Journal of Cardiovascular Magnetic Resonance最新文献

{"title":"Wide variation in shape of hypoplastic left ventricles undergoing recruitment and biventricular repair: A statistical shape modeling study.","authors":"Isabel R Barnet, Noah E Schulz, Sunil J Ghelani, David M Hoganson, Eric N Feins, Peter E Hammer, Sitaram M Emani, Lynn A Sleeper, Rebecca S Beroukhim","doi":"10.1016/j.jocmr.2024.101131","DOIUrl":"10.1016/j.jocmr.2024.101131","url":null,"abstract":"<p><strong>Background: </strong>Patients with hypoplastic left ventricles (LV) who undergo volume-loading procedures (recruitment, biventricular [BIV] repair) are at risk for adverse outcomes, including heart failure and death. We investigated pre-BIV LV shape as a predictor of outcome after BIV repair in patients with hypoplastic LVs.</p><p><strong>Methods: </strong>Baseline and post-recruitment cardiac magnetic resonance imaging and computed tomography data were analyzed in patients with hypoplastic LV (<50 mL/m²). Statistical shape modeling (SSM) was utilized to generate a model of the shape and variability of LVs. Traditional measures of LV sphericity and eccentricity were also measured. Major adverse cardiovascular events (MACE) included heart failure, transplant, and death.</p><p><strong>Results: </strong>Of 95 patients with baseline mean LV volume 29 ± 13 mL/m², 45/95 (47%) had a right dominant atrioventricular canal defect, 31/95 (33%) had a variant of hypoplastic left heart syndrome, and 18/95 (19%) had endocardial fibroelastosis (EFE). A wide variation in LV shape was found by SSM, and shape modes were associated with right ventricle (RV) and LV size, and diagnosis. BIV repair was achieved in 74/95 (78%) patients; 13/74 (18%) of BIV patients had MACE. Predictors of MACE following BIV repair included EFE, higher RV mass index, and higher RV end-diastolic volume index. No baseline or post-recruitment LV shape parameter was associated with the outcome after BIV repair.</p><p><strong>Conclusion: </strong>The shape model of hypoplastic LVs demonstrated a wide array of LV shapes. LVs gained sphericity and size and lost eccentricity with recruitment. Though the ventricles changed shape with recruitment, no specific LV shape characteristic at the baseline or post-recruitment stage was predictive of decision to proceed with BIV repair or outcome. Higher RV mass and volume may represent new biomarkers that predict outcomes following BIV repair in patients with hypoplastic LV. Further investigation could determine the reproducibility of these findings.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101131"},"PeriodicalIF":4.2,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142794766","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sex differences and determinants of coronary microvascular function in asymptomatic adults with type 2 diabetes.","authors":"Jian L Yeo, Abhishek Dattani, Joanna M Bilak, Alice L Wood, Lavanya Athithan, Aparna Deshpande, Anvesha Singh, J Ranjit Arnold, Emer M Brady, David Adlam, John D Biglands, Peter Kellman, Hui Xue, Thomas Yates, Melanie J Davies, Gaurav S Gulsin, Gerry P McCann","doi":"10.1016/j.jocmr.2024.101132","DOIUrl":"10.1016/j.jocmr.2024.101132","url":null,"abstract":"<p><strong>Background: </strong>Coronary microvascular dysfunction (CMD) is a significant complication in type 2 diabetes (T2D) and may be more common in women. We aimed to evaluate the sex differences and sex-specific clinical determinants of CMD in adults with T2D without prevalent cardiovascular disease.</p><p><strong>Methods: </strong>Single center pooled analysis of four prospective studies comparing asymptomatic people with T2D and controls. All subjects underwent comprehensive cardiovascular phenotyping with myocardial perfusion reserve (MPR) quantified with perfusion cardiovascular magnetic resonance (CMR). Participants with silent coronary disease were excluded. Multivariable linear regression was performed to identify determinants of MPR with an interaction term for sex.</p><p><strong>Results: </strong>Four hundred and seventy-nine T2D (age 57 ± 11 years, 42% [202/479] women) were compared with 116 controls (age 53 ± 11 years, 41% [48/116] women). Men with T2D, but not women, demonstrated worse systolic function and higher extracellular volume fraction than controls. MPR was significantly lower in T2D than controls (women, 2.6 ± 0.9 vs 3.3 ± 1.0, p < 0.001; men, 3.1 ± 0.9 vs 3.5 ± 1.0, p = 0.004), and lower in women than men with T2D (p < 0.001). More women than men with T2D had MPR <2.5 (46% [79/202] vs 26% [64/277], p < 0.001). There was a significant interaction between sex and body mass index (BMI) for MPR (p interaction <0.001). Following adjustment for clinical risk factors, inverse association with MPR were BMI in women (β = -0.17, p = 0.045) and systolic blood pressure in men (β = -0.14, p = 0.049).</p><p><strong>Conclusion: </strong>Among asymptomatic adults with T2D, women had a greater prevalence of CMD than men. Risk factors modestly but significantly associated with CMD in asymptomatic people with T2D were BMI among women and systolic blood pressure among men.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101132"},"PeriodicalIF":4.2,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11761338/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142794754","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mean pulmonary artery pressure prediction with explainable multi-view cardiovascular magnetic resonance cine series deep learning model.","authors":"Li-Hsin Cheng, Xiaowu Sun, Charlie Elliot, Robin Condliffe, David G Kiely, Samer Alabed, Andrew J Swift, Rob J van der Geest","doi":"10.1016/j.jocmr.2024.101133","DOIUrl":"10.1016/j.jocmr.2024.101133","url":null,"abstract":"<p><strong>Background: </strong>Pulmonary hypertension (PH) is a heterogeneous condition and regardless of etiology impacts negatively on survival. Diagnosis of PH is based on hemodynamic parameters measured invasively at right heart catheterization (RHC); however, a non-invasive alternative would be clinically valuable. Our aim was to estimate RHC parameters non-invasively from cardiac magnetic resonance (MR) data using deep learning models and to identify key contributing imaging features.</p><p><strong>Methods: </strong>We constructed an explainable convolutional neural network (CNN) taking cardiac MR cine series from four different views as input to predict mean pulmonary artery pressure (mPAP). The model was trained and evaluated on 1646 examinations. The model's attention weight and predictive performance associated with each frame, view, or phase were used to judge its importance. Additionally, the importance of each cardiac chamber was inferred by perturbing part of the input pixels.</p><p><strong>Results: </strong>The model achieved a Pearson correlation coefficient of 0.80 and R² of 0.64 in predicting mPAP and identified the right ventricle region on short-axis view to be especially informative.</p><p><strong>Conclusion: </strong>Hemodynamic parameters can be estimated non-invasively with a CNN, using MR cine series from four views, revealing key contributing features at the same time.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101133"},"PeriodicalIF":4.2,"publicationDate":"2024-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142791832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Feasibility of fetal cardiac function and anatomy assessment by real-time spiral balanced steady-state free precession magnetic resonance imaging at 0.55T.","authors":"Ye Tian, Jon Detterich, Jay D Pruetz, Ecrin Yagiz, John C Wood, Krishna S Nayak","doi":"10.1016/j.jocmr.2024.101130","DOIUrl":"10.1016/j.jocmr.2024.101130","url":null,"abstract":"<p><strong>Background: </strong>Contemporary 0.55T magnetic resonance imaging (MRI) is promising for fetal MRI, due to the larger bore, reduced safety concerns, lower acoustic noise, and improved fast imaging capability. In this work, we explore improved fetal cardiovascular magnetic resonance (CMR) without relying on any synchronizing devices, prospective, or retrospective gating, to determine the feasibility of real-time MRI evaluation of fetal cardiac function as well as cardiac and great vessel anatomies by using spiral balanced steady-state free precession (bSSFP) at 0.55T.</p><p><strong>Methods: </strong>A real-time spiral bSSFP pulse sequence for fetal CMR was implemented and optimized on a 0.55T whole-body MRI. Fetal CMR was prospectively performed between May 2022 and August 2023. The protocol included (1) real-time images at standard cardiac views, for 10-20 s/view and 40-43.6 ms/frame and (2) 4-9 stacks of slices at standard cardiac views that each cover the whole heart, with 15-30 slices/stack, and 2-5 s/slice, at 320-349 ms/frame. Images were evaluated by a fetal cardiologist. Quantitative measurements of cardiothoracic area ratio and cardiac axis were compared with previous reports. Diagnostic accuracy was compared against postnatal echocardiographic findings.</p><p><strong>Results: </strong>Twenty-nine participants were enrolled for 32 CMR exams, with mean maternal age 33.6 ± 5.8 years (range 22-44 years) and mean gestational age 32.8 ± 3.9 weeks (range 23-38 weeks). The proposed sequence enabled evaluation of the fetal heart in <30 min in all cases (average 22 min). Real-time MRI allowed easy adjustment of scan plan, automatic whole-heart volumetric sweeping, and flexible choice of reconstruction temporal resolution. For key cardiac anatomic features, 60% (315/527) were delineated well. Mean cardiothoracic area ratio and cardiac axis were 0.27 ± 0.04 and 45.8 ± 7.8 degrees. Diagnostic agreement with postnatal echocardiographic findings was 84% (26/31).</p><p><strong>Conclusion: </strong>A spiral real-time bSSFP pulse sequence at 0.55T can provide both low-framerate and high-framerate fetal heart images without relying on maternal breath-hold, specialized gating devices, or cardiac gating. The low-framerate images offer high diagnostic quality structural evaluations of the fetal heart, while the high-framerate images capture fetal heart motion and may enable functional assessments.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101130"},"PeriodicalIF":4.2,"publicationDate":"2024-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142785510","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improving the efficiency and accuracy of cardiovascular magnetic resonance with artificial intelligence-review of evidence and proposition of a roadmap to clinical translation.","authors":"Qiang Zhang, Anastasia Fotaki, Sona Ghadimi, Yu Wang, Mariya Doneva, Jens Wetzl, Jana G Delfino, Declan P O'Regan, Claudia Prieto, Frederick H Epstein","doi":"10.1016/j.jocmr.2024.101051","DOIUrl":"10.1016/j.jocmr.2024.101051","url":null,"abstract":"<p><strong>Background: </strong>Cardiovascular magnetic resonance (CMR) is an important imaging modality for the assessment of heart disease; however, limitations of CMR include long exam times and high complexity compared to other cardiac imaging modalities. Recently advancements in artificial intelligence (AI) technology have shown great potential to address many CMR limitations. While the developments are remarkable, translation of AI-based methods into real-world CMR clinical practice remains at a nascent stage and much work lies ahead to realize the full potential of AI for CMR.</p><p><strong>Methods: </strong>Herein we review recent cutting-edge and representative examples demonstrating how AI can advance CMR in areas such as exam planning, accelerated image reconstruction, post-processing, quality control, classification and diagnosis.</p><p><strong>Results: </strong>These advances can be applied to speed up and simplify essentially every application including cine, strain, late gadolinium enhancement, parametric mapping, 3D whole heart, flow, perfusion and others. AI is a unique technology based on training models using data. Beyond reviewing the literature, this paper discusses important AI-specific issues in the context of CMR, including (1) properties and characteristics of datasets for training and validation, (2) previously published guidelines for reporting CMR AI research, (3) considerations around clinical deployment, (4) responsibilities of clinicians and the need for multi-disciplinary teams in the development and deployment of AI in CMR, (5) industry considerations, and (6) regulatory perspectives.</p><p><strong>Conclusions: </strong>Understanding and consideration of all these factors will contribute to the effective and ethical deployment of AI to improve clinical CMR.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101051"},"PeriodicalIF":4.2,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11331970/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141442759","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Double aortic arch: a comparison of fetal cardiovascular magnetic resonance, postnatal computed tomography and surgical findings.","authors":"Milou P M van Poppel, David F A Lloyd, Johannes K Steinweg, Sujeev Mathur, James Wong, Vita Zidere, Simone Speggiorin, Haran Jogeesvaran, Reza Razavi, John M Simpson, Kuberan Pushparajah, Trisha V Vigneswaran","doi":"10.1016/j.jocmr.2024.101053","DOIUrl":"10.1016/j.jocmr.2024.101053","url":null,"abstract":"<p><strong>Background: </strong>In double aortic arch (DAA), one of the arches can demonstrate atretic portions postnatally, leading to diagnostic uncertainty due to overlap with isolated right aortic arch (RAA) variants. The main objective of this study is to demonstrate the morphological evolution of different DAA phenotypes from prenatal to postnatal life using three-dimensional (3D) fetal cardiac magnetic resonance (CMR) imaging and postnatal computed tomography (CT)/CMR imaging.</p><p><strong>Methods: </strong>Three-dimensional fetal CMR was undertaken in fetuses with suspected DAA over a 6-year period (January 2016-January 2022). All cases with surgical confirmation of DAA were retrospectively studied and morphology on fetal CMR was compared to postnatal CT/CMR and surgical findings.</p><p><strong>Results: </strong>Thirty-four fetuses with surgically confirmed DAA underwent fetal CMR. The RAA was dominant in 32/34 (94%). Postnatal CT/CMR was undertaken at a median age of 3.3 months (interquartile range 2.0-3.9) demonstrating DAA with patency of both arches in 10/34 (29%), with 7 showing signs of coarctation of the left aortic arch (LAA). The LAA isthmus was not present on CT/CMR in 22/34 (65%), and the transverse arch between left carotid and left subclavian artery was not present in 2 cases.</p><p><strong>Conclusion: </strong>Fetal CMR provides novel insights into perinatal evolution of DAA. The smaller LAA can develop coarctation or atresia related to postnatal constriction of the arterial duct, making diagnosis of DAA challenging with contrast-enhanced CT/CMR. This highlights the potentially important role for prenatal 3D vascular imaging and might improve the interpretation of postnatal imaging.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101053"},"PeriodicalIF":4.2,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11417329/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141498147","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Occult coronary microvascular dysfunction and ischemic heart disease in patients with diabetes and heart failure.","authors":"Noor Sharrack, Louise A E Brown, Jonathan Farley, Ali Wahab, Nicholas Jex, Sharmaine Thirunavukarasu, Amrit Chowdhary, Miroslawa Gorecka, Wasim Javed, Hui Xue, Eylem Levelt, Erica Dall'Armellina, Peter Kellman, Pankaj Garg, John P Greenwood, Sven Plein, Peter P Swoboda","doi":"10.1016/j.jocmr.2024.101073","DOIUrl":"10.1016/j.jocmr.2024.101073","url":null,"abstract":"<p><strong>Background: </strong>Patients with diabetes mellitus (DM) and heart failure (HF) have worse outcomes than normoglycemic HF patients. Cardiovascular magnetic resonance (CMR) can identify ischemic heart disease (IHD) and quantify coronary microvascular dysfunction (CMD) using myocardial perfusion reserve (MPR). We aimed to quantify the extent of silent IHD and CMD in patients with DM presenting with HF.</p><p><strong>Methods: </strong>Prospectively recruited outpatients undergoing assessment into the etiology of HF underwent in-line quantitative perfusion CMR for calculation of stress and rest myocardial blood flow (MBF) and MPR. Exclusions included angina or history of IHD. Patients were followed up (median 3.0 years) for major adverse cardiovascular events (MACE).</p><p><strong>Results: </strong>Final analysis included 343 patients (176 normoglycemic, 84 with pre-diabetes, and 83 with DM). Prevalence of silent IHD was highest in DM 31% ( 26/83), then pre-diabetes 20% (17/84) then normoglycemia 17%, ( 30/176). Stress MBF was lowest in DM (1.53 ± 0.52), then pre-diabetes (1.59 ± 0.54) then normoglycemia (1.83 ± 0.62). MPR was lowest in DM (2.37 ± 0.85) then pre-diabetes (2.41 ± 0.88) then normoglycemia (2.61 ± 0.90). During follow-up, 45 patients experienced at least one MACE. On univariate Cox regression analysis, MPR and presence of silent IHD were both associated with MACE. However, after correction for HbA1c, age, and left ventricular ejection fraction, the associations were no longer significant.</p><p><strong>Conclusion: </strong>Patients with DM and HF had higher prevalence of silent IHD, more evidence of CMD, and worse cardiovascular outcomes than their non-diabetic counterparts. These findings highlight the potential value of CMR for the assessment of silent IHD and CMD in patients with DM presenting with HF.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101073"},"PeriodicalIF":4.2,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11417243/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141889370","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Comparison of dual-bolus versus dual-sequence techniques for determining myocardial blood flow and myocardial perfusion reserve by cardiac magnetic resonance stress perfusion: From the Automated Quantitative analysis of myocardial perfusion cardiac Magnetic Resonance Consortium.","authors":"Emily Yin Sing Chong, Haonan Wang, Kwan Ho Gordon Leung, Paul Kim, Yuko Tada, Tsun Hei Sin, Chun Ka Wong, Kwong Yue Eric Chan, Chor Cheung Frankie Tam, Mitchel Benovoy, Andrew E Arai, Victor Goh, Martin A Janich, Amit R Patel, Ming-Yen Ng","doi":"10.1016/j.jocmr.2024.101085","DOIUrl":"10.1016/j.jocmr.2024.101085","url":null,"abstract":"<p><strong>Background: </strong>Quantitative stress cardiac magnetic resonance (CMR) can be performed using the dual-sequence (DS) technique or dual-bolus (DB) method. It is unknown if DS and DB produce similar results for myocardial blood flow (MBF) and myocardial perfusion reserve (MPR). The study objective is to investigate if there are any differences between DB- and DS-derived MBF and MPR.</p><p><strong>Methods: </strong>Retrospective observational study with 168 patients who underwent stress CMR. DB and DS methods were simultaneously performed on each patient on the same day. Global and segmental stress MBF and rest MBF values were collected.</p><p><strong>Results: </strong>Using Bland-Altman analysis, segmental and global stress MBF values were higher in DB than DS (0.22 ± 0.60 mL/g/min, p < 0.001 and 0.20 ± 0.48 mL/g/min, p = 0.005, respectively) with strong correlation (r = 0.81, p < 0.001 for segmental and r = 0.82, p < 0.001 for global). In rest MBF, segmental and global DB values were higher than by DS (0.15 ± 0.51 mL/g/min, p < 0.001 and 0.14 ± 0.36 mL/g/min, p = 0.011, respectively) with strong correlation (r = 0.81, p < 0.001 and r = 0.77, p < 0.001). Mean difference between MPR by DB and DS was -0.02 ± 0.68 mL/g/min (p = 0.758) for segmental values and -0.01 ± 0.49 mL/g/min (p = 0.773) for global values. MPR values correlated strongly as well in both segmental and global, both (r = 0.74, p < 0.001) and (r = 0.75, p < 0.001), respectively.</p><p><strong>Conclusion: </strong>There is a very good correlation between DB- and DS-derived MBF and MPR values. However, there are significant differences between DB- and DS-derived global stress and rest MBF. While MPR values did not show statistically significant differences between DB and DS methods.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101085"},"PeriodicalIF":4.2,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11422560/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142000050","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improved robustness for deep learning-based segmentation of multi-center myocardial perfusion cardiovascular MRI datasets using data-adaptive uncertainty-guided space-time analysis.","authors":"Dilek M Yalcinkaya, Khalid Youssef, Bobak Heydari, Janet Wei, C Noel Bairey Merz, Robert Judd, Rohan Dharmakumar, Orlando P Simonetti, Jonathan W Weinsaft, Subha V Raman, Behzad Sharif","doi":"10.1016/j.jocmr.2024.101082","DOIUrl":"10.1016/j.jocmr.2024.101082","url":null,"abstract":"<p><strong>Background: </strong>Fully automatic analysis of myocardial perfusion cardiovascular magnetic resonance imaging datasets enables rapid and objective reporting of stress/rest studies in patients with suspected ischemic heart disease. Developing deep learning techniques that can analyze multi-center datasets despite limited training data and variations in software (pulse sequence) and hardware (scanner vendor) is an ongoing challenge.</p><p><strong>Methods: </strong>Datasets from three medical centers acquired at 3T (n = 150 subjects; 21,150 first-pass images) were included: an internal dataset (inD; n = 95) and two external datasets (exDs; n = 55) used for evaluating the robustness of the trained deep neural network (DNN) models against differences in pulse sequence (exD-1) and scanner vendor (exD-2). A subset of inD (n = 85) was used for training/validation of a pool of DNNs for segmentation, all using the same spatiotemporal U-Net architecture and hyperparameters but with different parameter initializations. We employed a space-time sliding-patch analysis approach that automatically yields a pixel-wise \"uncertainty map\" as a byproduct of the segmentation process. In our approach, dubbed data-adaptive uncertainty-guided space-time (DAUGS) analysis, a given test case is segmented by all members of the DNN pool and the resulting uncertainty maps are leveraged to automatically select the \"best\" one among the pool of solutions. For comparison, we also trained a DNN using the established approach with the same settings (hyperparameters, data augmentation, etc.).</p><p><strong>Results: </strong>The proposed DAUGS analysis approach performed similarly to the established approach on the inD (Dice score for the testing subset of inD: 0.896 ± 0.050 vs 0.890 ± 0.049; p = n.s.) whereas it significantly outperformed on the exDs (Dice for exD-1: 0.885 ± 0.040 vs 0.849 ± 0.065, p < 0.005; Dice for exD-2: 0.811 ± 0.070 vs 0.728 ± 0.149, p < 0.005). Moreover, the number of image series with \"failed\" segmentation (defined as having myocardial contours that include bloodpool or are noncontiguous in ≥1 segment) was significantly lower for the proposed vs the established approach (4.3% vs 17.1%, p < 0.0005).</p><p><strong>Conclusion: </strong>The proposed DAUGS analysis approach has the potential to improve the robustness of deep learning methods for segmentation of multi-center stress perfusion datasets with variations in the choice of pulse sequence, site location, or scanner vendor.</p>","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101082"},"PeriodicalIF":4.2,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11663771/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"141982310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Diffusion tensor magnetic resonance imaging of the heart: Now feasible on your neighborhood scanner.","authors":"David E Sosnovik, Daniel B Ennis","doi":"10.1016/j.jocmr.2024.101101","DOIUrl":"10.1016/j.jocmr.2024.101101","url":null,"abstract":"","PeriodicalId":15221,"journal":{"name":"Journal of Cardiovascular Magnetic Resonance","volume":" ","pages":"101101"},"PeriodicalIF":4.2,"publicationDate":"2024-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11647480/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142347431","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}