Multicenter Quantification of Radiation Exposure and Associated Risks for Prostatic Artery Embolization in 1476 Patients.

Background Prostatic artery embolization (PAE) is a safe, minimally invasive angiographic procedure that effectively treats benign prostatic hyperplasia; however, PAE-related patient radiation exposure and associated risks are not completely understood. Purpose To quantify radiation dose and assess radiation-related adverse events in patients who underwent PAE at multiple centers. Materials and Methods This retrospective study included patients undergoing PAE for any indication performed by experienced operators at 10 high-volume international centers from January 2014 to May 2021. Patient characteristics, procedural and radiation dose data, and radiation-related adverse events were collected. Procedural radiation effective doses were calculated by multiplying kerma-area product values by an established conversion factor for abdominopelvic fluoroscopy-guided procedures. Relationships between cumulative air kerma (CAK) or effective dose and patient body mass index (BMI), fluoroscopy time, or radiation field area were assessed with linear regression. Differences in radiation dose stemming from radiopaque prostheses or fluoroscopy unit type were assessed using two-sample t tests and Wilcoxon rank sum tests. Results A total of 1476 patients (mean age, 69.9 years ± 9.0 [SD]) were included, of whom 1345 (91.1%) and 131 (8.9%) underwent the procedure with fixed interventional or mobile fluoroscopy units, respectively. Median procedure effective dose was 17.8 mSv for fixed interventional units and 12.3 mSv for mobile units. CAK and effective dose both correlated positively with BMI (R2 = 0.15 and 0.17; P < .001) and fluoroscopy time (R2 = 0.16 and 0.08; P < .001). No radiation-related 90-day adverse events were reported. Patients with radiopaque implants versus those without implants had higher median CAK (1452 mGy [range, 900-2685 mGy] vs 1177 mGy [range, 700-1959 mGy], respectively; P = .01). Median effective dose was lower for mobile than for fixed interventional systems (12.3 mSv [range, 8.5-22.0 mSv] vs 20.4 mSv [range, 13.8-30.6 mSv], respectively; P < .001). Conclusion Patients who underwent PAE performed with fixed interventional or mobile fluoroscopy units were exposed to a median effective radiation dose of 17.8 mSv or 12.3 mSv, respectively. No radiation-related adverse events at 90 days were reported. © RSNA, 2024 See also the editorial by Mahesh in this issue.

Multi-Task Learning for Motion Analysis and Segmentation in 3D Echocardiography.

Characterizing left ventricular deformation and strain using 3D+time echocardiography provides useful insights into cardiac function and can be used to detect and localize myocardial injury. To achieve this, it is imperative to obtain accurate motion estimates of the left ventricle. In many strain analysis pipelines, this step is often accompanied by a separate segmentation step; however, recent works have shown both tasks to be highly related and can be complementary when optimized jointly. In this work, we present a multi-task learning network that can simultaneously segment the left ventricle and track its motion between multiple time frames. Two task-specific networks are trained using a composite loss function. Cross-stitch units combine the activations of these networks by learning shared representations between the tasks at different levels. We also propose a novel shape-consistency unit that encourages motion propagated segmentations to match directly predicted segmentations. Using a combined synthetic and in-vivo 3D echocardiography dataset, we demonstrate that our proposed model can achieve excellent estimates of left ventricular motion displacement and myocardial segmentation. Additionally, we observe strong correlation of our image-based strain measurements with crystal-based strain measurements as well as good correspondence with SPECT perfusion mappings. Finally, we demonstrate the clinical utility of the segmentation masks in estimating ejection fraction and sphericity indices that correspond well with benchmark measurements.

Reliable Prostate Cancer Risk Mapping From MRI Using Targeted and Systematic Core Needle Biopsy Histopathology.

OBJECTIVE: To compute a dense prostate cancer risk map for the individual patient post-biopsy from magnetic resonance imaging (MRI) and to provide a more reliable evaluation of its fitness in prostate regions that were not identified as suspicious for cancer by a human-reader in pre- and intra-biopsy imaging analysis. METHODS: Low-level pre-biopsy MRI biomarkers from targeted and non-targeted biopsy locations were extracted and statistically tested for representativeness against biomarkers from non-biopsied prostate regions. A probabilistic machine learning classifier was optimized to map biomarkers to their core-level pathology, followed by extrapolation of pathology scores to non-biopsied prostate regions. Goodness-of-fit was assessed at targeted and non-targeted biopsy locations for the post-biopsy individual patient. RESULTS: Our experiments showed high predictability of imaging biomarkers in differentiating histopathology scores in thousands of non-targeted core-biopsy locations (ROC-AUCs: 0.85-0.88), but also high variability between patients (Median ROC-AUC [IQR]: 0.81-0.89 [0.29-0.40]). CONCLUSION: The sparseness of prostate biopsy data makes the validation of a whole gland risk mapping a non-trivial task. Previous studies i) focused on targeted-biopsy locations although biopsy-specimens drawn from systematically scattered locations across the prostate constitute a more representative sample to non-biopsied regions, and ii) estimated prediction-power across predicted instances (e.g., biopsy specimens) with no patient distinction, which may lead to unreliable estimation of model fitness to the individual patient due to variation between patients in instance count, imaging characteristics, and pathologies. SIGNIFICANCE: This study proposes a personalized whole-gland prostate cancer risk mapping post-biopsy to allow clinicians to better stage and personalize focal therapy treatment plans.

INTEGRATING PROSTATE SPECIFIC ANTIGEN DENSITY BIOMARKER INTO DEEP LEARNING PROSTATE MRI LESION SEGMENTATION MODELS.

Prostate cancer lesion segmentation in multi-parametric magnetic resonance imaging (mpMRI) is crucial for pre-biopsy diagnosis and targeted biopsy guidance. Deep convolution neural networks have been widely utilized for lesion segmentation. However, these methods fail to achieve a high Dice coefficient because of the large variations in lesion size and location within the gland. To address this problem, we integrate the clinically-meaningful prostate specific antigen density (PSAD) biomarker into the deep learning model using feature-wise transformations to condition the features in latent space, and thus control the size of lesion prediction. We tested our models on a public dataset with 214 annotated mpMRI scans and compared the segmentation performance to a baseline 3D U-Net model. Results demonstrate that integrating the PSAD biomarker significantly improves segmentation performance in both Dice coefficient and centroid distance metric.

A Multiclass Radiomics Method-Based WHO Severity Scale for Improving COVID-19 Patient Assessment and Disease Characterization From CT Scans.

OBJECTIVES: The aim of this study was to evaluate the severity of COVID-19 patients' disease by comparing a multiclass lung lesion model to a single-class lung lesion model and radiologists' assessments in chest computed tomography scans. MATERIALS AND METHODS: The proposed method, AssessNet-19, was developed in 2 stages in this retrospective study. Four COVID-19-induced tissue lesions were manually segmented to train a 2D-U-Net network for a multiclass segmentation task followed by extensive extraction of radiomic features from the lung lesions. LASSO regression was used to reduce the feature set, and the XGBoost algorithm was trained to classify disease severity based on the World Health Organization Clinical Progression Scale. The model was evaluated using 2 multicenter cohorts: a development cohort of 145 COVID-19-positive patients from 3 centers to train and test the severity prediction model using manually segmented lung lesions. In addition, an evaluation set of 90 COVID-19-positive patients was collected from 2 centers to evaluate AssessNet-19 in a fully automated fashion. RESULTS: AssessNet-19 achieved an F1-score of 0.76 ± 0.02 for severity classification in the evaluation set, which was superior to the 3 expert thoracic radiologists (F1 = 0.63 ± 0.02) and the single-class lesion segmentation model (F1 = 0.64 ± 0.02). In addition, AssessNet-19 automated multiclass lesion segmentation obtained a mean Dice score of 0.70 for ground-glass opacity, 0.68 for consolidation, 0.65 for pleural effusion, and 0.30 for band-like structures compared with ground truth. Moreover, it achieved a high agreement with radiologists for quantifying disease extent with Cohen κ of 0.94, 0.92, and 0.95. CONCLUSIONS: A novel artificial intelligence multiclass radiomics model including 4 lung lesions to assess disease severity based on the World Health Organization Clinical Progression Scale more accurately determines the severity of COVID-19 patients than a single-class model and radiologists' assessment.

Prediction of Adverse Pathology at Radical Prostatectomy in Grade Group 2 and 3 Prostate Biopsies Using Machine Learning.

PURPOSE: There is ongoing clinical need to improve estimates of disease outcome in prostate cancer. Machine learning (ML) approaches to pathologic diagnosis and prognosis are a promising and increasingly used strategy. In this study, we use an ML algorithm for prediction of adverse outcomes at radical prostatectomy (RP) using whole-slide images (WSIs) of prostate biopsies with Grade Group (GG) 2 or 3 disease. METHODS: We performed a retrospective review of prostate biopsies collected at our institution which had corresponding RP, GG 2 or 3 disease one or more cores, and no biopsies with higher than GG 3 disease. A hematoxylin and eosin-stained core needle biopsy from each site with GG 2 or 3 disease was scanned and used as the sole input for the algorithm. The ML pipeline had three phases: image preprocessing, feature extraction, and adverse outcome prediction. First, patches were extracted from each biopsy scan. Subsequently, the pre-trained Visual Geometry Group-16 convolutional neural network was used for feature extraction. A representative feature vector was then used as input to an Extreme Gradient Boosting classifier for predicting the binary adverse outcome. We subsequently assessed patient clinical risk using CAPRA score for comparison with the ML pipeline results. RESULTS: The data set included 361 WSIs from 107 patients (56 with adverse pathology at RP). The area under the receiver operating characteristic curves for the ML classification were 0.72 (95% CI, 0.62 to 0.81), 0.65 (95% CI, 0.53 to 0.79) and 0.89 (95% CI, 0.79 to 1.00) for the entire cohort, and GG 2 and GG 3 patients, respectively, similar to the performance of the CAPRA clinical risk assessment. CONCLUSION: We provide evidence for the potential of ML algorithms to use WSIs of needle core prostate biopsies to estimate clinically relevant prostate cancer outcomes.

Atlas-based Semantic Segmentation of Prostate Zones.

Segmentation of the prostate into specific anatomical zones is important for radiological assessment of prostate cancer in magnetic resonance imaging (MRI). Of particular interest is segmenting the prostate into two regions of interest: the central gland (CG) and peripheral zone (PZ). In this paper, we propose to integrate an anatomical atlas of prostate zone shape into a deep learning semantic segmentation framework to segment the CG and PZ in T2-weighted MRI. Our approach incorporates anatomical information in the form of a probabilistic prostate zone atlas and utilizes a dynamically controlled hyperparameter to combine the atlas with the semantic segmentation result. In addition to providing significantly improved segmentation performance, this hyperparameter is capable of being dynamically adjusted during the inference stage to provide users with a mechanism to refine the segmentation. We validate our approach using an external test dataset and demonstrate Dice similarity coefficient values (mean±SD) of 0.91±0.05 for the CG and 0.77±0.16 for the PZ that significantly improves upon the baseline segmentation results without the atlas. All code is publicly available on GitHub: https://github.com/OnofreyLab/prostate_atlas_segm_miccai2022.

Cumulative diagnostic imaging radiation exposure in premature neonates.

BACKGROUND: To date, there has been limited work evaluating the total cumulative effective radiation dose received by infants in the neonatal intensive care unit. Most previous publications report that the total radiation dose received falls within the safe limits but does not include all types of ionizing radiation studies typically performed on this vulnerable patient population. We aimed to provide an estimate of the cumulative effective ionizing radiation dose (cED) in microSieverts (µSv) received by premature infants ≤32 weeks from diagnostic studies performed throughout their NICU stay, and predictors of exposures. METHODS: Retrospective chart review from 2004-2011. Data included demographics, gestational age (GA), birth weight (BW), length of stay (LOS), clinical diagnosis, and radiological studies. RESULTS: 1045 charts were reviewed. Median GA = 30.0 weeks (SD 2.7, range 22.0-32.6). Median BW = 1340.0 grams (SD 445.4, range 420-2470). Median number of radiographic studies = 9 (SD 28.5, range 0-210). Median cED = 162µSv (range 0-9248). The cED was positively associated with LOS (p < 0.001) and inversely correlated with GA (p < 0.001) and BW (p < 0.001). Infants with intestinal perforation had the highest median cED 1661µSv compared to 162µSv for others (p < 0.001). CONCLUSION: Our results provide an estimate of the cumulative effective radiation dose received by premature infants in a level 4 neonatal intensive care unit from all radiological studies involving ionizing radiation and identifies risk factors and predictors of such exposure. Radiation exposure in NICU is highest among the most premature and among infants who suffer from intestinal perforation.

BrainGNN: Interpretable Brain Graph Neural Network for fMRI Analysis.

Understanding which brain regions are related to a specific neurological disorder or cognitive stimuli has been an important area of neuroimaging research. We propose BrainGNN, a graph neural network (GNN) framework to analyze functional magnetic resonance images (fMRI) and discover neurological biomarkers. Considering the special property of brain graphs, we design novel ROI-aware graph convolutional (Ra-GConv) layers that leverage the topological and functional information of fMRI. Motivated by the need for transparency in medical image analysis, our BrainGNN contains ROI-selection pooling layers (R-pool) that highlight salient ROIs (nodes in the graph), so that we can infer which ROIs are important for prediction. Furthermore, we propose regularization terms-unit loss, topK pooling (TPK) loss and group-level consistency (GLC) loss-on pooling results to encourage reasonable ROI-selection and provide flexibility to encourage either fully individual- or patterns that agree with group-level data. We apply the BrainGNN framework on two independent fMRI datasets: an Autism Spectrum Disorder (ASD) fMRI dataset and data from the Human Connectome Project (HCP) 900 Subject Release. We investigate different choices of the hyper-parameters and show that BrainGNN outperforms the alternative fMRI image analysis methods in terms of four different evaluation metrics. The obtained community clustering and salient ROI detection results show a high correspondence with the previous neuroimaging-derived evidence of biomarkers for ASD and specific task states decoded for HCP. Our code is available at https://github.com/xxlya/BrainGNN_Pytorch.

The cavernous carotid inferolateral trunk and persistent primitive maxillary arteries: analysis of dissected arterially injected fetal specimens and high-resolution micro-CT of the dog's anastomotic arteries.

PURPOSE: The presence of a persistent primitive maxillary artery is described in the literature dealing with the development of the cavernous carotid inferolateral trunk, and the relevant similarities of the cranial circulation of the human and dog. The literature includes no dissection photographs of the above-mentioned two human fetal arteries, only diagrammatic representations. This study's objectives were to analyze photographs of fetal dissections for the presence of these two arteries, and also investigate the possibility of obtaining, in preserved dog specimens, high-resolution micro-CT imaging of arteries homologous with the above-mentioned two human arteries. METHODS: The literature describing the embryologic development of the cavernous carotid inferolateral trunk, the persistent primitive maxillary arteries, and their homologies in the dog was reviewed. Relevant dissections of fetal specimens were analyzed. High-resolution micro-CT images of un-dissected dog arteries were produced and analyzed. RESULTS: Photographs of fetal specimen dissections demonstrate the cavernous carotid inferolateral trunk. A separate persistent primitive maxillary artery was not present in the dissected specimens. High-resolution micro-CT images of the dog demonstrate homologous arteries with segments of the human inferolateral trunk, and other skull base and brain arteries. CONCLUSION: This investigation provides the only photographs in the literature of dissected human fetal cavernous carotid inferolateral trunks. A persistent primitive maxillary artery was not present in the dissected specimens and is a non-existent structure, likely a previously misidentified carotid inferolateral trunk. High-resolution micro-CT images of the dog visualized arteries that are homologous to segments of the human cavernous carotid inferolateral trunk artery.

Learning-Based Regularization for Cardiac Strain Analysis via Domain Adaptation.

Reliable motion estimation and strain analysis using 3D+ time echocardiography (4DE) for localization and characterization of myocardial injury is valuable for early detection and targeted interventions. However, motion estimation is difficult due to the low-SNR that stems from the inherent image properties of 4DE, and intelligent regularization is critical for producing reliable motion estimates. In this work, we incorporated the notion of domain adaptation into a supervised neural network regularization framework. We first propose a semi-supervised Multi-Layered Perceptron (MLP) network with biomechanical constraints for learning a latent representation that is shown to have more physiologically plausible displacements. We extended this framework to include a supervised loss term on synthetic data and showed the effects of biomechanical constraints on the network's ability for domain adaptation. We validated the semi-supervised regularization method on in vivo data with implanted sonomicrometers. Finally, we showed the ability of our semi-supervised learning regularization approach to identify infarct regions using estimated regional strain maps with good agreement to manually traced infarct regions from postmortem excised hearts.

Use of Artificial Intelligence in Non-Oncologic Interventional Radiology: Current State and Future Directions.

The future of radiology is disproportionately linked to the applications of artificial intelligence (AI). Recent exponential advancements in AI are already beginning to augment the clinical practice of radiology. Driven by a paucity of review articles in the area, this article aims to discuss applications of AI in non-oncologic IR across procedural planning, execution, and follow-up along with a discussion on the future directions of the field. Applications in vascular imaging, radiomics, touchless software interactions, robotics, natural language processing, post-procedural outcome prediction, device navigation, and image acquisition are included. Familiarity with AI study analysis will help open the current 'black box' of AI research and help bridge the gap between the research laboratory and clinical practice.

Automated detection and delineation of hepatocellular carcinoma on multiphasic contrast-enhanced MRI using deep learning.

PURPOSE: Liver Imaging Reporting and Data System (LI-RADS) uses multiphasic contrast-enhanced imaging for hepatocellular carcinoma (HCC) diagnosis. The goal of this feasibility study was to establish a proof-of-principle concept towards automating the application of LI-RADS, using a deep learning algorithm trained to segment the liver and delineate HCCs on MRI automatically. METHODS: In this retrospective single-center study, multiphasic contrast-enhanced MRIs using T1-weighted breath-hold sequences acquired from 2010 to 2018 were used to train a deep convolutional neural network (DCNN) with a U-Net architecture. The U-Net was trained (using 70% of all data), validated (15%) and tested (15%) on 174 patients with 231 lesions. Manual 3D segmentations of the liver and HCC were ground truth. The dice similarity coefficient (DSC) was measured between manual and DCNN methods. Postprocessing using a random forest (RF) classifier employing radiomic features and thresholding (TR) of the mean neural activation was used to reduce the average false positive rate (AFPR). RESULTS: 73 and 75% of HCCs were detected on validation and test sets, respectively, using > 0.2 DSC criterion between individual lesions and their corresponding segmentations. Validation set AFPRs were 2.81, 0.77, 0.85 for U-Net, U-Net + RF, and U-Net + TR, respectively. Combining both RF and TR with the U-Net improved the AFPR to 0.62 and 0.75 for the validation and test sets, respectively. Mean DSC between automatically detected lesions using the DCNN + RF + TR and corresponding manual segmentations was 0.64/0.68 (validation/test), and 0.91/0.91 for liver segmentations. CONCLUSION: Our DCNN approach can segment the liver and HCCs automatically. This could enable a more workflow efficient and clinically realistic implementation of LI-RADS.

Multi-site fMRI analysis using privacy-preserving federated learning and domain adaptation: ABIDE results.

Deep learning models have shown their advantage in many different tasks, including neuroimage analysis. However, to effectively train a high-quality deep learning model, the aggregation of a significant amount of patient information is required. The time and cost for acquisition and annotation in assembling, for example, large fMRI datasets make it difficult to acquire large numbers at a single site. However, due to the need to protect the privacy of patient data, it is hard to assemble a central database from multiple institutions. Federated learning allows for population-level models to be trained without centralizing entities' data by transmitting the global model to local entities, training the model locally, and then averaging the gradients or weights in the global model. However, some studies suggest that private information can be recovered from the model gradients or weights. In this work, we address the problem of multi-site fMRI classification with a privacy-preserving strategy. To solve the problem, we propose a federated learning approach, where a decentralized iterative optimization algorithm is implemented and shared local model weights are altered by a randomization mechanism. Considering the systemic differences of fMRI distributions from different sites, we further propose two domain adaptation methods in this federated learning formulation. We investigate various practical aspects of federated model optimization and compare federated learning with alternative training strategies. Overall, our results demonstrate that it is promising to utilize multi-site data without data sharing to boost neuroimage analysis performance and find reliable disease-related biomarkers. Our proposed pipeline can be generalized to other privacy-sensitive medical data analysis problems. Our code is publicly available at: https://github.com/xxlya/Fed_ABIDE/.

Sparse Data-Driven Learning for Effective and Efficient Biomedical Image Segmentation.

Sparsity is a powerful concept to exploit for high-dimensional machine learning and associated representational and computational efficiency. Sparsity is well suited for medical image segmentation. We present a selection of techniques that incorporate sparsity, including strategies based on dictionary learning and deep learning, that are aimed at medical image segmentation and related quantification.

Regional myocardial strain analysis via 2D speckle tracking echocardiography: validation with sonomicrometry and correlation with regional blood flow in the presence of graded coronary stenoses and dobutamine stress.

BACKGROUND: Quantitative regional strain analysis by speckle tracking echocardiography (STE) may be particularly useful in the assessment of myocardial ischemia and viability, although reliable measurement of regional strain remains challenging, especially in the circumferential and radial directions. We present an acute canine model that integrates a complex sonomicrometer array with microsphere blood flow measurements to evaluate regional myocardial strain and flow in the setting of graded coronary stenoses and dobutamine stress. We apply this unique model to rigorously evaluate a commercial 2D STE software package and explore fundamental regional myocardial flow-function relationships. METHODS: Sonomicrometers (16 crystals) were implanted in epicardial and endocardial pairs across the anterior myocardium of anesthetized open chest dogs (n = 7) to form three adjacent cubes representing the ischemic, border, and remote regions, as defined by their relative locations to a hydraulic occluder on the mid-left anterior descending coronary artery (LAD). Additional cardiac (n = 3) and extra-cardiac (n = 3) reference crystals were placed to define the cardiac axes and aid image registration. 2D short axis echocardiograms, sonometric data, and microsphere blood flow data were acquired at baseline and in the presence of mild and moderate LAD stenoses, both before and during low-dose dobutamine stress (5 µg/kg/min). Regional end-systolic 2D STE radial and circumferential strains were calculated with commercial software (EchoInsight) and compared to those determined by sonomicrometry and to microsphere blood flow measurements. Post-systolic indices (PSIs) were also calculated for radial and circumferential strains. RESULTS: Low-dose dobutamine augmented both strain and flow in the presence of mild and moderate stenoses. Regional 2D STE strains correlated moderately with strains assessed by sonomicrometry (Rradial = 0.56, p < 0.0001; Rcirc = 0.55, p < 0.0001) and with regional flow quantities (Rradial = 0.61, Rcirc = 0.63). Overall, correspondence between 2D STE and sonomicrometry was better in the circumferential direction (Bias ± 1.96 SD: - 1.0 ± 8.2% strain, p = 0.06) than the radial direction (5.7 ± 18.3%, p < 0.0001). Mean PSI values were greatest in low flow conditions and normalized with low-dose dobutamine. CONCLUSIONS: 2D STE identifies changes in regional end-systolic circumferential and radial strain produced by mild and moderate coronary stenoses and low-dose dobutamine stress. Regional 2D STE end-systolic strain measurements correlate modestly with regional sonomicrometer strain and microsphere flow measurements.

Impact of Radiologist-Driven Change-Order Requests on Outpatient CT and MRI Examinations.

PURPOSE: To assess impact of electronic medical record-embedded radiologist-driven change-order request on outpatient CT and MRI examinations. METHODS: Outpatient CT and MRI requests where an order change was requested by the protocoling radiologist in our tertiary care center, from April 11, 2017, to January 3, 2018, were analyzed. Percentage and categorization of requested order change, provider acceptance of requested change, patient and provider demographics, estimated radiation exposure reduction, and cost were analyzed. P < .05 was used for statistical significance. RESULTS: In 79,310 outpatient studies in which radiologists determined protocol, change-order requests were higher for MRI (5.2%, 1,283 of 24,553) compared with CT (2.9%, 1,585 of 54,757; P < .001). Provider approval of requested change was equivalent for CT (82%, 1,299 of 1,585) and MRI (82%, 1,052 of 1,283). Change requests driven by improper contrast media utilization were most common and different between CT (76%, 992 of 1,299) and MRI (65%, 688 of 1,052; P < .001). Changing without and with intravenous contrast orders to with contrast only was most common for CT (39%, 505 of 1,299) and with and without intravenous contrast to without contrast only was most common for MRI (26%, 274 of 1,052; P < .001). Of approved changes in CT, 51% (661 of 1,299) resulted in lower radiation exposure. Approved changes frequently resulted in less costly examinations (CT 67% [799 of 1,198], MRI 48% [411 of 863]). CONCLUSION: Outpatient CT and MRI orders are deemed incorrect in 2.9% to 5% of cases. Radiologist-driven change-order request for CT and MRI are well accepted by ordering providers and decrease radiation exposure associated with imaging.

Scoring of Coronary Artery Disease Characteristics on Coronary CT Angiograms by Using Machine Learning.

Background Coronary CT angiography contains prognostic information but the best method to extract these data remains unknown. Purpose To use machine learning to develop a model of vessel features to discriminate between patients with and without subsequent death or cardiovascular events. Performance was compared with that of conventional scores. Materials and Methods Coronary CT angiography was analyzed by radiologists into four features for each of 16 coronary segments. Four machine learning model types were explored. Five conventional vessel scores were computed for comparison including the Coronary Artery Disease Reporting and Data System (CAD-RADS) score. The National Death Index was retrospectively queried from January 2004 through December 2015. Outcomes were all-cause mortality, coronary heart disease deaths, and coronary deaths or nonfatal myocardial infarctions. Score performance was assessed by using area under the receiver operating characteristic curve (AUC). Results Between February 2004 and November 2009, 6892 patients (4452 men [mean age ± standard deviation, 51 years ± 11] and 2440 women [mean age, 57 years ± 12]) underwent coronary CT angiography (median follow-up, 9.0 years; interquartile range, 8.2-9.8 years). There were 380 deaths of all causes, 70 patients died of coronary artery disease, and 43 patients reported nonfatal myocardial infarctions. For all-cause mortality, the AUC was 0.77 (95% confidence interval: 0.76, 0.77) for machine learning (k-nearest neighbors) versus 0.72 (95% confidence interval: 0.72, 0.72) for CAD-RADS (P < .001). For coronary artery heart disease deaths, AUC was 0.85 (95% confidence interval: 0.84, 0.85) for machine learning versus 0.79 (95% confidence interval: 0.78, 0.80) for CAD-RADS (P < .001). When deciding whether to start statins, if the choice is made to tolerate treating 45 patients to be sure to include one patient who will later die of coronary disease, the use of the machine learning score ensures that 93% of patients with events will be administered the drug; if CAD-RADS is used, only 69% will be treated. Conclusion Compared with Coronary Artery Disease Reporting and Data System and other scores, machine learning methods better discriminated patients who subsequently experienced an adverse event from those who did not. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Schoepf and Tesche in this issue.

Effect of Reader and Patient Parameters on the Performance of Last-Image-Hold for Fluoroscopic Grading of Vesicoureteric Reflux.

OBJECTIVE. The purpose of this study is to determine the effect of different reader and patient parameters on the degree of agreement and the rate of misclassification of vesicoureteric reflux grading on last-image-hold frames in relation to spot-exposed frames from voiding cystourethrography (VCUG) as well as to determine the nature of reflux misclassification on last-image-hold frames. MATERIALS AND METHODS. Blinded readers conducted a retrospective evaluation of last-image-hold and spot-exposed frames of the renal fossae from 191 sequential VCUG examinations performed during a five-year period. Kappa tests were used to determine the agreement between reflux gradings and to assess the impact of reader and patient parameters. Pearson product-moment correlations were used to evaluate the effect of patient parameters on reader level of certainty regarding reflux grading. RESULTS. We measured almost perfect overall agreement for more experienced readers and substantial overall agreement for less experienced readers. Point estimates of overall misclassification were less than 2% for more experienced readers and less than 4% for less experienced readers. The readers' level of certainty about reflux grading had a positive impact on agreement values and misclassification rates. Experienced readers' most common misclassification was assigning reflux a grade of 3 on a spot-exposed frame and a grade of 2 on an equivalent last-image-hold frame. Inexperienced readers' most common misclassification involved missing reflux altogether. CONCLUSION. Instances of grade 2 reflux on last-image-hold frames may warrant supplemental evaluation with spot-exposed frames. Otherwise, a reader's level of certainty regarding reflux grading on a last-image-hold frame may help determine whether a supplemental spot-exposed frame would be beneficial.