Advanced Techniques for Rapid High-Resolution First-Pass CMR Perfusion with Whole-Heart Coverage

Coronary artery disease (CAD) is a major public health concern. According to a report from the AHA in 2020, an estimated 18.2 million adult Americans have CAD. CAD is responsible for 1 in every 7 deaths in the United States. Cardiac magnetic resonance (CMR) quantitative myocardial first-pass perfusion imaging is a non-invasive and non-ionizing technique for diagnosing CAD which provides an accurate assessment of myocardial ischemia and a comprehensive evaluation of myocardial function and infarction.

Despite multiple potential advantages of CMR perfusion imaging, current clinically available techniques have limited in-plane spatial resolution (~2-3 mm) and incomplete heart coverage, which impede the assessment of transmural perfusion differences and underestimate the extent of ischemia. Furthermore, motion-induced dark-rim artifacts can significantly reduce image quality and limit evaluation of the sub-endocardium, which is most sensitive to myocardial ischemia. Recently, studies from our lab have demonstrated CMR quantitative spiral perfusion techniques for both interleaved single-slice (SS) and simultaneous multi-slice (SMS) acquisitions enabling whole-heart coverage (6-8 slices) with high spatial resolution (2 mm). Our lab has developed novel motion-compensated compressed-sensing (CS) L1-SPIRiT reconstruction techniques, that correct for breathing motion and enable free-breathing acquisition. Sampling efficiency can also be improved by using outer-volume suppression (OVS) technique to achieve a reduced field-of-view (rFOV) so that the sampling in k-space can be coarser. We have previously applied an OVS technique for single-shot spiral perfusion imaging and demonstrated that it produced superior image quality as compared with full-FOV acquisitions. We have also developed a quantification pipeline for spiral perfusion imaging to quantify myocardial blood flow and myocardial perfusion reserve (MPR). High diagnostic accuracy of the proposed techniques has been demonstrated.

With higher spatial resolution, there is an increased ability to detect transmural perfusion differences between the epicardium and the endocardium, which could improve the ability to detecting obstructive CAD as demonstrated in prior studies. Additionally, one significant barrier to clinical translation of these techniques is the need for off-line reconstruction and quantification which currently takes hours to complete, and thus cannot provide data to physician in a clinically acceptable time frame. Considering that greater than 10 million stress tests are performed in the US alone, improvements in the accuracy of non-invasive assessment of CAD could significantly reduce health care costs resulting from incorrect diagnoses. In this dissertation, we propose to develop advanced rapid and high-resolution imaging techniques for first-pass myocardial perfusion with whole-heart coverage.

Specific Aim #1 is to develop high spatial resolution spiral first-pass myocardial perfusion imaging with whole-heart coverage at 3 T. (a) Optimize spiral perfusion pulse sequences for both SS and SMS acquisitions with or without OVS to address the higher undersampling factors required to achieve 1.25 mm spatial resolution with high temporal resolution and whole-heart coverage. (b) Optimize the motion- compensated L1-SPIRiT image reconstruction technique and develop the motion-compensated SMS-Slice-L1-SPIRiT reconstruction technique that incorporates through-plane kernels for spiral SMS imaging that could reduce the slice leakage and improve image quality. (c) Evaluate image quality of the proposed technique in both healthy volunteers and patients undergoing clinically ordered CMR studies.

Specific Aim #2 is to develop DEep learning-based rapid Spiral Image REconstruction (DESIRE) for high-resolution spiral first-pass myocardial perfusion imaging for both 3 T and 1.5 T. (a) Develop a DEep learning-based rapid Spiral Image REconstruction technique (DESIRE) for high-resolution spiral first-pass myocardial perfusion imaging for both SS and SMS MB=2 acquisitions with whole-heart coverage. (b) Assess the image reconstruction network performance with varying factors including data type, convolutional units, etc. (c) Validate the proposed technique in both healthy volunteers and patients as compared to the CS-based L1-SPIRiT reconstructions.

Specific Aim #3 is to develop quantitative perfusion imaging with Cartesian acquisition and compare it to spiral perfusion imaging. (a) Develop the 2D and SMS Cartesian perfusion sequence with Poisson-disc acquisition pattern along k-t dimension. (b) Apply the k-t based image reconstruction technique for this Cartesian acquisition. (c) Develop the deep learning-based rapid image reconstruction techniques for Cartesian 2D and SMS perfusion imaging. (d) Validate the proposed technique in both healthy volunteers and patients.