About

Shreyas Vasanawala is the William R. Brody Professor of Pediatric Radiology and Child Health at Stanford University. He is associated with the Center for Artificial Intelligence in Medicine & Imaging (AIMI). His work focuses on the application of artificial intelligence in medicine and imaging, particularly in pediatric radiology. As a faculty member at Stanford, he contributes to advancing research in medical imaging and AI, aiming to improve healthcare outcomes through innovative technological solutions.

Research topics

• Computer Science
• Machine Learning
• Artificial Intelligence
• Medicine
• Radiology
• Anesthesia
• Cardiology
• Computer vision

Selected publications

• Semi-supervision for clinical contrast-weighted image synthesis from magnetic resonance fingerprinting

  Magnetic Resonance Materials in Physics Biology and Medicine · 2026-03-17

  articleOpen access

  To facilitate ease of data compilation across diverse populations for training models to synthesize clinical contrast-weighted images from magnetic resonance fingerprinting. We leverage a semi-supervised training framework using highly accelerated acquisitions of the target contrasts used as ground truths. We utilize complementary randomized data sampling masks across training subjects and contrasts for homogeneous learning in k-space, together with multi-task learning. Our experiments indicate that the proposed method achieves high-quality synthesis with networks trained on retrospectively and prospectively undersampled data of the contrast-weighted images, enabling undersampling up to 12–16 $$\times $$ . The proposed method enables semi-supervised learning for synthesis from MRF with an end-to-end, ultra-fast training data acquisition protocol that is easier to obtain across a large population in clinical settings.

  PublisherOA PDFDOI

• Consensus ADMM for Distributed, Constrained Reconstruction with Low-Rank Subspace and Phase Priors

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  article

  Motivation: Spatiotemporally-undersampled sequences help to accelerate scans and push resolution limits. However, reconstruction of this data is slow and memory-intensive. Goal(s): We aim to accelerate such reconstructions by splitting our objective into independent components and parallelizing the optimization across multiple compute devices. Approach: We implement fast, GPU-accelerated proximal operators for data consistency, Locally-Low-rank, and S-LORAKS. We use a technique called Consensus ADMM to fuse them together in a distributed fashion. Results: Our reconstruction enables reconstruction of high-resolution 3D magnetic resonance fingerprinting datasets and quantitative maps quickly and with good accuracy. Impact: By adding more sophisticated prior knowledge to the reconstruction, we can further accelerate the scan, shortening scan times while maintaining quality.Our techniques are also quite general and apply broadly to a wide variety of reconstruction problems.

  PublisherDOI

• DM-Net: a physics-model-independent direct mapping approach for calibration-free multi-coil MRI

  Research Square · 2025-09-01

  preprintOpen accessSenior author
  PublisherOA PDFDOI

• A Prototype Microwave Link for a Wireless MRI Coil Array

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  article

  Motivation: MRI coil array cables provide an obstacle to positioning the array and form-fitting it to the patient, increasing scan times and reducing patient comfort. Goal(s): We aim to remove the coil array cable by developing a scalable microwave link to wirelessly transmit digitized coil data from the array. Approach: Using off-the-shelf components, we constructed a custom microwave link consisting of a test transmitter that sends encoded pseudo-random data and a receiver that streams recovered data out over ethernet. Results: We present a prototype wireless link capable of transferring error-free data above at least 200 Mbps over distances up to several meters. Impact: This microwave link prototype brings us closer to a wireless MRI system, which will streamline the imaging process by enabling faster and more effective positioning of the coil array on the patient, reducing scan times and improving patient comfort.

  PublisherDOI

• MM-Wave Radar Vitals Sensing: Towards Non-Contact Cardiac and Respiratory Gating for MRI

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  article

  Motivation: Present day cardiac and respiratory gating require cabled sensors that impede patient workflow. A non-contact sensing method is needed. Goal(s): Our goal is to determine whether non-contact mm-wave radar neck vessel motion sensing can distinguish the cardiac cycle from respiratory motion with a long term aim of non-contact gating. Approach: We updated a 4 channel 24GHz CW radar system using dual Analog Discovery Pros for chest/neck data capture and integrated ECG monitoring with Python scripting. Results: Respiratory motion remains easily detected with high SNR. Neck vessel motion creates unique waveforms with carotid notch and jugular venous pulse contributions variably mixed. Impact: MM-wave radar is a promising candidate for non-contact sensing of cardiac and respiratory cycles, with potential applications in non-contact motion gating in pediatric MRI and synchronizing on pulsatile motion of the cerebrum in neuroimaging.

  PublisherDOI

• Fast Reconstruction of Time-Resolved 4D MRA with Unrolled Neural Networks

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  article

  Motivation: Recently an ASL-based accelerated 4D MRA was demonstrated with great potential in delineating dynamic blood flow patterns at high spatiotemporal resolution. However, the time-consuming reconstruction presents a bottleneck for wider clinical translation. Goal(s): To introduce DL-AngioNet, a ML-based framework that accelerates the reconstruction while providing improved SNR. Approach: The network was trained using historical data via a data-driven method with a physical model. The unrolled structure of the network provided a data consistency term to ensure validity of the results. Results: DL-AngioNet accelerated the reconstruction by ~30-fold while preserving good flow dynamic information. Results demonstrated superior SNR comparing to the conventional PICS method. Impact: DL-AngioNet significantly accelerates 4D MRA reconstruction by ~30-fold, which not only preserves good 4D MRA flow dynamics, but also provides improved SNR in the results. DL-AngioNet could facilitate 4D MRA into a wider clinical use.

  PublisherDOI

• Twenty-fold accelerated cardiovascular phase-contrast MRI using channel-shift CNN

  Journal of Cardiovascular Magnetic Resonance · 2025-01-01

  articleOpen access
  PublisherDOI

• Cardiac MRI Z-scores in children under 2 years age: Comparison between different models

  Journal of Cardiovascular Magnetic Resonance · 2025-01-01

  articleOpen access
  PublisherDOI

• Evaluation of Pediatric MRI with a High-Density Posterior Cushion Coil

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  article

  Motivation: Posterior coil arrays built into MRI tables are common, but provide suboptimal performance in pediatric imaging due to their distance from the imaged volume and lower coil density. Goal(s): Improve pediatric MRI by constructing a prototype 30 channel high density posterior cushion coil (HDPCC) incorporated into a padded cushion. Approach: 25 pediatric patients were scanned with HDPCC. 25 age-matched controls were scanned with standard coils. Two radiologists graded image quality on a 5-point scale. SNR was measured in the right paraspinal musculature and the liver. Results: All images obtained with HDPCC were diagnostic (mean score 4.13/5). Posterior SNR was higher than standard coils. Impact: HDPCC can improve SNR in pediatric MRI by conforming to the body and minimizing distance to the imaged volume.

  PublisherDOI

• Signal distortion characterization and correction using pilot chirp pulse train for free-running clock synchronization

  Proceedings on CD-ROM - International Society for Magnetic Resonance in Medicine. Scientific Meeting and Exhibition/Proceedings of the International Society for Magnetic Resonance in Medicine, Scientific Meeting and Exhibition · 2025-09-16

  article

  Motivation: to investigate a pilot signal that emulates acquisition bandwidth for reliable frequency offset tracking and corrections of an asynchronous receiver. Goal(s): to achieve successful on-coil digitizer operation over a wireless MRI data link. Approach: by creating synchronous and asynchronous time series using two software-defined radios with pre-defined chirp pulse train and characterizing and correcting distortions on images. Results: Chirp pulse train can emulate image acquisition bandwidth and correct offsets, exhibiting non-coherent residual artifacts when unmixing. Impact: Pilot chirp can improve the usability of tracking and correcting offset distortions associated with asynchronous free-running receivers.

  PublisherDOI

Recent grants

• MRI methods for high resolution imaging of the lung

  NIH · $2.0M · 2018–2024

• Development and Validation of Radiation-Free Pediatric Renal Function Quantification

  NIH · $2.7M · 2018–2023

• Rapid Robust Pediatric MRI

  NIH · $5.6M · 2010–2024

• MRI methods for high resolution imaging of the lung

  NIH · $730k · 2018–2022

• Enabling the Next Generation of High Performance Pediatric Whole Body MR Imaging

  NIH · $3.4M · 2020–2025

Frequent coauthors

• John M. Pauly

  142 shared

• Joseph Y. Cheng

  86 shared

• Marcus T. Alley

  67 shared

• Michael Lustig

  62 shared

• Christopher M. Sandino

  48 shared

• Brian A. Hargreaves

  Stanford University

  40 shared

• Xucheng Zhu

  38 shared

• Frank Ong

  Stanford University

  37 shared

Education

• M.D., Pediatric Radiology

  Stanford University

  2008

• B.S., Human Biology

  Stanford University

  2003