Validation of Deep Learning techniques for quality augmentation in diffusion MRI for clinical studies
|Title||Validation of Deep Learning techniques for quality augmentation in diffusion MRI for clinical studies|
|Publication Type||Conference Proceedings|
|Year of Conference||2023|
|Authors||Aja-Fernandez, S., C. Martin-Martin, T. Pieciak, Á. Planchuelo-Gómez, A. Faiyaz, N. Uddin, A. Tiwari, S. J. Shigwan, T. Zheng, Z. Cao, S. B. Blumberg, S. Sen, M. Yigit Avci, Z. Li, X. Wang, Z. Tang, A. Rauland, D. Merhof, R. Manzano Maria, V. P. Campos, SK. HashemiazadehKolowri, E. DiBella, C. Peng, Z. Chen, I. Ullah, M. Mani, S. Eckstrom, S. H. Baete, S. Scifitto, R. Kumar Singh, D. Wu, T. Goodwin-Allcock, P. J. Slator, B. Bilgic, Q. Tian, M. Cabezas, T. Santini, M. Andrade da Vieira, Z. Shen, H. Abdolmotalleby, P. Filipiak, A. Tristan-Vega, and R. de Luis-Garcia|
|Conference Name||2023 ISMRM & ISMRT Annual Meeting & Exhibition|
This work gathers the results of the QuadD22 challenge, held in MICCAI 2022. We evaluate whether Deep Learning (DL) Techniques are able to improve the quality of diffusion MRI data in clinical studies. To that end, we focused on a real study on migraine, where the differences between groups are drastically reduced when using 21 gradient directions instead of 61. Thus, we asked the participants to augment dMRI data acquired with only 21 directions to 61 via DL. The results were evaluated using a real clinical study with TBSS in which we statistically compared episodic migraine to chronic migraine.