2023 Survey on User Experience of Artificial Intelligence Software in Radiology by the Korean Society of Radiology.

OBJECTIVE: In Korea, radiology has been positioned towards the early adoption of artificial intelligence-based software as medical devices (AI-SaMDs); however, little is known about the current usage, implementation, and future needs of AI-SaMDs. We surveyed the current trends and expectations for AI-SaMDs among members of the Korean Society of Radiology (KSR). MATERIALS AND METHODS: An anonymous and voluntary online survey was open to all KSR members between April 17 and May 15, 2023. The survey was focused on the experiences of using AI-SaMDs, patterns of usage, levels of satisfaction, and expectations regarding the use of AI-SaMDs, including the roles of the industry, government, and KSR regarding the clinical use of AI-SaMDs. RESULTS: Among the 370 respondents (response rate: 7.7% [370/4792]; 340 board-certified radiologists; 210 from academic institutions), 60.3% (223/370) had experience using AI-SaMDs. The two most common use-case of AI-SaMDs among the respondents were lesion detection (82.1%, 183/223), lesion diagnosis/classification (55.2%, 123/223), with the target imaging modalities being plain radiography (62.3%, 139/223), CT (42.6%, 95/223), mammography (29.1%, 65/223), and MRI (28.7%, 64/223). Most users were satisfied with AI-SaMDs (67.6% [115/170, for improvement of patient management] to 85.1% [189/222, for performance]). Regarding the expansion of clinical applications, most respondents expressed a preference for AI-SaMDs to assist in detection/diagnosis (77.0%, 285/370) and to perform automated measurement/quantification (63.5%, 235/370). Most respondents indicated that future development of AI-SaMDs should focus on improving practice efficiency (81.9%, 303/370) and quality (71.4%, 264/370). Overall, 91.9% of the respondents (340/370) agreed that there is a need for education or guidelines driven by the KSR regarding the use of AI-SaMDs. CONCLUSION: The penetration rate of AI-SaMDs in clinical practice and the corresponding satisfaction levels were high among members of the KSR. Most AI-SaMDs have been used for lesion detection, diagnosis, and classification. Most respondents requested KSR-driven education or guidelines on the use of AI-SaMDs.

Loss of LPAR6 and CAB39L dysregulates the basal-to-luminal urothelial differentiation program, contributing to bladder carcinogenesis.

We describe a strategy that combines histologic and molecular mapping that permits interrogation of the chronology of changes associated with cancer development on a whole-organ scale. Using this approach, we present the sequence of alterations around RB1 in the development of bladder cancer. We show that RB1 is not involved in initial expansion of the preneoplastic clone. Instead, we found a set of contiguous genes that we term "forerunner" genes whose silencing is associated with the development of plaque-like field effects initiating carcinogenesis. Specifically, we identified five candidate forerunner genes (ITM2B, LPAR6, MLNR, CAB39L, and ARL11) mapping near RB1. Two of these genes, LPAR6 and CAB39L, are preferentially downregulated in the luminal and basal subtypes of bladder cancer, respectively. Their loss of function dysregulates urothelial differentiation, sensitizing the urothelium to N-butyl-N-(4-hydroxybutyl)nitrosamine-induced cancers, which recapitulate the luminal and basal subtypes of human bladder cancer.

ZOZI-Seg: A transformer and UNet cascade network with Zoom-Out and Zoom-In scheme for aortic dissection segmentation in enhanced CT images.

BACKGROUND & OBJECTIVE: Aortic dissection (AD) is a serious condition requiring rapid and accurate diagnosis. In this study, we aimed to improve the diagnostic accuracy of AD by presenting a novel method for aortic segmentation in computed tomography images that uses a combination of a transformer and a UNet cascade network with a Zoom-Out and Zoom-In scheme (ZOZI-seg). METHODS: The proposed method segments each compartment of the aorta, comprising the true lumen (TL), false lumen (FL), and thrombosis (TH) using a cascade strategy that captures both the global context (anatomical structure) and the local detail texture based on the dynamic patch size with ZOZI schemes. The ZOZI-seg model has a two-stage architecture using both a "3D transformer for panoptic context-awareness" and a "3D UNet for localized texture refinement." The unique ZOZI strategies for patching were demonstrated in an ablation study. The performance of our proposed ZOZI-seg model was tested using a dataset from Asan Medical Center and compared with those of existing models such as nnUNet and nnFormer. RESULTS: In terms of segmentation accuracy, our method yielded better results, with Dice similarity coefficients (DSCs) of 0.917, 0.882, and 0.630 for TL, FL, and TH, respectively. Furthermore, we indirectly compared our model with those in previous studies using an external dataset to evaluate its robustness and generalizability. CONCLUSIONS: This approach may help in the diagnosis and treatment of AD in different clinical situations and provide a strong basis for further research and clinical applications.

Deep Learning-Based Lumen and Vessel Segmentation of Intravascular Ultrasound Images in Coronary Artery Disease.

BACKGROUND AND OBJECTIVES: Intravascular ultrasound (IVUS) evaluation of coronary artery morphology is based on the lumen and vessel segmentation. This study aimed to develop an automatic segmentation algorithm and validate the performances for measuring quantitative IVUS parameters. METHODS: A total of 1,063 patients were randomly assigned, with a ratio of 4:1 to the training and test sets. The independent data set of 111 IVUS pullbacks was obtained to assess the vessel-level performance. The lumen and external elastic membrane (EEM) boundaries were labeled manually in every IVUS frame with a 0.2-mm interval. The Efficient-UNet was utilized for the automatic segmentation of IVUS images. RESULTS: At the frame-level, Efficient-UNet showed a high dice similarity coefficient (DSC, 0.93±0.05) and Jaccard index (JI, 0.87±0.08) for lumen segmentation, and demonstrated a high DSC (0.97±0.03) and JI (0.94±0.04) for EEM segmentation. At the vessel-level, there were close correlations between model-derived vs. experts-measured IVUS parameters; minimal lumen image area (r=0.92), EEM area (r=0.88), lumen volume (r=0.99) and plaque volume (r=0.95). The agreement between model-derived vs. expert-measured minimal lumen area was similarly excellent compared to the experts' agreement. The model-based lumen and EEM segmentation for a 20-mm lesion segment required 13.2 seconds, whereas manual segmentation with a 0.2-mm interval by an expert took 187.5 minutes on average. CONCLUSIONS: The deep learning models can accurately and quickly delineate vascular geometry. The artificial intelligence-based methodology may support clinicians' decision-making by real-time application in the catheterization laboratory.

Radiologist's Guide to Evaluating Publications of Clinical Research on AI: How We Do It.

Literacy in research studies of artificial intelligence (AI) has become an important skill for radiologists. It is required to make a proper assessment of the validity, reproducibility, and clinical applicability of AI studies. However, AI studies are generally perceived to be more difficult for clinician readers to evaluate than traditional clinical research studies. This special report-as an effective, concise guide for readers-aims to assist clinical radiologists in critically evaluating different types of clinical research articles involving AI. It does not intend to be a comprehensive checklist or methodological summary for complete clinical evaluation of AI or a reporting guideline. Ten key items for readers to check are described, regarding study purpose, function and clinical context of AI, training data, data preprocessing, AI modeling techniques, test data, AI performance, helpfulness and value of AI, interpretability of AI, and code sharing. The important aspects of each item are explained for readers to consider when reading publications on AI clinical research. Evaluating each item can help radiologists assess the validity, reproducibility, and clinical applicability of clinical research articles involving AI.

Fully automatic coronary calcium scoring in non-ECG-gated low-dose chest CT: comparison with ECG-gated cardiac CT.

OBJECTIVES: To validate an artificial intelligence (AI)-based fully automatic coronary artery calcium (CAC) scoring system on non-electrocardiogram (ECG)-gated low-dose chest computed tomography (LDCT) using multi-institutional datasets with manual CAC scoring as the reference standard. METHODS: This retrospective study included 452 subjects from three academic institutions, who underwent both ECG-gated calcium scoring computed tomography (CSCT) and LDCT scans. For all CSCT and LDCT scans, automatic CAC scoring (CAC_auto) was performed using AI-based software, and manual CAC scoring (CAC_man) was set as the reference standard. The reliability and agreement of CAC_auto was evaluated and compared with that of CAC_man using intraclass correlation coefficients (ICCs) and Bland-Altman plots. The reliability between CAC_auto and CAC_man for CAC severity categories was analyzed using weighted kappa (κ) statistics. RESULTS: CAC_auto on CSCT and LDCT yielded a high ICC (0.998, 95% confidence interval (CI) 0.998-0.999 and 0.989, 95% CI 0.987-0.991, respectively) and a mean difference with 95% limits of agreement of 1.3 ± 37.1 and 0.8 ± 75.7, respectively. CAC_auto achieved excellent reliability for CAC severity (κ = 0.918-0.972) on CSCT and good to excellent but heterogenous reliability among datasets (κ = 0.748-0.924) on LDCT. CONCLUSIONS: The application of an AI-based automatic CAC scoring software to LDCT shows good to excellent reliability in CAC score and CAC severity categorization in multi-institutional datasets; however, the reliability varies among institutions. KEY POINTS: • AI-based automatic CAC scoring on LDCT shows excellent reliability with manual CAC scoring in multi-institutional datasets. • The reliability for CAC score-based severity categorization varies among datasets. • Automatic scoring for LDCT shows a higher false-positive rate than automatic scoring for CSCT, and most common causes of a false-positive are image noise and artifacts for both CSCT and LDCT.

Methods for Clinical Evaluation of Artificial Intelligence Algorithms for Medical Diagnosis.

Adequate clinical evaluation of artificial intelligence (AI) algorithms before adoption in practice is critical. Clinical evaluation aims to confirm acceptable AI performance through adequate external testing and confirm the benefits of AI-assisted care compared with conventional care through appropriately designed and conducted studies, for which prospective studies are desirable. This article explains some of the fundamental methodological points that should be considered when designing and appraising the clinical evaluation of AI algorithms for medical diagnosis. The specific topics addressed include the following: (a) the importance of external testing of AI algorithms and strategies for conducting the external testing effectively, (b) the various metrics and graphical methods for evaluating the AI performance as well as essential methodological points to note in using and interpreting them, (c) paired study designs primarily for comparative performance evaluation of conventional and AI-assisted diagnoses, (d) parallel study designs primarily for evaluating the effect of AI intervention with an emphasis on randomized clinical trials, and (e) up-to-date guidelines for reporting clinical studies on AI, with an emphasis on guidelines registered in the EQUATOR Network library. Sound methodological knowledge of these topics will aid the design, execution, reporting, and appraisal of clinical evaluation of AI.

A deep learning-based automatic analysis of cardiovascular borders on chest radiographs of valvular heart disease: development/external validation.

OBJECTIVES: Cardiovascular border (CB) analysis is the primary method for detecting and quantifying the severity of cardiovascular disease using posterior-anterior chest radiographs (CXRs). This study aimed to develop and validate a deep learning-based automatic CXR CB analysis algorithm (CB_auto) for diagnosing and quantitatively evaluating valvular heart disease (VHD). METHODS: We developed CB_auto using 816 normal and 798 VHD CXRs. For validation, 640 normal and 542 VHD CXRs from three different hospitals and 132 CXRs from a public dataset were assigned. The reliability of the CB parameters determined by CB_auto was evaluated. To evaluate the differences between parameters determined by CB_auto and manual CB drawing (CB_hand), the absolute percentage measurement error (APE) was calculated. Pearson correlation coefficients were calculated between CB_hand and echocardiographic measurements. RESULTS: CB parameters determined by CB_auto yielded excellent reliability (intraclass correlation coefficient > 0.98). The 95% limits of agreement for the cardiothoracic ratio were 0.00 ± 0.04% without systemic bias. The differences between parameters determined by CB_auto and CB_hand as defined by the APE were < 10% for all parameters except for carinal angle and left atrial appendage. In the public dataset, all CB parameters were successfully drawn in 124 of 132 CXRs (93.9%). All CB parameters were significantly greater in VHD than in normal controls (all p < 0.05). All CB parameters showed significant correlations (p < 0.05) with echocardiographic measurements. CONCLUSIONS: The CB_auto system empowered by deep learning algorithm provided highly reliable CB measurements that could be useful not only in daily clinical practice but also for research purposes. KEY POINTS: • A deep learning-based automatic CB analysis algorithm for diagnosing and quantitatively evaluating VHD using posterior-anterior chest radiographs was developed and validated. • Our algorithm (CB_auto) yielded comparable reliability to manual CB drawing (CB_hand) in terms of various CB measurement variables, as confirmed by external validation with datasets from three different hospitals and a public dataset. • All CB parameters were significantly different between VHD and normal control measurements, and echocardiographic measurements were significantly correlated with CB parameters measured from normal control and VHD CXRs.

Fully Automatic Coronary Calcium Score Software Empowered by Artificial Intelligence Technology: Validation Study Using Three CT Cohorts.

OBJECTIVE: This study aimed to validate a deep learning-based fully automatic calcium scoring (coronary artery calcium [CAC]_auto) system using previously published cardiac computed tomography (CT) cohort data with the manually segmented coronary calcium scoring (CAC_hand) system as the reference standard. MATERIALS AND METHODS: We developed the CAC_auto system using 100 co-registered, non-enhanced and contrast-enhanced CT scans. For the validation of the CAC_auto system, three previously published CT cohorts (n = 2985) were chosen to represent different clinical scenarios (i.e., 2647 asymptomatic, 220 symptomatic, 118 valve disease) and four CT models. The performance of the CAC_auto system in detecting coronary calcium was determined. The reliability of the system in measuring the Agatston score as compared with CAC_hand was also evaluated per vessel and per patient using intraclass correlation coefficients (ICCs) and Bland-Altman analysis. The agreement between CAC_auto and CAC_hand based on the cardiovascular risk stratification categories (Agatston score: 0, 1-10, 11-100, 101-400, > 400) was evaluated. RESULTS: In 2985 patients, 6218 coronary calcium lesions were identified using CAC_hand. The per-lesion sensitivity and false-positive rate of the CAC_auto system in detecting coronary calcium were 93.3% (5800 of 6218) and 0.11 false-positive lesions per patient, respectively. The CAC_auto system, in measuring the Agatston score, yielded ICCs of 0.99 for all the vessels (left main 0.91, left anterior descending 0.99, left circumflex 0.96, right coronary 0.99). The limits of agreement between CAC_auto and CAC_hand were 1.6 ± 52.2. The linearly weighted kappa value for the Agatston score categorization was 0.94. The main causes of false-positive results were image noise (29.1%, 97/333 lesions), aortic wall calcification (25.5%, 85/333 lesions), and pericardial calcification (24.3%, 81/333 lesions). CONCLUSION: The atlas-based CAC_auto empowered by deep learning provided accurate calcium score measurement as compared with manual method and risk category classification, which could potentially streamline CAC imaging workflows.

A Portable Smartphone-Based Laryngoscope System for High-Speed Vocal Cord Imaging of Patients With Throat Disorders: Instrument Validation Study.

BACKGROUND: Currently, high-speed digital imaging (HSDI), especially endoscopic HSDI, is routinely used for the diagnosis of vocal cord disorders. However, endoscopic HSDI devices are usually large and costly, which limits access to patients in underdeveloped countries and in regions with inadequate medical infrastructure. Modern smartphones have sufficient functionality to process the complex calculations that are required for processing high-resolution images and videos with a high frame rate. Recently, several attempts have been made to integrate medical endoscopes with smartphones to make them more accessible to people in underdeveloped countries. OBJECTIVE: This study aims to develop a smartphone adaptor for endoscopes, which enables smartphone-based vocal cord imaging, to demonstrate the feasibility of performing high-speed vocal cord imaging via the high-speed imaging functions of a high-performance smartphone camera, and to determine the acceptability of the smartphone-based high-speed vocal cord imaging system for clinical applications in developing countries. METHODS: A customized smartphone adaptor optical relay was designed for clinical endoscopy using selective laser melting-based 3D printing. A standard laryngoscope was attached to the smartphone adaptor to acquire high-speed vocal cord endoscopic images. Only existing basic functions of the smartphone camera were used for HSDI of the vocal cords. Extracted still frames were observed for qualitative glottal volume and shape. For image processing, segmented glottal and vocal cord areas were calculated from whole HSDI frames to characterize the amplitude of the vibrations on each side of the glottis, including the frequency, edge length, glottal areas, base cord, and lateral phase differences over the acquisition time. The device was incorporated into a preclinical videokymography diagnosis routine to compare functionality. RESULTS: Smartphone-based HSDI with the smartphone-endoscope adaptor could achieve 940 frames per second and a resolution of 1280 by 720 frames, which corresponds to the detection of 3 to 8 frames per vocal cycle at double the spatial resolution of existing devices. The device was used to image the vocal cords of 4 volunteers: 1 healthy individual and 3 patients with vocal cord paralysis, chronic laryngitis, or vocal cord polyps. The resultant image stacks were sufficient for most diagnostic purposes. The cost of the device including the smartphone was lower than that of existing HSDI devices. The image processing and analytics demonstrated the successful calculation of relevant diagnostic variables from the acquired images. Patients with vocal pathologies were easily differentiable in the quantitative data. CONCLUSIONS: A smartphone-based HSDI endoscope system can function as a point-of-care clinical diagnostic device. The resulting analysis is of higher quality than that accessible by videostroboscopy and promises comparable quality and greater accessibility than HSDI. In particular, this system is suitable for use as an accessible diagnostic tool in underdeveloped areas with inadequate medical service infrastructure.

Impact of coronary calcium score and lesion characteristics on the diagnostic performance of machine-learning-based computed tomography-derived fractional flow reserve.

AIMS: To evaluate the impact of coronary artery calcium (CAC) score, minimal lumen area (MLA), and length of coronary artery stenosis on the diagnostic performance of the machine-learning-based computed tomography-derived fractional flow reserve (ML-FFR). METHODS AND RESULTS: In 471 patients with coronary artery disease, computed tomography angiography (CTA) and invasive coronary angiography were performed with fractional flow reserve (FFR) in 557 lesions at a single centre. Diagnostic performances of ML-FFR, computational fluid dynamics-based CT-FFR (CFD-FFR), MLA, quantitative coronary angiography (QCA), and visual stenosis grading were evaluated using invasive FFR as a reference standard. Diagnostic performances were analysed according to lesion characteristics including the MLA, length of stenosis, CAC score, and stenosis degree. ML-FFR was obtained by automated feature selection and model building from quantitative CTA. A total of 272 lesions showed significant ischaemia, defined by invasive FFR ≤0.80. There was a significant correlation between CFD-FFR and ML-FFR (r = 0.99, P < 0.001). ML-FFR showed moderate sensitivity and specificity in the per-patient analysis. Diagnostic performances of CFD-FFR and ML-FFR did not decline in patients with high CAC scores (CAC > 400). Sensitivities of CFD-FFR and ML-FFR showed a downward trend along with the increase in lesion length and decrease in MLA. The area under the curve (AUC) of ML-FFR (0.73) was higher than those of QCA and visual grading (AUC = 0.65 for both, P < 0.001) and comparable to those of MLA (AUC = 0.71, P = 0.21) and CFD-FFR (AUC = 0.73, P = 0.86). CONCLUSION: ML-FFR showed comparable results to MLA and CFD-FFR for the prediction of lesion-specific ischaemia. Specificities and accuracies of CFD-FFR and ML-FFR decreased with smaller MLA and long lesion length.

Prediction of Coronary Stent Underexpansion by Pre-Procedural Intravascular Ultrasound-Based Deep Learning.

OBJECTIVES: The aim of this study was to develop pre-procedural intravascular ultrasound (IVUS)-based models for predicting the occurrence of stent underexpansion. BACKGROUND: Although post-stenting IVUS has been used to optimize percutaneous coronary intervention, there are no pre-procedural guidelines to estimate the degree of stent expansion and provide preemptive management before stent deployment. METHODS: A total of 618 coronary lesions in 618 patients undergoing percutaneous coronary intervention were randomized into training and test sets in a 5:1 ratio. Following the coregistration of pre- and post-stenting IVUS images, the pre-procedural images and clinical information (stent diameter, length, and inflation pressure; balloon diameter; and maximal balloon pressure) were used to develop a regression model using a convolutional neural network to predict post-stenting stent area. To separate the frames with from those without the occurrence of underexpansion (stent area <5.5 mm2), binary classification models (XGBoost) were developed. RESULTS: Overall, the frequency of stent underexpansion was 15% (5,209 of 34,736 frames). At the frame level, stent areas predicted by the pre-procedural IVUS-based regression model significantly correlated with those measured on post-stenting IVUS (r = 0.802). To predict stent underexpansion, maximal accuracy of 94% (area under the curve = 0.94) was achieved when the convolutional neural network- and mask image-derived features were used for the classification model. At the lesion level, there were significant correlations between predicted and measured minimal stent area (r = 0.832) and between predicted and measured total stent volume (r = 0.958). CONCLUSIONS: Deep-learning algorithms accurately predicted incomplete stent expansion. A data-driven approach may assist clinicians in making treatment decisions to avoid stent underexpansion as a preventable cause of stent failure.

CycleMorph: Cycle consistent unsupervised deformable image registration.

Image registration is a fundamental task in medical image analysis. Recently, many deep learning based image registration methods have been extensively investigated due to their comparable performance with the state-of-the-art classical approaches despite the ultra-fast computational time. However, the existing deep learning methods still have limitations in the preservation of original topology during the deformation with registration vector fields. To address this issues, here we present a cycle-consistent deformable image registration, dubbed CycleMorph. The cycle consistency enhances image registration performance by providing an implicit regularization to preserve topology during the deformation. The proposed method is so flexible that it can be applied for both 2D and 3D registration problems for various applications, and can be easily extended to multi-scale implementation to deal with the memory issues in large volume registration. Experimental results on various datasets from medical and non-medical applications demonstrate that the proposed method provides effective and accurate registration on diverse image pairs within a few seconds. Qualitative and quantitative evaluations on deformation fields also verify the effectiveness of the cycle consistency of the proposed method.

Intravascular ultrasound-based deep learning for plaque characterization in coronary artery disease.

BACKGROUND AND AIMS: Although plaque characterization by intravascular ultrasound (IVUS) is important for risk stratification, frame-by-frame analysis of a whole vascular segment is time-consuming. The aim was to develop IVUS-based algorithms for classifying attenuation and calcified plaques. METHODS: IVUS image sets of 598 coronary arteries from 598 patients were randomized into training and test sets with 5:1 ratio. Each IVUS frame at a 0.4-mm interval was circumferentially labeled as one of three classes: attenuated plaque, calcified plaque, or plaque without attenuation or calcification. The model was trained on multi-class classification with 5-fold cross validation. By converting from Cartesian to polar coordinate images, the class corresponding to each array from 0 to 360° was plotted. RESULTS: At the angle-level, Dice similarity coefficients for identifying calcification vs. attenuation vs. none by using ensemble model were 0.79, 0.74 and 0.99, respectively. Also, the maximal accuracy was 98% to classify those groups in the test set. At the frame-level, the model identified the presence of attenuation with 80% sensitivity, 96% specificity, and 93% overall accuracy, and the presence of calcium with 86% sensitivity, 97% specificity, and 96% overall accuracy. In the per-vessel analysis, the attenuation and calcification burden index closely correlated with human measurements (r = 0.89 and r = 0.95, respectively), as did the maximal attenuation and calcification burden index over 4 mm (r = 0.82 and r = 0.91, respectively). The inference times were 0.05 s per frame and 7.8 s per vessel. CONCLUSIONS: Our deep learning algorithms for plaque characterization may assist clinicians in recognizing high-risk coronary lesions.

Deep learning-based algorithm to detect primary hepatic malignancy in multiphase CT of patients at high risk for HCC.

OBJECTIVES: To develop and evaluate a deep learning-based model capable of detecting primary hepatic malignancies in multiphase CT images of patients at high risk for hepatocellular carcinoma (HCC). METHODS: A total of 1350 multiphase CT scans of 1280 hepatic malignancies (1202 HCCs and 78 non-HCCs) in 1320 patients at high risk for HCC were retrospectively analyzed. Following the delineation of the focal hepatic lesions according to reference standards, the CT scans were categorized randomly into the training (568 scans), tuning (193 scans), and test (589 scans) sets. Multiphase CT information was subjected to multichannel integration, and livers were automatically segmented before model development. A deep learning-based model capable of detecting malignancies was developed using a mask region-based convolutional neural network. The thresholds of the prediction score and the intersection over union were determined on the tuning set corresponding to the highest sensitivity with < 5 false-positive cases per CT scan. The sensitivity and the number of false-positives of the proposed model on the test set were calculated. Potential causes of false-negatives and false-positives on the test set were analyzed. RESULTS: This model exhibited a sensitivity of 84.8% with 4.80 false-positives per CT scan on the test set. The most frequent potential causes of false-negatives and false-positives were determined to be atypical enhancement patterns for HCC (71.7%) and registration/segmentation errors (42.7%), respectively. CONCLUSIONS: The proposed deep learning-based model developed to automatically detect primary hepatic malignancies exhibited an 84.8% of sensitivity with 4.80 false-positives per CT scan in the test set. KEY POINTS: • Image processing, including multichannel integration of multiphase CT and automatic liver segmentation, enabled the application of a deep learning-based model to detect primary hepatic malignancy. • Our model exhibited a sensitivity of 84.8% with a false-positive rate of 4.80 per CT scan.

Machine Learning-Based Automatic Rating for Cardinal Symptoms of Parkinson Disease.

OBJECTIVE: We developed and investigated the feasibility of a machine learning-based automated rating for the 2 cardinal symptoms of Parkinson disease (PD): resting tremor and bradykinesia. METHODS: Using OpenPose, a deep learning-based human pose estimation program, we analyzed video clips for resting tremor and finger tapping of the bilateral upper limbs of 55 patients with PD (110 arms). Key motion parameters, including resting tremor amplitude and finger tapping speed, amplitude, and fatigue, were extracted to develop a machine learning-based automatic Unified Parkinson's Disease Rating Scale (UPDRS) rating using support vector machine (SVM) method. To evaluate the performance of this model, we calculated weighted κ and intraclass correlation coefficients (ICCs) between the model and the gold standard rating by a movement disorder specialist who is trained and certified by the Movement Disorder Society for UPDRS rating. These values were compared to weighted κ and ICC between a nontrained human rater and the gold standard rating. RESULTS: For resting tremors, the SVM model showed a very good to excellent reliability range with the gold standard rating (κ 0.791; ICC 0.927), with both values higher than that of nontrained human rater (κ 0.662; ICC 0.861). For finger tapping, the SVM model showed a very good reliability range with the gold standard rating (κ 0.700 and ICC 0.793), which was comparable to that for nontrained human raters (κ 0.627; ICC 0.797). CONCLUSION: Machine learning-based algorithms that automatically rate PD cardinal symptoms are feasible, with more accurate results than nontrained human ratings. CLASSIFICATION OF EVIDENCE: This study provides Class II evidence that machine learning-based automated rating of resting tremor and bradykinesia in people with PD has very good reliability compared to a rating by a movement disorder specialist.

Volumetric assessment of extrusion in medial meniscus posterior root tears through semi-automatic segmentation on 3-tesla magnetic resonance images.

BACKGROUND: Many reports have described the relationship between medial meniscus posterior root tears (MMPRTs) and meniscal extrusion on coronal magnetic resonance (MR) images. However, volumetric assessment of meniscal extrusion has not been performed, and the correlation between extrusion length and volume remains unclear. HYPOTHESIS: Extrusion in both length and volume would be greater in MMPRTs than that in the normal medial meniscus, and the extrusion length measured on coronal MR images would be correlated with the extrusion volume. PATIENTS AND METHODS: A total of 20 knees who underwent isolated MMPRTs without trauma history were included in the MMPRT group, and another 20 knees with normal medial meniscus were selected as the control group. All 40 knees underwent 3-tesla MR imaging. The extrusion length of the medial meniscus was measured using coronal MR images only. Volumetric assessments of the meniscus were performed and analyzed via a semi-automatic segmentation. Group-wise comparisons of the extrusion length and volumetric values were conducted, and the correlation between the two measures in both groups was evaluated. RESULTS: The mean extrusion length of the medial meniscus in the MMPRT group was significantly longer (2.60 vs. 0.63mm; p<0.001) than that in the control group. The mean extrusion volume was also significantly higher in the MMPRT than that in the control group (770.93 vs. 193.80 mm3; p<0.001). The extrusion length was significantly and positively correlated with the extrusion volume in both groups (R=0.64; p=0.002 in MMPRT, R=0.73; p<0.001 in the control group). DISCUSSION: Semi-automatic segmentation was used to measure the volume of meniscal extrusion, which had previously only been estimated indirectly with the extrusion length on coronal MR images. MMPRTs significantly increased the extrusion in both measures. The extrusion length measured on coronal MR images was positively correlated with the extrusion volume in both groups. LEVEL OF EVIDENCE: III, Case-control study.

Urothelial-to-Neural Plasticity Drives Progression to Small Cell Bladder Cancer.

We report a comprehensive molecular analysis of 34 cases of small cell carcinoma (SCC) and 84 cases of conventional urothelial carcinoma (UC), with The Cancer Genome Atlas cohort of 408 conventional UC bladder cancers used as the reference. SCCs showed mutational landscapes characterized by nearly uniform inactivation of TP53 and were dominated by Sanger mutation signature 3 associated with loss of BRCA1/2 function. SCCs were characterized by downregulation of luminal and basal markers and were referred to as double-negative. Transcriptome analyses indicated that SCCs displayed lineage plasticity driven by a urothelial-to-neural phenotypic switch with a dysregulated epithelial-to-mesenchymal transition network. SCCs were depleted of immune cells, and expressed high levels of the immune checkpoint receptor, adenosine receptor A2A (ADORA2A), which is a potent inhibitor of immune infiltration. Our observations have important implications for the prognostication and development of more effective therapies for this lethal bladder cancer variant.

Assessment of Luminal and Basal Phenotypes in Bladder Cancer.

Genomic profiling studies have demonstrated that bladder cancer can be divided into two molecular subtypes referred to as luminal and basal with distinct clinical behaviors and sensitivities to frontline chemotherapy. We analyzed the mRNA expressions of signature luminal and basal genes in bladder cancer tumor samples from publicly available and MD Anderson Cancer Center cohorts. We developed a quantitative classifier referred to as basal to luminal transition (BLT) score which identified the molecular subtypes of bladder cancer with 80-94% sensitivity and 83-93% specificity. In order to facilitate molecular subtyping of bladder cancer in primary care centers, we analyzed the protein expressions of signature luminal (GATA3) and basal (KRT5/6) markers by immunohistochemistry, which identified molecular subtypes in over 80% of the cases. In conclusion, we provide a tool for assessment of molecular subtypes of bladder cancer in routine clinical practice.

Automated Segmentation of Left Ventricular Myocardium on Cardiac Computed Tomography Using Deep Learning.

OBJECTIVE: To evaluate the accuracy of a deep learning-based automated segmentation of the left ventricle (LV) myocardium using cardiac CT. MATERIALS AND METHODS: To develop a fully automated algorithm, 100 subjects with coronary artery disease were randomly selected as a development set (50 training / 20 validation / 30 internal test). An experienced cardiac radiologist generated the manual segmentation of the development set. The trained model was evaluated using 1000 validation set generated by an experienced technician. Visual assessment was performed to compare the manual and automatic segmentations. In a quantitative analysis, sensitivity and specificity were calculated according to the number of pixels where two three-dimensional masks of the manual and deep learning segmentations overlapped. Similarity indices, such as the Dice similarity coefficient (DSC), were used to evaluate the margin of each segmented masks. RESULTS: The sensitivity and specificity of automated segmentation for each segment (1-16 segments) were high (85.5-100.0%). The DSC was 88.3 ± 6.2%. Among randomly selected 100 cases, all manual segmentation and deep learning masks for visual analysis were classified as very accurate to mostly accurate and there were no inaccurate cases (manual vs. deep learning: very accurate, 31 vs. 53; accurate, 64 vs. 39; mostly accurate, 15 vs. 8). The number of very accurate cases for deep learning masks was greater than that for manually segmented masks. CONCLUSION: We present deep learning-based automatic segmentation of the LV myocardium and the results are comparable to manual segmentation data with high sensitivity, specificity, and high similarity scores.