Weakly supervised deep learning for diagnosis of multiple vertebral compression fractures in CT.

OBJECTIVE: This study aims to develop a weakly supervised deep learning (DL) model for vertebral-level vertebral compression fracture (VCF) classification using image-level labelled data. METHODS: The training set included 815 patients with normal (n = 507, 62%) or VCFs (n = 308, 38%). Our proposed model was trained on image-level labelled data for vertebral-level classification. Another supervised DL model was trained with vertebral-level labelled data to compare the performance of the proposed model. RESULTS: The test set included 227 patients with normal (n = 117, 52%) or VCFs (n = 110, 48%). For a fair comparison of the two models, we compared sensitivities with the same specificities of the proposed model and the vertebral-level supervised model. The specificity for overall L1-L5 performance was 0.981. The proposed model may outperform the vertebral-level supervised model with sensitivities of 0.770 vs 0.705 (p = 0.080), respectively. For vertebral-level analysis, the specificities for each L1-L5 were 0.974, 0.973, 0.970, 0.991, and 0.995, respectively. The proposed model yielded the same or better sensitivity than the vertebral-level supervised model in L1 (0.750 vs 0.694, p = 0.480), L3 (0.793 vs 0.586, p < 0.05), L4 (0.833 vs 0.667, p = 0.480), and L5 (0.600 vs 0.600, p = 1.000), respectively. The proposed model showed lower sensitivity than the vertebral-level supervised model for L2, but there was no significant difference (0.775 vs 0.825, p = 0.617). CONCLUSIONS: The proposed model may have a comparable or better performance than the supervised model in vertebral-level VCF classification. CLINICAL RELEVANCE STATEMENT: Vertebral-level vertebral compression fracture classification aids in devising patient-specific treatment plans by identifying the precise vertebrae affected by compression fractures. KEY POINTS: • Our proposed weakly supervised method may have comparable or better performance than the supervised method for vertebral-level vertebral compression fracture classification. • The weakly supervised model could have classified cases with multiple vertebral compression fractures at the vertebral-level, even if the model was trained with image-level labels. • Our proposed method could help reduce radiologists' labour because it enables vertebral-level classification from image-level labels.

Deep-Learning-Aided Evaluation of Spondylolysis Imaged with Ultrashort Echo Time Magnetic Resonance Imaging.

Isthmic spondylolysis results in fracture of pars interarticularis of the lumbar spine, found in as many as half of adolescent athletes with persistent low back pain. While computed tomography (CT) is the gold standard for the diagnosis of spondylolysis, the use of ionizing radiation near reproductive organs in young subjects is undesirable. While magnetic resonance imaging (MRI) is preferable, it has lowered sensitivity for detecting the condition. Recently, it has been shown that ultrashort echo time (UTE) MRI can provide markedly improved bone contrast compared to conventional MRI. To take UTE MRI further, we developed supervised deep learning tools to generate (1) CT-like images and (2) saliency maps of fracture probability from UTE MRI, using ex vivo preparation of cadaveric spines. We further compared quantitative metrics of the contrast-to-noise ratio (CNR), mean squared error (MSE), peak signal-to-noise ratio (PSNR), and structural similarity index (SSIM) between UTE MRI (inverted to make the appearance similar to CT) and CT and between CT-like images and CT. Qualitative results demonstrated the feasibility of successfully generating CT-like images from UTE MRI to provide easier interpretability for bone fractures thanks to improved image contrast and CNR. Quantitatively, the mean CNR of bone against defect-filled tissue was 35, 97, and 146 for UTE MRI, CT-like, and CT images, respectively, being significantly higher for CT-like than UTE MRI images. For the image similarity metrics using the CT image as the reference, CT-like images provided a significantly lower mean MSE (0.038 vs. 0.0528), higher mean PSNR (28.6 vs. 16.5), and higher SSIM (0.73 vs. 0.68) compared to UTE MRI images. Additionally, the saliency maps enabled quick detection of the location with probable pars fracture by providing visual cues to the reader. This proof-of-concept study is limited to the data from ex vivo samples, and additional work in human subjects with spondylolysis would be necessary to refine the models for clinical use. Nonetheless, this study shows that the utilization of UTE MRI and deep learning tools could be highly useful for the evaluation of isthmic spondylolysis.

Development and validation of a hybrid deep learning-machine learning approach for severity assessment of COVID-19 and other pneumonias.

The Coronavirus Disease 2019 (COVID-19) is transitioning into the endemic phase. Nonetheless, it is crucial to remain mindful that pandemics related to infectious respiratory diseases (IRDs) can emerge unpredictably. Therefore, we aimed to develop and validate a severity assessment model for IRDs, including COVID-19, influenza, and novel influenza, using CT images on a multi-centre data set. Of the 805 COVID-19 patients collected from a single centre, 649 were used for training and 156 were used for internal validation (D1). Additionally, three external validation sets were obtained from 7 cohorts: 1138 patients with COVID-19 (D2), and 233 patients with influenza and novel influenza (D3). A hybrid model, referred to as Hybrid-DDM, was constructed by combining two deep learning models and a machine learning model. Across datasets D1, D2, and D3, the Hybrid-DDM exhibited significantly improved performance compared to the baseline model. The areas under the receiver operating curves (AUCs) were 0.830 versus 0.767 (p = 0.036) in D1, 0.801 versus 0.753 (p < 0.001) in D2, and 0.774 versus 0.668 (p < 0.001) in D3. This study indicates that the Hybrid-DDM model, trained using COVID-19 patient data, is effective and can also be applicable to patients with other types of viral pneumonia.

Deep learning referral suggestion and tumour discrimination using explainable artificial intelligence applied to multiparametric MRI.

OBJECTIVES: An appropriate and fast clinical referral suggestion is important for intra-axial mass-like lesions (IMLLs) in the emergency setting. We aimed to apply an interpretable deep learning (DL) system to multiparametric MRI to obtain clinical referral suggestion for IMLLs, and to validate it in the setting of nontraumatic emergency neuroradiology. METHODS: A DL system was developed in 747 patients with IMLLs ranging 30 diseases who underwent pre- and post-contrast T1-weighted (T1CE), FLAIR, and diffusion-weighted imaging (DWI). A DL system that segments IMLLs, classifies tumourous conditions, and suggests clinical referral among surgery, systematic work-up, medical treatment, and conservative treatment, was developed. The system was validated in an independent cohort of 130 emergency patients, and performance in referral suggestion and tumour discrimination was compared with that of radiologists using receiver operating characteristics curve, precision-recall curve analysis, and confusion matrices. Multiparametric interpretable visualisation of high-relevance regions from layer-wise relevance propagation overlaid on contrast-enhanced T1WI and DWI was analysed. RESULTS: The DL system provided correct referral suggestions in 94 of 130 patients (72.3%) and performed comparably to radiologists (accuracy 72.6%, McNemar test; p = .942). For distinguishing tumours from non-tumourous conditions, the DL system (AUC, 0.90 and AUPRC, 0.94) performed similarly to human readers (AUC, 0.81~0.92, and AUPRC, 0.88~0.95). Solid portions of tumours showed a high overlap of relevance, but non-tumours did not (Dice coefficient 0.77 vs. 0.33, p < .001), demonstrating the DL's decision. CONCLUSIONS: Our DL system could appropriately triage patients using multiparametric MRI and provide interpretability through multiparametric heatmaps, and may thereby aid neuroradiologic diagnoses in emergency settings. CLINICAL RELEVANCE STATEMENT: Our AI triages patients with raw MRI images to clinical referral pathways in brain intra-axial mass-like lesions. We demonstrate that the decision is based on the relative relevance between contrast-enhanced T1-weighted and diffusion-weighted images, providing explainability across multiparametric MRI data. KEY POINTS: • A deep learning (DL) system using multiparametric MRI suggested clinical referral to patients with intra-axial mass-like lesions (IMLLs) similar to radiologists (accuracy 72.3% vs. 72.6%). • In the differentiation of tumourous and non-tumourous conditions, the DL system (AUC, 0.90) performed similar with radiologists (AUC, 0.81-0.92). • The DL's decision basis for differentiating tumours from non-tumours can be quantified using multiparametric heatmaps obtained via the layer-wise relevance propagation method.

Intelligent noninvasive meningioma grading with a fully automatic segmentation using interpretable multiparametric deep learning.

OBJECTIVES: To establish a robust interpretable multiparametric deep learning (DL) model for automatic noninvasive grading of meningiomas along with segmentation. METHODS: In total, 257 patients with pathologically confirmed meningiomas (162 low-grade, 95 high-grade) who underwent a preoperative brain MRI, including T2-weighted (T2) and contrast-enhanced T1-weighted images (T1C), were included in the institutional training set. A two-stage DL grading model was constructed for segmentation and classification based on multiparametric three-dimensional U-net and ResNet. The models were validated in the external validation set consisting of 61 patients with meningiomas (46 low-grade, 15 high-grade). Relevance-weighted Class Activation Mapping (RCAM) method was used to interpret the DL features contributing to the prediction of the DL grading model. RESULTS: On external validation, the combined T1C and T2 model showed a Dice coefficient of 0.910 in segmentation and the highest performance for meningioma grading compared to the T2 or T1C only models, with an area under the curve (AUC) of 0.770 (95% confidence interval: 0.644-0.895) and accuracy, sensitivity, and specificity of 72.1%, 73.3%, and 71.7%, respectively. The AUC and accuracy of the combined DL grading model were higher than those of the human readers (AUCs of 0.675-0.690 and accuracies of 65.6-68.9%, respectively). The RCAM of the DL grading model showed activated maps at the surface regions of meningiomas indicating that the model recognized the features at the tumor margin for grading. CONCLUSIONS: An interpretable multiparametric DL model combining T1C and T2 can enable fully automatic grading of meningiomas along with segmentation. KEY POINTS: • The multiparametric DL model showed robustness in grading and segmentation on external validation. • The diagnostic performance of the combined DL grading model was higher than that of the human readers. • The RCAM interpreted that DL grading model recognized the meaningful features at the tumor margin for grading.

Ultrathin crystalline-silicon-based strain gauges with deep learning algorithms for silent speech interfaces.

A wearable silent speech interface (SSI) is a promising platform that enables verbal communication without vocalization. The most widely studied methodology for SSI focuses on surface electromyography (sEMG). However, sEMG suffers from low scalability because of signal quality-related issues, including signal-to-noise ratio and interelectrode interference. Hence, here, we present a novel SSI by utilizing crystalline-silicon-based strain sensors combined with a 3D convolutional deep learning algorithm. Two perpendicularly placed strain gauges with minimized cell dimension (<0.1 mm2) could effectively capture the biaxial strain information with high reliability. We attached four strain sensors near the subject's mouths and collected strain data of unprecedently large wordsets (100 words), which our SSI can classify at a high accuracy rate (87.53%). Several analysis methods were demonstrated to verify the system's reliability, as well as the performance comparison with another SSI using sEMG electrodes with the same dimension, which exhibited a relatively low accuracy rate (42.60%).

Small Bowel Detection for Wireless Capsule Endoscopy Using Convolutional Neural Networks with Temporal Filtering.

By automatically classifying the stomach, small bowel, and colon, the reading time of the wireless capsule endoscopy (WCE) can be reduced. In addition, it is an essential first preprocessing step to localize the small bowel in order to apply automated small bowel lesion detection algorithms based on deep learning. The purpose of the study was to develop an automated small bowel detection method from long untrimmed videos captured from WCE. Through this, the stomach and colon can also be distinguished. The proposed method is based on a convolutional neural network (CNN) with a temporal filtering on the predicted probabilities from the CNN. For CNN, we use a ResNet50 model to classify three organs including stomach, small bowel, and colon. The hybrid temporal filter consisting of a Savitzky-Golay filter and a median filter is applied to the temporal probabilities for the "small bowel" class. After filtering, the small bowel and the other two organs are differentiated with thresholding. The study was conducted on dataset of 200 patients (100 normal and 100 abnormal WCE cases), which was divided into a training set of 140 cases, a validation set of 20 cases, and a test set of 40 cases. For the test set of 40 patients (20 normal and 20 abnormal WCE cases), the proposed method showed accuracy of 99.8% in binary classification for the small bowel. Transition time errors for gastrointestinal tracts were only 38.8 ± 25.8 seconds for the transition between stomach and small bowel and 32.0 ± 19.1 seconds for the transition between small bowel and colon, compared to the ground truth organ transition points marked by two experienced gastroenterologists.

Fully automatic quantification of transient severe respiratory motion artifact of gadoxetate disodium-enhanced MRI during arterial phase.

PURPOSE: It is important to fully automate the evaluation of gadoxetate disodium-enhanced arterial phase images because the efficient quantification of transient severe motion artifacts can be used in a variety of applications. Our study proposes a fully automatic evaluation method of motion artifacts during the arterial phase of gadoxetate disodium-enhanced MR imaging. METHODS: The proposed method was based on the construction of quality-aware features to represent the motion artifact using MR image statistics and multidirectional filtered coefficients. Using the quality-aware features, the method calculated quantitative quality scores of gadoxetate disodium-enhanced images fully automatically. The performance of our proposed method, as well as two other methods, was acquired by correlating scores against subjective scores from radiologists based on the 5-point scale and binary evaluation. The subjective scores evaluated by two radiologists were severity scores of motion artifacts in the evaluation set on a scale of 1 (no motion artifacts) to 5 (severe motion artifacts). RESULTS: Pearson's linear correlation coefficient (PLCC) and Spearman's rank-ordered correlation coefficient (SROCC) values of our proposed method against the subjective scores were 0.9036 and 0.9057, respectively, whereas the PLCC values of two other methods were 0.6525 and 0.8243, and the SROCC values were 0.6070 and 0.8348. Also, in terms of binary quantification of transient severe respiratory motion, the proposed method achieved 0.9310 sensitivity, 0.9048 specificity, and 0.9200 accuracy, whereas the other two methods achieved 0.7586, 0.8996 sensitivities, 0.8098, 0.8905 specificities, and 0.9200, 0.9048 accuracies CONCLUSIONS: This study demonstrated the high performance of the proposed automatic quantification method in evaluating transient severe motion artifacts in arterial phase images.

Importance of CT image normalization in radiomics analysis: prediction of 3-year recurrence-free survival in non-small cell lung cancer.

OBJECTIVES: To analyze whether CT image normalization can improve 3-year recurrence-free survival (RFS) prediction performance in patients with non-small cell lung cancer (NSCLC) relative to the use of unnormalized CT images. METHODS: A total of 106 patients with NSCLC were included in the training set. For each patient, 851 radiomic features were extracted from the normalized and the unnormalized CT images, respectively. After the feature selection, random forest models were constructed with selected radiomic features and clinical features. The models were then externally validated in the test set consisting of 79 patients with NSCLC. RESULTS: The model using normalized CT images yielded better performance than the model using unnormalized CT images (with an area under the receiver operating characteristic curve of 0.802 vs 0.702, p = 0.01), with the model performing especially well among patients with adenocarcinoma (with an area under the receiver operating characteristic curve of 0.880 vs 0.720, p < 0.01). CONCLUSIONS: CT image normalization may improve prediction performance among patients with NSCLC, especially for patients with adenocarcinoma. KEY POINTS: • After CT image normalization, more radiomic features were able to be identified. • Prognostic performance in patients was improved significantly after CT image normalization compared with before the CT image normalization. • The improvement in prognostic performance following CT image normalization was superior in patients with adenocarcinoma.

Quantitative analysis of the mouth opening movement of temporomandibular joint disorder patients according to disc position using computer vision: a pilot study.

Background: Temporomandibular joint disorder (TMD), which is a broad category encompassing disc displacement, is a common condition with an increasing prevalence. This study aimed to develop an automated movement tracing algorithm for mouth opening and closing videos, and to quantitatively analyze the relationship between the results obtained using this developed system and disc position on magnetic resonance imaging (MRI). Methods: Mouth opening and closing videos were obtained with a digital camera from 91 subjects, who underwent MRI. Before video acquisition, an 8.0-mm-diameter circular sticker was attached to the center of the subject's upper and lower lips. The automated mouth opening tracing system based on computer vision was developed in two parts: (I) automated landmark detection of the upper and lower lips in acquired videos, and (II) graphical presentation of the tracing results for detected landmarks and an automatically calculated graph height (mouth opening length) and width (sideways values). The graph paths were divided into three types: straight, sideways-skewed, and limited-straight line graphs. All traced results were evaluated according to disc position groups determined using MRI. Graph height and width were compared between groups using analysis of variance (SPSS version 25.0; IBM Corp., Armonk, NY, USA). Results: Subjects with a normal disc position predominantly (85.72%) showed straight line graphs. The other two types (sideways-skewed or limited-straight line graphs) were found in 85.0% and 89.47% in the anterior disc displacement with reduction group and in the anterior disc displacement without reduction group, respectively, reflecting a statistically significant correlation (χ2=38.113, P<0.001). A statistically significant difference in graph height was found between the normal group and the anterior disc displacement without reduction group, 44.90±9.61 and 35.78±10.24 mm, respectively (P<0.05). Conclusions: The developed mouth opening tracing system was reliable. It presented objective and quantitative information about different trajectories from those associated with a normal disc position in mouth opening and closing movements. This system will be helpful to clinicians when it is difficult to obtain information through MRI.

Fat-saturated image generation from multi-contrast MRIs using generative adversarial networks with Bloch equation-based autoencoder regularization.

Obtaining multiple series of magnetic resonance (MR) images with different contrasts is useful for accurate diagnosis of human spinal conditions. However, this can be time consuming and a burden on both the patient and the hospital. We propose a Bloch equation-based autoencoder regularization generative adversarial network (BlochGAN) to generate a fat saturation T2-weighted (T2 FS) image from T1-weighted (T1-w) and T2-weighted (T2-w) images of human spine. To achieve this, our approach was to utilize the relationship between the contrasts using Bloch equation since it is a fundamental principle of MR physics and serves as a physical basis of each contrasts. BlochGAN properly generated the target-contrast images using the autoencoder regularization based on the Bloch equation to identify the physical basis of the contrasts. BlochGAN consists of four sub-networks: an encoder, a decoder, a generator, and a discriminator. The encoder extracts features from the multi-contrast input images, and the generator creates target T2 FS images using the features extracted from the encoder. The discriminator assists network learning by providing adversarial loss, and the decoder reconstructs the input multi-contrast images and regularizes the learning process by providing reconstruction loss. The discriminator and the decoder are only used in the training process. Our results demonstrate that BlochGAN achieved quantitatively and qualitatively superior performance compared to conventional medical image synthesis methods in generating spine T2 FS images from T1-w, and T2-w images.

Results of the 2020 fastMRI Challenge for Machine Learning MR Image Reconstruction.

Accelerating MRI scans is one of the principal outstanding problems in the MRI research community. Towards this goal, we hosted the second fastMRI competition targeted towards reconstructing MR images with subsampled k-space data. We provided participants with data from 7,299 clinical brain scans (de-identified via a HIPAA-compliant procedure by NYU Langone Health), holding back the fully-sampled data from 894 of these scans for challenge evaluation purposes. In contrast to the 2019 challenge, we focused our radiologist evaluations on pathological assessment in brain images. We also debuted a new Transfer track that required participants to submit models evaluated on MRI scanners from outside the training set. We received 19 submissions from eight different groups. Results showed one team scoring best in both SSIM scores and qualitative radiologist evaluations. We also performed analysis on alternative metrics to mitigate the effects of background noise and collected feedback from the participants to inform future challenges. Lastly, we identify common failure modes across the submissions, highlighting areas of need for future research in the MRI reconstruction community.

Robust performance of deep learning for automatic detection and segmentation of brain metastases using three-dimensional black-blood and three-dimensional gradient echo imaging.

OBJECTIVES: To evaluate whether a deep learning (DL) model using both three-dimensional (3D) black-blood (BB) imaging and 3D gradient echo (GRE) imaging may improve the detection and segmentation performance of brain metastases compared to that using only 3D GRE imaging. METHODS: A total of 188 patients with brain metastases (917 lesions) who underwent a brain metastasis MRI protocol including contrast-enhanced 3D BB and 3D GRE were included in the training set. DL models based on 3D U-net were constructed. The models were validated in the test set consisting of 45 patients with brain metastases (203 lesions) and 49 patients without brain metastases. RESULTS: The combined 3D BB and 3D GRE model yielded better performance than the 3D GRE model (sensitivities of 93.1% vs 76.8%, p < 0.001), and this effect was significantly stronger in subgroups with small metastases (p interaction < 0.001). For metastases < 3 mm, ≥ 3 mm and < 10 mm, and ≥ 10 mm, the sensitivities were 82.4%, 93.2%, and 100%, respectively. The combined 3D BB and 3D GRE model showed a false-positive per case of 0.59 in the test set. The combined 3D BB and 3D GRE model showed a Dice coefficient of 0.822, while 3D GRE model showed a lower Dice coefficient of 0.756. CONCLUSIONS: The combined 3D BB and 3D GRE DL model may improve the detection and segmentation performance of brain metastases, especially in detecting small metastases. KEY POINTS: • The combined 3D BB and 3D GRE model yielded better performance for the detection of brain metastases than the 3D GRE model (p < 0.001), with sensitivities of 93.1% and 76.8%, respectively. • The combined 3D BB and 3D GRE model showed a false-positive rate per case of 0.59 in the test set. • The combined 3D BB and 3D GRE model showed a Dice coefficient of 0.822, while the 3D GRE model showed a lower Dice coefficient of 0.756.

Deep model-based magnetic resonance parameter mapping network (DOPAMINE) for fast T1 mapping using variable flip angle method.

Quantitative tissue characteristics, which provide valuable diagnostic information, can be represented by magnetic resonance (MR) parameter maps using magnetic resonance imaging (MRI); however, a long scan time is necessary to acquire them, which prevents the application of quantitative MR parameter mapping to real clinical protocols. For fast MR parameter mapping, we propose a deep model-based MR parameter mapping network called DOPAMINE that combines a deep learning network with a model-based method to reconstruct MR parameter maps from undersampled multi-channel k-space data. DOPAMINE consists of two networks: 1) an MR parameter mapping network that uses a deep convolutional neural network (CNN) that estimates initial parameter maps from undersampled k-space data (CNN-based mapping), and 2) a reconstruction network that removes aliasing artifacts in the parameter maps with a deep CNN (CNN-based reconstruction) and an interleaved data consistency layer by an embedded MR model-based optimization procedure. We demonstrated the performance of DOPAMINE in brain T1 map reconstruction with a variable flip angle (VFA) model. To evaluate the performance of DOPAMINE, we compared it with conventional parallel imaging, low-rank based reconstruction, model-based reconstruction, and state-of-the-art deep-learning-based mapping methods for three different reduction factors (R = 3, 5, and 7) and two different sampling patterns (1D Cartesian and 2D Poisson-disk). Quantitative metrics indicated that DOPAMINE outperformed other methods in reconstructing T1 maps for all sampling patterns and reduction factors. DOPAMINE exhibited quantitatively and qualitatively superior performance to that of conventional methods in reconstructing MR parameter maps from undersampled multi-channel k-space data. The proposed method can thus reduce the scan time of quantitative MR parameter mapping that uses a VFA model.

Deep-learned spike representations and sorting via an ensemble of auto-encoders.

Spike sorting refers to the technique of detecting signals generated by single neurons from multi-neuron recordings and is a valuable tool for analyzing the relationships between individual neuronal activity patterns and specific behaviors. Since the precision of spike sorting affects all subsequent analyses, sorting accuracy is critical. Many semi-automatic to fully-automatic spike sorting algorithms have been developed. However, due to unsatisfactory classification accuracy, manual sorting is preferred by investigators despite the intensive time and labor costs. Thus, there still is a strong need for fully automatic spike sorting methods with high accuracy. Various machine learning algorithms have been developed for feature extraction but have yet to show sufficient accuracy for spike sorting. Here we describe a deep learning-based method for extracting features from spike signals using an ensemble of auto-encoders, each with a distinct architecture for distinguishing signals at different levels of resolution. By utilizing ensemble of auto-encoder ensemble, where shallow networks better represent overall signal structure and deep networks better represent signal details, extraction of high-dimensional representative features for improved spike sorting performance is achieved. The model was evaluated on publicly available simulated datasets and single-channel and 4-channel tetrode in vivo datasets. Our model not only classified single-channel spikes with varying degrees of feature similarities and signal to noise levels with higher accuracy, but also more precisely determined the number of source neurons compared to other machine learning methods. The model also demonstrated greater overall accuracy for spike sorting 4-channel tetrode recordings compared to single-channel recordings.

Deep-learned short tau inversion recovery imaging using multi-contrast MR images.

PURPOSE: To generate short tau, or short inversion time (TI), inversion recovery (STIR) images from three multi-contrast MR images, without additional scanning, using a deep neural network. METHODS: For simulation studies, we used multi-contrast simulation images. For in-vivo studies, we acquired knee MR images including 288 slices of T1 -weighted (T1 -w), T2 -weighted (T2 -w), gradient-recalled echo (GRE), and STIR images taken from 12 healthy volunteers. Our MR image synthesis method generates a new contrast MR image from multi-contrast MR images. We used a deep neural network to identify the complex relationships between MR images that show various contrasts for the same tissues. Our contrast-conversion deep neural network (CC-DNN) is an end-to-end architecture that trains the model to create one image from three (T1 -w, T2 -w, and GRE images). We propose a new loss function to take into account intensity differences, misregistration, and local intensity variations. The CC-DNN-generated STIR images were evaluated with four quantitative evaluation metrics, including mean squared error, peak signal-to-noise ratio (PSNR), structural similarity (SSIM), and multi-scale SSIM (MS-SSIM). Furthermore, a subjective evaluation was performed by musculoskeletal radiologists. RESULTS: Our method showed improved results in all quantitative evaluations compared with other methods and received the highest scores in subjective evaluations by musculoskeletal radiologists. CONCLUSION: This study suggests the feasibility of our method for generating STIR sequence images without additional scanning that offered a potential alternative to the STIR pulse sequence when additional scanning is limited or STIR artifacts are severe.

Accelerating Cartesian MRI by domain-transform manifold learning in phase-encoding direction.

This study developed a domain-transform framework comprising domain-transform manifold learning with an initial analytic transform to accelerate Cartesian magnetic resonance imaging (DOTA-MRI). The proposed method directly transforms undersampled Cartesian k-space data into a reconstructed image. In Cartesian undersampling, the k-space is fully or zero sampled in the data-acquisition direction (i.e., the frequency-encoding direction or the x-direction); one-dimensional (1D) inverse Fourier transform (IFT) along the x-direction on the undersampled k-space does not induce any aliasing. To exploit this, the algorithm first applies an analytic x-direction 1D IFT to the undersampled Cartesian k-space input, and subsequently transforms it into a reconstructed image using deep neural networks. The initial analytic transform (i.e., 1D IFT) allows the fully connected layers of the neural network to learn 1D global transform only in the phase-encoding direction (i.e., the y-direction) instead of 2D transform. This drastically reduces the number of parameters to be learned from O(N2) to O(N) compared with the existing manifold learning algorithm (i.e., automated transform by manifold approximation) (AUTOMAP). This enables DOTA-MRI to be applied to high-resolution MR datasets, which has previously proved difficult to implement in AUTOMAP because of the enormous memory requirements involved. After the initial analytic transform, the manifold learning phase uses a symmetric network architecture comprising three types of layers: front-end convolutional layers, fully connected layers for the 1D global transform, and back-end convolutional layers. The front-end convolutional layers take 1D IFT of the undersampled k-space (i.e., undersampled data in the intermediate domain or in the ky-x domain) as input and performs data-domain restoration. The following fully connected layers learn the 1D global transform between the ky-x domain and the image domain (i.e., the y-x domain). Finally, the back-end convolutional layers reconstruct the final image by denoising in the image domain. DOTA-MRI exhibited superior performance over nine other existing algorithms, including state-of-the-art deep learning-based algorithms. The generality of the algorithm was demonstrated by experiments conducted under various sampling ratios, datasets, and noise levels.

[The Latest Trends in Attention Mechanisms and Their Application in Medical Imaging]. / ì–´í ì ˜ 기법 및 의료 영상에의 ì ìš©ì— 관한 ìµœì‹ ë™í–¥.

Deep learning has recently achieved remarkable results in the field of medical imaging. However, as a deep learning network becomes deeper to improve its performance, it becomes more difficult to interpret the processes within. This can especially be a critical problem in medical fields where diagnostic decisions are directly related to a patient's survival. In order to solve this, explainable artificial intelligence techniques are being widely studied, and an attention mechanism was developed as part of this approach. In this paper, attention techniques are divided into two types: post hoc attention, which aims to analyze a network that has already been trained, and trainable attention, which further improves network performance. Detailed comparisons of each method, examples of applications in medical imaging, and future perspectives will be covered.

Megahertz-wave-transmitting conducting polymer electrode for device-to-device integration.

The ideal combination of high optical transparency and high electrical conductivity, especially at very low frequencies of less than the gigahertz (GHz) order, such as the radiofrequencies at which electronic devices operate (tens of kHz to hundreds of GHz), is fundamental incompatibility, which creates a barrier to the realization of enhanced user interfaces and 'device-to-device integration.' Herein, we present a design strategy for preparing a megahertz (MHz)-transparent conductor, based on a plasma frequency controlled by the electrical conductivity, with the ultimate goal of device-to-device integration through electromagnetic wave transmittance. This approach is verified experimentally using a conducting polymer, poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS), the microstructure of which is manipulated by employing a solution process. The use of a transparent conducting polymer as an electrode enables the fabrication of a fully functional touch-controlled display device and magnetic resonance imaging (MRI)-compatible biomedical monitoring device, which would open up a new paradigm for transparent conductors.

Parallel imaging in time-of-flight magnetic resonance angiography using deep multistream convolutional neural networks.

PURPOSE: To develop and evaluate a method of parallel imaging time-of-flight (TOF) MRA using deep multistream convolutional neural networks (CNNs). METHODS: A deep parallel imaging network ("DPI-net") was developed to reconstruct 3D multichannel MRA from undersampled data. It comprises 2 deep-learning networks: a network of multistream CNNs for extracting feature maps of multichannel images and a network of reconstruction CNNs for reconstructing images from the multistream network output feature maps. The images were evaluated using normalized root mean square error (NRMSE), peak signal-to-noise ratio (PSNR), and structural similarity (SSIM) values, and the visibility of blood vessels was assessed by measuring the vessel sharpness of middle and posterior cerebral arteries on axial maximum intensity projection (MIP) images. Vessel sharpness was compared using paired t tests, between DPI-net, 2 conventional parallel imaging methods (SAKE and ESPIRiT), and a deep-learning method (U-net). RESULTS: DPI-net showed superior performance in reconstructing vessel signals in both axial slices and MIP images for all reduction factors. This was supported by the quantitative metrics, with DPI-net showing the lowest NRMSE, the highest PSNR and SSIM (except R = 3.8 on sagittal MIP images, and R = 5.7 on axial slices and sagittal MIP images), and significantly higher vessel sharpness values than the other methods. CONCLUSION: DPI-net was effective in reconstructing 3D TOF MRA from highly undersampled multichannel MR data, achieving superior performance, both quantitatively and qualitatively, over conventional parallel imaging and other deep-learning methods.