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.

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.

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.

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.

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.

Volume Doubling Times of Lung Adenocarcinomas: Correlation with Predominant Histologic Subtypes and Prognosis.

Background The volume doubling time (VDT) is a key parameter in the differentiation of aggressive tumors from slow-growing tumors. How different histologic subtypes of primary lung adenocarcinomas vary in their VDT and the prognostic value of this measurement is unknown. Purpose To investigate differences in VDT between the predominant histologic subtypes of primary lung adenocarcinomas and to assess the correlation between VDT and prognosis. Materials and Methods This retrospective study included patients who underwent at least two serial CT examinations before undergoing operation between July 2010 and December 2018. Three-dimensional tumor segmentation was performed on two CT images and VDTs were calculated. VDTs were compared between predominant histologic subtypes and lesion types by using Kruskal-Wallis tests. Disease-free survival (DFS) was obtained in patients undergoing surgical procedures before July 2017. Univariable and multivariable Cox proportional hazards regression analyses were performed to determine predictors of DFS. Results Among 268 patients (mean age, 64 years ± 8 [standard deviation]; 143 men), there were 30 lepidic, 87 acinar, 109 papillary, and 42 solid or micropapillary predominant subtypes. The median VDT was 529 days (interquartile range, 278-872 days) for lung adenocarcinomas. VDTs differed across subtypes (P < .001) and were shortest in solid or micropapillary subtypes (229 days; interquartile range, 77-530 days). Solid lesions (VDT, 248 days) had shorter VDTs than subsolid lesions (part-solid lesions, 665 days; nonsolid lesions, 648 days) (P < .001). In the 148 patients (mean age, 64 years ± 8; 89 men) included in the survival analysis, 35 patients had disease recurrence and 17 patients died. VDT (<400 days) was an independent risk factor for poor DFS (hazard ratio, 2.6; P = .01) and higher TNM stage. Adding VDT to TNM stage improved model performance (C-index, 0.69 for TNM stage vs 0.77 for combined VDT class and TNM stage; P = .002). Conclusion Volume doubling times varied significantly according to the predominant histologic subtypes of lung adenocarcinoma and had additional prognostic value for disease-free survival. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Ko in this issue.

Outcome prediction in resectable lung adenocarcinoma patients: value of CT radiomics.

OBJECTIVES: Lung adenocarcinoma shows broad spectrum of prognosis and histologic heterogeneity. This study was to investigate the prognostic value of CT radiomics in resectable lung adenocarcinoma patients and assess its incremental value over clinical-pathologic risk factors. METHODS: This retrospective analysis evaluated 1058 patients who underwent curative surgery for lung adenocarcinoma (training cohort: N = 754; temporal validation cohort: N = 304). Radiomics features were extracted from preoperative contrast-enhanced CT. Radiomics signature to predict disease-free survival (DFS) and overall survival (OS) was generated. Association between the radiomics signature and prognosis were evaluated using univariable and multivariable Cox proportional hazards regression analyses. Incremental value of the radiomics signature beyond clinical-pathologic risk factors was assessed using concordance index (C-index). RESULTS: The radiomics signatures were independently associated with DFS (hazard ratio [HR], 1.920; p < 0.001) and OS (HR, 2.079; p < 0.001). The radiomics signature showed performance comparable to stage in estimation of DFS (C-index, 0.724 vs 0.685) and OS (0.735 vs 0.703). The radiomics added prognostic value to clinical-pathologic models (stage and histologic subtype) in predicting DFS (C-index, 0.764 vs 0.713; p < 0.001), which was also shown in the validation cohort (0.782 vs 0.734; p = 0.016). In terms of OS, including radiomics led to significant improvement in prognostic performance of the clinical-pathologic model (stage and age) in the training cohort (0.784 vs 0.737; p < 0.001), but the improvement was not significant in the validation cohort (0.805 vs 0.734; p = 0.149). CONCLUSIONS: CT radiomics was effective in predicting prognosis in lung adenocarcinoma patients, providing additional prognostic information beyond clinical-pathologic risk factors. KEY POINTS: • CT radiomics signature was an independent prognostic factor predicting disease-free and overall survival along with clinical risk factors of lung adenocarcinoma (stage, histologic subtype, and age). • CT radiomics added prognostic value to clinical-pathologic models (stage and subtype) in predicting disease-free survival (C-index for integrated model and clinical-pathologic model, 0.764 vs 0.713; p < 0.001), which was also proven in the validation cohort (0.782 vs 0.734; p = 0.016). • Integrated model incorporating radiomics signature can successfully stratify patients into high-risk, intermediate-, or low-risk groups in patients with resectable lung adenocarcinoma.

Deep Learning-based Image Conversion of CT Reconstruction Kernels Improves Radiomics Reproducibility for Pulmonary Nodules or Masses.

Background Intratumor heterogeneity in lung cancer may influence outcomes. CT radiomics seeks to assess tumor features to provide detailed imaging features. However, CT radiomic features vary according to the reconstruction kernel used for image generation. Purpose To investigate the effect of different reconstruction kernels on radiomic features and assess whether image conversion using a convolutional neural network (CNN) could improve reproducibility of radiomic features between different kernels. Materials and Methods In this retrospective analysis, patients underwent non-contrast material-enhanced and contrast material-enhanced axial chest CT with soft kernel (B30f) and sharp kernel (B50f) reconstruction using a single CT scanner from April to June 2017. To convert different kernels without sinogram, the CNN model was developed using residual learning and an end-to-end way. Kernel-converted images were generated, from B30f to B50f and from B50f to B30f. Pulmonary nodules or masses were semiautomatically segmented and 702 radiomic features (tumor intensity, texture, and wavelet features) were extracted. Measurement variability in radiomic features was evaluated using the concordance correlation coefficient (CCC). Results A total of 104 patients were studied, including 54 women and 50 men, with pulmonary nodules or masses (mean age, 63.2 years ± 10.5). The CCC between two readers using the same kernel was 0.92, and 592 of 702 (84.3%) of the radiomic features were reproducible (CCC ≥ 0.85); using different kernels, the CCC was 0.38 and only 107 of 702 (15.2%) of the radiomic features were reliable. Texture features and wavelet features were predominantly affected by reconstruction kernel (CCC, from 0.88 to 0.61 for texture features and from 0.92 to 0.35 for wavelet features). After applying image conversion, CCC improved to 0.84 and 403 of 702 (57.4%) radiomic features were reproducible (CCC, 0.85 for texture features and 0.84 for wavelet features). Conclusion Chest CT image conversion using a convolutional neural network effectively reduced the effect of two different reconstruction kernels and may improve the reproducibility of radiomic features in pulmonary nodules or masses. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Park in this issue.

Prognostic value of radiomic analysis of iodine overlay maps from dual-energy computed tomography in patients with resectable lung cancer.

OBJECTIVES: To investigate whether radiomics on iodine overlay maps from dual-energy computed tomography (DECT) can predict survival outcomes in patients with resectable lung cancer. METHODS: Ninety-three lung cancer patients eligible for curative surgery were examined with DECT at the time of diagnosis. The median follow-up was 60.4 months. Radiomic features of the entire primary tumour were extracted from iodine overlay maps generated by DECT. A Cox proportional hazards regression model was used to determine independent predictors of overall survival (OS) and disease-free survival (DFS), respectively. RESULTS: Forty-two patients (45.2%) had disease recurrence and 39 patients (41.9%) died during the follow-up period. The mean DFS was 49.8 months and OS was 55.2 months. Univariate analysis revealed that significant predictors of both OS and DFS were stage and radiomic parameters, including histogram energy, histogram entropy, grey-level co-occurrence matrix (GLCM) angular second moment, GLCM entropy and homogeneity. The multivariate analysis identified stage and entropy as independent risk factors predicting both OS (stage, hazard ratio (HR) = 2.020 [95% CI 1.014-4.026], p = 0.046; entropy, HR = 1.543 [95% CI 1.069-2.228], p = 0.021) and DFS (stage, HR = 2.132 [95% CI 1.060-4.287], p = 0.034; entropy, HR = 1.497 [95% CI 1.031-2.173], p = 0.034). The C-index showed that adding entropy improved prediction of OS compared to stage only (0.720 and 0.667, respectively; p = 0.048). CONCLUSIONS: Radiomic features extracted from iodine overlay map reflecting heterogeneity of tumour perfusion can add prognostic information for patients with resectable lung cancer. KEY POINTS: • Radiomic feature (histogram entropy) from DECT iodine overlay maps was an independent risk factor predicting both overall survival and disease-free survival. • Adding histogram entropy to clinical stage improved prediction of overall survival compared to stage only (0.720 and 0.667, respectively; p = 0.048). • DECT can be a good option for comprehensive pre-operative evaluation in cases of resectable lung cancer.

Hybrid Airway Segmentation Using Multi-Scale Tubular Structure Filters and Texture Analysis on 3D Chest CT Scans.

Airway diseases are frequently related to morphological changes that may influence lung physiology. Accurate airway region segmentation may be useful for quantitative evaluation of disease prognosis and therapy efficacy. The information can also be applied to understand the fundamental mechanisms of various lung diseases. We present a hybrid method to automatically segment the airway regions on 3D volume chest computed tomography (CT) scans. This method uses multi-scale filtering and support vector machine (SVM) classification. The proposed scheme is comprised of two hybrid steps. First, a tubular structure-based multi-scale filter is applied to find the initial candidate airway regions. Second, for identifying candidate airway regions using the fuzzy connectedness technique, the small and disconnected branches of airway regions are detected using SVM classification trained to differentiate between airway and non-airway regions through texture analysis of user-defined landmark points. For development and evaluation of the method, two datasets were incorporated: (1) 55 lung-CT volumes from the Korean Obstructive Lung Disease (KOLD) Cohort Study and (2) 20 cases from the publicly open database (EXACT'09). The average tree-length detection rates of EXACT'09 and KOLD were 56.9 ± 11.0 and 70.5 ± 8.98, respectively. Comparison of the results for the EXACT'09 data set between the presented method and other methods revealed that our approach was a high performer. The method limitations were higher false-positive rates than those of the other methods and risk of leakage. In future studies, application of a convolutional neural network will help overcome these shortcomings.

Machine learning assessment of myocardial ischemia using angiography: Development and retrospective validation.

BACKGROUND: Invasive fractional flow reserve (FFR) is a standard tool for identifying ischemia-producing coronary stenosis. However, in clinical practice, over 70% of treatment decisions still rely on visual estimation of angiographic stenosis, which has limited accuracy (about 60%-65%) for the prediction of FFR < 0.80. One of the reasons for the visual-functional mismatch is that myocardial ischemia can be affected by the supplied myocardial size, which is not always evident by coronary angiography. The aims of this study were to develop an angiography-based machine learning (ML) algorithm for predicting the supplied myocardial volume for a stenosis, as measured using coronary computed tomography angiography (CCTA), and then to build an angiography-based classifier for the lesions with an FFR < 0.80 versus ≥ 0.80. METHODS AND FINDINGS: A retrospective study was conducted using data from 1,132 stable and unstable angina patients with 1,132 intermediate lesions who underwent invasive coronary angiography, FFR, and CCTA at the Asan Medical Center, Seoul, Korea, between 1 May 2012 and 30 November 2015. The mean age was 63 ± 10 years, 76% were men, and 72% of the patients presented with stable angina. Of these, 932 patients (assessed before 31 January 2015) constituted the training set for the algorithm, and 200 patients (assessed after 1 February 2015) served as a test cohort to validate its diagnostic performance. Additionally, external validation with 79 patients from two centers (CHA University, Seongnam, Korea, and Ajou University, Suwon, Korea) was conducted. After automatic contour calibration using the caliber of guiding catheter, quantitative coronary angiography was performed using the edge-detection algorithms (CAAS-5, Pie-Medical). Clinical information was provided by the Asan BiomedicaL Research Environment (ABLE) system. The CCTA-based myocardial segmentation (CAMS)-derived myocardial volume supplied by each vessel (right coronary artery [RCA], left anterior descending [LAD], left circumflex [LCX]) and the myocardial volume subtended to a stenotic segment (CAMS-%Vsub) were measured for labeling. The ML for (1) predicting vessel territories (CAMS-%LAD, CAMS-%LCX, and CAMS-%RCA) and CAMS-%Vsub and (2) identifying the lesions with an FFR < 0.80 was constructed. Angiography-based ML, employing a light gradient boosting machine (GBM), showed mean absolute errors (MAEs) of 5.42%, 8.57%, and 4.54% for predicting CAMS-%LAD, CAMS-%LCX, and CAMS-%RCA, respectively. The percent myocardial volumes predicted by ML were used to predict the CAMS-%Vsub. With 5-fold cross validation, the MAEs between ML-predicted percent myocardial volume subtended to a stenotic segment (ML-%Vsub) and CAMS-%Vsub were minimized by the elastic net (6.26% ± 0.55% for LAD, 5.79% ± 0.68% for LCX, and 2.95% ± 0.14% for RCA lesions). Using all attributes (age, sex, involved vessel segment, and angiographic features affecting the myocardial territory and stenosis degree), the ML classifiers (L2 penalized logistic regression, support vector machine, and random forest) predicted an FFR < 0.80 with an accuracy of approximately 80% (area under the curve [AUC] = 0.84-0.87, 95% confidence intervals 0.71-0.94) in the test set, which was greater than that of diameter stenosis (DS) > 53% (66%, AUC = 0.71, 95% confidence intervals 0.65-0.78). The external validation showed 84% accuracy (AUC = 0.89, 95% confidence intervals 0.83-0.95). The retrospective design, single ethnicity, and the lack of clinical outcomes may limit this prediction model's generalized application. CONCLUSION: We found that angiography-based ML is useful to predict subtended myocardial territories and ischemia-producing lesions by mitigating the visual-functional mismatch between angiographic and FFR. Assessment of clinical utility requires further validation in a large, prospective cohort study.

Intravascular ultrasound-derived morphological predictors of myocardial ischemia assessed by stress myocardial perfusion computed tomography.

BACKGROUND: Although ischemia-guided revascularization improves clinical outcomes, morphological determinants of clinically relevant myocardial ischemia have not been studied. To identify intravascular ultrasound (IVUS)-derived anatomical parameters for predicting myocardial perfusion defect and its extent. METHODS: A total of 103 patients (88 stable and 15 unstable angina) with 153 lesions (angiographic diameter stenosis of 30-80%) underwent stress myocardial perfusion computed tomography (CT) and IVUS pre-procedure. The volume of CT perfusion defect and %CT perfusion defect in the target vessel territories were measured. RESULTS: The CT perfusion defect was seen in 76 (50%) lesions. The independent determinants for the presence of CT perfusion defect were IVUS-minimal lumen area (MLA) (adjusted OR = 0.56, 95% CI = 0.38-0.82), plaque burden (adjusted OR = 1.07, 95% CI = 1.02-1.11) and involvement of left main or left anterior descending artery (adjusted OR = 4.13, 95% CI = 1.75-9.78, all P < 0.05). The CT perfusion defect was predicted by IVUS-MLA <2.28mm2 (sensitivity 74%, specificity 82%) and plaque burden >77% (sensitivity 79%, specificity 78%) as thresholds. The independent determinants for the volume of CT perfusion defect were the involvement of left main or left anterior descending artery (ß = 16.43, standard errors = 4.387, P = 0.020) and a greater plaque burden (ß = 0.56, standard errors = 0.163, P = 0.026). CONCLUSIONS: IVUS-derived morphological parameters were useful to predict the presence of CT perfusion defect and the size of myocardial ischemia that were primarily determined by lesion severity and subtended myocardial territory. © 2016 Wiley Periodicals, Inc.

Novel methodology to evaluate renal cysts in polycystic kidney disease.

AIM: To develop and assess a semiautomated method for segmenting and counting individual renal cysts from mid-slice MR images in patients with autosomal dominant polycystic kidney disease (ADPKD). METHODS: A semiautomated method was developed to segment and count individual renal cysts from mid-slice MR images in 241 subjects with ADPKD from the Consortium for Radiologic Imaging Studies of Polycystic Kidney Disease. For each subject, a mid-slice MR image was selected from each set of coronal T2-weighted MR images covering the entire kidney. The selected mid-slice image was processed with the semiautomated method to segment and count individual renal cysts. The number of cysts from the mid-slice image of each kidney was also measured by manual counting. The level of agreement between the semiautomated and manual cyst counts was compared using intraclass correlation (ICC) and a Bland-Altman plot. RESULTS: Individual renal cysts were successfully segmented using the semiautomated method in all 241 cases. The number of cysts in each kidney measured with the semiautomated and manual counting methods correlated well (ICC = 0.96 for the right or left kidney), with a small average difference (-0.52, with higher semiautomated counts, for the right kidney, and 0.13, with higher manual counts, for the left kidney) in the semiautomated method. However, there was substantial variation in a small number of subjects; 6 of 241 participants (2.5%) had a difference in the total cyst count of more than 15. CONCLUSION: We have developed a semiautomated method to segment individual renal cysts from mid-slice MR images in ADPKD kidneys as a quantitative indicator of characterization and disease progression of ADPKD.

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.

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.

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.

Informatics in radiology: dual-energy electronic cleansing for fecal-tagging CT colonography.

Electronic cleansing (EC) is an emerging technique for the removal of tagged fecal materials at fecal-tagging computed tomographic (CT) colonography. However, existing EC methods may generate various types of artifacts that severely impair the quality of the cleansed CT colonographic images. Dual-energy fecal-tagging CT colonography is regarded as a next-generation imaging modality. EC that makes use of dual-energy fecal-tagging CT colonographic images promises to be effective in reducing cleansing artifacts by means of applying the material decomposition capability of dual-energy CT. The dual-energy index (DEI), which is calculated from the relative change in the attenuation values of a material at two different photon energies, is a reliable and effective indicator for differentiating tagged fecal materials from various types of tissues on fecal-tagging CT colonographic images. A DEI-based dual-energy EC scheme uses the DEI to help differentiate the colonic lumen-including the luminal air, tagged fecal materials, and air-tagging mixture-from the colonic soft-tissue structures, and then segments the entire colonic lumen for cleansing of the tagged fecal materials. As a result, dual-energy EC can help identify partial-volume effects in the air-tagging mixture and inhomogeneous tagging in residual fecal materials, the major causes of EC artifacts. This technique has the potential to significantly improve the quality of EC and promises to provide images of a cleansed colon that are free of the artifacts commonly observed with conventional single-energy EC methods.

Dual-energy index value of luminal air in fecal-tagging computed tomography colonography: findings and impact on electronic cleansing.

PURPOSE: The purpose of our study was to measure the dual-energy index (DEI) value of colonic luminal air in both phantom and clinical fecal-tagging dual-energy computed tomography (CT) colonography (DE-CTC) images and to demonstrate its impact on dual-energy electronic cleansing. METHODS: For the phantom study, a custom-ordered colon phantom was scanned by a dual-energy CT scanner (SOMATON Definition Flash; Siemens Healthcare, Forchheim, Germany) at two photon energies: 80 and 140 kVp. Before imaging, the phantom was filled with a 300-mL mixture of simulated fecal materials tagged by a nonionic iodinated contrast agent at three contrast concentrations: 20, 40, and 60 mg/mL. Ten regions-of-interest (ROIs) were randomly placed in each of the colonic luminal air, abdominal fat, bony structure, and tagged material in each scan. For the clinical study, 22 DE-CTC (80 and 140 kVp) patient cases were collected, who underwent a low-fiber, low-residue diet bowel preparation and orally administered iodine-based fecal tagging. Twenty ROIs were randomly placed in each of the colonic luminal air, abdominal fat, abdominal soft tissue, and tagged fecal material in each scan. For each ROI, the mean CT values in both 80- and 140-kVp images were measured, and then its DEI was calculated. RESULTS: In the phantom study, the mean DEI values of luminal air were 0.270, 0.298, 0.386, and 0.402 for the four groups of tagging conditions: no tagged material and tagged with three groups of contrast concentrations at 20, 40, and 60 mg/mL. In the clinical study, the mean DEI values were 0.341, -0.012, -0.002, and 0.188 for colonic luminal air, abdominal fat, abdominal soft tissue, and tagged fecal material, respectively. CONCLUSIONS: In our study, we observed that the DEI values of colonic luminal air in DE-CTC images (>0.10) were substantially higher than the theoretical value of 0.0063. In addition, the observed DEI values of colonic luminal air were significantly higher than those of soft tissue. These findings have an important impact on electronic cleansing: it may provide an effective means of differentiating colonic soft-tissue structures from the air-tagging mixture caused by the partial volume effect and thus of minimizing the cleansing artifacts.

Evaluation of MRI resolution affecting trabecular bone parameters: determination of acceptable resolution.

The objective of this study is to evaluate the effect of MR image resolution on trabecular bone parameters and to determine the acceptable resolution that can be accurately analyzed to assess structural parameters. Ten distal femoral condyle specimens of 1 × 1 × 1 cm(3) were scanned with a 4.7-T Bruker BioSpec MRI scanner using a three-dimensional fast large-angle spin-echo sequence with various iso-cubic voxels sizes (65, 130, 160, 196, 230, and 260 µm). Otsu thresholding was applied to identify voxels containing bone. Conventional bone parameters, structural bone parameters, and skeleton-based local trabecular thickness (slTB.Th) were evaluated. The Bland-Altman method and correlation indicated that the conventional and structural bone parameters were preserved with an iso-cubic voxel size up to 230 µm (r > 0.932 and r > 0.843, respectively). In addition, slTB.Th derived from the highest resolution images (65 µm iso-cubic voxel size), correlated well (r > 0.833) with the values computed from lower resolution images, up to 230 µm, which is twice typical human trabecular thickness range (100-150 µm). The outcome of this study suggests that the various bone parameters were well preserved up to 230 µm images.