Accelerated and quantitative three‐dimensional molecular MRI using a generative adversarial network

Abstract

Purpose To substantially shorten the acquisition time required for quantitative three‐dimensional (3D) chemical exchange saturation transfer (CEST) and semisolid magnetization transfer (MT) imaging and allow for rapid chemical exchange parameter map reconstruction. Methods Three‐dimensional CEST and MT magnetic resonance fingerprinting (MRF) datasets of L‐arginine phantoms, whole‐brains, and calf muscles from healthy volunteers, cancer patients, and cardiac patients were acquired using 3T clinical scanners at three different sites, using three different scanner models and coils. A saturation transfer‐oriented generative adversarial network (GAN‐ST) supervised framework was then designed and trained to learn the mapping from a reduced input data space to the quantitative exchange parameter space, while preserving perceptual and quantitative content. Results The GAN‐ST 3D acquisition time was 42–52 s, 70% shorter than CEST‐MRF. The quantitative reconstruction of the entire brain took 0.8 s. An excellent agreement was observed between the ground truth and GAN‐based L‐arginine concentration and pH values (Pearson's r > 0.95, ICC > 0.88, NRMSE < 3%). GAN‐ST images from a brain‐tumor subject yielded a semi‐solid volume fraction and exchange rate NRMSE of 3.8prefix±1.3%$$ 3.8\pm 1.3\% $$ and 4.6prefix±1.3%$$ 4.6\pm 1.3\% $$, respectively, and SSIM of 96.3prefix±1.6%$$ 96.3\pm 1.6\% $$ and 95.0prefix±2.4%$$ 95.0\pm 2.4\% $$, respectively. The mapping of the calf‐muscle exchange parameters in a cardiac patient, yielded NRMSE < 7% and SSIM > 94% for the semi‐solid exchange parameters. In regions with large susceptibility artifacts, GAN‐ST has demonstrated improved performance and reduced noise compared to MRF. Conclusion GAN‐ST can substantially reduce the acquisition time for quantitative semi‐solid MT/CEST mapping, while retaining performance even when facing pathologies and scanner models that were not available during training.

Molecular MRI-Based Monitoring of Cancer Immunotherapy Treatment Response

“…To account for the exponential growth in dictionary size, which translates into very long molecular parameter matching, a variety of deep-learning architectures have been suggested. These works enable rapid image acquisition (e.g., <5 min) and whole brain molecular parameter quantification in a matter of seconds [ 151 , 193 , 194 , 195 , 196 , 197 ]. Importantly, in a virotherapy treatment monitoring animal study, a comparison of conventional CEST-MRF to deep-learning-based CEST-MRF pointed to the clear superiority of the latter in terms of quantification accuracy and speed [ 151 ].…”

AI in MRI: Computational Frameworks for a Faster, Optimized, and Automated Imaging Workflow

“…However, recently proposed frameworks for AI-based acquisition and quantification have rendered the rapid extraction of these parameters a viable option. A few such examples include the mapping of T 1 and T 2 relaxation times [ 54 , 55 , 56 , 57 , 58 ], semisolid magnetization transfer (MT) and chemical exchange saturation transfer (CEST) proton volume fraction and exchange rate [ 59 , 60 , 61 , 62 , 63 ], and susceptibility [ 64 ].…”

In Vivo Mapping of the Chemical Exchange Relayed Nuclear Overhauser Effect using Deep Magnetic Resonance Fingerprinting (rNOE-MRF)

“…While the rst CEST-MRF reports used correlation-based pattern recognition for quantitative image reconstruction, the complexity of the in-vivo multi-pool environment, which translates into impractical parameter quanti cation times, has motivated the pursuit of alternative and faster reconstruction methods. Speci cally, a variety of neural network architectures were designed and validated for ultrashort (~ a few seconds-long or less) reconstruction of the proton exchange parameter maps [25][26][27][28][29][30] .…”