Journal of Magnetic Resonance Imaging最新文献

{"title":"Diffusion-Based Virtual MR Elastography: Association With Pancreatic Fibrosis and Identification of Postoperative Pancreatic Fistula After Pancreaticoduodenectomy.","authors":"Wei Cai, Yongjian Zhu, Dengfeng Li, Bingzhi Wang, Xiaohong Ma, Xinming Zhao","doi":"10.1002/jmri.70214","DOIUrl":"10.1002/jmri.70214","url":null,"abstract":"<p><strong>Background: </strong>Preoperative identification of clinically relevant postoperative pancreatic fistula (CR-POPF) after pancreaticoduodenectomy holds substantial clinical importance. The significance of virtual MR elastography (vMRE) in identifying CR-POPF risk remains unclear.</p><p><strong>Purpose: </strong>To explore the application of vMRE in assessing pancreatic fibrosis and preoperatively identifying CR-POPF with inter-device validation.</p><p><strong>Study type: </strong>Retrospective.</p><p><strong>Population: </strong>Three hundred twenty patients (212 male; mean age = 56.93 ± 9.34 years) receiving pancreatoduodenectomy were assessed and divided into two cohorts (cohort 1: 228; cohort 2: 92) based on two MRI scanners.</p><p><strong>Field strength/sequence: </strong>3 T, diffusion-weighted spin echo planar imaging and MRE using spin-echo echo planar imaging sequence.</p><p><strong>Assessment: </strong>Optimal combination of b values was identified for calculating shifted apparent diffusion coefficient (sADC). The diffusion-based tissue shear modulus (μ_diff) was constructed using optimal sADC. Association between μ_diff and pancreatic fibrosis grades was evaluated. CR-POPF was diagnosed postoperatively.</p><p><strong>Statistical tests: </strong>Spearman correlation analysis, linear regression analysis, univariate and multivariate logistic regression analysis, and receiver operating characteristic (ROC) analysis were used. Significance was set at p < 0.05.</p><p><strong>Results: </strong>Seventy-six patients developed (23.75%) CR-POPF. sADC generated from b values of 400 and 1500 s/mm² was identified for fitting μ_diff. μ_diff was significantly associated with histopathologic fibrosis grade (Spearman ρ = 0.763) and achieved performance for identifying CR-POPF with area under the curve (AUC) of 0.765. μ_diff, body mass index (BMI), and main pancreatic duct (MPD) were revealed as independent risk factors of CR-POPF. Their combination significantly improved the performance to AUC of 0.923 (AUC of μ_diff = 0.769; AUC of BMI = 0.673; AUC of MPD = 0.666). Results were confirmed in cohort 2.</p><p><strong>Data conclusion: </strong>Diffusion-based vMRE could effectively assess pancreatic fibrosis grade and provide a noninvasive biomarker for identifying CR-POPF preoperatively. The combination of μ_diff, BMI, and MPD demonstrated superior discriminative performance.</p><p><strong>Evidence level: </strong>3.</p><p><strong>Technical efficacy: </strong>Stage 2.</p>","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":"1448-1463"},"PeriodicalIF":3.5,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145810316","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Editorial for \"Analysis of Pulsatile Vessel Expansion in Healthy, COPD- and PH-Patients Using Dynamic Vessel Segmentation in Free-Breathing Lung MRI\".","authors":"Z J T Peggs, S T Francis","doi":"10.1002/jmri.70259","DOIUrl":"10.1002/jmri.70259","url":null,"abstract":"","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":"1433-1434"},"PeriodicalIF":3.5,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146118817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multidisciplinary Protocol for 1.5T MRI in Adult Patients With Active Implantable Medical Devices: Safety and Efficacy in a Five-Year Single-Center Experience.","authors":"Silvano Filice, Antonio Pavarani, Angelo Placci, Maurizio Falcioni, Maria Elena Manferdini","doi":"10.1002/jmri.70188","DOIUrl":"10.1002/jmri.70188","url":null,"abstract":"<p><strong>Background: </strong>The growing population of patients with active implantable medical devices (AIMDs) presents challenges for MRI access. Safety concerns regarding device malfunction and patient risk, demand standardized management strategies.</p><p><strong>Purpose: </strong>To evaluate the safety and feasibility of a structured multidisciplinary safety protocol for 1.5T MRI in AIMD carriers.</p><p><strong>Study type: </strong>Retrospective.</p><p><strong>Population: </strong>514 consecutive patients with AIMDs (cardiac implantable electronic devices (CIEDs) (n = 347), auditory implants (n = 85), neurostimulators (n = 61), and drug delivery systems (n = 21)) scheduled for clinical MRI (head (34.3%), spine (19.0%), prostate (10.5%), abdomen/pelvis (9.5%), knee (8.6%), heart (3.8%), and other sites (14.3%)) between 2019 and 2025.</p><p><strong>Field strength/sequence: </strong>MRI was performed on 1.5T scanners using predefined protocols, with automatic compliance to device-specific SAR limits through the Philips ScanWise Implant software.</p><p><strong>Assessment: </strong>The safety protocol included device verification, MRI-mode programming, compliance with manufacturer safety conditions, physiologic monitoring, and post-MRI device interrogation. MRI completion rates, exclusion frequencies, and adverse events were recorded, and exclusion proportions were compared across AIMD categories.</p><p><strong>Statistical tests: </strong>Descriptive statistics.</p><p><strong>Results: </strong>Of 514 patients, 50 (9.7%) with MR-unsafe, MR-nonconditional, or artifact-prone devices were not scanned: MR-unsafe (n = 23), MR-nonconditional (n = 22), or anticipated severe artifacts (n = 3). Four hundred and sixty-four (90.3%) patients successfully completed diagnostic MRI. Two brain examinations were terminated early due to implant-site pain. No device resets, malfunctions, or significant changes in sensing/pacing thresholds occurred. Internal magnet displacement was observed in three (3.5%) auditory implants, with two requiring surgical repositioning. Exclusion rates varied by device type, ranging from 11% to 14% for CIEDs to 29% for specific neurostimulators.</p><p><strong>Data conclusion: </strong>A structured, multidisciplinary protocol enables safe MRI in the majority of patients with MR-conditional AIMDs. Standardized pre-MRI screening and management support safe implementation, and improve MRI accessibility for this growing and complex patient population.</p><p><strong>Level of evidence: 4: </strong></p><p><strong>Technical efficacy stage: </strong>5.</p>","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":"1392-1401"},"PeriodicalIF":3.5,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145604491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Editorial for \"Large Language Models for Cardiac MRI Diagnosis Based on Standardized Text Descriptions\".","authors":"Mohammad Abd Alkhalik Basha, Mohamed M A Zaitoun","doi":"10.1002/jmri.70324","DOIUrl":"https://doi.org/10.1002/jmri.70324","url":null,"abstract":"","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-04-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773797","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Editorial for \"5 T-Based Glutamate Chemical Exchange Saturation Transfer Imaging for Adult Diffuse Glioma Stratification\".","authors":"Niels Bergsland","doi":"10.1002/jmri.70355","DOIUrl":"https://doi.org/10.1002/jmri.70355","url":null,"abstract":"","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-04-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimizing Cine Cardiac MRI: Technical Advances and Clinical Applications.","authors":"Sadegh Dehghani, Ali Rajabi, Amirhossein Rahmati, Negar Omidi","doi":"10.1002/jmri.70310","DOIUrl":"https://doi.org/10.1002/jmri.70310","url":null,"abstract":"<p><p>Cine cardiac magnetic resonance imaging (MRI) is the gold standard for cardiac function assessment, offering exceptional spatial and temporal resolution in a non-invasive manner. Recently, there have been attempts to improve signal-to-noise ratio (SNR) and temporal resolution by optimizing pulse sequence parameters such as repetition time (TR), echo time (TE), flip angle (FA), and bandwidth (BW). Although the usual technique is balanced steady-state free precession (bSSFP), its performance is limited due to off-resonance effects, specific absorption rate (SAR) constraints, and susceptibility artifacts, particularly at ≥ 3 T scanners. Due to the lack of breath-hold capacity or arrhythmias in some patients, faster acquisition and motion robustness can be addressed such as wideband bSSFP and real-time imaging through compressed sensing (CS) and artificial intelligence (AI). For patients with severe metal artifacts from cardiac implantable electronic devices (CIEDs), the spoiled gradient recalled-echo (SPGR) sequence is an effective alternative due to its lower sensitivity to susceptibility artifacts. Additionally, innovations in B₀/B₁ shimming, AI-driven reconstruction, and 7 T MRI promise exceptional image quality but require careful validation and safety management. This review summarizes advances in parameter optimization, artifact mitigation, and sequence selection across magnetic field strengths to enable faster, safer, and more personalized cine cardiac MRI.</p>","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-04-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773766","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mapping Neurodegenerative Changes in Clinically Uncertain Parkinsonian Syndrome Patients Using Fast MR Spin TomogrAphy in Time-Domain (MR-STAT) Relaxometry at 3T.","authors":"Martin B Schilder, Elon D Wallert, Stefano Mandija, Oscar van der Heide, Hongyan Liu, Miha Fuderer, Edwin Versteeg, Jan Booij, Rob M A de Bie, Martijn Beudel, Henk W Berendse, Tom van Mierlo, Jeroen Blankevoort, Cornelis A T van den Berg, Elsmarieke van de Giessen, Alessandro Sbrizzi","doi":"10.1002/jmri.70349","DOIUrl":"https://doi.org/10.1002/jmri.70349","url":null,"abstract":"&lt;p&gt;&lt;strong&gt;Background: &lt;/strong&gt;MR Spin TomogrAphy in Time-domain (MR-STAT) enables accelerated multiparametric relaxometry (T&lt;sub&gt;1&lt;/sub&gt;/T&lt;sub&gt;2&lt;/sub&gt;). Previous relaxometry studies predominantly compared Parkinson's disease patients with healthy controls (HC). The potential of relaxometry to distinguish neurodegenerative from non-neurodegenerative parkinsonism in clinically uncertain parkinsonian syndrome (CUPS) patients is unclear.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Purpose: &lt;/strong&gt;To investigate T&lt;sub&gt;1&lt;/sub&gt;-/T&lt;sub&gt;2&lt;/sub&gt;-differences between neurodegenerative and non-neurodegenerative parkinsonism in CUPS patients.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Study type: &lt;/strong&gt;Prospective cross-sectional study.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Population: &lt;/strong&gt;52 patients with neurodegenerative and 57 patients with non-neurodegenerative parkinsonism, diagnosed via dopamine transporter single photon emission computed tomography (DAT SPECT) and neurologist review, and 10 HC.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Fieldstrength/sequence: &lt;/strong&gt;MP-RAGE (magnetization-prepared rapid acquisition with gradient-echoes) and MR-STAT, a 2D Cartesian-encoded gradient-spoiled gradient-echo sequence with time-varying flip-angle preceded by a non-selective inversion pulse, at 3T.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Assessment: &lt;/strong&gt;Repeatability of T&lt;sub&gt;1&lt;/sub&gt;-/T&lt;sub&gt;2&lt;/sub&gt;-values was evaluated for cortical gray matter/cerebral white matter/thalamus/putamen/caudate nucleus/globus pallidus (GP) in HC. T&lt;sub&gt;1&lt;/sub&gt;-&lt;sub&gt;/&lt;/sub&gt;T&lt;sub&gt;2&lt;/sub&gt;-values of the parkinsonism groups were compared in the same regions per most/less affected hemisphere (MAH/LAH), determined by the putaminal uptake ratio on DAT SPECT.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Statistical tests: &lt;/strong&gt;Regional coefficients of variation (CoV) were computed to assess the repeatability of T&lt;sub&gt;1&lt;/sub&gt;-/T&lt;sub&gt;2&lt;/sub&gt;-values in HC. T-tests (α = 0.05) were used to compare T&lt;sub&gt;1&lt;/sub&gt;-/T&lt;sub&gt;2&lt;/sub&gt;-values between parkinsonism groups, and Cohen's D values were computed with bootstrapping to measure effect sizes with 95% confidence intervals (95% CI).&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Results: &lt;/strong&gt;CoVs ranged from 0.5% to 1.7% (T&lt;sub&gt;1&lt;/sub&gt;) to 1.5% to 2.7% (T&lt;sub&gt;2&lt;/sub&gt;). In the MAH, significant T&lt;sub&gt;1&lt;/sub&gt;-differences were found in the thalamus (Cohen's D = 0.635, 95% CI = [0.251, 1.016]); GP (Cohen's D = 0.508, 95% CI = [0.129, 0.887]); internal GP (Cohen's D = 0.603, 95% CI = [0.220, 0.983]); external GP (Cohen's D = 0.411, 95% CI = [0.033, 0.787]); and centromedial putamen (Cohen's D = 0.447, 95% CI = [0.069, 0.824]). In the LAH, significant T&lt;sub&gt;1&lt;/sub&gt;-differences were found in the thalamus (Cohen's D = 0.476, 95% CI = [0.097, 0.853]); GP (Cohen's D = 0.415, 95% CI = [0.037, 0.791]); anteromedial putamen (Cohen's D = 0.388, 95% CI = [0.011, 0.764]); and external GP (Cohen's D = 0.416, 95% CI = [0.038, 0.792]). T&lt;sub&gt;2&lt;/sub&gt;-differences were non-significant.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;Data conclusion: &lt;/strong&gt;MR-STAT showed high repeatability and showed potential to differentiate neurodegenerative from non-neurodegenerative par","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-04-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773780","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Editorial for \"Comparison of Diagnostic Performance Between Standard Susceptibility Map-Weighted Imaging and Susceptibility Map-Weighted Imaging Reconstructed From Clinical SWI and Neuromelanin MRI in Early Parkinson's Disease\".","authors":"Gábor Perlaki, Norbert Kovács","doi":"10.1002/jmri.70331","DOIUrl":"https://doi.org/10.1002/jmri.70331","url":null,"abstract":"","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimal Selection of the Effective Echo Time (TE_eff) for T2-Weighted Fast-Spin-Echo MRI of the Prostate at 3.0 T: Effect on Lesion Conspicuity.","authors":"Dahan Kim, Ronard Pabi, Eric A Borisch, Adam T Froemming, Roger C Grimm, Sara Hassanzadeh, Akira Kawashima, Ayumu Kido, Naoki Takahashi, John V Thomas, Stephen J Riederer","doi":"10.1002/jmri.70351","DOIUrl":"https://doi.org/10.1002/jmri.70351","url":null,"abstract":"<p><strong>Background: </strong>Clinical T₂-weighted (T2W) MRI of prostate cancer (PCa) usually implements a 2D fast-spin-echo (FSE) sequence, but understanding the echo-time-dependent contrast behavior is not straightforward due to the complicated FSE signal evolution. Consequently, consensus-based recommendations on the optimal effective echo time (TE_eff) to maximize contrast are lacking.</p><p><strong>Purpose: </strong>To investigate the effect of TE_eff on PCa lesion contrast in prostate T2W MRI.</p><p><strong>Study type: </strong>Phantom and in vivo prospective study.</p><p><strong>Subjects: </strong>A standard MRI phantom and 53 prostate cancer patients (29 for image evaluation and 24 for quantitative lesion analysis).</p><p><strong>Field strength and sequence: </strong>Multi-slice T2W FSE sequence at 3 T acquired at multiple TE_eff values.</p><p><strong>Assessment: </strong>Three readers (13, 25, and 27 years' experience) evaluated TE_eff = 100 ms versus 150 ms on diagnostic quality and reader preference. The effective T₂ relaxation rates (T_2,eff) were measured using FSE sequences in the phantom and used to predict the optimal TE_eff maximizing T2W contrasts. From multi-TE_eff scans of subjects with suspected PCa, T_2,eff values of PCa lesions and healthy tissue were measured and used to calculate the optimal TE_eff maximizing the lesion contrast. In vivo contrast metrics were measured as a function of TE_eff.</p><p><strong>Statistical tests: </strong>Wilcoxon test of reader scores, with p < 0.05 defining statistical significance.</p><p><strong>Results: </strong>No statistically significant difference was found in diagnostic quality between TE_eff = 150 ms and 100 ms images, but all three readers significantly favored TE_eff = 150 ms. Phantom experiments demonstrated that TE_eff-dependent FSE signal progression can be characterized by T_2,eff values typically 40% longer than actual T₂ values, allowing accurate prediction of optimal TE_eff within 2%-5%. Applied to PCa imaging, in vivo measurements indicated that effective T₂ relaxation times of the prostate are on average 1.5× longer than published T₂ values.</p><p><strong>Data conclusion: </strong>The benefits of TE_eff = 150 ms versus 100 ms in prostate T2W FSE were supported by reader preference, phantom experiments, in vivo measurements, and analytic optimizations.</p><p><strong>Evidence level: </strong>2.</p><p><strong>Technical efficacy: </strong>2.</p>","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Large Language Models for Cardiac MRI Diagnosis Based on Standardized Text Descriptions.","authors":"Hongbo Zhang, Junjie Zhou, Chen Zhang, Guanyu Lu, Zhihui Lu, Lanling Wang, Lili Wang, Huiwen Gong, Lei Zhao, Xiaohai Ma","doi":"10.1002/jmri.70327","DOIUrl":"https://doi.org/10.1002/jmri.70327","url":null,"abstract":"<p><strong>Background: </strong>MRI is important for cardiac disease evaluation, but accurate diagnosis remains challenging in less experienced centers. Although large language models (LLMs) have shown promise in medical imaging diagnosis, their application in cardiac MRI is limited.</p><p><strong>Hypothesis: </strong>LLMs may be effective in achieving cardiac MRI diagnosis based on standardized descriptions.</p><p><strong>Study type: </strong>Retrospective.</p><p><strong>Population: </strong>A total of 203 hypertrophic cardiomyopathy, 186 dilated cardiomyopathy, 46 hypertensive heart disease, 198 ischemic cardiomyopathy, 38 constrictive pericarditis, 45 cardiac amyloidosis, 91 myocarditis, and 144 normal controls.</p><p><strong>Field strength/sequences: </strong>Balanced steady-state free-precession, short tau inversion recovery, and breath-hold inversion-recovery segmented gradient-echo sequences at 3.0 T.</p><p><strong>Assessment: </strong>Clinical and cardiac MRI information from each subject was converted into standardized descriptions and input into Generative Pre-trained Transformer-4.5 (GPT-4.5), GPT-4 Omni (GPT-4o), Deepseek-V3, and Deepseek-R1 LLMs. Cardiac MRI information included LV function, wall thickness and motion, and abnormalities on T2WI, perfusion and late gadolinium enhancement sequences. Each model was asked to generate an imaging diagnosis. In addition, a medical student (8 months experience) and three radiologists (junior, mid-level and senior: with 3, 6, and 10 years' experience, respectively) provided diagnoses based on cardiac MRI images and clinical information.</p><p><strong>Statistic tests: </strong>Frequency-weighted sensitivity and specificity were calculated. The diagnostic performances of the LLMs and human readers were compared using the McNemar test with Bonferroni correction. A p value < 0.05 was considered significant.</p><p><strong>Results: </strong>All LLMs showed excellent frequency-weighted specificity (0.973-0.983). The frequency-weighted sensitivities of all LLMs were not significantly different from that of the junior radiologist, were significantly higher than that of the medical student, and significantly inferior to those of the senior radiologist (GPT-4.5: 0.863, GPT-4o: 0.821, Deepseek-V3: 0.843, and Deepseek-R1: 0.851 vs. junior radiologist: 0.850, all adjusted p = 1.000; vs. medical student: 0.731, all adjusted p < 0.001; vs. senior radiologist: 0.942, all adjusted p < 0.001). Additionally, the mid-level radiologist achieved a frequency-weighted sensitivity of 0.895, outperforming all LLMs except GPT-4.5.</p><p><strong>Data conclusion: </strong>LLMs may generate accurate diagnoses from standardized cardiac MRI descriptions, potentially benefiting less experienced physicians.</p><p><strong>Level of evidence: 4: </strong></p><p><strong>Technical efficacy: </strong>Stage 5.</p>","PeriodicalId":16140,"journal":{"name":"Journal of Magnetic Resonance Imaging","volume":" ","pages":""},"PeriodicalIF":3.5,"publicationDate":"2026-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147773761","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}