Generative T2-weighted images as a substitute for true T2-weighted images on brain MRI in patients with acute stroke - 20/03/25

Doi : 10.1016/j.diii.2025.03.004

Antoine Pelcat ^a, Alice Le Berre ^a,^b, Wagih Ben Hassen ^a,^b, Clement Debacker ^a,^b, Sylvain Charron ^a, Bertrand Thirion ^c, Laurence Legrand ^a,^b, Guillaume Turc ^d,^e, Catherine Oppenheim ^a,^b, Joseph Benzakoun ^a,^b,^⁎
^a Université Paris Cité, Institute of Psychiatry and Neuroscience of Paris (IPNP), INSERM U1266, IMA-BRAIN, 75014 Paris, France
^b GHU Paris Psychiatrie et Neurosciences, Hôpital Sainte Anne, Department of Neuroradiology, 75014 Paris, France
^c INRIA, CEA, Université Paris-Saclay, MIND Team, 91400 Palaiseau, France
^d Université Paris Cité, Institute of Psychiatry and Neuroscience of Paris (IPNP), INSERM U1266, Stroke Team, 75014 Paris, France
^e GHU Paris Psychiatrie et Neurosciences, Hôpital Sainte Anne, Department of Neurology, 75014 Paris, France

^⁎Corresponding author.

Sous presse. Épreuves corrigées par l'auteur. Disponible en ligne depuis le Thursday 20 March 2025

Highlights

•	Generative T2*-weighted images of the brain can be produced using diffusion-weighted images only.
•	Generative T2*-weighted images of the brain have high intra- and inter-observer concordance for hemorrhage detection on MRI in the context of acute stroke (κ = 0.95 and 0.92).
•	Generative T2-weighted images of the brain are highly concordant with true T2-weighted images for hemorrhage detection on MRI in the context of acute stroke (κ = 0.92).

Le texte complet de cet article est disponible en PDF.

Abstract

Purpose

The purpose of this study was to validate a deep learning algorithm that generates T2*-weighted images from diffusion-weighted (DW) images and to compare its performance with that of true T2*-weighted images for hemorrhage detection on MRI in patients with acute stroke.

Materials and methods

This single-center, retrospective study included DW- and T2*-weighted images obtained less than 48 hours after symptom onset in consecutive patients admitted for acute stroke. Datasets were divided into training (60 %), validation (20 %), and test (20 %) sets, with stratification by stroke type (hemorrhagic/ischemic). A generative adversarial network was trained to produce generative T2*-weighted images using DW images. Concordance between true T2*-weighted images and generative T2*-weighted images for hemorrhage detection was independently graded by two readers into three categories (parenchymal hematoma, hemorrhagic infarct or no hemorrhage), and discordances were resolved by consensus reading. Sensitivity, specificity and accuracy of generative T2*-weighted images were estimated using true T2*-weighted images as the standard of reference.

Results

A total of 1491 MRI sets from 939 patients (487 women, 452 men) with a median age of 71 years (first quartile, 57; third quartile, 81; range: 21–101) were included. In the test set (n = 300), there were no differences between true T2*-weighted images and generative T2*-weighted images for intraobserver reproducibility (κ = 0.97 [95 % CI: 0.95–0.99] vs. 0.95 [95 % CI: 0.92–0.97]; P = 0.27) and interobserver reproducibility (κ = 0.93 [95 % CI: 0.90–0.97] vs. 0.92 [95 % CI: 0.88–0.96]; P = 0.64). After consensus reading, concordance between true T2*-weighted images and generative T2*-weighted images was excellent (κ = 0.92; 95 % CI: 0.91–0.96). Generative T2*-weighted images achieved 90 % sensitivity (73/81; 95 % CI: 81–96), 97 % specificity (213/219; 95 % CI: 94–99) and 95 % accuracy (286/300; 95 % CI: 92–97) for the diagnosis of any cerebral hemorrhage (hemorrhagic infarct or parenchymal hemorrhage).

Conclusion

Generative T2*-weighted images and true T2*-weighted images have non-different diagnostic performances for hemorrhage detection in patients with acute stroke and may be used to shorten MRI protocols.

Le texte complet de cet article est disponible en PDF.

Keywords : Deep-learning, Intracerebral hemorrhage, Magnetic resonance imaging, Stroke, T2*-weighted images

Abbreviations : AI, CT, DW, Ea-GAN, EPI, FLAIR, HI, MCA, MR, NIHSS, NoH, PH, ROC, SD