Fat-saturated T2-weighted Images Research Articles

Using MRI appendicitis scale and DWI for the diagnosis of acute appendicitis in pregnant women.

To assess the performance of MRI scale for the diagnosis of acute appendicitis in pregnant women and to determine the added diagnostic value of diffusion-weighted imaging (DWI). From January 2018 to December 2020, 80 patients were included. All MRI were performed with a 1.5-Tesla scanner with anterior array body coil. This analysis included (1) T2-weighted imaging (T2WI), (2) fat-saturated T2WI, and (3) DWI. Two radiologists blinded to the diagnosis recorded their assessment of four findings: appendiceal diameter, appendiceal wall thickness, luminal mucus, and periappendiceal inflammation. The MRI scale of acute appendicitis which ranged from 0 to 4 was determined from these factors. An additional one point was added to the MRI appendicitis scale in those patients with evidence of appendiceal restricted diffusion on DWI. The diagnostic values and predictive factors were computed. Multivariate analysis demonstrated that the calculated MRI appendicitis scale was a significant independent predictor of acute appendicitis with a sensitivity of 96.6%, specificity of 90.2%, and PPV of 84.8%. The odds ratio of appendicitis is increased by 22.3 times for every increase in one point on the MRI appendicitis scale. Therefore, the addition of one point for restricted diffusion in the appendix on DWI imaging can add substantial value, both positive and negative predictive value, towards making an accurate diagnosis of acute appendicitis. MRI appendicitis scale is an objective and significant independent predictive factor for acute appendicitis in pregnant women. Incorporation of diffusion weighted imaging to MRI can improve diagnosis of acute appendicitis. MRI appendicitis scale is an objective and significant independent predictor of acute appendicitis in pregnant women. Incorporation of DWI/ADC map to MRI examinations can improve diagnosis of acute appendicitis in pregnant women. • MRI appendicitis scale is an objective and significant independent predictive factor for acute appendicitis in pregnant women. • The odds ratio of appendicitis can be increased by 22.3 times for every increase of one unit in MRI scale. • Incorporation of diffusion-weighted imaging to MRI examinations can add value to the scale (4.2 ± 0.7 vs. 0.7 ± 1.1; p < 0.001) among pregnant women with appendicitis versus pregnant women without appendicitis.

Segmentation-Based Analysis of T2- and T1-Weighted Dynamic Magnetic Resonance Images Provides Adequate Observer Agreement in the Evaluation of Interstitial Lung Disease.

The aim of the study is to quantify observer agreement in the magnetic resonance imaging (MRI) classification of inflammatory or fibrotic interstitial lung disease (ILD). Our study is a preliminary analysis of a larger prospective cohort. The MRI images of 18 patients with ILD (13 females; mean age, 65 years) were acquired in a 1.5 T scanner and included axial fat-saturated T2-weighted (T2-WI, n = 18) and coronal fat-saturated T1-weighted images before and 1, 3, 5, and 10 minutes after gadolinium administration (n = 16). The MRI studies were evaluated with 2 different methods: a qualitative evaluation (visual assessment and measurement of few regions of interest; evaluations were performed independently by 5 radiologists and 3 times by 1 radiologist) and a segmentation-based analysis with software extraction of signal intensity values (evaluations were performed independently by 2 radiologists and twice by 1 radiologist). Interstitial lung disease was classified as inflammatory or fibrotic, based on previously described imaging criteria. Regarding the qualitative evaluation, intraobserver agreement was excellent (κ = 0.92, P < 0.05) for T2-WI and fair (κ = 0.29, P < 0.05) for T1 dynamic study, while interobserver agreement was moderate (κ = 0.56, P < 0.05) and poor (κ = 0.11, P = 0.18), respectively. In contrast, upon segmentation-based analysis, intraobserver and interobserver agreement were excellent for T2-WI (κ = 0.886, P < 0.001; κ = 1.00, P < 0.001; respectively); for T1-WI, intraobserver agreement was excellent (κ = 0.87, P < 0.05) and interobserver agreement was good (κ = 0.75, P < 0.05). Segmentation-based MRI analysis is more reproducible than a qualitative evaluation with visual assessment and measurement of few regions of interest.

Intramuscular lesions in musculoskeletal MRI as a favourable prognostic sign in patients with anti-MDA5 antibody-positive dermatomyositis

ObjectivesAnti-differentiation-associated gene 5 (MDA5) antibody-positive dermatomyositis, which has been described as clinically amyopathic dermatomyositis, complicates rapidly progressive interstitial lung disease (ILD). Owing to the absence of significant muscle symptoms, musculoskeletal...

Implementation of GAN-Based, Synthetic T2-Weighted Fat Saturated Images in the Routine Radiological Workflow Improves Spinal Pathology Detection.

(1) Background and Purpose: In magnetic resonance imaging (MRI) of the spine, T2-weighted (T2-w) fat-saturated (fs) images improve the diagnostic assessment of pathologies. However, in the daily clinical setting, additional T2-w fs images are frequently missing due to time constraints or motion artifacts. Generative adversarial networks (GANs) can generate synthetic T2-w fs images in a clinically feasible time. Therefore, by simulating the radiological workflow with a heterogenous dataset, this study's purpose was to evaluate the diagnostic value of additional synthetic, GAN-based T2-w fs images in the clinical routine. (2) Methods: 174 patients with MRI of the spine were retrospectively identified. A GAN was trained to synthesize T2-w fs images from T1-w, and non-fs T2-w images of 73 patients scanned in our institution. Subsequently, the GAN was used to create synthetic T2-w fs images for the previously unseen 101 patients from multiple institutions. In this test dataset, the additional diagnostic value of synthetic T2-w fs images was assessed in six pathologies by two neuroradiologists. Pathologies were first graded on T1-w and non-fs T2-w images only, then synthetic T2-w fs images were added, and pathologies were graded again. Evaluation of the additional diagnostic value of the synthetic protocol was performed by calculation of Cohen's ĸ and accuracy in comparison to a ground truth (GT) grading based on real T2-w fs images, pre- or follow-up scans, other imaging modalities, and clinical information. (3) Results: The addition of the synthetic T2-w fs to the imaging protocol led to a more precise grading of abnormalities than when grading was based on T1-w and non-fs T2-w images only (mean ĸ GT versus synthetic protocol = 0.65; mean ĸ GT versus T1/T2 = 0.56; p = 0.043). (4) Conclusions: The implementation of synthetic T2-w fs images in the radiological workflow significantly improves the overall assessment of spine pathologies. Thereby, high-quality, synthetic T2-w fs images can be virtually generated by a GAN from heterogeneous, multicenter T1-w and non-fs T2-w contrasts in a clinically feasible time, which underlines the reproducibility and generalizability of our approach.

Combination Use of Compressed Sensing and Deep Learning for Shoulder Magnetic Resonance Imaging With Various Sequences.

For compressed sensing (CS) to become widely used in routine magnetic resonance imaging (MRI), it is essential to improve image quality. This study aimed to evaluate the usefulness of combining CS and deep learning-based reconstruction (DLR) for various sequences in shoulder MRI. This retrospective study included 37 consecutive patients who underwent undersampled shoulder MRIs, including T1-weighted (T1WI), T2-weighted (T2WI), and fat-saturation T2-weighted (FS-T2WI) images. Images were reconstructed using the conventional wavelet-based denoising method (wavelet method) and a combination of wavelet and DLR-based denoising methods (hybrid-DLR method) for each sequence. The signal-to-noise ratio and contrast-to-noise ratio of the bone, muscle, and fat and the full width at half maximum of the shoulder joint were compared between the 2 image types. In addition, 2 board-certified radiologists scored the image noise, contrast, sharpness, artifacts, and overall image quality of the 2 image types on a 4-point scale. The signal-to-noise ratios and contrast-to-noise ratios of the bone, muscle, and fat in T1WI, T2WI, and FS-T2WI obtained from the hybrid-DLR method were significantly higher than those of the conventional wavelet method ( P < 0.001). However, there were no significant differences in the full width at half maximum of the shoulder joint in any of the sequences ( P > 0.05). Furthermore, in all sequences, the mean scores of the image noise, sharpness, artifacts, and overall image quality were significantly higher in the hybrid-DLR method than in the wavelet method ( P < 0.001), but there were no significant differences in contrast among the sequences ( P > 0.05). The DLR denoising method can improve the image quality of CS in T1-weighted images, T2-weighted images, and fat-saturation T2-weighted images of the shoulder compared with the wavelet denoising method alone.

Shoulder MRI-based radiomics for diagnosis and severity staging assessment of surgically treated supraspinatus tendon tears.

To develop and validate MRI-based radiomics models capable of evaluating supraspinatus tendon tears within the shoulder joints by using arthroscopy as the reference standard. A total of 432 patients (332 in the training set and 100 in the external validation set) with intact supraspinatus tendon (n = 202) and supraspinatus tendon tear (n = 230, 130 full-thickness tears and 100 partial-thickness tears) were enrolled. Radiomics features were extracted from fat-saturated T2-weighted coronal images. Two radiomics signature models for detecting supraspinatus tendon abnormalities (tear or not), and stage lesion severity (full- or partial-thickness tear) and radiomics scores (Rad-score), were constructed and calculated using multivariate logistic regression analysis. The diagnostic performance of the two models was validated using ROC curves on the training and validation datasets. For the radiomics model of no tears or tears, thirteen features from MR images were used to build the radiomics signature with an AUC value of 0.98 in the training set, 0.97 in the internal validation set, and 0.98 in the external validation set. For the radiomics model of full- or partial-thickness tears, thirteen features from MR images were used to build the radiomics signature with an AUC value of 0.79 in the training set, 0.69 in the internal validation set, and 0.77 in the external validation set. The proposed radiomics models in this study can accurately rule out supraspinatus tendon tears and are capable of assessing the severity staging of tears with moderate accuracy based on shoulder MR images. • The radiomics model of no tears or tears achieved a high overall accuracy of 93.6%, sensitivity of 91.6%, and specificity of 95.2% for supraspinatus tendon tears. • The radiomics model of full- or partial-thickness tears displayed moderate performance with an accuracy of 76.4%, a sensitivity of 79.2%, and a specificity of 74.3% for supraspinatus tendon tears severity staging.

Intraobserver and Interobserver Reproducibility of Breast Diffusion-Weighted Imaging Quantitative Parameters: Readout-Segmented vs. Single-Shot Echo-Planar Imaging.

The recommended technique for breast diffusion-weighted imaging (DWI) acquisitions is not sufficiently standardized in clinical practice. To investigate the intraobserver and interobserver reproducibility of DWI measurements, diffusion-kurtosis imaging (DKI) parameters, and image quality evaluation in breast lesions between single-shot echo-planar imaging (ss-EPI) and readout-segmented echo-planar imaging (rs-EPI). Prospective. A total of 295 women with 209 malignant and 86 benign breast lesions. A 3-T; fat-saturated T2-weighted MR imaging (T2WI); multi-b-value DWI with both ss-EPI and rs-EPI readouts; T1-weighted dynamic contrast-enhanced MRI (DCE-MRI). Mean kurtosis (MK), mean diffusion (MD), and apparent diffusion coefficient (ADC) values were measured for each lesion on ss-EPI and rs-EPI, respectively. Image quality was visually evaluated regarding image sharpness, geometric distortion, lesion conspicuity, visualization of anatomic structures, and overall quality. Quantitative and qualitative analyses were performed twice with a time interval of 2 weeks. Intraobserver and interobserver reproducibility wereevaluated using intra-class correlation coefficients (ICC), within-subject coefficient of variation (wCV), and Bland-Altman plots. MK, MD, and ADC quantitative parameters for breast lesions showed excellent intraobserver and interobserver reproducibility, with ICCs >0.75 and wCV values ranging from 2.51% to 7.08% for both sequences. The wCV values in both intraobserver and interobserver measurements were higher in the ss-EPI sequence (3.63%-7.08%) than that of the rs-EPI sequence (2.51%-3.62%). The wCV values differed in subgroups with different histopathological types of lesions, breast density, lesion morphology, and lesion sizes, respectively. Furthermore, rs-EPI (ICCs, 0.76-0.97; wCV values, 2.41%-6.04%) had better intraobserver and interobserver reproducibility than ss-EPI (ICCs, 0.54-0.90; wCV values, 6.18%-13.69%) with regard to image quality. Compared to the ss-EPI, the rs-EPI sequence showed higher intraobserver and interobserver reproducibility for quantitative diffusion-related parameters and image quality assessments measured in breast DWI and DKI. 2. Stage 2.

MRI-based radiomics nomogram for differentiation of solitary metastasis and solitary primary tumor in the spine

BackgroundDifferentiating between solitary spinal metastasis (SSM) and solitary primary spinal tumor (SPST) is essential for treatment decisions and prognosis. The aim of this study was to develop and validate an MRI-based radiomics nomogram for discriminating SSM from SPST.MethodsOne hundred and thirty-five patients with solitary spinal tumors were retrospectively studied and the data set was divided into two groups: a training set (n = 98) and a validation set (n = 37). Demographics and MRI characteristic features were evaluated to build a clinical factors model. Radiomics features were extracted from sagittal T1-weighted and fat-saturated T2-weighted images, and a radiomics signature model was constructed. A radiomics nomogram was established by combining radiomics features and significant clinical factors. The diagnostic performance of the three models was evaluated using receiver operator characteristic (ROC) curves on the training and validation sets. The Hosmer–Lemeshow test was performed to assess the calibration capability of radiomics nomogram, and we used decision curve analysis (DCA) to estimate the clinical usefulness.ResultsThe age, signal, and boundaries were used to construct the clinical factors model. Twenty-six features from MR images were used to build the radiomics signature. The radiomics nomogram achieved good performance for differentiating SSM from SPST with an area under the curve (AUC) of 0.980 in the training set and an AUC of 0.924 in the validation set. The Hosmer–Lemeshow test and decision curve analysis demonstrated the radiomics nomogram outperformed the clinical factors model.ConclusionsA radiomics nomogram as a noninvasive diagnostic method, which combines radiomics features and clinical factors, is helpful in distinguishing between SSM and SPST.

Orbital Lipolysis.

HomeRadiologyVol. 307, No. 2 PreviousNext Reviews and CommentaryFree AccessImages in RadiologyOrbital LipolysisTuba Kalelioglu , Sohil H. PatelTuba Kalelioglu , Sohil H. PatelAuthor AffiliationsFrom the Department of Radiology and Medical Imaging, University of Virginia Health System, 1215 Lee St, 1st Floor, Charlottesville, VA 22903.Address correspondence to T.K. (email: [email protected]).Tuba Kalelioglu Sohil H. PatelPublished Online:Jan 24 2023https://doi.org/10.1148/radiol.221947MoreSectionsPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In A 47-year-old man with a history of squamous cell carcinoma of the oral tongue, alcohol-induced chronic pancreatitis with gastrostomy tube dependence, and chronic diarrhea presented to the emergency department with altered mental status. The patient was cachectic, with a body mass index of 14.3 kg/m2.MRI of the brain demonstrated diffuse loss of expected fat signal intensity with T1-weighted imaging, including within the orbits, bone marrow, and subcutaneous scalp (Figure). This appearance is similar to a fat-saturated T1-weighted image; hence, it is called “pseudo fat-saturated” (1). It is due to excessive lipolysis and lipid peroxidation induced by cachexia from chronic malnutrition in the setting of anorexia nervosa or systemic cancer. Furthermore, endothelial injury results in edema and contrast enhancement in the involved areas, conspicuous in the postseptal orbit (2).Images in a 47-year-old man with cachectic state. (A) T1-weighted image before cachectic state for comparison shows expected normal high signal intensity of the orbital, scalp, and bone marrow fat. (B) Axial T1-weighted image without fat suppression shows lack of expected hyperintense fat in the orbits, subcutaneous scalp, and bone marrow. The imaging appearance mimics a fat-saturated T1-weighted image. The volume of subcutaneous fat in temporal regions is decreased. (C) Postcontrast T1-weighted image shows diffuse contrast enhancement of the affected areas secondary to increased vessel wall permeability. (D) Noncontrast CT scan shows increased attenuation in the retrobulbar fat, consistent with edema.Download as PowerPointOpen in Image Viewer Disclosures of conflicts of interest: T.K. No relevant relationships. S.H.P. No relevant relationships.

Classification tree analysis to evaluate the most useful magnetic resonance image type in the differentiation between early and progressed hepatocellular carcinoma.

Using classification tree analysis, we evaluated the most useful magnetic resonance (MR) image type in the differentiation between early and progressed hepatocellular carcinoma (eHCC and pHCC). We included pathologically proven 214 HCCs (28 eHCCs and 186 pHCCs) in 144 patients. The signal intensity of HCCs was assessed on in-phase (T1in) and opposed-phase T1-weighted images (T1op), ultrafast T2-weighted images (ufT2WI), fat-saturated T2-weighted images (fsT2WI), diffusion-weighted images (DWI), contrast enhanced T1-weighted images in the arterial phase (AP), portal venous phase (PVP), and the hepatobiliary phase. Fat content and washout were also evaluated. Fisher's exact test was performed to evaluate usefulness for the differentiation. Then, we chose MR images using binary logistic regression analysis and performed classification and regression tree analysis with them. Diagnostic performances of the classification tree were evaluated using a stratified 10-fold cross-validation method. T1in, ufT2WI, fsT2WI, DWI, AP, PVP, fat content, and washout were all useful for the differentiation (p < 0.05), and AP and T1in were finally chosen for creating classification trees (p < 0.05). AP appeared in the first node in the tree. The area under the curve, sensitivity and specificity for eHCC, and balanced accuracy of the classification tree were 0.83 (95% CI 0.74-0.91), 0.64 (18/28, 95% CI 0.46-0.82), 0.94 (174/186, 95% CI 0.90-0.97), and 0.79 (95% CI 0.70-0.87), respectively. AP is the most useful MR image type and T1in the second in the differentiation between eHCC and pHCC.

Synthetic MR image generation of macrotrabecular-massive hepatocellular carcinoma using generative adversarial networks

PurposeThe purpose of this study was to develop a method for generating synthetic MR images of macrotrabecular-massive hepatocellular carcinoma (MTM-HCC). Materials and methodsA set of abdominal MR images including fat-saturated T1-weighted images obtained during the arterial and portal venous phases of enhancement and T2-weighted images of 91 patients with MTM-HCC, and another set of MR abdominal images from 67 other patients were used. Synthetic images were obtained using a 3-step pipeline that consisted in: (i), generating a synthetic MTM-HCC tumor on a neutral background; (ii), randomly selecting a background among the 67 patients and a position inside the liver; and (iii), merging the generated tumor in the background at the specified location. Synthetic images were qualitatively evaluated by three radiologists and quantitatively assessed using a mix of 1-nearest neighbor classifier metric and Fréchet inception distance. ResultsA set of 1000 triplets of synthetic MTM-HCC images with consistent contrasts were successfully generated. Evaluation of selected synthetic images by three radiologists showed that the method gave realistic, consistent and diversified images. Qualitative and quantitative evaluation led to an overall score of 0.64. ConclusionThis study shows the feasibility of generating realistic synthetic MR images with very few training data, by leveraging the wide availability of liver backgrounds. Further studies are needed to assess the added value of those synthetic images for automatic diagnosis of MTM-HCC.

MRI Fat-Saturated T2-Weighted Radiomics Model for Identifying the Ki-67 Index of Soft Tissue Sarcomas.

Ki-67 expression has been shown to be an important risk factor associated with prognosis in patients with soft tissue sarcomas (STSs). Its assessment requires fine-needle biopsy and its accuracy can be influenced by tumor heterogeneity. To develop and test an MRI-based radiomics nomogram for identifying the Ki-67 status of STSs. Retrospective. A total of 149 patients at two independent institutions (training cohort [high Ki-67/low ki-67]: 102 [52/50], external validation cohort [high Ki-67/low ki-67]: 47 [28/19]) with STSs. Fat-saturated T2-weighted imaging (FS-T2WI) with a fat-suppressed fast spin/turbo spin echo sequence at 1.5T or 3T. After radiomics feature extraction, logistic regression (LR), random forest (RF), support vector machine (SVM), and k-nearest neighbor (KNN) were used to construct radiomics models to distinguish between high and low Ki-67 status. Clinical-MRI characteristics included age, gender, location, size, margin, and MRI morphological features (size, margin, signal intensity, and peritumoral hyperintensity) were assessed. Univariate and multivariate logistic regression analysis were applied for screening significant risk factors. Radiomics nomogram was constructed by radiomics signature and risk factors. Model performances (discrimination, calibration, and clinical usefulness) were validated in the validation cohort. The nomogram was assessed using the Harrell index of concordance (C-index), calibration curve analysis. The clinical utility of the model was assessed by decision curve analysis (DCA). LR, RF, SVM, and KNN models represented AUCs of 0.789, 0.755, 0.726, and 0.701 in the validation cohort (P> 0.05). The nomogram had a C-index of 0.895 (95% CI: 0.837-0.953) in the training cohort and 0.852 (95% CI: 0.796-0.957) in the validation cohort and it demonstrated good calibration and clinical utility (P=0.972 for the training cohort and P=0.727 for the validation cohort). This MRI-based radiomics nomogram developed showed good performance in identifying Ki-67 expression status in STSs. 2.

Radiomics-based machine learning approach in differentiating fibro-adipose vascular anomaly from venous malformation.

As a complex vascular malformation, fibro-adipose vascular anomaly was first proposed in 2014. Its overlap with other vascular malformations regarding imaging and clinical features often leads to misdiagnosis and improper management. To construct a radiomics-based machine learning model to help radiologists differentiate fibro-adipose vascular anomaly from common venous malformations. We retrospectively analyzed 178 children, adolescents and young adults with vascular malformations (41 fibro-adipose vascular anomaly and 137 common vascular malformation cases) who underwent MRI before surgery between May 2012 to January 2021. We extracted radiomics features from T1-weighted images and fat-saturated (FS) T2-weighted images and further selected features through least absolute shrinkage and selection operator (LASSO) and Boruta methods. We established eight weighted logistic regression classification models based on various combinations of feature-selection strategies (LASSO or Boruta) and sequence types (single- or multi-sequence). Finally, we evaluated the performance of each model by the mean area under the receiver operating characteristics curve (ROC-AUC), sensitivity and specificity in 10 runs of repeated k-fold (k = 10) cross-validation. Two multi-sequence models based on axial FS T2-W, coronal FS T2-W and axial T1-W images showed promising performance. The LASSO-based multi-sequence model achieved an AUC of 97%±3.8, a sensitivity of 94%±12.4 and a specificity of 89%±9.0. The Boruta-based multi-sequence model achieved an AUC of 97%±3.7, a sensitivity of 95%±10.5 and a specificity of 87%±9.0. The radiomics-based machine learning model can provide a promising tool to help distinguish fibro-adipose vascular anomaly from common venous malformations.

Differentiation of Benign From Malignant Parotid Gland Tumors Using Conventional MRI Based on Radiomics Nomogram.

ObjectivesWe aimed to develop and validate radiomic nomograms to allow preoperative differentiation between benign- and malignant parotid gland tumors (BPGT and MPGT, respectively), as well as between pleomorphic adenomas (PAs) and Warthin tumors (WTs).Materials and MethodsThis retrospective study enrolled 183 parotid gland tumors (68 PAs, 62 WTs, and 53 MPGTs) and divided them into training (n = 128) and testing (n = 55) cohorts. In total, 2553 radiomics features were extracted from fat-saturated T2-weighted images, apparent diffusion coefficient maps, and contrast-enhanced T1-weighted images to construct single-, double-, and multi-sequence combined radiomics models, respectively. The radiomics score (Rad-score) was calculated using the best radiomics model and clinical features to develop the radiomics nomogram. The receiver operating characteristic curve and area under the curve (AUC) were used to assess these models, and their performances were compared using DeLong’s test. Calibration curves and decision curve analysis were used to assess the clinical usefulness of these models.ResultsThe multi-sequence combined radiomics model exhibited better differentiation performance (BPGT vs. MPGT, AUC=0.863; PA vs. MPGT, AUC=0.929; WT vs. MPGT, AUC=0.825; PA vs. WT, AUC=0.927) than the single- and double sequence radiomics models. The nomogram based on the multi-sequence combined radiomics model and clinical features attained an improved classification performance (BPGT vs. MPGT, AUC=0.907; PA vs. MPGT, AUC=0.961; WT vs. MPGT, AUC=0.879; PA vs. WT, AUC=0.967).ConclusionsRadiomics nomogram yielded excellent diagnostic performance in differentiating BPGT from MPGT, PA from MPGT, and PA from WT.

Multiparametric magnetic resonance imaging for differentiating active from inactive thyroid-associated ophthalmopathy: Added value from magnetization transfer imaging

PurposeTo investigate the feasibility of using magnetization transfer (MT) magnetic resonance imaging for evaluating patients with thyroid-associated ophthalmopathy (TAO), and determine its added value for differentiating active from inactive TAO and predicting clinical activity score (CAS), compared with conventional fat-saturated T2-weighted and diffusion-weighted imaging. MethodsOrbital MT, fat-saturated T2-weighted, and diffusion-weighted imaging of 60 prospectively enrolled consecutive patients with TAO was analyzed. Simplified histogram parameters (mean, max, min) of magnetization transfer ratio (MTR), signal intensity ratio (SIR), and apparent diffusion coefficient (ADC) at extraocular muscles were calculated for each orbit and compared between the active and inactive groups. ResultsIntraclass correlation coefficients of MTRs and SIRs were similar (0.802–0.963 vs 0.812–0.974, respectively), followed by those of ADCs (0.714–0.855). Patients with active TAO showed significantly lower MTRs and higher SIRs and ADCs than those with inactive TAO (P < 0.05). MTRmean achieved the highest area under the curve (AUC) of 0.868 for differentiating active from inactive group, followed by SIRmax (AUC, 0.836). MTRmean also demonstrated a higher and negative correlation with CAS (r = –0.614, P < 0.001) than MTRmax and MTRmin (r = –0.495, P < 0.001; r = –0.243, P = 0.007; respectively). Support vector machine-based analysis revealed that uniting MTRs could prosper concurrently added performance for disease activity differentiation and CAS prediction, compared with merely combining SIRs and ADCs (AUC, 0.933 vs 0.901; r = 0.703 vs. 0.673). ConclusionsMT imaging could potentially be used as a noninvasive method for differentiating the activity of TAO and predicting CAS, thereby offering added value to conventional SIR and ADC.

Accelerated Two-Point Dixon MR Angiography Improves Diagnostic Performance for Cervical Artery Diseases.

Nonenhanced MR angiography (MRA) studies are often used to manage acute and chronic large cervical artery disease, but lengthy scan times limit their clinical usefulness. To develop an accelerated cervical MRA and test its diagnostic performance. Prospective. Patients with cervical artery disease (n= 32, 17 males). 3.0 T; accelerated two-point Dixon three-dimensional Cartesian spoiled gradient-echo (FLEXA) and conventional time-of-flight MRA (cMRA) sequences. All patients underwent FLEXA (1'28″) and cMRA (6'47″) acquisitions. Quantitative evaluation (artery-to-background signal ratio and a blur metric) and qualitative evaluation using diagnostic performance measured by the sensitivity, specificity, and positive/negative predictive values (PPV/NPV), and vessel and plaque visualization scores from three board-certified radiologists' (with 10, 11, and 12 years of experience) independent readings using maximum intensity projection (MIP) for luminal diseases and axial images for plaque. The reference standards were contrast-enhanced angiography and fat-saturated T1-weighted images, respectively. All measures were compared between FLEXA and cMRA using the paired t, Wilcoxon signed-rank, McNemar's, or chi-squared test, as appropriate. Interreader agreement was assessed using Cohen's κ. P< 0.05 was considered statistically significant. The artery-to-background signal ratio was significantly higher for FLEXA (FLEXA: 7.20 ± 1.63 [fat]; 4.26 ± 0.52 [muscle]; cMRA: 2.57 ± 0.49 [fat]), while image blurring was significantly less (FLEXA: 0.24 ± 0.016; cMRA: 0.30 ± 0.029). In luminal disease detection, sensitivity (FLEXA: 0.97/0.91/0.91; cMRA:0.71/0.69/0.63), specificity (FLEXA: 0.98/0.93/0.98; cMRA:0.93/0.85/0.92), PPV (FLEXA: 0.92/0.86/0.86; cMRA: 0.64/0.5/0.58), and NPV (FLEXA: 0.99/0.98/0.98; cMRA: 0.92/0.91/0.9) were significantly higher for FLEXA. interreader agreement was substantial to almost perfect for FLEXA (κ= 0.82/0.86/0.78) and moderate to substantial for cMRA (κ= 0.67/0.56/0.57). MIP visualization scores were significantly higher for FLEXA, with substantial to almost perfect interreader agreement (FLEXA: κ= 0.83/0.86/0.82; cMRA: κ= 0.89/0.79/0.79). In plaque detection, sensitivity (FLEXA: 0.9/0.9/0.7; cMRA: 0.3/0.6/0.2) and specificity (FLEXA: 1/0.87/1; cMRA: 0.93/0.63/0.97) were significantly higher for FLEXA in two of three readers. The interreader plaque detection agreement was fair to substantial (FLEXA: κ= 0.63/0.69/0.48; cMRA: κ= 0.21/0.45/0.20). Side-by-side plaque and vessel wall visualization was superior for FLEXA in all readers, with moderate to substantial interreader agreement (plaque: κ= 0.73/0.73/0.77; vessel wall: κ= 0.57/0.40/0.39). FLEXA enhanced visualization of the cervical arterial system and improved diagnostic performance for luminal abnormalities and plaques in patients with cervical artery diseases. 1 TECHNICAL EFFICACY STAGE: 2.

Breast MRI Background Parenchymal Enhancement Categorization Using Deep Learning: Outperforming the Radiologist.

Background parenchymal enhancement (BPE) is assessed on breast MRI reports as mandated by the Breast Imaging Reporting and Data System (BI-RADS) but is prone to inter and intrareader variation. Semiautomated and fully automated BPE assessment tools have been developed but none has surpassed radiologist BPE designations. To develop a deep learning model for automated BPE classification and to compare its performance with current standard-of-care radiology report BPE designations. Retrospective. Consecutive high-risk patients (i.e. &gt;20% lifetime risk of breast cancer) who underwent contrast-enhanced screening breast MRI from October 2013 to January 2019. The study included 5224 breast MRIs, divided into 3998 training, 444 validation, and 782 testing exams. On radiology reports, 1286 exams were categorized as high BPE (i.e., marked or moderate) and 3938 as low BPE (i.e., mild or minimal). A 1.5 T or 3T system; one precontrast and three postcontrast phases of fat-saturated T1-weighted dynamic contrast-enhanced imaging. Breast MRIs were used to develop two deep learning models (Slab artificial intelligence (AI); maximum intensity projection [MIP] AI) for BPE categorization using radiology report BPE labels. Models were tested on a heldout test sets using radiology report BPE and three-reader averaged consensus as the reference standards. Model performance was assessed using receiver operating characteristic curve analysis. Associations between high BPE and BI-RADS assessments were evaluated using McNemar's chi-square test (α*=0.025). The Slab AI model significantly outperformed the MIP AI model across the full test set (area under the curve of 0.84 vs. 0.79) using the radiology report reference standard. Using three-reader consensus BPE labels reference standard, our AI model significantly outperformed radiology report BPE labels. Finally, the AI model was significantly more likely than the radiologist to assign "high BPE" to suspicious breast MRIs and significantly less likely than the radiologist to assign "high BPE" to negative breast MRIs. Fully automated BPE assessments for breast MRIs could be more accurate than BPE assessments from radiology reports. 4 TECHNICAL EFFICACY STAGE: 3.

Differentiation of carcinosarcoma from endometrial carcinoma on magnetic resonance imaging using deep learning.

PurposeTo verify whether deep learning can be used to differentiate between carcinosarcomas (CSs) and endometrial carcinomas (ECs) using several magnetic resonance imaging (MRI) sequences.Material and methodsThis retrospective study included 52 patients with CS and 279 patients with EC. A deep-learning model that uses convolutional neural networks (CNN) was trained with 572 T2-weighted images (T2WI) from 42 patients, 488 apparent diffusion coefficient of water maps from 33 patients, and 539 fat-saturated contrast-enhanced T1-weighted images from 40 patients with CS, as well as 1612 images from 223 patients with EC for each sequence. These were tested with 9-10 images of 9-10 patients with CS and 56 images of 56 patients with EC for each sequence, respectively. Three experienced radiologists independently interpreted these test images. The sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) for each sequence were compared between the CNN models and the radiologists.ResultsThe CNN model of each sequence had sensitivity 0.89-0.93, specificity 0.44-0.70, accuracy 0.83-0.89, and AUC 0.80-0.94. It also showed an equivalent or better diagnostic performance than the 3 readers (sensitivity 0.43-0.91, specificity 0.30-0.78, accuracy 0.45-0.88, and AUC 0.49-0.92). The CNN model displayed the highest diagnostic performance on T2WI (sensitivity 0.93, specificity 0.70, accuracy 0.89, and AUC 0.94).ConclusionsDeep learning provided diagnostic performance comparable to or better than experienced radiologists when distinguishing between CS and EC on MRI.

Classification of MR-Detected Additional Lesions in Patients With Breast Cancer Using a Combination of Radiomics Analysis and Machine Learning.

ObjectiveThis study was conducted in order to investigate the feasibility of using radiomics analysis (RA) with machine learning algorithms based on breast magnetic resonance (MR) images for discriminating malignant from benign MR-detected additional lesions in patients with primary breast cancer.Materials and MethodsOne hundred seventy-four MR-detected additional lesions (benign, n = 86; malignancy, n = 88) from 158 patients with ipsilateral primary breast cancer from a tertiary medical center were included in this retrospective study. The entire data were randomly split to training (80%) and independent test sets (20%). In addition, 25 patients (benign, n = 21; malignancy, n = 15) from another tertiary medical center were included for the external test. Radiomics features that were extracted from three regions-of-interest (ROIs; intratumor, peritumor, combined) using fat-saturated T1-weighted images obtained by subtracting pre- from postcontrast images (SUB) and T2-weighted image (T2) were utilized to train the support vector machine for the binary classification. A decision tree method was utilized to build a classifier model using clinical imaging interpretation (CII) features assessed by radiologists. Area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, and specificity were used to compare the diagnostic performance.ResultsThe RA models trained using radiomics features from the intratumor-ROI showed comparable performance to the CII model (accuracy, AUROC: 73.3%, 69.6% for the SUB RA model; 70.0%, 75.1% for the T2 RA model; 73.3%, 72.0% for the CII model). The diagnostic performance increased when the radiomics and CII features were combined to build a fusion model. The fusion model that combines the CII features and radiomics features from multiparametric MRI data demonstrated the highest performance with an accuracy of 86.7% and an AUROC of 91.1%. The external test showed a similar pattern where the fusion models demonstrated higher levels of performance compared with the RA- or CII-only models. The accuracy and AUROC of the SUB+T2 RA+CII model in the external test were 80.6% and 91.4%, respectively.ConclusionOur study demonstrated the feasibility of using RA with machine learning approach based on multiparametric MRI for quantitatively characterizing MR-detected additional lesions. The fusion model demonstrated an improved diagnostic performance over the models trained with either RA or CII alone.

Diagnostic abdominal MR imaging on a prototype low-field 0.55T scanner operating at two different gradient strengths.

To develop a protocol for abdominal imaging on a prototype 0.55T scanner and to benchmark the image quality against conventional 1.5T exam. In this prospective IRB-approved HIPAA-compliant study, 10 healthy volunteers were recruited and imaged. A commercial MRI system was modified to operate at 0.55T (LF) with two different gradient performance levels. Each subject underwent non-contrast abdominal examinations on the 0.55T scanner utilizing higher gradients (LF-High), lower adjusted gradients (LF-Adjusted), and a conventional 1.5T scanner. The following pulse sequences were optimized: fat-saturated T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and Dixon T1-weighted imaging (T1WI). Three readers independently evaluated image quality in a blinded fashion on a 5-point Likert scale, with a score of 1 being non-diagnostic and 5 being excellent. An exact paired sample Wilcoxon signed-rank test was used to compare the image quality. Diagnostic image quality (overall image quality score ≥ 3) was achieved at LF in all subjects for T2WI, DWI, and T1WI with no more than one unit lower score than 1.5T. The mean difference in overall image quality score was not significantly different between LF-High and LF-Adjusted for T2WI (95% CI -0.44 to 0.44; p = 0.98), DWI (95% CI -0.43 to 0.36; p = 0.92), and for T1 in- and out-of-phase imaging (95%C I -0.36 to 0.27; p = 0.91) or T1 fat-sat (water only) images (95% CI -0.24 to 0.18; p = 1.0). Diagnostic abdominal MRI can be performed on a prototype 0.55T scanner, either with conventional or with reduced gradient performance, within an acquisition time of 10min or less.