Artificial Intelligence Quantification of Enhanced Synovium Throughout the Entire Hand in Rheumatoid Arthritis on Dynamic Contrast-Enhanced MRI.

BACKGROUND: Challenges persist in achieving automatic and efficient inflammation quantification using dynamic contrast-enhanced (DCE) MRI in rheumatoid arthritis (RA) patients. PURPOSE: To investigate an automatic artificial intelligence (AI) approach and an optimized dynamic MRI protocol for quantifying disease activity in RA in whole hands while excluding arterial pixels. STUDY TYPE: Retrospective. SUBJECTS: Twelve RA patients underwent DCE-MRI with 27 phases for creating the AI model and tested on images with a variable number of phases from 35 RA patients. FIELD STRENGTH/SEQUENCE: 3.0 T/DCE T1-weighted gradient echo sequence (mDixon, water image). ASSESSMENT: The model was trained with various DCE-MRI time-intensity number of phases. Evaluations were conducted for similarity between AI segmentation and manual outlining in 51 ROIs with synovitis. The relationship between synovial volume via AI segmentation with rheumatoid arthritis magnetic resonance imaging scoring (RAMRIS) across whole hands was then evaluated. The reference standard was determined by an experienced musculoskeletal radiologist. STATISTICAL TEST: Area under the curve (AUC) of receiver operating characteristic (ROC), Dice and Spearman's rank correlation coefficients, and interclass correlation coefficient (ICC). A P-value <0.05 was considered statistically significant. RESULTS: A minimum of 15 phases (acquisition time at least 2.5 minutes) was found to be necessary. AUC ranged from 0.941 ± 0.009 to 0.965 ± 0.009. The Dice score was 0.557-0.615. Spearman's correlation coefficients between the AI model and ground truth were 0.884-0.927 and 0.736-0.831, for joint ROIs and whole hands, respectively. The Spearman's correlation coefficient for the additional test set between the model trained with 15 phases and RAMRIS was 0.768. CONCLUSION: The AI-based classification model effectively identified synovitis pixels while excluding arteries. The optimal performance was achieved with at least 15 phases, providing a quantitative assessment of inflammatory activity in RA while minimizing acquisition time. EVIDENCE LEVEL: 3 TECHNICAL EFFICACY: Stage 2.

Estimating Body Weight From Measurements From Different Single-Slice Computed Tomography Levels: An Evaluation of Total Cross-Sectional Body Area Measurements and Deep Learning.

OBJECTIVE: This study aimed to evaluate the correlation between the estimated body weight obtained from 2 easy-to-perform methods and the actual body weight at different computed tomography (CT) levels and determine the best reference site for estimating body weight. METHODS: A total of 862 patients from a public database of whole-body positron emission tomography/CT studies were retrospectively analyzed. Two methods for estimating body weight at 10 single-slice CT levels were evaluated: a linear regression model using total cross-sectional body area and a deep learning-based model. The accuracy of body weight estimation was evaluated using the mean absolute error (MAE), root mean square error (RMSE), and Spearman rank correlation coefficient ( ρ ). RESULTS: In the linear regression models, the estimated body weight at the T5 level correlated best with the actual body weight (MAE, 5.39 kg; RMSE, 7.01 kg; ρ = 0.912). The deep learning-based models showed the best accuracy at the L5 level (MAE, 6.72 kg; RMSE, 8.82 kg; ρ = 0.865). CONCLUSIONS: Although both methods were feasible for estimating body weight at different single-slice CT levels, the linear regression model using total cross-sectional body area at the T5 level as an input variable was the most favorable method for single-slice CT analysis for estimating body weight.

Deep learning to assess right ventricular ejection fraction from two-dimensional echocardiograms in precapillary pulmonary hypertension.

BACKGROUND: Precapillary pulmonary hypertension (PH) is characterized by a sustained increase in right ventricular (RV) afterload, impairing systolic function. Two-dimensional (2D) echocardiography is the most performed cardiac imaging tool to assess RV systolic function; however, an accurate evaluation requires expertise. We aimed to develop a fully automated deep learning (DL)-based tool to estimate the RV ejection fraction (RVEF) from 2D echocardiographic videos of apical four-chamber views in patients with precapillary PH. METHODS: We identified 85 patients with suspected precapillary PH who underwent cardiac magnetic resonance imaging (MRI) and echocardiography. The data was divided into training (80%) and testing (20%) datasets, and a regression model was constructed using 3D-ResNet50. Accuracy was assessed using five-fold cross validation. RESULTS: The DL model predicted the cardiac MRI-derived RVEF with a mean absolute error of 7.67%. The DL model identified severe RV systolic dysfunction (defined as cardiac MRI-derived RVEF < 37%) with an area under the curve (AUC) of .84, which was comparable to the AUC of RV fractional area change (FAC) and tricuspid annular plane systolic excursion (TAPSE) measured by experienced sonographers (.87 and .72, respectively). To detect mild RV systolic dysfunction (defined as RVEF ≤ 45%), the AUC from the DL-predicted RVEF also demonstrated a high discriminatory power of .87, comparable to that of FAC (.90), and significantly higher than that of TAPSE (.67). CONCLUSION: The fully automated DL-based tool using 2D echocardiography could accurately estimate RVEF and exhibited a diagnostic performance for RV systolic dysfunction comparable to that of human readers.

Assessment of changes in vessel area during needle manipulation in microvascular anastomosis using a deep learning-based semantic segmentation algorithm: A pilot study.

Appropriate needle manipulation to avoid abrupt deformation of fragile vessels is a critical determinant of the success of microvascular anastomosis. However, no study has yet evaluated the area changes in surgical objects using surgical videos. The present study therefore aimed to develop a deep learning-based semantic segmentation algorithm to assess the area change of vessels during microvascular anastomosis for objective surgical skill assessment with regard to the "respect for tissue." The semantic segmentation algorithm was trained based on a ResNet-50 network using microvascular end-to-side anastomosis training videos with artificial blood vessels. Using the created model, video parameters during a single stitch completion task, including the coefficient of variation of vessel area (CV-VA), relative change in vessel area per unit time (ΔVA), and the number of tissue deformation errors (TDE), as defined by a ΔVA threshold, were compared between expert and novice surgeons. A high validation accuracy (99.1%) and Intersection over Union (0.93) were obtained for the auto-segmentation model. During the single-stitch task, the expert surgeons displayed lower values of CV-VA (p < 0.05) and ΔVA (p < 0.05). Additionally, experts committed significantly fewer TDEs than novices (p < 0.05), and completed the task in a shorter time (p < 0.01). Receiver operating curve analyses indicated relatively strong discriminative capabilities for each video parameter and task completion time, while the combined use of the task completion time and video parameters demonstrated complete discriminative power between experts and novices. In conclusion, the assessment of changes in the vessel area during microvascular anastomosis using a deep learning-based semantic segmentation algorithm is presented as a novel concept for evaluating microsurgical performance. This will be useful in future computer-aided devices to enhance surgical education and patient safety.

Deep learning-based video-analysis of instrument motion in microvascular anastomosis training.

PURPOSE: Attaining sufficient microsurgical skills is paramount for neurosurgical trainees. Kinematic analysis of surgical instruments using video offers the potential for an objective assessment of microsurgical proficiency, thereby enhancing surgical training and patient safety. The purposes of this study were to develop a deep-learning-based automated instrument tip-detection algorithm, and to validate its performance in microvascular anastomosis training. METHODS: An automated instrument tip-tracking algorithm was developed and trained using YOLOv2, based on clinical microsurgical videos and microvascular anastomosis practice videos. With this model, we measured motion economy (procedural time and path distance) and motion smoothness (normalized jerk index) during the task of suturing artificial blood vessels for end-to-side anastomosis. These parameters were validated using traditional criteria-based rating scales and were compared across surgeons with varying microsurgical experience (novice, intermediate, and expert). The suturing task was deconstructed into four distinct phases, and parameters within each phase were compared between novice and expert surgeons. RESULTS: The high accuracy of the developed model was indicated by a mean Dice similarity coefficient of 0.87. Deep learning-based parameters (procedural time, path distance, and normalized jerk index) exhibited correlations with traditional criteria-based rating scales and surgeons' years of experience. Experts completed the suturing task faster than novices. The total path distance for the right (dominant) side instrument movement was shorter for experts compared to novices. However, for the left (non-dominant) side, differences between the two groups were observed only in specific phases. The normalized jerk index for both the right and left sides was significantly lower in the expert than in the novice groups, and receiver operating characteristic analysis showed strong discriminative ability. CONCLUSION: The deep learning-based kinematic analytic approach for surgical instruments proves beneficial in assessing performance in microvascular anastomosis. Moreover, this methodology can be adapted for use in clinical settings.

Deep learning-based body weight from scout images can be an alternative to actual body weight in CT radiation dose management.

PURPOSE: Accurate body weight measurement is essential to promote computed tomography (CT) dose optimization; however, body weight cannot always be measured prior to CT examination, especially in the emergency setting. The aim of this study was to investigate whether deep learning-based body weight from chest CT scout images can be an alternative to actual body weight in CT radiation dose management. METHODS: Chest CT scout images and diagnostic images acquired for medical checkups were collected from 3601 patients. A deep learning model was developed to predict body weight from scout images. The correlation between actual and predicted body weight was analyzed. To validate the use of predicted body weight in radiation dose management, the volume CT dose index (CTDIvol ) and the dose-length product (DLP) were compared between the body weight subgroups based on actual and predicted body weight. Surrogate size-specific dose estimates (SSDEs) acquired from actual and predicted body weight were compared to the reference standard. RESULTS: The median actual and predicted body weight were 64.1 (interquartile range: 56.5-72.4) and 64.0 (56.3-72.2) kg, respectively. There was a strong correlation between actual and predicted body weight (ρ = 0.892, p < 0.001). The CTDIvol and DLP of the body weight subgroups were similar based on actual and predicted body weight (p < 0.001). Both surrogate SSDEs based on actual and predicted body weight were not significantly different from the reference standard (p = 0.447 and 0.410, respectively). CONCLUSION: Predicted body weight can be an alternative to actual body weight in managing dose metrics and simplifying SSDE calculation. Our proposed method can be useful for CT radiation dose management in adult patients with unknown body weight.

Acquisition time reduction in pediatric 99m Tc-DMSA planar imaging using deep learning.

PURPOSE: Given the potential risk of motion artifacts, acquisition time reduction is desirable in pediatric 99m Tc-dimercaptosuccinic acid (DMSA) scintigraphy. The aim of this study was to evaluate the performance of predicted full-acquisition-time images from short-acquisition-time pediatric 99m Tc-DMSA planar images with only 1/5th acquisition time using deep learning in terms of image quality and quantitative renal uptake measurement accuracy. METHODS: One hundred and fifty-five cases that underwent pediatric 99m Tc-DMSA planar imaging as dynamic data for 10 min were retrospectively collected for the development of three deep learning models (DnCNN, Win5RB, and ResUnet), and the generation of full-time images from short-time images. We used the normalized mean squared error (NMSE), peak signal-to-noise ratio (PSNR), and structural similarity index metrics (SSIM) to evaluate the accuracy of the predicted full-time images. In addition, the renal uptake of 99m Tc-DMSA was calculated, and the difference in renal uptake from the reference full-time images was assessed using scatter plots with Pearson correlation and Bland-Altman plots. RESULTS: The predicted full-time images from the deep learning models showed a significant improvement in image quality compared to the short-time images with respect to the reference full-time images. In particular, the predicted full-time images obtained by ResUnet showed the lowest NMSE (0.4 [0.4-0.5] %) and the highest PSNR (55.4 [54.7-56.1] dB) and SSIM (0.997 [0.995-0.997]). For renal uptake, an extremely high correlation was achieved in all short-time and three predicted full-time images (R2  > 0.999 for all). The Bland-Altman plots showed the lowest bias (-0.10) of renal uptake in ResUnet, while short-time images showed the lowest variance (95% confidence interval: -0.14, 0.45) of renal uptake. CONCLUSIONS: Our proposed method is capable of producing images that are comparable to the original full-acquisition-time images, allowing for a reduction of acquisition time/injected dose in pediatric 99m Tc-DMSA planar imaging.

Estimation of Left and Right Ventricular Ejection Fractions from cine-MRI Using 3D-CNN.

Cardiac function indices must be calculated using tracing from short-axis images in cine-MRI. A 3D-CNN (convolutional neural network) that adds time series information to images can estimate cardiac function indices without tracing using images with known values and cardiac cycles as the input. Since the short-axis image depicts the left and right ventricles, it is unclear which motion feature is captured. This study aims to estimate the indices by learning the short-axis images and the known left and right ventricular ejection fractions and to confirm the accuracy and whether each index is captured as a feature. A total of 100 patients with publicly available short-axis cine images were used. The dataset was divided into training:test = 8:2, and a regression model was built by training with the 3D-ResNet50. Accuracy was assessed using a five-fold cross-validation. The correlation coefficient, MAE (mean absolute error), and RMSE (root mean squared error) were determined as indices of accuracy evaluation. The mean correlation coefficient of the left ventricular ejection fraction was 0.80, MAE was 9.41, and RMSE was 12.26. The mean correlation coefficient of the right ventricular ejection fraction was 0.56, MAE was 11.35, and RMSE was 14.95. The correlation coefficient was considerably higher for the left ventricular ejection fraction. Regression modeling using the 3D-CNN indicated that the left ventricular ejection fraction was estimated more accurately, and left ventricular systolic function was captured as a feature.

Predicting the response to pulmonary vasodilator therapy in systemic sclerosis with pulmonary hypertension by using quantitative chest CT.

OBJECTIVES: Systemic sclerosis (SSc) is associated with pulmonary vascular disease and interstitial lung disease, making it difficult to differentiate pulmonary arterial hypertension and pulmonary hypertension (PH) due to lung diseases and/or hypoxia and to decide treatments. We aimed to predict the response to pulmonary vasodilators in patients with SSc and PH. METHODS: Eighty-four SSc patients were included with 47 having PH. Chest computed tomography was evaluated using software to calculate the abnormal lung volume (ALV). To define the response to vasodilators, Δ mean pulmonary artery pressure (mPAP)/basal mPAP was used (cut-off value: 10%). The predictive value was evaluated by using the receiver operating characteristic curve. RESULTS: The mean (±standard deviation) value of ALV was 26.8 (±32.2) %. A weak correlation was observed between ALV and forced vital capacity (FVC) (R = -0.46). The predictive value of ALV [area under curve (AUC) = 0.74] was superior to that of FVC (AUC = 0.62) for the response to vasodilators. No hemodynamic parameters differed between patients with high and low ALV, whereas survival was worse in high ALV. CONCLUSIONS: Quantitative chest computed tomography well predicted the response to vasodilators in patients with SSc and PH. Our results suggest its utility in differentiating the dominance of pulmonary vascular disease or interstitial lung disease.

Intravoxel incoherent motion MRI for discrimination of synovial proliferation in the hand arthritis: A prospective proof-of-concept study.

BACKGROUND: Postcontrast-enhanced MRI is currently the reference standard for synovial proliferation in rheumatoid arthritis (RA). However, the technique is somewhat invasive due to the use of gadolinium contrast agents, which may cause severe adverse/side effects. Intravoxel incoherent motion (IVIM) simultaneously permits quantification of perfusion as well as diffusion using a single imaging scan. PURPOSE/HYPOTHESIS: To test the capability of IVIM MRI for noninvasive discrimination of synovial proliferation in hand arthritis. STUDY TYPE: Prospective. SUBJECTS: Seven suspected RA patients (three women and four men; mean age, 61 years; range, 26-74 years). FIELD STRENGTH/SEQUENCE: 3 T/short tau inversion recovery (STIR), IVIM, postcontrast-enhanced MRI. ASSESSMENT: Region of interest (ROI) was identified based on STIR. Contrast-enhanced MRI was evaluated using a 5-point grading scale of 0 (water) to 4 (synovial proliferation) according to the degree of contrast enhancement within the ROI. For each ROI, we calculated the apparent diffusion coefficient (ADC) and IVIM parameters (molecular diffusion coefficient [D], perfusion fraction [f], and perfusion-related diffusion coefficient [D*]). These parameters were subsequently compared with ROI contrast enhancement grades. STATISTICAL TESTS: Spearman's rank correlation test and a receiver operating characteristic (ROC) curve. RESULTS: A total of 90 ROIs of suspected synovial proliferation and/or joint effusion were identified. ROI grades were correlated with ADC and D values (r S = -0.385, P < 0.001, r S = -0.458, P < 0.0001, respectively), but not with the f and D* values (r S = -0.010, P = 0.936, r S = -0.084, P = 0.505, respectively). The area under the curves (AUCs) of D values (0.708-0.888, P = 0.002-0.0002) were slightly larger than those of ADC values (0.692-0.791, P = 0.013-0.001) when comparing low- vs. high-contrast enhancement grades. DATA CONCLUSION: The IVIM parameter D and ADC may be useful for the noninvasive identification of synovial proliferation in hand arthritis. LEVEL OF EVIDENCE: 2 Technical Efficacy Stage: 2 J. Magn. Reson. Imaging 2019;50:1199-1206.

Computed diffusion-weighted imaging for differentiating synovial proliferation from joint effusion in hand arthritis.

The objective of this study is to investigate computed DWI (cDWI) as an alternative method to contrast-enhanced MRI in comparison with directory measured DWI (mDWI) and apparent diffusion coefficient (ADC) for differentiating synovial proliferation from joint effusion. Nine patients suspected with RA (5 women) were included in this study. A radiologist identified region of interest (ROI) based on STIR, and evaluated using a 5-point grading scale of 0 (fluid) to 4 (synovial proliferation) according to the degree of contrast enhancement within the ROI. cDWI was synthesized for b values from 1000 to 2000 at 200 s/mm2 intervals using the combination of b values at mDWI. In addition to ADC values, contrast ratios were calculated using signal intensity for each ROI on the mDWI and cDWI. Visual assessment by a radiologist was conducted between pairs of STIR image and mDWI or cDWI. ROI grades were most significantly correlated with cDWI2000 based on b values of 400-1000 s/mm2 (rs = 0.405, p < 0.01). The area under the curve of cDWI2000 based on b values of 400-1000 s/mm2 (0.762) was larger than that of ADC values (0.570-0.608) when comparing low versus high contrast enhancement grades. Both cDWI1800 (200-1000) and cDWI2000 (400-1000) demonstrated high sensitivity and specificity in visual assessment (84.6% and 66.7%, respectively). The cDWI2000 based on b values of 400-1000 s/mm2 may be useful for noninvasive differentiation of synovial proliferation from joint effusion in hand arthritis.

Acceleration of ASL-based time-resolved MR angiography by acquisition of control and labeled images in the same shot (ACTRESS).

PURPOSE: Noncontrast 4D-MR-angiography (MRA) using arterial spin labeling (ASL) is beneficial because high spatial and temporal resolution can be achieved. However, ASL requires acquisition of labeled and control images for each phase. The purpose of this study is to present a new accelerated 4D-MRA approach that requires only a single control acquisition, achieving similar image quality in approximately half the scan time. METHODS: In a multi-phase Look-Locker sequence, the first phase was used as the control image and the labeling pulse was applied before the second phase. By acquiring the control and labeled images within a single Look-Locker cycle, 4D-MRA was generated in nearly half the scan time of conventional ASL. However, this approach potentially could be more sensitive to off-resonance and magnetization transfer (MT) effects. To counter this, careful optimizations of the labeling pulse were performed by Bloch simulations. In in-vivo studies arterial visualization was compared between the new and conventional ASL approaches. RESULTS: Optimization of the labeling pulse successfully minimized off-resonance effects. Qualitative assessment showed that residual MT effects did not degrade visualization of the peripheral arteries. CONCLUSION: This study demonstrated that the proposed approach achieved similar image quality as conventional ASL-MRA approaches in just over half the scan time. Magn Reson Med 79:224-233, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Quantification of hand synovitis in rheumatoid arthritis: Arterial mask subtraction reinforced with mutual information can improve accuracy of pixel-by-pixel time-intensity curve shape analysis in dynamic MRI.

BACKGROUND: Synovitis, which is a hallmark of rheumatoid arthritis (RA), needs to be precisely quantified to determine the treatment plan. Time-intensity curve (TIC) shape analysis is an objective assessment method for characterizing the pixels as artery, inflamed synovium, or other tissues using dynamic contrast-enhanced MRI (DCE-MRI). PURPOSE/HYPOTHESIS: To assess the feasibility of our original arterial mask subtraction method (AMSM) with mutual information (MI) for quantification of synovitis in RA. STUDY TYPE: Prospective study. SUBJECTS: Ten RA patients (nine women and one man; mean age, 56.8 years; range, 38-67 years). FIELD STRENGTH/SEQUENCE: 3T/DCE-MRI. ASSESSMENT: After optimization of TIC shape analysis to the hand region, a combination of TIC shape analysis and AMSM was applied to synovial quantification. The MI between pre- and postcontrast images was utilized to determine the arterial mask phase objectively, which was compared with human subjective selection. The volume of objectively measured synovitis by software was compared with that of manual outlining by an experienced radiologist. Simple TIC shape analysis and TIC shape analysis combined with AMSM were compared in slices without synovitis according to subjective evaluation. STATISTICAL TESTS: Pearson's correlation coefficient, paired t-test and intraclass correlation coefficient (ICC). RESULTS: TIC shape analysis was successfully optimized in the hand region with a correlation coefficient of 0.725 (P < 0.01) with the results of manual assessment regarded as ground truth. Objective selection utilizing MI had substantial agreement (ICC = 0.734) with subjective selection. Correlation of synovial volumetry in combination with TIC shape analysis and AMSM with manual assessment was excellent (r = 0.922, P < 0.01). In addition, negative predictive ability in slices without synovitis pixels was significantly increased (P < 0.01). DATA CONCLUSIONS: The combination of TIC shape analysis and image subtraction reinforced with MI can accurately quantify synovitis of RA in the hand by eliminating arterial pixels. LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018.

FDG PET/CT diagnostic criteria may need adjustment based on MRI to estimate the presurgical risk of extrapelvic infiltration in patients with uterine endometrial cancer.

PURPOSE: The staging of endometrial cancer requires surgery which carries the risk of morbidity. FDG PET/CT combined with anatomical imaging may reduce the number of unnecessary lymphadenectomies by demonstrating the risk of extrapelvic infiltration. The purpose of this study was to optimize FDG PET/CT diagnostic criteria for risk assessment in endometrial cancer after first-line risk triage with MRI. METHODS: The study population comprised 37 patients who underwent curative surgery for the treatment of endometrial cancer. First, the risk of extrapelvic infiltration was triaged using MRI. Second, multiple glucose metabolic profiles of the primary lesion were assessed with FDG PET/CT, and these were correlated with the histopathological risk of extrapelvic infiltration including lymphovascular space invasion (LVSI) and high-grade malignancy (grades 2 and 3). The results of histological correlation were used to adjust FDG PET/CT diagnostic criteria. RESULTS: Presurgical assessment using MRI was positive for deep (>50 %) myometrial invasion in 17 patients. The optimal FDG PET/CT diagnostic criteria vary depending on the results of MRI. Specifically, SUVmax (≥16.0) was used to indicate LVSI risk with an overall diagnostic accuracy of 88.2 % in patients with MRI findings showing myometrial invasion. High-grade malignancy did not correlate with any of metabolic profiles in this patient group. In the remaining patients without myometrial invasion, lesion glycolysis (LG) or metabolic volume were better indicators of LVSI than SUVmax with the same diagnostic accuracy of 80.0 %. In addition, LG (≥26.9) predicted high-grade malignancy with an accuracy of 72.2 %. Using the optimized cut-off criteria for LVSI, glucose metabolic profiling of primary lesions correctly predicted lymph node metastasis with an accuracy of 73.0 %, which was comparable with the accuracy of visual assessment for lymph node metastasis using MRI and FDG PET/CT. CONCLUSION: FDG PET/CT diagnostic criteria may need adjustment based on the anatomical information provided by MRI. The optimized criteria can predict the risk of pathology-proven LVSI correctly in 83.8 % of patients before surgery, and thus would improve presurgical treatment planning.

Quantification of myocardial blood flow with dynamic perfusion 3.0 Tesla MRI: Validation with (15) O-water PET.

BACKGROUND: To develop and validate a method for quantifying myocardial blood flow (MBF) using dynamic perfusion magnetic resonance imaging (MBFMRI ) at 3.0 Tesla (T) and compare the findings with those of (15) O-water positron emission tomography (MBFPET ). METHODS: Twenty healthy male volunteers underwent magnetic resonance imaging (MRI) and (15) O-water positron emission tomography (PET) at rest and during adenosine triphosphate infusion. The single-tissue compartment model was used to estimate the inflow rate constant (K1). We estimated the extraction fraction of Gd-DTPA using K1 and MBF values obtained from (15) O-water PET for the first 10 subjects. For validation, we calculated MBFMRI values for the remaining 10 subjects and compared them with the MBFPET values. In addition, we compared MBFMRI values of 10 patients with coronary artery disease with those of healthy subjects. RESULTS: The mean resting and stress MBFMRI values were 0.76 ± 0.10 and 3.04 ± 0.82 mL/min/g, respectively, and showed excellent correlation with the mean MBFPET values (r = 0.96, P < 0.01). The mean stress MBFMRI value was significantly lower for the patients (1.92 ± 0.37) than for the healthy subjects (P < 0.001). CONCLUSION: The use of dynamic perfusion MRI at 3T is useful for estimating MBF and can be applied for patients with coronary artery disease.

The montage method improves the classification of suspected acute ischemic stroke using the convolution neural network and brain MRI.

This study investigated the usefulness of the montage method that combines four different magnetic resonance images into one images for automatic acute ischemic stroke (AIS) diagnosis with deep learning method. The montage image was consisted from diffusion weighted image (DWI), fluid attenuated inversion recovery (FLAIR), arterial spin labeling (ASL), and apparent diffusion coefficient (ASL). The montage method was compared with pseudo color map (pCM) which was consisted from FLAIR, ASL and ADC. 473 AIS patients were classified into four categories: mechanical thrombectomy, conservative therapy, hemorrhage, and other diseases. The results showed that the montage image significantly outperformed pCM in terms of accuracy (montage image = 0.76 ± 0.01, pCM = 0.54 ± 0.05) and the area under the curve (AUC) (montage image = 0.94 ± 0.01, pCM = 0.76 ± 0.01). This study demonstrates the usefulness of the montage method and its potential for overcoming the limitations of pCM.

Deep learning-based correction for time truncation in cerebral computed tomography perfusion.

Cerebral computed tomography perfusion (CTP) imaging requires complete acquisition of contrast bolus inflow and washout in the brain parenchyma; however, time truncation undoubtedly occurs in clinical practice. To overcome this issue, we proposed a three-dimensional (two-dimensional + time) convolutional neural network (CNN)-based approach to predict missing CTP image frames at the end of the series from earlier acquired image frames. Moreover, we evaluated three strategies for predicting multiple time points. Seventy-two CTP scans with 89 frames and eight slices from a publicly available dataset were used to train and test the CNN models capable of predicting the last 10 image frames. The prediction strategies were single-shot prediction, recursive multi-step prediction, and direct-recursive hybrid prediction.Single-shot prediction predicted all frames simultaneously, while recursive multi-step prediction used prior predictions as input for subsequent steps, and direct-recursive hybrid prediction employed separate models for each step with prior predictions as input for the next step. The accuracies of the predicted image frames were evaluated in terms of image quality, bolus shape, and clinical perfusion parameters. We found that the image quality metrics were superior when multiple CTP images were predicted simultaneously rather than recursively. The bolus shape also showed the highest correlation (r = 0.990, p < 0.001) and the lowest variance (95% confidence interval, -453.26-445.53) in the single-shot prediction. For all perfusion parameters, the single-shot prediction had the smallest absolute differences from ground truth. Our proposed approach can potentially minimize time truncation errors and support the accurate quantification of ischemic stroke.

Energy spectrum measurement of scattered X-rays during IVR procedure.

With the increase of the number of interventional radiology (IVR) procedures, the occupational exposure of operators and medical staff has attracted keen attention. The energy of scattered radiation in medical clinical sites is important for estimating the biological effects of occupational exposure. Recent years have seen many reports on the dose of scattered radiation by IVR, but few on the energy spectrum. In this study, the energy spectrum of scattered X-rays was measured by using a cadmium telluride (CdTe) semiconductor detector during IVR on several neurosurgical and cardiovascular cases. The cumulated spectra in each case were compared. The spectra showed little changes among neurosurgical cases and relatively large changes among cardiovascular cases. This was assumed to be due to the change of X-ray tube voltage and tube angle was larger in cardiovascular cases. The resulting energy spectra will be essential for the assessment of detailed biological effects of occupational exposure.

Fully automatic quantification for hand synovitis in rheumatoid arthritis using pixel-classification-based segmentation network in DCE-MRI.

PURPOSE: A classification-based segmentation method is proposed to quantify synovium in rheumatoid arthritis (RA) patients using a deep learning (DL) method based on time-intensity curve (TIC) analysis in dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). MATERIALS AND METHODS: This retrospective study analyzed a hand MR dataset of 28 RA patients (six males, mean age 53.7 years). A researcher, under expert guidance, used in-house software to delineate regions of interest (ROIs) for hand muscles, bones, and synovitis, generating a dataset with 27,255 pixels with corresponding TICs (muscle: 11,413, bone: 8502, synovitis: 7340). One experienced musculoskeletal radiologist performed ground truth segmentation of enhanced pannus in the joint bounding box on the 10th DCE phase, or around 5 min after contrast injection. Data preprocessing included median filtering for noise reduction, phase-only correlation algorithm for motion correction, and contrast-limited adaptive histogram equalization for improved image contrast and noise suppression. TIC intensity values were normalized using zero-mean normalization. A DL model with dilated causal convolution and SELU activation function was developed for enhanced pannus segmentation, tested using leave-one-out cross-validation. RESULTS: 407 joint bounding boxes were manually segmented, with 129 synovitis masks. On the pixel-based level, the DL model achieved sensitivity of 85%, specificity of 98%, accuracy of 99% and precision of 84% for enhanced pannus segmentation, with a mean Dice score of 0.73. The false-positive rate for predicting cases without synovitis was 0.8%. DL-measured enhanced pannus volume strongly correlated with ground truth at both pixel-based (r = 0.87, p < 0.001) and patient-based levels (r = 0.84, p < 0.001). Bland-Altman analysis showed the mean difference for hand joints at the pixel-based and patient-based levels were -9.46 mm3 and -50.87 mm3, respectively. CONCLUSION: Our DL-based DCE-MRI TIC shape analysis has the potential for automatic segmentation and quantification of enhanced synovium in the hands of RA patients.

Imaging of 17O-labeled Water Using Fast T2 Mapping with T2-preparation: A Phantom Study.

17O-labeled water is a T2-shortening contrast agent used in proton MRI and is a promising method for visualizing cerebrospinal fluid (CSF) dynamics because it provides long-term tracking of water molecules. However, various external factors reduce the accuracy of 17O-concentration measurements using conventional signal-intensity-based methods. In addition, T2 mapping, which is expected to provide a stable assessment, is generally limited to temporal-spatial resolution. We developed the T2-prepared based on T2 mapping used in cardiac imaging to adapt to long T2 values and tested whether it could accurately measure 17O-concentration in the CSF using a phantom. The results showed that 17O-concentration in a fluid mimicking CSF could be evaluated with an accuracy comparable to conventional T2-mapping (Carr-Purcell-Meiboom-Gill multi-echo spin-echo method). This method allows 17O-imaging with a high temporal resolution and stability in proton MRI. This imaging technique may be promising for visualizing CSF dynamics using 17O-labeled water.