Explore the vast range of titles available.

Objectives To develop a deep learning–based method for contrast-enhanced breast lesion detection in ultrafast screening MRI. Materials and methods A total of 837 breast MRI exams of 488 consecutive patients were included. Lesion’s location was independently annotated in the maximum intensity projection (MIP) image of the last time-resolved angiography with stochastic trajectories (TWIST) sequence for each individual breast, resulting in 265 lesions (190 benign, 75 malignant) in 163 breasts (133 women). YOLOv5 models were fine-tuned using training sets containing the same number of MIP images with and without lesions. A long short-term memory (LSTM) network was employed to help reduce false positive predictions. The integrated system was then evaluated on test sets containing enriched uninvolved breasts during cross-validation to mimic the performance in a screening scenario. Results In five-fold cross-validation, the YOLOv5x model showed a sensitivity of 0.95, 0.97, 0.98, and 0.99, with 0.125, 0.25, 0.5, and 1 false positive per breast, respectively. The LSTM network reduced 15.5% of the false positive prediction from the YOLO model, and the positive predictive value was increased from 0.22 to 0.25. Conclusions A fine-tuned YOLOv5x model can detect breast lesions on ultrafast MRI with high sensitivity in a screening population, and the output of the model could be further refined by an LSTM network to reduce the amount of false positive predictions. Clinical relevance statement The proposed integrated system would make the ultrafast MRI screening process more effective by assisting radiologists in prioritizing suspicious examinations and supporting the diagnostic workup. Key Points •  Deep convolutional neural networks could be utilized to automatically pinpoint breast lesions in screening MRI with high sensitivity . •  False positive predictions significantly increased when the detection models were tested on highly unbalanced test sets with more normal scans . •  Dynamic enhancement patterns of breast lesions during contrast inflow learned by the long short-term memory networks helped to reduce false positive predictions .

Objectives To compare image quality of diffusion-weighted imaging (DWI) and contrast-enhanced breast MRI (DCE-T1) stratified by the amount of fibroglandular tissue (FGT) as a measure of breast density. Methods Retrospective, multi-reader, bicentric visual grading analysis study on breast density (A–D) and overall image and fat suppression quality of DWI and DCE-T1, scored on a standard 5-point Likert scale. Cross tabulations and visual grading characteristic (VGC) curves were calculated for fatty breasts (A/B) versus dense breasts (C/D). Results Image quality of DWI was higher in the case of increased breast density, with good scores (score 3–5) in 85.9% (D) and 88.4% (C), compared to 61.6% (B) and 53.5% (A). Overall image quality of DWI was in favor of dense breasts (C/D), with an area under the VGC curve of 0.659 ( p < 0.001). Quality of DWI and DCE-T1 fat suppression increased with higher breast density, with good scores (score 3–5) for 86.9% and 45.7% of density D, and 90.2% and 42.9% of density C cases, compared to 76.0% and 33.6% for density B and 54.7% and 29.6% for density A (DWI and DCE-T1 respectively). Conclusions Dense breasts show excellent fat suppression and substantially higher image quality in DWI images compared with non-dense breasts. These results support the setup of studies exploring DWI-based MR imaging without IV contrast for additional screening of women with dense breasts. Clinical relevance statement Our findings demonstrate that image quality of DWI is robust in women with an increased amount of fibroglandular tissue, technically supporting the feasibility of exploring applications such as screening of women with mammographically dense breasts. Key Points • Image and fat suppression quality of diffusion-weighted imaging are dependent on the amount of fibroglandular tissue (FGT) which is closely connected to breast density. • Fat suppression quality in diffusion-weighted imaging of the breast is best in women with a high amount of fibroglandular tissue. • High image quality of diffusion-weighted imaging in women with a high amount of FGT in MRI supports that the technical feasibility of DWI can be explored in the additional screening of women with mammographically dense breasts. Graphical Abstract

Purpose: To investigate the feasibility of using deep learning methods to differentiate benign from malignant breast lesions in ultrafast MRI with both temporal and spatial information. Methods: A total of 173 single breasts of 122 women (151 examinations) with lesions above 5 mm were retrospectively included. A total of 109 out of 173 lesions were benign. Maximum intensity projection (MIP) images were generated from each of the 14 contrast-enhanced T1-weighted acquisitions in the ultrafast MRI scan. A 2D convolutional neural network (CNN) and a long short-term memory (LSTM) network were employed to extract morphological and temporal features, respectively. The 2D CNN model was trained with the MIPs from the last four acquisitions to ensure the visibility of the lesions, while the LSTM model took MIPs of an entire scan as input. The performance of each model and their combination were evaluated with 100-times repeated stratified four-fold cross-validation. Those models were then compared with models developed with standard DCE-MRI which followed the same data split. Results: In the differentiation between benign and malignant lesions, the ultrafast MRI-based 2D CNN achieved a mean AUC of 0.81 ± 0.06, and the LSTM network achieved a mean AUC of 0.78 ± 0.07; their combination showed a mean AUC of 0.83 ± 0.06 in the cross-validation. The mean AUC values were significantly higher for ultrafast MRI-based models than standard DCE-MRI-based models. Conclusion: Deep learning models developed with ultrafast breast MRI achieved higher performances than standard DCE-MRI for malignancy discrimination. The improved AUC values of the combined models indicate an added value of temporal information extracted by the LSTM model in breast lesion characterization.

MRI, as a completely noninvasive technique, can provide quantitative assessment of perfusion, diffusion, viscoelasticity and metabolism, yielding diverse information about liver function. Furthermore, pathological accumulations of iron and lipids can be quantified. Perfusion MRI with various contrast agents is commonly used for the detection and characterization of focal liver disease and the quantification of blood flow parameters. An extended new application is the evaluation of the therapeutic effect of antiangiogenic drugs on liver tumours. Novel, but already widespread, is a histologically validated relaxometry method using five gradient echo sequences for quantifying liver iron content elevation, a measure of inflammation, liver disease and cancer. Because of the high perfusion fraction in the liver, the apparent diffusion coefficients strongly depend on the gradient factors used in diffusion-weighted MRI. While complicating analysis, this offers the opportunity to study perfusion without contrast injection. Another novel method, MR elastography, has already been established as the only technique able to stage fibrosis or diagnose mild disease. Liver fat content is accurately determined with multivoxel MR spectroscopy (MRS) or by faster MRI methods that are, despite their widespread use, prone to systematic error. Focal liver disease characterisation will be of great benefit once multivoxel methods with fat suppression are implemented in proton MRS, in particular on high-field MR systems providing gains in signal-to-noise ratio and spectral resolution.

It has been demonstrated that hepatic apparent diffusion coefficients (ADC) are decreasing in patients with a Fontan circulation. It remains however unclear whether this is a true decrease of molecular diffusion, or rather reflects decreased microperfusion due to decreased portal blood flow. The purpose of this study was therefore to differentiate diffusion and microperfusion using intravoxel incoherent motion (IVIM) modeled diffusion-weighted imaging (DWI) for different liver segments in patients with a Fontan circulation, compare to a control group, and relate with liver function, chronic hepatic congestion and hepatic disease. For that purpose, livers of 59 consecutively included patients with Fontan circulation (29 men; mean-age, 19.1 years) were examined (Oct 2012─Dec 2013) with 1.5T MRI and DWI (b = 0,50,100,250,500,750,1500,1750 s/mm2). IVIM (Dslow, Dfast, ffast) and ADC were calculated for eight liver segments, compared to a control group (19 volunteers; 10 men; mean-age, 32.9 years), and correlated to follow-up duration, clinical variables, and laboratory measurements associated with liver function. The results demonstrated that microperfusion was reduced (p<0.001) in Fontan livers compared to controls with ─38.1% for Dfast and ─32.6% for ffast. Molecular diffusion (Dslow) was similar between patients and controls, while ADC was significantly lower (─14.3%) in patients (p<0.001). ADC decreased significantly with follow-up duration after Fontan operation (r = ─0.657). Dslow showed significant inverse correlations (r = ─0.591) with follow-up duration whereas Dfast and ffast did not. From these results it was concluded that the decreasing ADC values in Fontan livers compared with controls reflect decreases in hepatic microperfusion rather than any change in molecular diffusion. However, with the time elapsed since the Fontan operation molecular diffusion and ADC decreased while microperfusion remained stable. This indicates that after Fontan operation initial blood flow effects on the liver are followed by intracellular changes preceding the formation of fibrosis and cirrhosis.

Objectives To evaluate the effect of the choice of b values and prior use of contrast medium on apparent diffusion coefficients (ADCs) of breast lesions derived from diffusion-weighted imaging (DWI), and on the discrimination between benign and malignant lesions. Methods A literature search of relevant DWI studies was performed. The accuracy of DWI to characterize lesions by using b value ≤600 s/mm 2 and b value >600 s/mm 2 was presented as pooled sensitivity and specificity, and the ADC was calculated for both groups. Lesions were pooled as pre- or post-contrast DWI. Results Of 198 articles, 26 met the inclusion criteria. Median ADCs were significantly higher (13.2–35.1 %, p  < 0.001) for the group of b values ≤600 s/mm 2 compared to >600 s/mm 2 . The sensitivity in both groups was similar (91 % and 89 %, p  = 0.495) as well as the specificity (75 % and 84 %, p  = 0.237). Contrast medium had no significant effects on the ADCs ( p  ≥ 0.08). The differentiation between benign and malignant lesions was optimal (58.4 %) for the combination of b  = 0 and 1,000 s/mm 2 . Conclusions The wide variety of b value combinations applied in different studies significantly affects the ADC of breast lesions and confounds quantitative DWI. If only a couple of b values are used, those of b  = 0 and 1,000 s/mm 2 are recommended for the best improvement of differentiating between benign and malignant lesions. Key Points • The choice of b values significantly affects the ADC of breast lesions. • Sensitivity and specificity are not affected by the choice of b values. • b values 0 and 1,000 s/mm 2 are recommended for optimal differentiation between benign and malignant lesions. • Contrast medium prior to DWI does not significantly affect the ADC.

Objective Clinical hepatic diffusion weighted imaging (DWI) generally relies on mono-exponential diffusion. The aim was to demonstrate that mono-exponential diffusion in the liver is contaminated by microperfusion and that the bi-exponential model is required. Methods Nineteen fasting healthy volunteers were examined with DWI (seven b -values) using fat suppression and respiratory triggering (1.5 T). Five different regions in the liver were analysed regarding the mono-exponentially fitted apparent diffusion coefficient (ADC), and the bi-exponential model: molecular diffusion ( D slow ), microperfusion ( D fast ) and the respective fractions ( f slow/fast ). Data were compared using ANOVA and Kruskal–Wallis tests. Simulations were performed by repeating our data analyses, using just the DWI series acquired with b -values approximating those of previous studies. Results Median mono-exponentially fitted ADCs varied significantly ( P  < 0.001) between 1.107 and 1.423 × 10 −3  mm 2 /s for the five regions. Bi-exponential fitted D slow varied between 0.923 and 1.062 × 10 −3  mm 2 /s without significant differences ( P  = 0.140). D fast varied significantly, between 17.8 and 46.8 × 10 −3  mm 2 /s ( P  < 0.001). F-tests showed that the diffusion data fitted the bi-exponential model significantly better than the mono-exponential model ( F  > 21.4, P  < 0.010). These results were confirmed by the simulations. Conclusion ADCs of normal liver tissue are significantly dependent on the measurement location because of substantial microperfusion contamination; therefore the bi-exponential model should be used. Key Points Diffusion weighted MR imaging helps clinicians to differentiate tumours by diffusion properties Fast moving water molecules experience microperfusion, slow molecules diffusion Hepatic diffusion should be measured by bi-exponential models to avoid microperfusion contamination Mono-exponential models are contaminated with microperfusion, resulting in apparent regional diffusion differences Bi-exponential models are necessary to measure diffusion and microperfusion in the liver

Diffusion-weighted imaging (DWI) is an important diagnostic tool in the assessment of focal liver lesions and diffuse liver diseases such as cirrhosis and fibrosis. Quantitative DWI parameters such as molecular diffusion, microperfusion and their fractions, are known to be affected when hepatic fat fractions (HFF) are higher than 5.5% (steatosis). However, less is known about the effect on DWI for HFF in the normal non-steatotic range below 5.5%, which can be found in a large part of the population. The aim of this study was therefore to evaluate the diagnostic implications of non-steatotic HFF on quantitative DWI parameters in eight liver segments. For this purpose, eleven healthy volunteers (2 men, mean-age 31.0) were prospectively examined with DWI and three series of in-/out-of-phase dual-echo spoiled gradient-recalled MRI sequences to obtain the HFF and T2*. DWI data were analyzed using the intravoxel incoherent motion (IVIM) model. Four circular regions (ø22.3 mm) were drawn in each of eight liver segments and averaged. Measurements were divided in group 1 (HFF ≤ 2.75%), group 2 (2.75< HFF ≤ 5.5%) and group 3 (HFF>5.5%). DWI parameters and T2* were compared between the three groups and between the segments. It was observed that the molecular diffusion (0.85, 0.72 and 0.49 × 10(-3) mm(2)/s) and T2* (32.2, 27.2 and 21.0 ms) differed significantly between the three groups of increasing HFF (2.18, 3.50 and 19.91%). Microperfusion and its fraction remained similar for different HFF. Correlations with HFF were observed for the molecular diffusion (r = -0.514, p<0.001) and T2* (-0.714, p<0.001). Similar results were obtained for the majority of individual liver segments. It was concluded that fat significantly decreases molecular diffusion in the liver, also in absence of steatosis (HFF ≤ 5.5%). Also, it was confirmed that fat influences T2*. Determination of HFF prior to quantitative DWI is therefore crucial.

In phenylketonuria, elevated plasma phenylalanine concentrations may disturb blood-to-brain large neutral amino acid (LNAA) transport and cerebral protein synthesis (CPS). We investigated the associations between these processes, using data obtained by positron emission tomography with l-[1-11C]-tyrosine (11C-Tyr) as a tracer. Blood-to-brain transport of non-Phe LNAAs was modeled by the rate constant for 11C-Tyr transport from arterial plasma to brain tissue (K1), while CPS was modeled by the rate constant for 11C-Tyr incorporation into cerebral protein (k3). Brain phenylalanine concentrations were measured by magnetic resonance spectroscopy in three volumes of interest (VOIs): supraventricular brain tissue (VOI 1), ventricular brain tissue (VOI 2), and fluid-containing ventricular voxels (VOI 3). The associations between k3 and each predictor variable were analyzed by multiple linear regression. The rate constant k3 was inversely associated with brain phenylalanine concentrations in VOIs 2 and 3 (adjusted R2=0.826, F=19.936, P=0.021). Since brain phenylalanine concentrations in these VOIs highly correlated with each other, the specific associations of each predictor with k3 could not be determined. The associations between k3 and plasma phenylalanine concentration, K1, and brain phenylalanine concentrations in VOI 1 were nonsignificant. In conclusion, our study shows an inverse association between k3 and increased brain phenylalanine concentrations.

Objectives To investigate the feasibility of automatically identifying normal scans in ultrafast breast MRI with artificial intelligence (AI) to increase efficiency and reduce workload. Methods In this retrospective analysis, 837 breast MRI examinations performed on 438 women from April 2016 to October 2019 were included. The left and right breasts in each examination were labelled normal (without suspicious lesions) or abnormal (with suspicious lesions) based on final interpretation. Maximum intensity projection (MIP) images of each breast were then used to train a deep learning model. A high sensitivity threshold was calculated based on the detection trade - off (DET) curve on the validation set. The performance of the model was evaluated by receiver operating characteristic analysis of the independent test set. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) with the high sensitivity threshold were calculated. Results The independent test set consisted of 178 examinations of 149 patients (mean age, 44 years ± 14 [standard deviation]). The trained model achieved an AUC of 0.81 (95% CI: 0.75–0.88) on the independent test set. Applying a threshold of 0.25 yielded a sensitivity of 98% (95% CI: 90%; 100%), an NPV of 98% (95% CI: 89%; 100%), a workload reduction of 15.7%, and a scan time reduction of 16.6%. Conclusion This deep learning model has a high potential to help identify normal scans in ultrafast breast MRI and thereby reduce radiologists’ workload and scan time. Key Points • Deep learning in TWIST may eliminate the necessity of additional sequences for identifying normal breasts during MRI screening. • Workload and scanning time reductions of 15.7% and 16.6%, respectively, could be achieved with the cost of 1 (1 of 55) false negative prediction.