstructed data. Caution is recommended when using DL for quality augmentation on clinical studies with MRI. The objective of this study is to evaluate the efficacy of deep learning (DL) techniques in improving the quality of diffusion MRI (dMRI) data in clinical applications. The study aims to determine whether the use of artificial intelligence (AI) methods in medical images may result in the loss of critical clinical information and/or the appearance of false information. To assess this, the focus was on the angular resolution of dMRI and a clinical trial was conducted on migraine, specifically between episodic and chronic migraine patients. The number of gradient directions had an impact on white matter analysis results, with statistically significant differences between groups being drastically reduced when using 21 gradient directions instead of the original 61. Fourteen teams from different institutions were tasked to use DL to enhance three diffusion metrics (FA, AD and MD) calculated from data acquired with 21 gradient directions and a b-value of 1000 s/mm2. The goal was to produce results that were comparable to those calculated from 61 gradient directions. The results were evaluated using both standard image quality metrics and Tract-Based Spatial Statistics (TBSS) to compare episodic and chronic migraine patients. The study results suggest that while most DL techniques improved the ability to detect statistical differences between groups, they also led to an increase in false positive. The results showed that there was a constant growth rate of false positives linearly proportional to the new true positives, which highlights the risk of generalization of AI-based tasks when assessing diverse clinical cohorts and training using data from a single group. The methods also showed divergent performance when replicating the original distribution of the data and some exhibited significant bias. In conclusion, extreme caution should be exercised when using AI methods for harmonization or synthesis in clinical studies when processing heterogeneous data in clinical studies, as important information may be altered, even when global metrics such as structural similarity or peak signal-to-noise ratio appear to suggest otherwise.
登录
查看更多内容
影响因子：
3.5
作者：
Chen G;Dong B;Zhang Y;Lin W;Shen D;Yap PT
通讯作者：
Yap PT
影响因子：
5.7
作者：
Daducci, Alessandro;Canales-Rodriguez, Erick J.;Thiran, Jean-Philippe
通讯作者：
Thiran, Jean-Philippe
影响因子：
2.9
作者：
Faiyaz, Abrar;Doyley, Marvin;Uddin, Md Nasir
通讯作者：
Uddin, Md Nasir
DOI：
10.1016/j.compmedimag.2022.102151
发表时间：
2023-01
期刊：
Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society
影响因子：
--
作者：
通讯作者：
--
影响因子：
3.7
作者：
Barrio-Arranz G;de Luis-García R;Tristán-Vega A;Martín-Fernández M;Aja-Fernández S
通讯作者：
Aja-Fernández S