Free-breathing Accelerated Cardiac MRI Using Deep Learning: Validation in Children and Young Adults

Feasibility of the application of deep learning-reconstructed ultra-fast respiratory-triggered T2-weighted imaging at 3 T in liver imaging

OBJECTIVE

The evaluate the feasibility of a novel deep learning-reconstructed ultra-fast respiratory-triggered T2WI sequence (DL-RT-T2WI) In liver imaging, compared with respiratory-triggered Arms-T2WI (Arms-RT-T2WI) and respiratory-triggered FSE-T2WI (FSE-RT-T2WI) sequences.

METHODS

71 patients with liver lesions underwent 3-T MRI and were prospectively enrolled. Two readers independently analyzed images acquired with DL-RT-T2WI, Arms-RT-T2WI, and FSE-RT-T2WI. The qualitative evaluation indicators, including overall image quality (OIQ), sharpness, noise, artifacts, lesion detectability (LC), lesion characterization (LD), cardiacmotion-related signal loss (CSL), and diagnostic confidence (DC), were evaluated in two readers, and further statistically compared using paired Wilcoxon rank-sum test among three sequences.

RESULTS

176 lesions were detected in DL-RT-T2W and Arms-RT-T2WI, and 175 were detected in FSE-RT-T2WI. The acquisition time of DL-RT-T2WI was improved by 4.8-7.9 folds compared to the other two sequences. The OIQ was scored highest for DL-RT-T2WI (R1, 4.61 ± 0.52 and R2, 4.62 ± 0.49), was significantly superior to Arms-RT-T2WI (R1, 4.30 ± 0.66 and R2, 4.34 ± 0.69) and FSE-RT-T2WI (R1, 3.65 ± 1.08 and R2, 3.75 ± 1.01). Artifacts and sharpness scored highest for DL-RT-T2WI, followed by Arms-RT-T2WI, and were lowest for FSE-RT-T2WI in both two readers. Noise and CSL for DL-RT-T2WI scored similar to Arms-RT-T2WI (P > 0.05) and were significantly superior to FSE-RT-T2WI (P < 0.001). Both LD and LC for DL-RT-T2WI were significantly superior to Arms-RT-T2WI and FSE-RT-T2WI in two readers (P < 0.001). DC for DL-RT-T2WI scored best, significantly superior to Arms-RT-T2WI (P < 0.010) and FSE-RT-T2WI (P < 0.001).

CONCLUSIONS

The novel ultra-fast DL-RT-T2WI is feasible for liver imaging and lesion characterization and diagnosis, not only offers a significant improvement in acquisition time but also outperforms Arms-RT-T2WI and FSE-RT-T2WI concerning image quality and DC.

Evaluation of deep learning-based reconstruction late gadolinium enhancement images for identifying patients with clinically unrecognized myocardial infarction

BACKGROUND

The presence of infarction in patients with unrecognized myocardial infarction (UMI) is a critical feature in predicting adverse cardiac events. This study aimed to compare the detection rate of UMI using conventional and deep learning reconstruction (DLR)-based late gadolinium enhancement (LGEO and LGEDL, respectively) and evaluate optimal quantification parameters to enhance diagnosis and management of suspected patients with UMI.

METHODS

This prospective study included 98 patients (68 men; mean age: 55.8 ± 8.1 years) with suspected UMI treated at our hospital from April 2022 to August 2023. LGEO and LGEDL images were obtained using conventional and commercially available inline DLR algorithms. The myocardial signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and percentage of enhanced area (Parea) employing the signal threshold versus reference mean (STRM) approach, which correlates the signal intensity (SI) within areas of interest with the average SI of normal regions, were analyzed. Analysis was performed using the standard deviation (SD) threshold approach (2SD-5SD) and full width at half maximum (FWHM) method. The diagnostic efficacies based on LGEDL and LGEO images were calculated.

RESULTS

The SNRDL and CNRDL were two times better than the SNRO and CNRO, respectively (P < 0.05). Parea-DL was elevated compared to Parea-O using the threshold methods (P < 0.05); however, no intergroup difference was found based on the FWHM method (P > 0.05). The Parea-DL and Parea-O also differed except between the 2SD and 3SD and the 4SD/5SD and FWHM methods (P < 0.05). The receiver operating characteristic curve analysis revealed that each SD method exhibited good diagnostic efficacy for detecting UMI, with the Parea-DL having the best diagnostic efficacy based on the 5SD method (P < 0.05). Overall, the LGEDL images had better image quality. Strong diagnostic efficacy for UMI identification was achieved when the STRM was ≥ 4SD and ≥ 3SD for the LGEDL and LGEO, respectively.

CONCLUSIONS

STRM selection for LGEDL magnetic resonance images helps improve clinical decision-making in patients with UMI. This study underscored the importance of STRM selection for analyzing LGEDL images to enhance diagnostic accuracy and clinical decision-making for patients with UMI, further providing better cardiovascular care.

Unveiling the next generation of MRI contrast agents: current insights and perspectives on ferumoxytol-enhanced MRI

Contrast-enhanced magnetic resonance imaging (CE-MRI) is a pivotal tool for global disease diagnosis and management. Since its clinical availability in 2009, the off-label use of ferumoxytol for ferumoxytol-enhanced MRI (FE-MRI) has significantly reshaped CE-MRI practices. Unlike MRI that is enhanced by gadolinium-based contrast agents, FE-MRI offers advantages such as reduced contrast agent dosage, extended imaging windows, no nephrotoxicity, higher MRI time efficiency and the capability for molecular imaging. As a leading superparamagnetic iron oxide contrast agent, ferumoxytol is heralded as the next generation of contrast agents. This review delineates the pivotal clinical applications and inherent technical superiority of FE-MRI, providing an avant-garde medical-engineering interdisciplinary lens, thus bridging the gap between clinical demands and engineering innovations. Concurrently, we spotlight the emerging imaging themes and new technical breakthroughs. Lastly, we share our own insights on the potential trajectory of FE-MRI, shedding light on its future within the medical imaging realm.

Prospective Comparison of Free-Breathing Accelerated Cine Deep Learning Reconstruction Versus Standard Breath-Hold Cardiac MRI Sequences in Patients With Ischemic Heart Disease

BACKGROUND. Cine cardiac MRI sequences require repeated breath-holds, which can be difficult for patients with ischemic heart disease (IHD). OBJECTIVE. The purpose of the study was to compare a free-breathing accelerated cine sequence using deep learning (DL) reconstruction and a standard breath-hold cine sequence in terms of image quality and left ventricular (LV) measurements in patients with IHD undergoing cardiac MRI. METHODS. This prospective study included patients undergoing 1.5- or 3-T cardiac MRI for evaluation of IHD between March 15, 2023, and June 21, 2023. Examinations included an investigational free-breathing cine short-axis sequence with DL reconstruction (hereafter, cine-DL sequence). Two radiologists (reader 1 [R1] and reader 2 [R2]), in blinded fashion, independently assessed left ventricular ejection fraction (LVEF), left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), and subjective image quality for the cine-DL sequence and a standard breath-hold balanced SSFP sequence; R1 assessed artifacts. RESULTS. The analysis included 26 patients (mean age, 64.3 ± 11.7 [SD] years; 14 men, 12 women). Acquisition was shorter for the cine-DL sequence than the standard sequence (mean ± SD, 0.6 ± 0.1 vs 2.4 ± 0.6 minutes; p < .001). The cine-DL sequence, in comparison with the standard sequence, showed no significant difference for LVEF for R1 (mean ± SD, 51.7% ± 14.3% vs 51.3% ± 14.7%; p = .56) or R2 (53.4% ± 14.9% vs 52.8% ± 14.6%; p = .53); significantly greater LVEDV for R2 (mean ± SD, 171.9 ± 51.9 vs 160.6 ± 49.4 mL; p = .01) but not R1 (171.8 ± 53.7 vs 165.5 ± 52.4 mL; p = .16); and no significant difference in LVESV for R1 (mean ± SD, 88.1 ± 49.3 vs 86.0 ± 50.5 mL; p = .45) or R2 (85.2 ± 48.1 vs 81.3 ± 48.2 mL; p = .10). The mean bias between the cine-DL and standard sequences by LV measurement was as follows: LVEF, 0.4% for R1 and 0.7% for R2; LVEDV, 6.3 mL for R1 and 11.3 mL for R2; and LVESV, 2.1 mL for R1 and 3.9 mL for R2. Subjective image quality was better for cine-DL sequence than the standard sequence for R1 (mean ± SD, 2.3 ± 0.5 vs 1.9 ± 0.8; p = .02) and R2 (2.2 ± 0.4 vs 1.9 ± 0.7; p = .02). R1 reported no significant difference between the cine-DL and standard sequences for off-resonance artifacts (3.8% vs 23.1% examinations; p = .10) and parallel imaging artifacts (3.8% vs 19.2%; p = .19); blurring artifacts were more frequent for the cine-DL sequence than the standard sequence (42.3% vs 7.7% examinations; p = .008). CONCLUSION. A free-breathing cine-DL sequence, in comparison with a standard breath-hold cine sequence, showed very small bias for LVEF measurements and better subjective quality. The cine-DL sequence yielded greater LV volumes than the standard sequence. CLINICAL IMPACT. A free-breathing cine-DL sequence may yield reliable LVEF measurements in patients with IHD unable to repeatedly breath-hold. TRIAL REGISTRATION. ClinicalTrials.gov NCT05105984.

From Compressed-Sensing to Deep Learning MR: Comparative Biventricular Cardiac Function Analysis in a Patient Cohort

BACKGROUND

Conventional segmented, retrospectively gated cine (Conv-cine) is challenged in patients with breath-hold difficulties. Compressed sensing (CS) has shown values in cine imaging but generally requires long reconstruction time. Recent artificial intelligence (AI) has demonstrated potential in fast cine imaging.

PURPOSE

To compare CS-cine and AI-cine with Conv-cine in quantitative biventricular functions, image quality, and reconstruction time.

STUDY TYPE

Prospective human studies.

SUBJECTS

70 patients (age, 39 ± 15 years, 54.3% male).

FIELD STRENGTH/SEQUENCE

3T; balanced steady state free precession gradient echo sequences.

ASSESSMENT

Biventricular functional parameters of CS-, AI-, and Conv-cine were measured by two radiologists independently and compared. The scan and reconstruction time were recorded. Subjective scores of image quality were compared by three radiologists.

STATISTICAL TESTS

Paired t-test and two related-samples Wilcoxon sign test were used to compare biventricular functional parameters between CS-, AI-, and Conv-cine. Intraclass correlation coefficient (ICC), Bland-Altman analysis, and Kendall's W method were applied to evaluate agreement of biventricular functional parameters and image quality of these three sequences. A P-value <0.05 was considered statistically significant, and standardized mean difference (SMD) < 0. 100 was considered no significant difference.

RESULTS

Compared to Conv-cine, no statistically significant differences were identified in CS- and AI-cine function results (all P > 0.05), except for very small differences in left ventricle end-diastole volumes of 2.5 mL (SMD = 0.082) and 4.1 mL (SMD = 0.096), respectively. Bland-Altman scatter plots revealed that biventricular function results were mostly distributed within the 95% confidence interval. All parameters had acceptable to excellent interobserver agreements (ICC: 0.748-0.989). Compared with Conv-cine (84 ± 13 sec), both CS (14 ± 2 sec) and AI (15 ± 2 sec) techniques reduced scan time. Compared with CS-cine (304 ± 17 sec), AI-cine (24 ± 4 sec) reduced reconstruction time. CS-cine demonstrated significantly lower quality scores than Conv-cine, while AI-cine demonstrated similar scores (P = 0.634).

CONCLUSION

CS- and AI-cine can achieve whole-heart cardiac cine imaging in a single breath-hold. Both CS- and AI-cine have the potential to supplement the gold standard Conv-cine in studying biventricular functions and benefit patients having difficulties with breath-holds.

LEVEL OF EVIDENCE

1 TECHNICAL EFFICACY STAGE: 1.

Evaluation of left ventricular and left atrial volumetric function from native MR multislice 4D flow magnitude data

OBJECTIVES

To assess the feasibility, precision, and accuracy of left ventricular (LV) and left atrial (LA) volumetric function evaluation from native magnetic resonance (MR) multislice 4D flow magnitude images.

MATERIALS & METHODS

In this prospective study, 60 subjects without signs or symptoms of heart failure underwent 3T native cardiac MR multislice 4D flow and bSSFP-cine realtime imaging. LV and LA volumetric function parameters were evaluated from 4D flow magnitude (4D flow-cine) and bSSFP-cine data using standard software to obtain end-diastolic volume (EDV), end-systolic volume (ESV), ejection-fraction (EF), stroke-volume (SV), LV muscle mass (LVM), LA maximum volume, LA minimum volume, and LA total ejection fraction (LATEF). Stroke volumes derived from both imaging methods were further compared to 4D pulmonary artery flow-derived net forward volumes (NFV). Methods were compared by correlation and Bland-Altman analysis.

RESULTS

Volumetric function parameters from 4D flow-cine and bSSFP-cine showed high to very high correlations (r = 0.83-0.98). SV, LA volumes and LATEF did not differ between methods. LV end-diastolic and end-systolic volumes were slightly underestimated (EDV: -2.9 ± 5.8 mL; ESV: -2.3 ± 3.8 mL), EF was slightly overestimated (EF: 0.9 ± 2.6%), and LV mass was considerably overestimated (LVM: 39.0 ± 11.4 g) by 4D flow-cine imaging. SVs from both methods correlated very highly with NFV (r = 0.91 in both cases) and did not differ from NFV.

CONCLUSION

Native multislice 4D flow magnitude data allows precise evaluation of LV and LA volumetric parameters; however, apart from SV, LV volumetric parameters demonstrate bias and need to be referred to their respective normal values.

CLINICAL RELEVANCE STATEMENT

Volumetric function assessment from native multislice 4D flow magnitude images can be performed with routinely used clinical software, facilitating the application of 4D flow as a one-stop-shop functional cardiac MR exam, providing consistent, simultaneously acquired, volume and flow data.

KEY POINTS

• Native multislice 4D flow imaging allows evaluation of volumetric left ventricular and atrial function parameters. • Left ventricular and left atrial function parameters derived from native multislice 4D flow data correlate highly with corresponding standard cine-derived parameters. • Multislice 4D flow-derived volumetric stroke volume and net forward volume do not differ.

Ultrafast Brain MRI with Deep Learning Reconstruction for Suspected Acute Ischemic Stroke

Background Deep learning (DL)-accelerated MRI can substantially reduce examination times. However, studies prospectively evaluating the diagnostic performance of DL-accelerated MRI reconstructions in acute suspected stroke are lacking. Purpose To investigate the interchangeability of DL-accelerated MRI with conventional MRI in patients with suspected acute ischemic stroke at 1.5 T. Materials and Methods In this prospective study, 211 participants with suspected acute stroke underwent clinically indicated MRI at 1.5 T between June 2022 and March 2023. For each participant, conventional MRI (including T1-weighted, T2-weighted, T2*-weighted, T2 fluid-attenuated inversion-recovery, and diffusion-weighted imaging; 14 minutes 18 seconds) and DL-accelerated MRI (same sequences; 3 minutes 4 seconds) were performed. The primary end point was the interchangeability between conventional and DL-accelerated MRI for acute ischemic infarction detection. Secondary end points were interchangeability regarding the affected vascular territory and clinically relevant secondary findings (eg, microbleeds, neoplasm). Three readers evaluated the overall occurrence of acute ischemic stroke, affected vascular territory, clinically relevant secondary findings, overall image quality, and diagnostic confidence. For acute ischemic lesions, size and signal intensities were assessed. The margin for interchangeability was chosen as 5%. For interrater agreement analysis and interrater reliability analysis, multirater Fleiss κ and the intraclass correlation coefficient, respectively, was determined. Results The study sample consisted of 211 participants (mean age, 65 years ± 16 [SD]); 123 male and 88 female). Acute ischemic stroke was confirmed in 79 participants. Interchangeability was demonstrated for all primary and secondary end points. No individual equivalence indexes (IEIs) exceeded the interchangeability margin of 5% (IEI, -0.002 [90% CI: -0.007, 0.004]). Almost perfect interrater agreement was observed (P > .91). DL-accelerated MRI provided higher overall image quality (P < .001) and diagnostic confidence (P < .001). The signal properties of acute ischemic infarctions were similar in both techniques and demonstrated good to excellent interrater reliability (intraclass correlation coefficient, ≥0.8). Conclusion Despite being four times faster, DL-accelerated brain MRI was interchangeable with conventional MRI for acute ischemic lesion detection. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Haller in this issue.

The future of CMR: All-in-one vs. real-time CMR (Part 2)

Cardiac Magnetic Resonance (CMR) protocols can be lengthy and complex, which has driven the research community to develop new technologies to make these protocols more efficient and patient-friendly. Two different approaches to improving CMR have been proposed, specifically "all-in-one" CMR, where several contrasts and/or motion states are acquired simultaneously, and "real-time" CMR, in which the examination is accelerated to avoid the need for breathholding and/or cardiac gating. The goal of this two-part manuscript is to describe these two different types of emerging rapid CMR protocols. To this end, the vision of all-in-one and real-time imaging are described, along with techniques which have been devised and tested along the pathway of clinical implementation. The pros and cons of the different methods are presented, and the remaining open needs of each are detailed. Part 1 tackles the "All-in-One" approaches, and Part 2 focuses on the "Real-Time" approaches along with an overall summary of these emerging methods.

A new compressed sensing cine cardiac MRI sequence with free-breathing real-time acquisition and fully automated motion-correction: A comprehensive evaluation

PURPOSE

The purpose of this study was to compare a new free-breathing compressed sensing cine (FB-CS) cardiac magnetic resonance imaging (CMR) to the standard reference multi-breath-hold segmented cine (BH-SEG) CMR in an unselected population.

MATERIALS AND METHODS

From January to April 2021, 52 consecutive adult patients who underwent both conventional BH-SEG CMR and new FB-CS CMR with fully automated respiratory motion correction were retrospectively enrolled. There were 29 men and 23 women with a mean age of 57.7 ± 18.9 (standard deviation [SD]) years (age range: 19.0-90.0 years) and a mean cardiac rate of 74.6 ± 17.9 (SD) bpm. For each patient, short-axis stacks were acquired with similar parameters providing a spatial resolution of 1.8 × 1.8 × 8.0 mm3 and 25 cardiac frames. Acquisition and reconstruction times, image quality (Likert scale from 1 to 4), left and right ventricular volumes and ejection fractions, left ventricular mass, and global circumferential strain were assessed for each sequence.

RESULTS

FB-CS CMR acquisition time was significantly shorter (123.8 ± 28.4 [SD] s vs. 267.2 ± 39.3 [SD] s for BH-SEG CMR; P < 0.0001) at the penalty of a longer reconstruction time (271.4 ± 68.7 [SD] s vs. 9.9 ± 2.1 [SD] s for BH-SEG CMR; P < 0.0001). In patients without arrhythmia or dyspnea, FB-CS CMR provided subjective image quality that was not different from that of BH-SEG CMR (P = 0.13). FB-CS CMR improved image quality in patients with arrhythmia (n = 18; P = 0.002) or dyspnea (n = 7; P = 0.02), and the edge sharpness was improved at end-systole and end-diastole (P = 0.0001). No differences were observed between the two techniques in ventricular volumes and ejection fractions, left ventricular mass or global circumferential strain in patients in sinus rhythm or with cardiac arrhythmia.

CONCLUSION

This new FB-CS CMR addresses respiratory motion and arrhythmia-related artifacts without compromising the reliability of ventricular functional assessment.

Accelerated Cine Cardiac MRI Using Deep Learning-Based Reconstruction: A Systematic Evaluation

BACKGROUND

Breath-holding (BH) for cine balanced steady state free precession (bSSFP) imaging is challenging for patients with impaired BH capacity. Deep learning-based reconstruction (DLR) of undersampled k-space promises to shorten BHs while preserving image quality and accuracy of ventricular assessment.

PURPOSE

To perform a systematic evaluation of DLR of cine bSSFP images from undersampled k-space over a range of acceleration factors.

STUDY TYPE

Retrospective.

SUBJECTS

Fifteen pectus excavatum patients (mean age 16.8 ± 5.4 years, 20% female) with normal cardiac anatomy and function and 12-second BH capability.

FIELD STRENGTH/SEQUENCE

1.5-T, cine bSSFP.

ASSESSMENT

Retrospective DLR was conducted by applying compressed sensitivity encoding (C-SENSE) acceleration to systematically undersample fully sampled k-space cine bSSFP acquisition data over an acceleration/undersampling factor (R) considering a range of 2 to 8. Quality imperceptibility (QI) measures, including structural similarity index measure, were calculated using images reconstructed from fully sampled k-space as a reference. Image quality, including contrast and edge definition, was evaluated for diagnostic adequacy by three readers with varying levels of experience in cardiac MRI (>4 years, >18 years, and 1 year). Automated DL-based biventricular segmentation was performed commercially available software by cardiac radiologists with more than 4 years of experience.

STATISTICAL TESTS

Tukey box plots, linear mixed effects model, analysis of variance (ANOVA), weighted kappa, Kruskal-Wallis test, and Wilcoxon signed-rank test were employed as appropriate. A P-value <0.05 was considered statistically significant.

RESULTS

There was a significant decrease in the QI values and edge definition scores as R increased. Diagnostically adequate image quality was observed up to R = 5. The effect of R on all biventricular volumetric indices was non-significant (P = 0.447).

DATA CONCLUSION

The biventricular volumetric indices obtained from the reconstruction of fully sampled cine bSSFP acquisitions and DLR of the same k-space data undersampled by C-SENSE up to R = 5 may be comparable.

EVIDENCE LEVEL

3 TECHNICAL EFFICACY: Stage 1.

Assessment of Right Ventricular Function-a State of the Art

PURPOSE OF REVIEW

The right ventricle (RV) has a complex geometry and physiology which is distinct from the left. RV dysfunction and failure can be the aftermath of volume- and/or pressure-loading conditions, as well as myocardial and pericardial diseases.

RECENT FINDINGS

Echocardiography, magnetic resonance imaging and right heart catheterisation can assess RV function by using several qualitative and quantitative parameters. In pulmonary hypertension (PH) in particular, RV function can be impaired and is related to survival. An accurate assessment of RV function is crucial for the early diagnosis and management of these patients. This review focuses on the different modalities and indices used for the evaluation of RV function with an emphasis on PH.

Ngo L, J. Manning W, Nezafat R. Accelerated Cardiac MRI Cine with Use of Resolution Enhancement Generative Adversarial Inline Neural Network

Background Cardiac cine can benefit from deep learning-based image reconstruction to reduce scan time and/or increase spatial and temporal resolution. Purpose To develop and evaluate a deep learning model that can be combined with parallel imaging or compressed sensing (CS). Materials and Methods The deep learning model was built on the enhanced super-resolution generative adversarial inline neural network, trained with use of retrospectively identified cine images and evaluated in participants prospectively enrolled from September 2021 to September 2022. The model was applied to breath-hold electrocardiography (ECG)-gated segmented and free-breathing real-time cine images collected with reduced spatial resolution with use of generalized autocalibrating partially parallel acquisitions (GRAPPA) or CS. The deep learning model subsequently restored spatial resolution. For comparison, GRAPPA-accelerated cine images were collected. Diagnostic quality and artifacts were evaluated by two readers with use of Likert scales and compared with use of Wilcoxon signed-rank tests. Agreement for left ventricle (LV) function, volume, and strain was assessed with Bland-Altman analysis. Results The deep learning model was trained on 1616 patients (mean age ± SD, 56 years ± 16; 920 men) and evaluated in 181 individuals, 126 patients (mean age, 57 years ± 16; 77 men) and 55 healthy subjects (mean age, 27 years ± 10; 15 men). In breath-hold ECG-gated segmented cine and free-breathing real-time cine, the deep learning model and GRAPPA showed similar diagnostic quality scores (2.9 vs 2.9, P = .41, deep learning vs GRAPPA) and artifact score (4.4 vs 4.3, P = .55, deep learning vs GRAPPA). Deep learning acquired more sections per breath-hold than GRAPPA (3.1 vs one section, P < .001). In free-breathing real-time cine, the deep learning showed a similar diagnostic quality score (2.9 vs 2.9, P = .21, deep learning vs GRAPPA) and lower artifact score (3.9 vs 4.3, P < .001, deep learning vs GRAPPA). For both sequences, the deep learning model showed excellent agreement for LV parameters, with near-zero mean differences and narrow limits of agreement compared with GRAPPA. Conclusion Deep learning-accelerated cardiac cine showed similarly accurate quantification of cardiac function, volume, and strain to a standardized parallel imaging method. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Vannier and Wang in this issue.

Rapid 3D breath-hold MR cholangiopancreatography using deep learning-constrained compressed sensing reconstruction

OBJECTIVES

To compare the image quality of three-dimensional breath-hold magnetic resonance cholangiopancreatography with deep learning-based compressed sensing reconstruction (3D DL-CS-MRCP) to those of 3D breath-hold MRCP with compressed sensing (3D CS-MRCP), 3D breath-hold MRCP with gradient and spin-echo (3D GRASE-MRCP) and conventional 2D single-shot breath-hold MRCP (2D MRCP).

METHODS

In total, 102 consecutive patients who underwent MRCP at 3.0 T, including 2D MRCP, 3D GRASE-MRCP, 3D CS-MRCP, and 3D DL-CS-MRCP, were prospectively included. Two radiologists independently analyzed the overall image quality, background suppression, artifacts, and visualization of pancreaticobiliary ducts using a five-point scale. The signal-to-noise ratio (SNR) of the common bile duct (CBD), contrast-to-noise ratio (CNR) of the CBD and liver, and contrast ratio between the periductal tissue and CBD were measured. The Friedman test was performed to compare the four protocols.

RESULTS

3D DL-CS-MRCP resulted in improved SNR and CNR values compared with those in the other three protocols, and better contrast ratio compared with that in 3D CS-MRCP and 3D GRASE-MRCP (all, p < 0.05). Qualitative image analysis showed that 3D DL-CS-MRCP had better performance for second-level intrahepatic ducts and distal main pancreatic ducts compared with 3D CS-MRCP (all, p < 0.05). Compared with 2D MRCP, 3D DL-CS-MRCP demonstrated better performance for the second-order left intrahepatic duct but was inferior in assessing the main pancreatic duct (all, p < 0.05). Moreover, the image quality was significantly higher in 3D DL-CS-MRCP than in 3D GRASE-MRCP.

CONCLUSION

3D DL-CS-MRCP has superior performance compared with that of 3D CS-MRCP or 3D GRASE-MRCP. Deep learning reconstruction also provides a comparable image quality but with inferior main pancreatic duct compared with that revealed by 2D MRCP.

KEY POINTS

• 3D breath-hold MRCP with deep learning reconstruction (3D DL-CS-MRCP) demonstrated improved image quality compared with that of 3D MRCP with compressed sensing or GRASE. • Compared with 2D MRCP, 3D DL-CS-MRCP had superior performance in SNR and CNR, better visualization of the left second-level intrahepatic bile ducts, and comparable overall image quality, but an inferior main pancreatic duct.

Deep Learning-Based Reconstruction for Cardiac MRI: A Review

Cardiac magnetic resonance (CMR) is an essential clinical tool for the assessment of cardiovascular disease. Deep learning (DL) has recently revolutionized the field through image reconstruction techniques that allow unprecedented data undersampling rates. These fast acquisitions have the potential to considerably impact the diagnosis and treatment of cardiovascular disease. Herein, we provide a comprehensive review of DL-based reconstruction methods for CMR. We place special emphasis on state-of-the-art unrolled networks, which are heavily based on a conventional image reconstruction framework. We review the main DL-based methods and connect them to the relevant conventional reconstruction theory. Next, we review several methods developed to tackle specific challenges that arise from the characteristics of CMR data. Then, we focus on DL-based methods developed for specific CMR applications, including flow imaging, late gadolinium enhancement, and quantitative tissue characterization. Finally, we discuss the pitfalls and future outlook of DL-based reconstructions in CMR, focusing on the robustness, interpretability, clinical deployment, and potential for new methods.

Memory-efficient model-based deep learning with convergence and robustness guarantees

Computational imaging has been revolutionized by compressed sensing algorithms, which offer guaranteed uniqueness, convergence, and stability properties. Model-based deep learning methods that combine imaging physics with learned regularization priors have emerged as more powerful alternatives for image recovery. The main focus of this paper is to introduce a memory efficient model-based algorithm with similar theoretical guarantees as CS methods. The proposed iterative algorithm alternates between a gradient descent involving the score function and a conjugate gradient algorithm to encourage data consistency. The score function is modeled as a monotone convolutional neural network. Our analysis shows that the monotone constraint is necessary and sufficient to enforce the uniqueness of the fixed point in arbitrary inverse problems. In addition, it also guarantees the convergence to a fixed point, which is robust to input perturbations. We introduce two implementations of the proposed MOL framework, which differ in the way the monotone property is imposed. The first approach enforces a strict monotone constraint, while the second one relies on an approximation. The guarantees are not valid for the second approach in the strict sense. However, our empirical studies show that the convergence and robustness of both approaches are comparable, while the less constrained approximate implementation offers better performance. The proposed deep equilibrium formulation is significantly more memory efficient than unrolled methods, which allows us to apply it to 3D or 2D+time problems that current unrolled algorithms cannot handle.

Reliability of respiratory-gated real-time two-dimensional cine incorporating deep learning reconstruction for the assessment of ventricular function in an adult population

This study aimed to assess the image quality and accuracy of respiratory-gated real-time two-dimensional (2D) cine incorporating deep learning reconstruction (DLR) for the quantification of biventricular volumes and function compared with those of the standard reference, that is, breath-hold 2D balanced steady-state free precession (bSSFP) cine, in an adult population. Twenty-four patients (15 men, mean age 50.7 ± 16.5 years) underwent cardiac magnetic resonance for clinical indications, and 2D DLR and bSSFP cine were acquired on the short-axis view. The image quality scores were based on three main criteria: blood-to-myocardial contrast, endocardial edge delineation, and presence of motion artifacts throughout the cardiac cycle. Biventricular end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume (SV), ejection fraction (EF), and left ventricular mass (LVM) were analyzed. The 2D DLR cine had significantly shorter scan time than bSSFP (41.0 ± 11.3 s vs. 327.6 ± 65.8 s; p < 0.0001). Despite an analysis of endocardial edge definition and motion artifacts showed significant impairment using DLR cine compared with bSSFP (p < 0.01), the two sequences demonstrated no significant difference in terms of biventricular EDV, ESV, SV, and EF (p > 0.05). Moreover, the linear regression yielded good agreement between the two techniques (r ≥ 0.76). However, the LVM was underestimated for DLR cine (109.8 ± 34.6 g) compared with that for bSSFP (116.2 ± 40.2 g; p = 0.0291). Respiratory-gated 2D DLR cine is a reliable technique that could be used in the evaluation of biventricular volumes and function in an adult population.

Respiratory Motion Management in Abdominal MRI: Radiology In Training

A 96-year-old woman had a suboptimal evaluation of liver observations at abdominal MRI due to significant respiratory motion. State-of-the-art strategies to minimize respiratory motion during clinical abdominal MRI are discussed.

Accelerated two-dimensional phase-contrast for cardiovascular MRI using deep learning-based reconstruction with complex difference estimation

PURPOSE

To develop and validate a deep learning-based reconstruction framework for highly accelerated two-dimensional (2D) phase contrast (PC-MRI) data with accurate and precise quantitative measurements.

METHODS

We propose a modified DL-ESPIRiT reconstruction framework for 2D PC-MRI, comprised of an unrolled neural network architecture with a Complex Difference estimation (CD-DL). CD-DL was trained on 155 fully sampled 2D PC-MRI pediatric clinical datasets. The fully sampled data ( n = 29 $$ n=29 $$ ) was retrospectively undersampled (6-11 × $$ \times $$ ) and reconstructed using CD-DL and a parallel imaging and compressed sensing method (PICS). Measurements of peak velocity and total flow were compared to determine the highest acceleration rate that provided accuracy and precision within ± 5 % $$ \pm 5\% $$ . Feasibility of CD-DL was demonstrated on prospectively undersampled datasets acquired in pediatric clinical patients ( n = 5 $$ n=5 $$ ) and compared to traditional parallel imaging (PI) and PICS.

RESULTS

The retrospective evaluation showed that 9 × $$ \times $$ accelerated 2D PC-MRI images reconstructed with CD-DL provided accuracy and precision (bias, [95 % $$ \% $$ confidence intervals]) within ± 5 % $$ \pm 5\% $$ . CD-DL showed higher accuracy and precision compared to PICS for measurements of peak velocity (2.8 % $$ \% $$ [ - 2 . 9 $$ -2.9 $$ , 4.5] vs. 3.9 % $$ \% $$ [ - 11 . 0 $$ -11.0 $$ , 4.9]) and total flow (1.8 % $$ \% $$ [ - 3 . 9 $$ -3.9 $$ , 3.4] vs. 2.9 % $$ \% $$ [ - 7 . 1 $$ -7.1 $$ , 6.9]). The prospective feasibility study showed that CD-DL provided higher accuracy and precision than PICS for measurements of peak velocity and total flow.

CONCLUSION

In a retrospective evaluation, CD-DL produced quantitative measurements of 2D PC-MRI peak velocity and total flow with ≤ 5 % $$ \le 5\% $$ error in both accuracy and precision for up to 9 × $$ \times $$ acceleration. Clinical feasibility was demonstrated using a prospective clinical deployment of our 8 × $$ \times $$ undersampled acquisition and CD-DL reconstruction in a cohort of pediatric patients.

Compact pediatric cardiac magnetic resonance imaging protocols

Cardiac MRI is in many respects an ideal modality for pediatric cardiovascular imaging, enabling a complete noninvasive assessment of anatomy, morphology, function and flow in one radiation-free and potentially non-contrast exam. Nonetheless, traditionally lengthy and complex imaging acquisition strategies have often limited its broader use beyond specialized centers. In this review, the author presents practical cardiac MRI imaging protocols to facilitate the performance of succinct yet successful exams that provide the most salient clinical data for the majority of congenital and acquired pediatric cardiac disease. In addition, the author reviews newer and evolving techniques that permit more rapid but similarly diagnostic MRI, including compressed sensing and artificial intelligence/machine learning reconstruction, four-dimensional flow acquisition and blood pool contrast agents. With the modern armamentarium of cardiac MRI methods, the goal of compact yet comprehensive exams in children can now be realized.