How young radiologists use contrast media and manage adverse reactions: an international survey.

OBJECTIVES: To collect real-world data about the knowledge and self-perception of young radiologists concerning the use of contrast media (CM) and the management of adverse drug reactions (ADR). METHODS: A survey (29 questions) was distributed to residents and board-certified radiologists younger than 40 years to investigate the current international situation in young radiology community regarding CM and ADRs. Descriptive statistics analysis was performed. RESULTS: Out of 454 respondents from 48 countries (mean age: 31.7 ± 4 years, range 25-39), 271 (59.7%) were radiology residents and 183 (40.3%) were board-certified radiologists. The majority (349, 76.5%) felt they were adequately informed regarding the use of CM. However, only 141 (31.1%) received specific training on the use of CM and 82 (18.1%) about management ADR during their residency. Although 266 (58.6%) knew safety protocols for handling ADR, 69.6% (316) lacked confidence in their ability to manage CM-induced ADRs and 95.8% (435) expressed a desire to enhance their understanding of CM use and handling of CM-induced ADRs. Nearly 300 respondents (297; 65.4%) were aware of the benefits of contrast-enhanced ultrasound, but 249 (54.8%) of participants did not perform it. The preferred CM injection strategy in CT parenchymal examination and CT angiography examination was based on patient's lean body weight in 318 (70.0%) and 160 (35.2%), a predeterminate fixed amount in 79 (17.4%) and 116 (25.6%), iodine delivery rate in 26 (5.7%) and 122 (26.9%), and scan time in 31 (6.8%) and 56 (12.3%), respectively. CONCLUSION: Training in CM use and management ADR should be implemented in the training of radiology residents. CRITICAL RELEVANCE STATEMENT: We highlight the need for improvement in the education of young radiologists regarding contrast media; more attention from residency programs and scientific societies should be focused on training about contrast media use and the management of adverse drug reactions. KEY POINTS: • This survey investigated training of young radiologists about use of contrast media and management adverse reactions. • Most young radiologists claimed they did not receive dedicated training. • An extreme heterogeneity of responses was observed about contrast media indications/contraindications and injection strategy.

Localization of contrast-enhanced breast lesions in ultrafast screening MRI using deep convolutional neural networks.

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.

Image quality of DWI at breast MRI depends on the amount of fibroglandular tissue: implications for unenhanced screening.

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.

Automated Breast Density Assessment in MRI Using Deep Learning and Radiomics: Strategies for Reducing Inter-Observer Variability.

BACKGROUND: Accurate breast density evaluation allows for more precise risk estimation but suffers from high inter-observer variability. PURPOSE: To evaluate the feasibility of reducing inter-observer variability of breast density assessment through artificial intelligence (AI) assisted interpretation. STUDY TYPE: Retrospective. POPULATION: Six hundred and twenty-one patients without breast prosthesis or reconstructions were randomly divided into training (N = 377), validation (N = 98), and independent test (N = 146) datasets. FIELD STRENGTH/SEQUENCE: 1.5 T and 3.0 T; T1-weighted spectral attenuated inversion recovery. ASSESSMENT: Five radiologists independently assessed each scan in the independent test set to establish the inter-observer variability baseline and to reach a reference standard. Deep learning and three radiomics models were developed for three classification tasks: (i) four Breast Imaging-Reporting and Data System (BI-RADS) breast composition categories (A-D), (ii) dense (categories C, D) vs. non-dense (categories A, B), and (iii) extremely dense (category D) vs. moderately dense (categories A-C). The models were tested against the reference standard on the independent test set. AI-assisted interpretation was performed by majority voting between the models and each radiologist's assessment. STATISTICAL TESTS: Inter-observer variability was assessed using linear-weighted kappa (κ) statistics. Kappa statistics, accuracy, and area under the receiver operating characteristic curve (AUC) were used to assess models against reference standard. RESULTS: In the independent test set, five readers showed an overall substantial agreement on tasks (i) and (ii), but moderate agreement for task (iii). The best-performing model showed substantial agreement with reference standard for tasks (i) and (ii), but moderate agreement for task (iii). With the assistance of the AI models, almost perfect inter-observer variability was obtained for tasks (i) (mean κ = 0.86), (ii) (mean κ = 0.94), and (iii) (mean κ = 0.94). DATA CONCLUSION: Deep learning and radiomics models have the potential to help reduce inter-observer variability of breast density assessment. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY: Stage 1.

Comparison of conventional and higher-resolution reduced-FOV diffusion-weighted imaging of breast tissue.

OBJECTIVE: Reduced FOV-diffusion-weighted imaging (rFOV-DWI) allows for acquisition of a tissue region without back-folding, and may have better fat suppression than conventional DWI imaging (c-DWI). The aim was to compare the ADCs obtained with c-DWI bilateral-breast imaging with single-breast rFOV-DWI. MATERIALS AND METHODS: Breasts of 38 patients were scanned at 3 T. The mean ADC values obtained for 38 lesions, and fibro-glandular (N = 35) and adipose (N = 38) tissue ROIs were compared between c-DWI and higher-resolution rFOV-DWI (Wilcoxon rank test). Also, the ADCs were compared between the two acquisitions for an oil-only phantom and a combined water/oil phantom. Furthermore, ghost artifacts were assessed. RESULTS: No significant difference in mean ADC was found between the acquisitions for lesions (c-DWI: 1.08 × 10-3 mm2/s, rFOV-DWI: 1.13 × 10-3 mm2/s) and fibro-glandular tissue. For adipose tissue, the ADC using rFOV-DWI (0.31 × 10-3 mm2/s) was significantly higher than c-DWI (0.16 × 10-3 mm2/s). For the oil-only phantom, no difference in ADC was found. However, for the water/oil phantom, the ADC of oil was significantly higher with rFOV-DWI compared to c-DWI. DISCUSSION: Although ghost artifacts were observed for both acquisitions, they appeared to have a greater impact for rFOV-DWI. However, no differences in mean lesions' ADC values were found, and therefore this study suggests that rFOV can be used diagnostically for single-breast DWI imaging.

Using deep learning to safely exclude lesions with only ultrafast breast MRI to shorten acquisition and reading time.

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.

Breast Tumor Identification in Ultrafast MRI Using Temporal and Spatial Information.

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.

Quantitative DWI implemented after DCE-MRI yields increased specificity for BI-RADS 3 and 4 breast lesions.

PURPOSE: To assess if specificity can be increased when semiautomated breast lesion analysis of quantitative diffusion-weighted imaging (DWI) is implemented after dynamic contrast-enhanced (DCE-) magnetic resonance imaging (MRI) in the workup of BI-RADS 3 and 4 breast lesions larger than 1 cm. MATERIALS AND METHODS: In all, 120 consecutive patients (mean-age, 48 years; age range, 23-75 years) with 139 breast lesions (≥1 cm) were examined (2010-2014) with 1.5T DCE-MRI and DWI (b = 0, 50, 200, 500, 800, 1000 s/mm2 ) and the BI-RADS classification and histopathology were obtained. For each lesion malignancy was excluded using voxelwise semiautomated breast lesion analysis based on previously defined thresholds for the apparent diffusion coefficient (ADC) and the three intravoxel incoherent motion (IVIM) parameters: molecular diffusion (Dslow ), microperfusion (Dfast ), and the fraction of Dfast (ffast ). The sensitivity (Se), specificity (Sp), and negative predictive value (NPV) based on only IVIM parameters combined in parallel (Dslow , Dfast , and ffast ), or the ADC or the BI-RADS classification by DCE-MRI were compared. Subsequently, the Se, Sp, and NPV of the combination of the BI-RADS classification by DCE-MRI followed by the IVIM parameters in parallel (or the ADC) were compared. RESULTS: In all, 23 of 139 breast lesions were benign. Se and Sp of DCE-MRI was 100% and 30.4% (NPV = 100%). Se and Sp of IVIM parameters in parallel were 92.2% and 52.2% (NPV = 57.1%) and for the ADC 95.7% and 17.4%, respectively (NPV = 44.4%). In all, 26 of 139 lesions were classified as BI-RADS 3 (n = 7) or BI-RADS 4 (n = 19). DCE-MRI combined with ADC (Se = 99.1%, Sp = 34.8%) or IVIM (Se = 99.1%, Sp = 56.5%) did significantly improve (P = 0.016) Sp of DCE-MRI alone for workup of BI-RADS 3 and 4 lesions (NPV = 92.9%). CONCLUSION: Quantitative DWI has a lower NPV compared to DCE-MRI for evaluation of breast lesions and may therefore not be able to replace DCE-MRI; when implemented after DCE-MRI as problem solver for BI-RADS 3 and 4 lesions, the combined specificity improves significantly. J. Magn. Reson. Imaging 2016;44:1642-1649.

Semi-automated quantitative intravoxel incoherent motion analysis and its implementation in breast diffusion-weighted imaging.

BACKGROUND: To optimize and validate intravoxel incoherent motion (IVIM) modeled diffusion-weighted imaging (DWI) compared with the apparent diffusion coefficient (ADC) for semi-automated analysis of breast lesions using a multi-reader setup. MATERIALS AND METHODS: Patients (n = 176) with breast lesions (≥1 cm) and known pathology were prospectively examined (1.5 Tesla) with DWI (b = 0, 50, 200, 500, 800, 1000 s/mm(2) ) between November 2008 and July 2014 and grouped into a training and test set. Three independent readers applied a semi-automated procedure for setting regions-of-interest for each lesion and recorded ADC and IVIM parameters: molecular diffusion (Dslow ), microperfusion (Dfast ), and the fraction of Dfast (ffast ). In the training set (24 lesions, 12 benign), a semi-automated method was optimized to yield maximum true negatives (TN) with minimal false negatives (FN): only the optimal fraction (Fo) of voxels in the lesions was used and optimal thresholds were determined. The optimal Fo and thresholds were then applied to a consecutive test set (139 lesions, 23 benign) to obtain specificity and sensitivity. RESULTS: In the training set, optimal thresholds were 1.44 × 10(-3) mm(2) /s (Dslow ), 18.55 × 10(-3) mm(2) /s (Dfast ), 0.247 (ffast ) and 2.00 × 10(-3) mm(2) /s (ADC) with Fo set to 0.61, 0.85, 1.0, and 1.0, respectively, this resulted in TN = 5 (IVIM) and TN = 1 (ADC), with FN = 0. In the test set, sensitivity and specificity among the readers were 90.5-93.1% and 43.5-52.2%, respectively, for IVIM, and 94.8-95.7% and 13.0-21.7% for ADC (P ≤ 0.0034) without inter-reader differences (P = 1.000). CONCLUSION: The presented semi-automated method for breast lesion evaluation is reader independent and yields significantly higher specificity for IVIM compared with the ADC.