Journal of Cardiovascular Magnetic Resonance最新文献

Background: Acute aortic syndromes in Marfan syndrome (MFS) often occur before reaching the surgical diameter threshold, highlighting the need for new imaging biomarkers.

Objectives: Aim was to compare cardiovascular magnetic resonance (CMR)-derived aortic three-dimensional (3D) distensibility and displacement in MFS patients with or without a history of aortic root surgery (RR or native) and healthy volunteers.

Methods: The participants underwent 3T CMR of the thoracic aorta using an accelerated non-contrast-enhanced, free breathing, 3D cine balanced steady state free precession sequence, with spatiotemporal resolution: (1.0 mm)³/∼33ms. A deep learning-based algorithm was used to obtain aorta segmentations. Non-rigid registration of these segmentations was subsequently used to calculate 3D distensibility and its separate components: 2-dimensional distensibility, longitudinal strain, and displacement in the ascending (AAo) and descending aorta (DAo).

Results: Forty-seven volunteers, 51 native, and 33 RR MFS patients were included. AAo and DAo distensibility (10^-3*mmHg^-1) were different for healthy volunteers vs native vs RR patients (AAo: 5.1±1.4 vs 3.6±1.4 vs. 1.4±0.7, p<0.001, DAo: 3.2±1.1 vs. 2.5±0.9 vs 2.4±1.0, p=0.001). Sinotubular junction displacement (mm) was significantly higher for healthy volunteers vs native MFS vs RR MFS patients (10.3±1.3 vs 8.7±2.1 vs 5.7±1.6, p<0.001). In native patients, age (β=-0.06 (95% CI:-0.10 to -0.01), p=0.014) and root diameter (β=-0.1 (95% CI: -0.19 to -0.02), p=0.018) were negatively associated with AAo 3D distensibility, independent of male sex, body surface area, and aortic tortuosity index.

Conclusion: Aortic 3D distensibility and displacement, derived from 4-dimensional CMR, were significantly diminished in MFS compared to volunteers and should be investigated longitudinally to assess their potential value in predicting aortic events and guiding therapy.

Background: Cardiac magnetic resonance imaging (CMR) studies contain a wealth of information on a patient's cardiovascular status. The ability to extract this data from free-text reports could serve to automate clinical decision support tools and generate data for retrospective clinical knowledge discovery, and clinical operational purposes. Few studies have examined the automatic extraction of data from free-text CMR reports, and the existing studies that do have key limitations, including small sample size and disease-specific data extraction. Existing studies also fail to extract features associated with the cardiovascular conditions that reflect nuances in natural language, such as uncertainty, severity, subtype, and anatomical locations of the condition. The goal of this study was to build a broad named entity recognition model to automatically extract a broad variety of common CMR findings and their associated attributes from CMR reports.

Methods: We fine-tuned a Large Language Model Meta AI (LLaMA) model trained to identify 34 cardiovascular conditions and their associated attributes, including certainty, severity, location, and subtype of the condition. This model was trained on 1778 MRI reports and tested on 397 reports in an held-out test set and another 428 reports from another site in our hospital system with independent radiology practice and scanners.

Results: Our model shows robust performance in predicting the mention of the 31 cardiovascular conditions (average F1=0.85). It also showed strong performance predicting attributes, including certainty (average F1=0.97) and severity (average F1=0.97). Model performance on the external validation set was generally slightly lower than the internal validation set, but performance was still strong (average F1=0.78 for mention, 0.97 for certainty, and 0.96 for severity).

Conclusion: CMR-LLaMA has strong performance identifying a variety of concept mentions and moderate accuracies in extracting a selection of other associated attributes. NLP models can be used to automate the extraction of data from CMR reports to potentially assist with clinical and research workflow.

Background: First-pass stress-perfusion cardiovascular magnetic resonance (CMR) imaging is the guidelines-recommended non-invasive test for the detection of obstructive coronary artery disease (CAD). Recently developed quantitative perfusion CMR (QP CMR) allows quantification of myocardial blood flow. Moreover, the latest developments established several methods of CAD assessment without the need for a contrast agent, including stress T1 mapping reactivity (∆T1) and oxygenation-sensitive CMR (OS-CMR). These methods might eliminate the need for contrast administration in clinical practice, reducing time, invasiveness, and costs, thereby simplifying the evaluation of patients with suspected obstructive CAD. The ADVOCATE-CMR study aims to validate QP CMR, ∆T1, and OS-CMR imaging against invasive fractional flow reserve (FFR) for the detection of obstructive CAD. The study also aims to head-to-head compare the diagnostic accuracy of these CMR techniques with the conventional visual assessment of stress-perfusion CMR and to correlate them to short- and long-term clinical outcomes.

Study design and methodology: ADVOCATE-CMR is a single-center, observational, prospective, cross-sectional cohort study. The study will enroll 182 symptomatic patients with suspected obstructive CAD scheduled for invasive coronary angiography (ICA). Before ICA, all participants will undergo CMR imaging, including OS-CMR with breathing maneuvers, rest, and adenosine stress T1 mapping and rest and adenosine stress first-pass perfusion. Subsequently, ICA will be performed, including FFR, instantaneous wave-free ratio, resting Pd/Pa, coronary flow reserve, and index of microvascular resistance measurements in all main coronary arteries. A follow-up CMR scan with the same protocol will be performed at 3 months after ICA. Clinical follow-up will be performed at 3, 6 months, 1 and 3 years after ICA.

Conclusion: The ADVOCATE-CMR will be the first study comprehensively evaluating and comparing head-to-head the diagnostic performance of a range of contrast- and non-contrast agent-based CMR imaging methods (including QP CMR, ∆T1, and OS-CMR) for the detection of FFR-defined obstructive CAD. We expect to establish a validated and time-efficient diagnostic workflow available to a wide range of general CMR services. Finally, these improvements may enable CMR to become an effective non-invasive, radiation-free gatekeeper for ICA in patients with suspected obstructive CAD, potentially without the need for a contrast agent.

Background: Cardiovascular magnetic resonance (CMR) has a growing role in the diagnosis and management of cardiac disease. However, there is little recent data on the availability of CMR physicians (readers) in the United States (US).

Objective: To demonstrate the geographic proximity and accessibility of patients to CMR services and CMR physicians across the US.

Methods: Using Medicare Part B data in 2022, we analyzed the number and characteristics of CMR readers, their geographical location, and the volume of CMR scans between 2013 and 2022. CMR procedure types were identified using healthcare common procedure coding system (HCPCS) codes 75557, 75559, 75561, and 75563.

Results: Among Medicare beneficiaries in 2022, there were 48,622 CMR scans, up from 17,944 in 2013 (170.9% increase). The lowest scans and reader density were in West Virginia (125.8 procedures and 2.2 readers per million beneficiaries, respectively) and the highest in the District of Columbia (4566.5 procedures and 52.9 readers per million beneficiaries, respectively). No CMR scans were billed in Puerto Rico. Among states and territories that billed for CMR, 50.8 million U.S. citizens were located more than 50 miles from CMR readers and 18.1 million were located more than 100 miles away. Out of 991 readers, 51.9% were radiologists and 48.1% were cardiologists. The median number of scans interpreted by cardiologists was higher than radiologists across all graduation year intervals, and male and female readers interpreted a similar median number of scans. The relative proportion of female readers increased markedly when assessing physicians who graduated after 2010.

Conclusion: This study highlights significant geographic disparities and barriers to accessing CMR in the US.

Background: Radiomics has been proven to be an important method for the quantitative assessment of atherosclerotic plaques. Therefore, we aimed to evaluate a radiomics approach based on 7.0T high-resolution vessel wall imaging (HR-VWI) to identify culprit middle cerebral artery (MCA) plaques associated with subcortical infarctions.

Methods: One hundred patients with MCA plaques were prospectively enrolled. Among these patients, 145 plaques (74 culprit plaques and 71 non-culprit plaques) were included. A traditional model was constructed by recording the conventional radiological plaque characteristics of HR-VWI. Radiomics features from HR-VWI images were utilized to construct a radiomics model. A combined model was built using both conventional radiological and radiomics features. Receiver operating characteristic (ROC) curves and area under curve (AUC) were used to compare the performance of these models.

Results: Plaque surface irregularity and superior wall location of MCA plaques were independently associated with subcortical infarctions. The traditional model had AUCs of 0.744 and 0.700 in the training and test sets, respectively. The radiomics and the combined model showed improved AUCs: 0.860 and 0.896 in the training sets and 0.795 and 0.833 in the test sets, respectively. The radiomics model was superior to the traditional model (p = 0.042) in the training set. The combined model outperformed the traditional model (training p < 0.001, test p = 0.048).

Conclusion: The radiomics approach based on 7.0T HR-VWI can accurately identify culprit plaques associated with subcortical infarctions, potentially better than conventional HR-VWI features.

Background: Stress perfusion cardiovascular magnetic resonance (CMR) in the presence of atrial fibrillation (AF) has long been challenging due to electrocardiogram (ECG) mis-triggering. However, non-invasive ischemia imaging is important due to an increased risk of myocardial infarction in patients with AF, which has been attributed to underlying microvascular dysfunction. Myocardial blood flow (MBF) in patients with AF is poorly understood, and few studies have attempted to quantify this through non-invasive imaging.

Methods: Patients were recruited for stress perfusion CMR using a research sequence at 3-Tesla. Image acquisition occurred during both vasodilator-induced hyperemia and at rest. Stress and rest MBF maps were automatically generated. Analysis of perfusion maps included assessment of myocardial perfusion reserve (MPR) and endocardial-to-epicardial MBF ratios.

Results: Around 442 patients were analyzed; 63 of whom had a history of AF and were in AF during the scan. Both MBF during hyperemia (stress MBF) and MPR were reduced in patients with AF compared to those in sinus rhythm (median stress MBF 1.85 [1.52-2.24] vs. 2.35 [1.98-2.77] mL/min/g, p<0.001; median MPR 1.95 [1.62-2.19] vs. 2.37 [2.05-2.80], p<0.001). No significant difference was seen between the two groups at rest (p=0.451). When considering co-factors affecting MBF, multivariate linear regression analysis identified the presence of AF as a significant independent contributor to stress MBF and MPR values. Both endocardial and epicardial stress MBF and MPR were reduced in AF compared with sinus rhythm (both p<0.001) and endocardial/epicardial ratios were similar between the groups.

Conclusion: Automated quantitative MBF assessment can be performed in patients with AF. At hyperemia, MBF is reduced in AF compared to sinus rhythm.

Background: Whole-heart coronary magnetic resonance angiography (CMRA) enables noninvasive and accurate detection of coronary artery stenosis. Nevertheless, the visual interpretation of CMRA is constrained by the observer's experience, necessitating substantial training. The purposes of this study were to develop a deep learning (DL) algorithm using a deep convolutional neural network to accurately detect significant coronary artery stenosis in CMRA and to investigate the effectiveness of this DL algorithm as a tool for assisting in accurate detection of coronary artery stenosis.

Methods: Nine hundred and fifty-one coronary segments from 75 patients who underwent both CMRA and invasive coronary angiography (ICA) were studied. Significant stenosis was defined as a reduction in luminal diameter of >50% on quantitative ICA. A DL algorithm was proposed to classify CMRA segments into those with and without significant stenosis. A four-fold cross-validation method was used to train and test the DL algorithm. An observer study was then conducted using 40 segments with stenosis and 40 segments without stenosis. Three radiology experts and three radiology trainees independently rated the likelihood of the presence of stenosis in each coronary segment with a continuous scale from 0 to 1, first without the support of the DL algorithm, then using the DL algorithm.

Results: Significant stenosis was observed in 84 (8.8%) of the 951 coronary segments. Using the DL algorithm trained by the four-fold cross-validation method, the area under the receiver operating characteristic curve (AUC) for the detection of segments with significant coronary artery stenosis was 0.890, with 83.3% sensitivity, 83.6% specificity, and 83.6% accuracy. In the observer study, the average AUC of trainees was significantly improved using the DL algorithm (0.898) compared to that without the algorithm (0.821, p < 0.001). The average AUC of experts tended to be higher with the DL algorithm (0.897), but not significantly different from that without the algorithm (0.879, p = 0.082).

Conclusion: We developed a DL algorithm offering high diagnostic accuracy for detecting significant coronary artery stenosis on CMRA. Our proposed DL algorithm appears to be an effective tool for assisting inexperienced observers to accurately detect coronary artery stenosis in whole-heart CMRA.

Background: Women and men have been found to display differences in their cardiovascular anatomy and physiology, including differences in their cellular composition. While studies have shown cellular and molecular changes across sexes, few have performed sex-based studies of myocardial microstructure for healthy subjects. The purpose of this study was to quantify the myocardial microstructure in large healthy cohorts across sexes using in-vivo cardiac diffusion tensor imaging (cDTI) based on a second-order motion-compensated (M2) single-shot spin-echo sequence performed on a commercial ultra-high-performance gradient system.

Methods: In this single-center and cross-sectional study, free-breathing cDTI with a M2 spin-echo diffusion-weighted imaging scheme was evaluated in 103 healthy adult subjects (mean age 33.0 years, 52 women) scanned using an MR system with maximum gradient strength of 200mT/m. The diffusion tensor model was fit to obtain cDTI parameters, including mean diffusivity (MD), fractional anisotropy (FA), and helix angle transmurality (HAT).

Results: Women and men did not show significantly different distributions of cDTI parameters (MD, FA, and HAT). Healthy subjects scanned with cDTI protocols performed on an MR system with ultra-high performance gradients have an average of 1.51±0.08 µm²/ms for MD, 0.30±0.02 for FA, and -0.77±0.09°/% for HAT. Furthermore, women were reported to have an average MD 1.52±0.08 µm²/ms, FA 0.30±0.02, HAT -0.76±0.09°/%. Men presented an average of MD 1.50±0.08 µm²/ms, FA 0.30±0.02, and HAT -0.77±0.09°/% (p>0.05 for all cDTI parameters between sexes).

Conclusion: This is the first and largest single-center study to investigate cDTI in a large cohort (N>100) of healthy subjects performed with an ultra-high-performance gradient MR system. No significant difference was discovered in MD, FA, and HAT between men and women, suggesting biological sex does not impact myocardial microstructure in healthy subjects. Future work using ultra-high-performance systems should focus on the evaluation of microstructural changes in patients with cardiovascular disease.

Background: Quantitative myocardial mapping is critical for tissue characterization in non-ischemic cardiomyopathy (NICM). However, conventional techniques require separate breath-hold acquisitions, prolonging scan time and impairing co-registration. This study aimed to assess the feasibility and diagnostic performance of a novel free-breathing multimap (FBmultimap) sequence enabling simultaneous T1, T2, and T1ρ mapping in a single acquisition.

Methods: Onehundred-nine participants were prospectively enrolled, including 48 with hypertrophic cardiomyopathy (HCM), 28 with dilated cardiomyopathy (DCM), and 33 healthy controls. All underwent cardiac MRI with both FBmultimap and conventional mapping sequences (modified Look-Locker inversion recovery (MOLLI) T1, T2-prepared balanced steady-state free precession (bSSFP), and T1ρ-prepared bSSFP). Image quality was assessed using subjective (four-point Likert scale) and objective (edge sharpness) methods. Myocardial relaxation times were analyzed in the following two subgroups: (1) HCM and DCM vs. controls, and (2) late gadolinium enhancement (LGE)-positive and LGE-negative patients vs. controls. Combined diagnostic indices (T1 + T1ρ) were derived using logistic regression. Diagnostic performance was evaluated using receiver operating characteristic analysis across the following six models: FBmultimap (T1 + T1ρ), FBmultimap T1, FBmultimap T1ρ, conventional (T1 + T1ρ), MOLLI T1, and T1ρ-prepared bSSFP, with area under the curve (AUC) calculated.

Results: FBmultimap significantly reduced total scan time for T1 + T2 + T1ρ mapping to 66±6 s, compared with 195±10 s using conventional methods (p<0.001), while maintaining comparable image quality (all p>0.05). T1 and T1ρ values measured by FBmultimap were significantly elevated in HCM and DCM groups compared to controls, regardless of LGE status (all p<0.05), whereas T2 values showed no significant differences. FBmultimap (T1 + T1ρ) achieved higher AUCs for distinguishing LGE-positive (0.904) and LGE-negative (0.859) patients from controls than FBmultimap T1 (0.877 and 0.829), FBmultimap T1ρ (0.608 and 0.764), MOLLI T1 (0.770 and 0.671), T1ρ-prepared bSSFP (0.734 and 0.778), and the conventional (T1 + T1ρ) model (0.801 and 0.819).

Conclusion: FBmultimap enables rapid, co-registered, free-breathing mapping of myocardial T1, T2, and T1ρ with high reproducibility and improved diagnostic performance over conventional single-parameter methods. It holds promise as a clinically applicable tool for myocardial fibrosis detection, risk stratification, and longitudinal monitoring in patients with HCM and DCM.

Background: Myocardial blood flow (MBF) and myocardial perfusion reserve (MPR) can be quantified using vasodilator stress cardiovascular magnetic resonance (CMR). Exercise stress CMR (Ex-CMR) offers a more physiological assessment of cardiac functional reserve. While visual interpretation of Ex-CMR perfusion has been successfully applied, the feasibility of quantitative Ex-CMR perfusion remains unproven. We aimed to assess the feasibility of quantitative Ex-CMR perfusion imaging for characterizing exercise-induced perfusion responses and to perform a pilot study comparing MBF and MPR among patients with hypertrophic cardiomyopathy (HCM), heart failure with preserved ejection fraction (HFpEF), and non-cardiac dyspnea (NCD).

Methods: In this prospective study, patients with HCM, HFpEF, or NCD underwent Ex-CMR at 3T using a supine ergometer. Exercise was performed outside the scanner bore, followed by stress perfusion imaging 45-60 s post-exercise and rest perfusion 5-7 min later. A dual-sequence protocol with inline pixel-wise quantification was used to calculate MBF and MPR. Image quality and feasibility were visually assessed. Group comparisons were performed using analysis of variance and t-tests; linear regression was used to explore clinical associations.

Results: Of 108 patients enrolled, 9 were excluded due to obstructive coronary artery disease or reduced ejection fraction. Quantitative Ex-CMR was successful (at least one analyzable paired rest and post-exercise slice) in 90% (10/99) of cases. Most frequent quality issues were inadequate gating or arrhythmias and slice misalignment. The final cohort included 89 patients: 34 HCM, 34 HFpEF, and 21 NCD. Patients with HCM showed significantly lower MBF and MPR than HFpEF and NCD (MBF: 1.03 ± 0.27 vs 1.25 ± 0.40 and 1.13 ± 0.25 mL/min/g; MPR: 1.27 ± 0.21 vs 1.41 ± 0.29 and 1.44 ± 0.22; all p < 0.05). Peak exercise heart rate was the strongest independent predictor of MBF (β = 0.009, p < 0.001) and MPR (β = 0.004, p = 0.022).

Conclusion: Ex-CMR quantitative MBF and MPR assessment is feasible in most patients after image quality control. While the increase in MBF was limited during low-to-moderate exercise intensity in this pilot study, Ex-CMR revealed distinct perfusion response patterns among studied cohorts.