Validation of deep learning techniques for quality augmentation in diffusion MRI for clinical studies
NeuroImage Clinical, vol. 39, pp. 103483
Abstract
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.
Authors 44
-
Santiago Aja‐Fernández corresponding
Affiliation as printed
Laboratorio de Procesado de Imagen (LPI), ETSI Telecomunicación, Universidad de Valladolid, Spain. Electronic address: sanaja@tel.uva.es
Laboratorio de Procesado de Imagen (LPI), ETSI Telecomunicación, Universidad de Valladolid, Spain
-
Affiliation as printed
Laboratorio de Procesado de Imagen (LPI), ETSI Telecomunicación, Universidad de Valladolid, Spain
-
Universidad de Valladolid · Cardiff University
Affiliation as printed
Laboratorio de Procesado de Imagen (LPI), ETSI Telecomunicación, Universidad de Valladolid, Spain; Cardiff University Brain Research Imaging Centre (CUBRIC), School of Psychology, Cardiff University, Cardiff, UK
Cardiff University Brain Research Imaging Centre (CUBRIC), School of Psychology, Cardiff University, Cardiff, UK
Laboratorio de Procesado de Imagen (LPI), ETSI Telecomunicación, Universidad de Valladolid, Spain
-
Affiliation as printed
University of Rochester, USA
-
Affiliation as printed
University of Rochester, USA
-
Affiliation as printed
University of Rochester, USA
-
Affiliation as printed
Shiv Nadar Institution of Eminence, India
-
Affiliation as printed
Shiv Nadar Institution of Eminence, India
-
Affiliation as printed
Shiv Nadar Institution of Eminence, India
-
Affiliation as printed
Zhejiang University, China
-
Affiliation as printed
Zhejiang University, China
-
Affiliation as printed
University College London, UK
-
Affiliation as printed
University College London, UK
-
Affiliation as printed
University College London, UK
-
Affiliation as printed
University College London, UK
-
Athinoula A. Martinos Center for Biomedical Imaging
Affiliation as printed
Athinoula A. Martinos Center for Biomedical Imaging, USA
-
Athinoula A. Martinos Center for Biomedical Imaging
Affiliation as printed
Athinoula A. Martinos Center for Biomedical Imaging, USA
-
Athinoula A. Martinos Center for Biomedical Imaging
Affiliation as printed
Athinoula A. Martinos Center for Biomedical Imaging, USA
-
Athinoula A. Martinos Center for Biomedical Imaging
Affiliation as printed
Athinoula A. Martinos Center for Biomedical Imaging, USA
-
Affiliation as printed
University of Sydney, Australia
-
Affiliation as printed
University of Sydney, Australia
-
Affiliation as printed
University of Sydney, Australia
-
Amelie Rauland Aachen
Affiliation as printed
RWTH Aachen University, Germany
-
Affiliation as printed
University of Regensburg, Germany
-
Affiliation as printed
University of Sao Paulo, Brazil
-
Affiliation as printed
University of Sao Paulo, Brazil
-
Affiliation as printed
Western University, Canada
-
Affiliation as printed
University of Sao Paulo, Brazil
-
Affiliation as printed
University of Utah, USA
-
Affiliation as printed
University of Utah, USA
-
Zhejiang University of Technology
Affiliation as printed
Zhejiang University of Technology, China
-
Zhejiang University of Technology
Affiliation as printed
Zhejiang University of Technology, China
-
Zhejiang University of Technology
Affiliation as printed
Zhejiang University of Technology, China
-
Affiliation as printed
University of Iowa, USA
-
Affiliation as printed
University of Iowa, USA
-
Affiliation as printed
University of Iowa, USA
-
Affiliation as printed
New York University, USA
-
Affiliation as printed
New York University, USA
-
Affiliation as printed
New York University, USA
-
Affiliation as printed
Tianjin University, China
-
Tianjin University · University College London
Affiliation as printed
Tianjin University, China
University College London, UK
-
Affiliation as printed
Laboratorio de Procesado de Imagen (LPI), ETSI Telecomunicación, Universidad de Valladolid, Spain
-
Affiliation as printed
Laboratorio de Procesado de Imagen (LPI), ETSI Telecomunicación, Universidad de Valladolid, Spain
-
Affiliation as printed
Laboratorio de Procesado de Imagen (LPI), ETSI Telecomunicación, Universidad de Valladolid, Spain
Cited by 15 stored of 15
15 results
No patents citing this paper on Lens.org (checked 2026-10-06).