Work

Leveraging large scale data sets: a transfer learning approach for 7T super resolution

Year: 2021
Type: dissertation
Abstract: Brain morphometry on data from multi-scanner and multi-site studies can suffer from nonbiologicalvariance due to scanner and acquisition differences. Harmonization methods, such as ComBat,\nhave been introduced to remove unwanted variance in structural neuroimaging data. Statistical methods\nfor harmonizing structural data however operate on derived morphological measurements to remove site\nrelated effects rather than operating at the voxel level to remove scanner related effects. This study works\ntowards a deep learning-based image harmonization method by training and evaluating a generative\nadversarial model for transforming 3T images to a standard 7T-like image quality. 7T MRI can achieve\nbetter tissue contrast and tissue segmentation results but lacks the widespread availability of 3T MRI,\nresulting in limited dataset sizes for deep learning. Transfer learning from a 3T synthesis task to a 7T\nsynthesis task was hypothesized to improve synthesis results by greatly increasing dataset size and\ndiversity with multi-site longitudinal data. The 7T synthesis dataset was comprised of 9 subjects each\nwith a 3T MPRAGE and 7T MP2RAGE T1-weighted scan. Leave one out cross validation was used and\nperformance evaluation metrics were reported as the mean across all validation cross folds. The transfer\nlearning dataset consisted of 419 total subjects and 1124 T1 weighted images with a wide variety of sites,\nscanners, and acquisition sequences. An independent testing set of 17 subjects with paired 1.5T and 3T\nscans from the transfer learning dataset were used for evaluating the 3T synthesis task. Image similarity\nmetrics such as Structural Similarity Index Measure (SSIM) and Peak Signal to Noise Ratio (PSNR) were\nused to evaluate synthesis performance. Dice Similarity Coefficient (DSC) and Jaccard Similarity\nCoefficient (JSC) were used to evaluate the synthesized and 3T segmentations results using 7T\nsegmentation as ground truth. The 7T synthesis network with transfer learning weights achieved an SSIM\nof 0.950 ± 0.02 and PSNR of 25.44 ± 0.61, improved over the 3T image which had SSIM of 0.909 ± 0.01\nand PSNR of 21.83 ± 0.92. The DSC for grey matter regions of interest was 0.810 ± 0.02 and 0.916 ±\n0.004 for white matter regions of interest for the synthesized validation images, an improvement of 0.053\nDSC (p = 0.011) and 0.017 DSC (p = 0.0039) over the 3T results respectively. The JSC for grey matter\nregions of interest was 0.693 ± 0.03 and 0.842 ± 0.01 for white matter regions of interest, an improvement\nof 0.066 DSC (p = 0.011) and 0.026 DSC (p = 0.0039) respectively. Future work will evaluate the ability\nof the 7T synthesis models at removing non-biological variance, particularly in longitudinal studies where\nimaging protocol or scanners were updated.
Author: Ellis, Ryan
Language: English
Field: Medicine
Open Access status: green
OpenAlex ID: W7001765526
Cites: 31
Cited by: 1
Related to: 10