Magnetic resonance imaging is subject to slow acquisition times due to the inherent limitations in data sampling. Recently, supervised deep learning has emerged as a promising technique for reconstructing sub-sampled MRI. However, supervised deep learning requires a large dataset of fully-sampled data. Although unsupervised or self-supervised deep learning methods have emerged to address the limitations of supervised deep learning approaches, they still require a database of images. In contrast, scan-specific deep learning methods learn and reconstruct using only the sub-sampled data from a single scan. Here, we introduce Scan-Specific Self-Supervised Bayesian Deep Non-Linear Inversion (DNLINV) that does not require an auto calibration scan region. DNLINV utilizes a Deep Image Prior-type generative modeling approach and relies on approximate Bayesian inference to regularize the deep convolutional neural network. We demonstrate our approach on several anatomies, contrasts, and sampling patterns and show improved performance over existing approaches in scan-specific calibrationless parallel imaging and compressed sensing.
Despite the proliferation of deep learning techniques for accelerated MRI acquisition and enhanced image reconstruction, the construction of large and diverse MRI datasets continues to pose a barrier to effective clinical translation of these technologies. One major challenge is in collecting the MRI raw data (required for image reconstruction) from clinical scanning, as only magnitude images are typically saved and used for clinical assessment and diagnosis. The image phase and multi-channel RF coil information are not retained when magnitude-only images are saved in clinical imaging archives. Additionally, preprocessing used for data in clinical imaging can lead to biased results. While several groups have begun concerted efforts to collect large amounts of MRI raw data, current databases are limited in the diversity of anatomy, pathology, annotations, and acquisition types they contain. To address this, we present a method for synthesizing realistic MR data from magnitude-only data, allowing for the use of diverse data from clinical imaging archives in advanced MRI reconstruction development. Our method uses a conditional GAN-based framework to generate synthetic phase images from input magnitude images. We then applied ESPIRiT to derive RF coil sensitivity maps from fully sampled real data to generate multi-coil data. The synthetic data generation method was evaluated by comparing image reconstruction results from training Variational Networks either with real data or synthetic data. We demonstrate that the Variational Network trained on synthetic MRI data from our method, consisting of GAN-derived synthetic phase and multi-coil information, outperformed Variational Networks trained on data with synthetic phase generated using current state-of-the-art methods. Additionally, we demonstrate that the Variational Networks trained with synthetic k-space data from our method perform comparably to image reconstruction networks trained on undersampled real k-space data.
Background/Objective: Longitudinal in vivo studies of murine xenograft models are widely utilized in oncology to study cancer biology and develop therapies. Magnetic resonance imaging (MRI) of these tumors is an invaluable tool for monitoring tumor growth and characterizing the tumors as well. Methods: In this work, a pipeline for automating the segmentation of xenografts in mouse models was developed. T2-weighted (T2-wt) MRI images from mice implanted with six different prostate cancer patient-derived xenografts (PDX) in the kidneys, liver, and tibia were used. The segmentation pipeline included a slice classifier to identify the slices that had tumors and subsequent training and validation using several U-Net-based segmentation architectures. Multiple combinations of the algorithm and training images for different sites were evaluated for inference quality. Results and Conclusions: The slice classifier network achieved 90% accuracy in identifying slices containing tumors. Among the various segmentation architectures tested, the dense residual recurrent U-Net achieved the highest performance in kidney tumors. When evaluated across the kidneys, tibia, and liver, this architecture performed the best when trained on all data as compared to training on only data from a single site (and inferring on a multi-site tumor images), achieving a Dice score of 0.924 across the test set.
Purpose: To evaluate methods for quantification of pulmonary ventilation with ultrashort echo time (UTE) MRI. Methods: We performed a reproducibility study, acquiring two free-breathing 1 H UTE lung MRIs on the same day for six healthy volunteers. The 1) 3D + t cyclic b-spline and 2) symmetric image normalization (SyN) methods for image registration were applied after respiratory phase-resolved image reconstruction. Ventilation maps were calculated using 1) Jacobian determinant of the deformation fields minus one, termed regional ventilation, and 2) intensity percentage difference between the registered and fixed image, termed specific ventilation. We compared the reproducibility of all four method combinations via statistical analysis. Results: Split violin plots and Bland-Altman plots are shown for whole lungs and lung sections. The cyclic b-spline registration and Jacobian determinant regional ventilation quantification provide total ventilation volumes that match the segmentation tidal volume, smooth and uniform ventilation maps. The cyclic b-spline registration and specific ventilation combination yields the smallest standard deviation in the Bland-Altman plot. Conclusion: Cyclic registration performs better than SyN for respiratory phase-resolved 1 H UTE MRI ventilation quantification. Regional ventilation correlates better with segmentation lung volume, while specific ventilation is more reproducible.
PurposeDeep learning–based magnetic resonance imaging (MRI) reconstruction can enable faster imaging and improved image quality; however, clinical adoption is limited by the lack of diverse raw multicoil training datasets. In routine clinical workflows, multicoil k-space data needed for training are often discarded, and only magnitude images are retained. In addition, phase is not deterministically recoverable from magnitude, and storing full k-space is impractical.ApproachWe address these limitations by (1) using a deep generative model to produce multiple plausible synthetic phase images conditioned on a single magnitude image and (2) introducing a radiofrequency (RF) coil library to synthesize multicoil k-space from the generated complex-valued images.ResultsThe synthetic phase images resemble ground-truth phase as assessed by Fréchet inception distance, and the generated multicoil k-space effectively trains unrolled MRI reconstruction networks.ConclusionsThese results indicate that generative phase synthesis coupled with an RF coil library provides a practical alternative to real multicoil k-space for training reconstruction models.