Biederer J, Bergmann LL, Ackman JB
… +8 more, Hochhegger B, Azour L, Triphan SMF, Dinkel J, Ohno Y, Ozawa Y, van Beek EJR, Wucherpfennig L
J Magn Reson Imaging
· 2026 Jun · PMID 41928658
·
Full text
Management of lung cancer (LC) encompasses screening, diagnosis, staging, radiotherapy planning and guidance, therapy monitoring and surveillance. Across these domains, magnetic resonance imaging (MRI) offers a range of...Management of lung cancer (LC) encompasses screening, diagnosis, staging, radiotherapy planning and guidance, therapy monitoring and surveillance. Across these domains, magnetic resonance imaging (MRI) offers a range of morphological and functional imaging capabilities-including diffusion-weighted imaging (DWI), dynamic contrast-enhanced (DCE) imaging, and whole-body MRI-to complement established imaging modalities. Recent technical advances have substantially improved the feasibility of lung MRI, enabling more reliable image acquisition and lesion assessment under controlled conditions. In LC screening, meta-analyses and prospective studies indicate that MRI can detect solid pulmonary nodules above clinically actionable size thresholds with moderate to high sensitivity and a low false-positive rate. However, the available evidence is largely derived from pilot studies, selected cohorts, and modeling-based analyses. MRI should therefore be regarded as technically feasible for screening but not yet a validated alternative to low-dose computed tomography in population-based programs. For staging, whole-body MRI incorporating DWI has demonstrated comparable diagnostic performance to standard multimodality pathways in prospective and randomized studies, with potential advantages including reduced radiation exposure and streamlined imaging workflows. In radiotherapy planning, DCE, DWI, and motion-resolved MRI techniques can improve target delineation and treatment adaptation, but their use remains largely confined to specialized centers. MRI shows promise for therapy response assessment and prognostication through quantitative DCE- and DWI-derived biomarkers, although reported parameters remain heterogeneous and insufficiently standardized for routine clinical decision-making. Overall, MRI has established clinical utility in selected aspects of LC management, while broader adoption is currently limited by availability, standardization, and validation gaps. Further technical refinement and large-scale prospective trials are required to define its role in routine clinical practice. LEVEL OF EVIDENCE: 5. TECHNICAL EFFICACY: Stage 2.
Cai S, Simonsson C, Karlsson M
… +9 more, Balkhed W, Tellman J, Ignatova S, Nasr P, Ekstedt M, Kechagias S, Bartholomä WC, Dahlström N, Lundberg P
J Magn Reson Imaging
· 2026 Jul · PMID 41924972
·
Full text
BACKGROUND: Three-dimensional (3D) MR elastography (MRE) derives viscoelastic parameters that may reflect inflammation, but their frequency dependence and the influence of steatosis on inflammation grading and fibrosis s...BACKGROUND: Three-dimensional (3D) MR elastography (MRE) derives viscoelastic parameters that may reflect inflammation, but their frequency dependence and the influence of steatosis on inflammation grading and fibrosis staging remain unclear. PURPOSE: To investigate 3D multifrequency MRE for assessing hepatic inflammation, fibrosis stage across frequencies, and the influence of steatosis. STUDY TYPE: Prospective. POPULATION: Sixty-four (40 men, median age: 58 years) participants with chronic liver disease (CLD); 21 (8 men, median age: 28 years) healthy volunteers. FIELD STRENGTH/SEQUENCE: 3-T; gradient-echo sequence with mechanical vibrations at low (16.7 and 18 Hz), medium (33.4 and 36 Hz), and high (50.1 and 54 Hz) frequencies. ASSESSMENT: In CLD participants, MRE-derived viscoelastic parameters, shear stiffness, storage modulus, loss modulus, and damping ratio were compared with histologically assessed fibrosis, inflammation, and steatosis. MRE test-retest repeatability over 10 min was evaluated in healthy volunteers. STATISTICAL TESTS: Wilcoxon rank sum test, Spearman's correlation, multivariable regression analysis, and area under the receiver operating curve (AUROC). A p value of < 0.05 was considered statistically significant. RESULTS: Inflammation was significantly independently associated with damping ratio at medium frequency, which showed moderate performance for grading inflammation (AUROC = 0.76-0.83, sensitivity = 0.83-0.84, specificity = 0.70-0.79). Fibrosis staging using shear stiffness and moduli showed high diagnostic performance (AUROC = 0.82-0.95), with comparable accuracy between medium and high frequencies (p = 0.327-0.896). Steatosis was not significantly correlated with MRE overall (p = 0.212-0.459), but was significantly associated with 19% higher stiffness and 20% higher loss modulus at medium frequency in CLD participants without fibrosis or inflammation. DATA CONCLUSION: Medium frequency 3D MRE demonstrated an independent association with inflammation while preserving accurate fibrosis assessment. Steatosis seemed not to confound MRE-based evaluation. TECHNICAL EFFICACY: Stage 2.
Kaandorp MPT, Jakab A, Federau C
… +1 more, While PT
J Magn Reson Imaging
· 2026 Jul · PMID 41913432
·
Full text
BACKGROUND: Intravoxel incoherent motion (IVIM) analysis of diffusion-weighted MRI (DWI) provides microvascular perfusion and diffusion information. However, parameter estimation is limited by noise sensitivity, variabil...BACKGROUND: Intravoxel incoherent motion (IVIM) analysis of diffusion-weighted MRI (DWI) provides microvascular perfusion and diffusion information. However, parameter estimation is limited by noise sensitivity, variability across fitting methods, and lack of standardization. Deep-learning (DL)-based approaches, particularly spatially-aware transformers, may improve robustness, but their clinical utility remains unexplored. PURPOSE: To evaluate conventional, Bayesian, and DL-based IVIM fitting methods in glioma patients, focusing on tumor grading accuracy. STUDY TYPE: Retrospective. POPULATION: Fractal-noise-based simulations and preoperative DWI from 20 glioma patients (5 Grade-2, 3 Grade-3, 12 Grade-4). SEQUENCE: Spin-echo echo-planar DWI, 16 b values (0-900 s/mm), three orthogonal directions, 3 T. ASSESSMENT: IVIM parameter maps were compared across least squares (LSQ), segmented (SEG), Bayesian shrinkage (BSP), and spatial-homogeneity (FBM) priors, and DL-based methods, including IVIM-NET, novel spatially-aware transformers (NATTEN-17), and a refined version (SA-17). Simulation accuracy was evaluated using median absolute percentage error (MDAPE) and bias using median percentage error. In vivo data were visually assessed by the authors for noise suppression and structural preservation. Whole-tumor diffusion coefficient (D), pseudo-diffusion coefficient (D*), and signal fraction (f) values were evaluated across tumor grades and for differentiating Grade-4 from Grade-2/3 tumors. STATISTICAL TESTS: Mann-Whitney U tests for group comparisons; tumor grading performance using receiver operating characteristic-area under the curve (ROC-AUC), and pairwise AUC differences using the DeLong test. SIGNIFICANCE: p < 0.05. RESULTS: Transformer-based methods achieved superior simulation accuracy, with significantly lower MDAPE for f and D* than all other approaches: NATTEN-17 (5.91%, 13.31%), SA-17 (7.73%, 13.66%), LSQ (21.95%, 54.34%), SEG (17.10%, 21.27%), BSP (12.35%, 22.79%), FBM (16.32%, 20.67%). In vivo, they provided superior visual quality and tumor delineation in f and D* maps, producing seemingly denoised versions of LSQ, while preserving tumor heterogeneity. The spatially-aware transformers yielded consistently the highest ROC-AUC values, particularly for D* (SA-17: 0.78), significantly outperforming LSQ (0.62), SEG (0.58), FBM (0.62), and IVIM-NET (0.71). DATA CONCLUSION: Transformer-based model fitting has the potential to provide clinically valuable IVIM parameter estimates and improved tumor grading accuracy. EVIDENCE LEVEL: 3. TECHNICAL EFFICACY: Stage 2.
BACKGROUND: The stria of Gennari (SoG) is a densely myelinated band within layer IVb of the primary visual cortex (V1) and represents the only cortical laminar structure visible macroscopically in vivo. Ultrahigh-field M...BACKGROUND: The stria of Gennari (SoG) is a densely myelinated band within layer IVb of the primary visual cortex (V1) and represents the only cortical laminar structure visible macroscopically in vivo. Ultrahigh-field MRI may improve its detection and conspicuity. PURPOSE: To quantitatively and qualitatively compare the in vivo appearance of the SoG at 3 versus 7 T using matched high-resolution T-weighted fast spin-echo MRI. STUDY TYPE: Prospective. POPULATION: A total of 12 subjects (6 female, 6 male; median age, 17 years; range, 15-24 years) with pediatric-onset epilepsy. FIELD STRENGTH/SEQUENCE: 3 and 7 T MRI, T-weighted fast spin-echo, acquired within 1 month of each other. ASSESSMENT: Cortical line profiles per subject were extracted orthogonally to the calcarine sulcus in V1. Quantitative metrics included peak-valley distance, Δ-signal (peak-valley difference), and contrast ratio (CR). Two blinded neuroradiologists and a radiologist independently rated SoG conspicuity using a 5-point Likert scale. STATISTICAL TESTS: McNemar's test compared detection rates; paired t-tests or Wilcoxon signed-rank tests compared quantitative metrics; reader preferences were analyzed using Wilcoxon tests; inter-reader agreement was assessed using weighted Cohen's κ. No multiple-comparison correction was applied (α = 0.05). RESULTS: SoG-consistent valleys were detected in 31% of profiles at 3 T and 65% at 7 T, significantly. Peak-valley distance remained stable across field strengths (0.45 ± 0.06 depth units; p = 0.37). Mean Δ-signal (33.5 vs. 36.1; p = 0.81) and contrast ratio (0.013 vs. 0.000; p = 0.54) did not differ significantly. Both readers demonstrated a strong preference for 7 T images (pseudo-median +1 to +1.5; significant), with fair inter-reader agreement (κ = 0.36). DATA CONCLUSION: The SoG can be visualized in vivo at both 3 and 7 T, with higher detection frequency and greater subjective conspicuity at 7 T. Quantitative laminar metrics remained stable across field strengths, suggesting that improved detectability at 7 T likely reflects enhanced spatial definition rather than measurable changes in signal contrast. TECHNICAL EFFICACY: Stage 2.
Huang J, Ge R, Li C
… +23 more, Li Y, Wen Z, Huang C, Xu A, Huang M, Lin J, Yuan H, Shi H, Li C, Dai L, Chen W, Zhang X, Tu Y, Liu C, Liang S, Bi Y, Liu W, Su S, Zhang X, Li X, Wen Z, Duan C, Feng X
BACKGROUND: Inflammation is a critical driver of intracranial aneurysm (IA) instability. The systemic immune-inflammation index (SII), a novel composite inflammatory biomarker, may reflect the local mural changes observe...BACKGROUND: Inflammation is a critical driver of intracranial aneurysm (IA) instability. The systemic immune-inflammation index (SII), a novel composite inflammatory biomarker, may reflect the local mural changes observed on vessel wall imaging (VWI). PURPOSE/HYPOTHESIS: To investigate the association between the SII and aneurysmal wall enhancement (AWE) on high-resolution vessel wall imaging (HRVWI), and to evaluate SII as a potential biomarker for aneurysm instability. STUDY TYPE: Prospective cross-sectional study. POPULATION: Four hundred and eighteen intracranial aneurysms in 311 patients (65.3% female, median age 58.0 years) underwent HRVWI. A sub-cohort of 67 patients with 84 aneurysms was included for longitudinal analysis of aneurysm growth (median follow-up 7.0 months). FIELD STRENGTH/SEQUENCE: Field strength/sequence: A 3.0T MR scanner; 3D time-of-flight magnetic resonance angiography (TOF-MRA); 3D black-blood T1-weighted volumetric turbo spin echo acquisition (T1-VISTA); and post-contrast vessel wall imaging using a 3D fast field echo sequence. ASSESSMENT: AWE defined as aneurysm-to-pituitary stalk contrast ratio (CR) ≥ 0.60. SII was calculated from admission blood counts as (platelet count * neutrophil count)/(lymphocyte count * 1000). STATISTICAL TESTS: Multivariable logistic regression and propensity score matching evaluated the association between SII and AWE. Longitudinal analysis was performed using Firth's logistic regression. Subgroup analyses were stratified by clinical and aneurysmal characteristics. p < 0.05 was considered significant. RESULTS: AWE was present in 142/418 (34.0%) aneurysms. In multivariable analysis, elevated SII was independently associated with AWE (odds ratio [OR] = 1.93). This association remained significant after PSM (median SII: 0.60 in AWE group vs. 0.42 in non-AWE group). SII was also independently associated with aneurysm growth in the longitudinal sub-cohort (OR = 20.519). Subgroup analysis showed a significant interaction with aneurysm location, with the strongest association in the internal carotid artery (OR = 3.55; 95% CI: 1.63-7.71). DATA CONCLUSION: Elevated SII is independently associated with the presence of AWE and aneurysm growth. EVIDENCE LEVEL: 3. TECHNICAL EFFICACY: Stage 2.
BACKGROUND: The neurostructural underpinnings of premenstrual dysphoric disorder (PMDD), particularly integrated white matter and network alteration, remain unclear. PURPOSE: To identify a core structural network in PMDD...BACKGROUND: The neurostructural underpinnings of premenstrual dysphoric disorder (PMDD), particularly integrated white matter and network alteration, remain unclear. PURPOSE: To identify a core structural network in PMDD by integrating multiple diffusion tensor imaging (DTI)-derived metrics and to develop a predictive model. STUDY TYPE: Prospective case-control study. SUBJECTS: Forty-two PMDD patients (age: 23.86 ± 1.32 years), diagnosed according to the American Psychiatric Association DSM-5, and 42 healthy controls (age: 23.79 ± 1.72 years). FIELD STRENGTH/SEQUENCE: 3.0 T, T1-weighted three-dimensional gradient-echo and echo planar imaging DTI sequences. ASSESSMENT: Microstructural and connectivity features were extracted from DTI using tract-based spatial statistics (TBSS), network-based statistics (NBS), and graph theory analyses. A combined predictive model was constructed by integrating the most stable features from the three single-modality models via least absolute shrinkage and selection operator (LASSO) regression. STATISTICAL TESTS: Group comparisons were performed using two-sample t-tests or Mann-Whitney U tests, with false discovery rate correction. Features were selected using LASSO and integrated to construct a combined model. Model performance was evaluated by the area under the receiver operating characteristic curve (AUC) using leave-one-out cross-validation. p < 0.05 was considered significant. RESULTS: PMDD patients exhibited widespread microstructural and connectivity alterations, including elevated axial diffusivity in the right posterior limb of the internal capsule, enhanced edge connectivity, and altered network topology. The combined model achieved significantly superior predictive performance (AUC = 0.855) compared with the TBSS-based model (AUC = 0.699) and the network-based model (AUC = 0.727), and a higher AUC than the graph-based model (AUC = 0.790). Key predictive features included two enhanced edges originating from the left inferior frontal gyrus and reduced degree centrality of the left inferior occipital gyrus and sulcus. DATA CONCLUSION: Our DTI-based predictive model showed alterations in brain connections and network properties in the left inferior frontal and inferior occipital regions of PMDD patients. TECHNICAL EFFICACY: Stage 2.
Boyacioglu R, Kluge T, Buonincontri G
… +9 more, Lo WC, Kannengiesser S, Nittka M, Kaur A, Dupuis A, Ma D, Badve C, A Griswold M, Chen Y
J Magn Reson Imaging
· 2026 Jul · PMID 41903184
·
Full text
BACKGROUND: Magnetic resonance fingerprinting (MRF) is an emerging quantitative imaging technique that enables multiparametric tissue characterization, but its adoption has been hindered by the complexity of data acquisi...BACKGROUND: Magnetic resonance fingerprinting (MRF) is an emerging quantitative imaging technique that enables multiparametric tissue characterization, but its adoption has been hindered by the complexity of data acquisition and post-processing. These technical and implementation challenges have limited its broader clinical deployment. PURPOSE: To develop a modular MRF Development Kit (MRFDK) that enables efficient sequence design, streamlined implementation, and real-time image reconstruction. STUDY TYPE: Prospective. POPULATION: T and T relaxation phantom, nine volunteers (seven males and two females), five metastatic brain cancer patients. FIELD STRENGTH/SEQUENCE: 3 T, MR Fingerprinting. ASSESSMENT: Accuracy of T and T quantification was estimated from phantom experiments. Manual ROIs were drawn on brain lesions and contralateral white matter for metastatic cancer patients. STATISTICAL TESTS: t-test, in vivo repeatability was calculated with Bland-Altman analysis on healthy volunteer scan-rescan data, significance level p < 0.01. RESULTS: Phantom results showed high accuracy in T and T assessment, with absolute percentage differences of 3% for T and 5% for T compared to offline MATLAB reconstruction. In vivo scans of eight healthy subjects further demonstrated excellent repeatability (bias and agreement: 0.95% ± 1.85% for T; 1.78% ± 5.08% for T). In patients, metastatic lesions showed significantly higher T and T values (T, 1474 ms; T, 61 ms) compared to normal white matter (T, 913 ms; T, 38 ms). With integrated B correction, all T and T maps were available for visualization within 1 min post-MRF scan, enabling immediate image assessment. DATA CONCLUSION: A modular MRF development package enabling efficient 3D acquisition and rapid inline reconstruction was developed and evaluated in this study. TECHNICAL EFFICACY: Stage 2.
BACKGROUND: While pericardial adipose tissue (PEAT) volume is linked to cardiac risk, the prognostic value of its radiomic features for major adverse cardiac events (MACE: a composite of cardiac death, myocardial infarct...BACKGROUND: While pericardial adipose tissue (PEAT) volume is linked to cardiac risk, the prognostic value of its radiomic features for major adverse cardiac events (MACE: a composite of cardiac death, myocardial infarction, heart failure hospitalization, or stroke) in end-stage renal disease (ESRD) patients is unknown. PURPOSE: To evaluate the association of PEAT radiomics with MACE in patients with ESRD. STUDY TYPE: Retrospective. POPULATION: Two hundred forty-eight ESRD patients (160 males) from three centers were included and divided into development (198 patients) and validation cohorts (50 patients) by centers. FIELD STRENGTH/SEQUENCE: 3.0 T/cine imaging, T2 and T2* mappings, all gradient echo. ASSESSMENT: A triple-stage Unet model (3SUnet) was used for the segmentation of PEAT in cine images. Myocardial T2 and T2* values were extracted from native T2 and T2* images. Demographic and laboratory data were collected within 1 week of the MRI scan. Participants were followed up from the MRI scanning every 2 weeks until July 2023 until MACE occurred. STATISTICAL TESTS: Principal component analysis (PCA) was used to reduce radiomic features to principal radiomics (PrPEAT-rads). Cox regression analyses were used to investigate the prognostic value of extracted PrPEAT-rads. p-value ≤ 0.05 was considered statistically significant. RESULTS: During a median follow-up time of 35.6 months, 63 patients (25.4% of 248; 15 in the external validation cohort) experienced MACE. The combined use of selected PrPEAT-rads with conventional MRI risk scores significantly improved risk stratification of MACE in both the internal (C-index: 0.715 vs. 0.768; integrated Brier score [IBS]: 0.095 vs. 0.090) and the external validation cohort (C-index: 0.737 vs. 0.774; IBS: 0.130 vs. 0.070). PrPEAT-rad1 was a significant independent factor associated with MACE after adjusting for both clinical (HR: 0.937) and MRI (HR: 0.916) covariates. DATA CONCLUSION: The cardiac MRI cine-based PEAT radiomics features in ESRD patients demonstrated a stronger association with MACE compared to conventional clinical and MRI parameters. EVIDENCE LEVEL: 3. STAGE OF TECHNOLOGY EFFICACY: Stage 2.
Babaeipour R, Fox MS, Parraga G
… +1 more, Ouriadov A
J Magn Reson Imaging
· 2026 Jul · PMID 41891379
·
Full text
BACKGROUND: Hyperpolarized Xe MRI faces technical challenges including low signal-to-noise ratio and breath-hold constraints. Current literature focuses on proprietary deep learning methods or image-domain enhancements....BACKGROUND: Hyperpolarized Xe MRI faces technical challenges including low signal-to-noise ratio and breath-hold constraints. Current literature focuses on proprietary deep learning methods or image-domain enhancements. PURPOSE: To present a comprehensive evaluation of transformer and hybrid CNN-transformer architectures integrating dual-domain (k-space and image) processing for HP Xe MRI reconstruction. STUDY TYPE: Retrospective. POPULATION: Two hundred five participants (22 healthy [male and female, 18-85 years], 26 COPD [male and female, 50-85 years], 90 asthma [male and female, 18-70 years], 67 long-COVID [male and female, 18-70 years]) yielding 1640 2D slices. Dataset split: 80% training (1312 slices), 10% validation (164 slices), 10% test (164 slices). FIELD STRENGTH/SEQUENCE: 3 T; 3D fast gradient-recalled echo. ASSESSMENT: Five architectures were compared: KTMR (hybrid transformer-CNN), KIKI-net (pure CNN), ReconFormer, SwinMR, and MR-IPT (pure transformer) at acceleration factors of 3, 7, and 10. Performance was assessed using peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and normalized mean squared error (NMSE). Ventilation defect percentage (VDP) agreement with semi-automated analysis was evaluated. STATISTICAL TESTS: Friedman test with post hoc Dunn's test and Benjamini-Hochberg correction for multiple comparisons. Significance level: p < 0.05. RESULTS: At 10-fold acceleration, KTMR produced PSNR of 36.4 ± 2.8 dB and SSIM of 0.88 ± 0.12, significantly outperforming KIKI-net (32.5 ± 3.4 dB, 0.81 ± 0.12), ReconFormer (29.7 ± 2.6 dB, 0.76 ± 0.12), SwinMR (30.5 ± 2.8 dB, 0.76 ± 0.09), and MR-IPT (28.8 ± 2.4 dB, 0.74 ± 0.11). VDP measurements showed mean bias of 1.94% at 3-fold, 2.12% at 7-fold, and 2.69% at 10-fold acceleration. DATA CONCLUSION: KTMR demonstrated superior performance for HP Xe MRI reconstruction at high acceleration factors. EVIDENCE LEVEL: 3. TECHNICAL EFFICACY: Stage 1.
BACKGROUND: In addition to intracranial plaques, extracranial carotid plaques have also been linked to stroke recurrence. However, the association of combined intracranial and extracranial plaque characteristics with str...BACKGROUND: In addition to intracranial plaques, extracranial carotid plaques have also been linked to stroke recurrence. However, the association of combined intracranial and extracranial plaque characteristics with stroke recurrence remains unclear. PURPOSE: To evaluate the association between intracranial plaque features and extracranial carotid Plaque Reporting and Data System (Plaque-RADS) with stroke recurrence utilizing high-resolution vessel wall imaging (HR-VWI). STUDY TYPE: Retrospective. POPULATION: The 388 intracranial atherosclerotic ischemic stroke patients (mean age, 57.3 ± 11.5 years, 274 males). FIELD STRENGTH/SEQUENCE: 3T, three-dimensional T1 weighted 3D fast spin echo. ASSESSMENT: HR-VWI was performed in all patients within 7 days of the stroke onset. The imaging features assessed included intracranial plaque characteristics (degree of stenosis, plaque burden, enhancement ratio, remodeling index, and intraplaque hemorrhage [IPH]) and extracranial carotid Plaque-RADS. Patients enrolled between May 2022 and July 2024 were included in the study. All patients were followed for a minimum of 12 months. STATISTICAL TESTS: Mann-Whitney U test or χ test, univariate and multivariate Cox regression analyses, time-dependent ROC and AUC curves, and Kaplan-Meier survival curves. p-values less than 0.05 were regarded as statistically significant. RESULTS: During a median follow-up period of 24 months, 55 patients experienced recurrent stroke. Intracranial plaque enhancement ratio (HR, 1.62; 95% CI, 1.11-2.38), IPH (HR, 2.55; 95% CI, 1.47-4.40), and extracranial carotid Plaque-RADS (HR, 2.23; 95% CI, 1.29-3.86) were significantly associated with stroke recurrence. Time-dependent ROC indicated that the maximum AUCs for enhancement ratio, IPH, and Plaque-RADS were 0.75, 0.69, and 0.68, respectively, while the Cox model including all three reached a maximum AUC of 0.79. DATA CONCLUSION: The combination of intracranial plaque characteristics and extracranial carotid Plaque-RADS can be used to assess ischemic stroke recurrence. EVIDENCE LEVEL: 3. TECHNICAL EFFICACY: Stage 2.
Nevalainen MT, Nykänen O, Järvinen J
… +7 more, Kemppainen A, Räsänen L, Casula V, Cloos M, Lattanzi R, Nissi MJ, Nieminen MT
J Magn Reson Imaging
· 2026 Jul · PMID 41882935
·
Full text
BACKGROUND: Magnetic Resonance Fingerprinting (MRF) enables rapid quantitative parameter mapping from which synthetic clinical contrast images can be derived using deep learning (DL). PURPOSE: This study evaluates the re...BACKGROUND: Magnetic Resonance Fingerprinting (MRF) enables rapid quantitative parameter mapping from which synthetic clinical contrast images can be derived using deep learning (DL). PURPOSE: This study evaluates the reliability and interchangeability of MRF-derived synthetic knee MRI relative to conventional MRI in patients with osteoarthritis. STUDY TYPE: Prospective single-center comparative study. SUBJECTS: Between March 2022 and 2023, 78 participants (54 females, mean age 57.2, range 33-78 years) with knee osteoarthritis. FIELD STRENGTH/SEQUENCE: 3.0 T; proton density weighted (PDw) imaging, T2-weighted fat-saturated (T2w fs) imaging, and MRF. ASSESSMENT: U-Nets were trained to produce synthetic contrasts from MRF data. Three musculoskeletal radiologists performed MRI OsteoArthritis Knee Score (MOAKS) assessments and image quality evaluation using a Likert scale (1-5). STATISTICAL TESTS: The inter-rater and inter-method reliability were evaluated using prevalence-and-bias-adjusted kappa (PABAK), and percentages of exact matches. Image quality scores were compared using the Wilcoxon test. The limit of statistical significance was set at p < 0.05 and no multiple comparisons corrections were applied. RESULTS: The inter-rater reliability for synthetic MR images varied between 0.980-0.994 (CI 0.975-0.997; exact matches 77.%-89.2%) and for conventional MR images between 0.979 and 0.994 (CI 0.973-0.997; exact matches 75.3%-89.5%). Inter-method reliability between synthetic and conventional MR images was near-perfect: mean PABAK-values and exact matches were 0.927 and 77.9% for cartilage, 0.915 and 89.7% for bone marrow lesions, 0.922 and 65.2% for osteophytes, 0.950 and 72.9% for meniscus pathology, 0.934 and 66.6% for effusion, and 0.857 and 92.5% for Baker's cyst. Average Likert scores were significantly better for conventional than synthetic images: 4.5 vs. 3.9 for PDw and 4.1 vs. 3.2 for T2w fs. DATA CONCLUSION: MRF-derived DL-based synthetic clinical contrasts provide excellent inter-rater reliability and interchangeability against conventional MR sequences in knee OA; however, image quality needs further development. EVIDENCE LEVEL: 2. STAGE OF TECHNICAL EFFICACY: 2.
BACKGROUND: Reliable quantification of perivascular spaces (PVS) in the basal ganglia (BG) is of growing interest for understanding the glymphatic system but remains challenging in infants. PURPOSE: To develop an automat...BACKGROUND: Reliable quantification of perivascular spaces (PVS) in the basal ganglia (BG) is of growing interest for understanding the glymphatic system but remains challenging in infants. PURPOSE: To develop an automated deep learning method for BG and BG-PVS segmentation in infant brain MRI using an anatomy-informed pseudo-labeling approach. STUDY TYPE: Retrospective, multi-cohort technical development, and validation study. POPULATION: Three cohorts: 150 neonates from the Developing Human Connectome Project (dHCP, 37-44 weeks of gestational age (GA); 76 males, 74 females), 133 infants from the Baby Connectome Project (BCP; ≤ 24 months; 70 males, 63 females) and 70 infants from an in-house dataset (30-41 weeks of GA; 36 males, 34 females). Manual ground-truth labels were generated by a trained researcher (dHCP, n = 150; BCP, n = 8; in-house, n = 10) and validated by a radiologist with 15 years of experience. FIELD STRENGTH/SEQUENCE: Data included 3 T MRI with T1- and T2-weighted sequences: dHCP (inversion recovery turbo spin-echo [IR-TSE] and turbo spin-echo [TSE]), BCP (magnetization-prepared rapid gradient-echo [MPRAGE] and TSE), and in-house (MPRAGE and variable-flip-angle TSE). ASSESSMENT: The proposed approach was compared with alternative automated approaches trained with different labeling strategies. Training/validation/test splits were 100/25/25 (dHCP), 100/25/8 (BCP), and 50/10/10 (in-house). STATISTICAL TESTS: Dice similarity coefficient (DSC), recall, positive predictive value, and Hausdorff distance were calculated for BG and BG-PVS quantification. Statistical significance was assessed using Wilcoxon signed-rank tests (p < 0.05), and quantification agreement was evaluated using Pearson's correlation, intraclass correlation coefficient (ICC), and mean absolute error (MAE). RESULTS: The proposed method improved accuracy (dHCP: BG DSC = 0.91 ± 0.03 and BG-PVS DSC = 0.78 ± 0.09; external datasets with fine-tuning: BG DSC = 0.86-0.89) and high agreement in PVS quantification with reference measurements (r = 0.90-0.99, ICC ≥ 0.96, MAE = 0.10). DATA CONCLUSION: The proposed method seems to enable robust and annotation-efficient BG and BG-PVS segmentation in infants. EVIDENCE LEVEL: 3. TECHNICAL EFFICACY: 1.