K-space Optimized Spatial Temporal Architecture for Low Latency in Interventional Cardiac MRI
- 1University of California, San Diego
Abstract
PurposeTo develop and evaluate a recurrent neural network that operates directly on undersampled golden-angle radial k-space, enabling low-latency cardiac cine reconstruction for interventional cardiac MRI (iCMR).
MethodsKOSTAL (K-space Optimized Spatial Temporal Architecture for Low latency) combines a per-coil k-space convolutional gated recurrent unit (convGRU), which learns a mask-conditioned view-sharing rule over the accumulated k-space, with an image-space recurrent module. It was trained and tested on the open-access multi-coil OCMR dataset using 10 simulated golden-angle spokes per frame against a 1000-spoke reference. Ablations varied recurrent module architecture and placement; temporal behavior was assessed using a imaging-plane transition test. Generalization was evaluated without retraining in 23 pediatric patients prospectively scanned during free-breathing using a radial golden-angle bSSFP sequence with binned XD-GRASP and clinical breath-held cine serving as references.
ResultsAt 10 spokes per frame, KOSTAL reached a median SSIM of 0.907 (gridding: 0.622) in under 34 ms of processing time per frame. Image quality (IQ) matched gridded results formed from 7-11 times more spokes. Recurrence in k-space rather than image space accounted for most of this gain (0.891 versus [≤]0.650). IQ recovered from an abrupt plane transition in 75 ms, versus 303 ms for 200-spoke gridding. KOSTAL had the best agreement with the binned reference and the highest long-axis CNR. KOSTAL accurately tracked left-ventricular blood-pool volume (end-diastolic volume bias-0.7 mL) with a shorter lag (71 ms) than iterative reconstruction (239-287 ms).
ConclusionLearned, mask-conditioned recurrence in k-space yields high image quality at a latency compatible with interventional guidance.
Related
- An EfficientNet hybrid deep learning detection framework of network intrusion with an optimized mixture-of-experts architecture
- Spatial robustness of P100 latency compared with amplitude: a data-driven rationale for visual pathway monitoring during neuroendovascular procedures
- Traffic surveillance video-based a unified road accident detection, classification, and vehicle tracking framework using an optimized spatio-temporal attention mechanism
- Preserving RNA integrity and morphology in whole rat allograft hindlimb sections: an optimized protocol for spatial transcriptomics and immunofluorescence
- Spatial and Temporal Evaluation of Violent and Total Crime Following Community-Led Urban Greening
BibTeX
@misc{buckup2026kspace,
title = {{K-space Optimized Spatial Temporal Architecture for Low Latency in Interventional Cardiac MRI}},
author = {M. Buckup and N. Mahajan and A. E. Rodriguez-Soto and J. Blood and S. Hegde and E. L. Schuchardt and B. Printz and H. K. Narayan and B. L. Gordon and F. Contijoch},
howpublished = {medRxiv},
year = {2026},
month = sep,
doi = {10.64898/2026.09.21.26363257},
url = {https://www.medrxiv.org/content/10.64898/2026.09.21.26363257v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - K-space Optimized Spatial Temporal Architecture for Low Latency in Interventional Cardiac MRI
AU - M. Buckup
AU - N. Mahajan
AU - A. E. Rodriguez-Soto
AU - J. Blood
AU - S. Hegde
AU - E. L. Schuchardt
AU - B. Printz
AU - H. K. Narayan
AU - B. L. Gordon
AU - F. Contijoch
PY - 2026
DA - 2026/09/21
PB - medRxiv
DO - 10.64898/2026.09.21.26363257
UR - https://www.medrxiv.org/content/10.64898/2026.09.21.26363257v1
AB - PurposeTo develop and evaluate a recurrent neural network that operates directly on undersampled golden-angle radial k-space, enabling low-latency cardiac cine reconstruction for interventional cardiac MRI (iCMR). MethodsKOSTAL (K-space Optimized Spatial Temporal Architecture for Low latency) combines a per-coil k-space convolutional gated recurrent unit (convGRU), which learns a mask-conditioned view-sharing rule over the accumulated k-space, with an image-space recurrent module. It was trained and tested on the open-access multi-coil OCMR dataset using 10 simulated golden-angle spokes per frame against a 1000-spoke reference. Ablations varied recurrent module architecture and placement; temporal behavior was assessed using a imaging-plane transition test. Generalization was evaluated without retraining in 23 pediatric patients prospectively scanned during free-breathing using a radial golden-angle bSSFP sequence with binned XD-GRASP and clinical breath-held cine serving as references. ResultsAt 10 spokes per frame, KOSTAL reached a median SSIM of 0.907 (gridding: 0.622) in under 34 ms of processing time per frame. Image quality (IQ) matched gridded results formed from 7-11 times more spokes. Recurrence in k-space rather than image space accounted for most of this gain (0.891 versus [≤]0.650). IQ recovered from an abrupt plane transition in 75 ms, versus 303 ms for 200-spoke gridding. KOSTAL had the best agreement with the binned reference and the highest long-axis CNR. KOSTAL accurately tracked left-ventricular blood-pool volume (end-diastolic volume bias-0.7 mL) with a shorter lag (71 ms) than iterative reconstruction (239-287 ms). ConclusionLearned, mask-conditioned recurrence in k-space yields high image quality at a latency compatible with interventional guidance.
ER -