Low blood flow ratio is associated with hemorrhagic transformation secondary to mechanical thrombectomy in patients with acute ischemic stroke.

BACKGROUND AND PURPOSE: A significant decrease of cerebral blood flow (CBF) is a risk factor for hemorrhagic transformation (HT) in acute ischemic stroke (AIS). This study aimed to ascertain whether the ratio of different CBF thresholds derived from computed tomography perfusion (CTP) is an independent risk factor for HT after mechanical thrombectomy (MT). METHODS: A retrospective single center cohort study was conducted on patients with AIS undergoing MT at the First Affiliated Hospital of Wenzhou Medical University from August 2018 to December 2023. The perfusion parameters before thrombectomy were obtained according to CTP automatic processing software. The low blood flow ratio (LFR) was defined as the ratio of brain volume with relative CBF <20 % over volume with relative CBF <30 %. HT was evaluated on the follow-up CT images. Binary logistic regression was used to analyze the correlation between parameters that differ between the two groups with regards to HT occurrence. The predictive efficacy was assessed utilizing the receiver operating characteristic curve. RESULTS: In total, 243 patients met the inclusion criteria. During the follow-up, 46.5 % of the patients (113/243) developed HT. Compared with the Non-HT group, the HT group had a higher LFR (0.47 (0.34-0.65) vs. 0.32 (0.07-0.56); P < 0.001). According to the binary logistic regression analysis, the LFR (aOR: 6.737; 95 % CI: 1.994-22.758; P = 0.002), Hypertension history (aOR: 2.231; 95 % CI: 1.201-4.142; P = 0.011), plasma FIB levels before MT (aOR: 0.641; 95 % CI: 0.456-0.902; P = 0.011), and the mismatch ratio (aOR: 0.990; 95 % CI: 0.980-0.999; P = 0.030) were independently associated with HT secondary to MT. The area under the curve of the regression model for predicting HT was 0.741. CONCLUSION: LFR, a ratio quantified via CTP, demonstrates potential as an independent risk factor of HT secondary to MT.

MEF-UNet: An end-to-end ultrasound image segmentation algorithm based on multi-scale feature extraction and fusion.

Ultrasound image segmentation is a challenging task due to the complexity of lesion types, fuzzy boundaries, and low-contrast images along with the presence of noises and artifacts. To address these issues, we propose an end-to-end multi-scale feature extraction and fusion network (MEF-UNet) for the automatic segmentation of ultrasound images. Specifically, we first design a selective feature extraction encoder, including detail extraction stage and structure extraction stage, to precisely capture the edge details and overall shape features of the lesions. In order to enhance the representation capacity of contextual information, we develop a context information storage module in the skip-connection section, responsible for integrating information from adjacent two-layer feature maps. In addition, we design a multi-scale feature fusion module in the decoder section to merge feature maps with different scales. Experimental results indicate that our MEF-UNet can significantly improve the segmentation results in both quantitative analysis and visual effects.

MR-guided percutaneous microwave coagulation of small breast tumors.

BACKGROUND: To evaluate the technical success and patient safety of magnetic resonance-guided percutaneous microwave coagulation (MR-guided PMC) for breast malignancies. METHODS: From May 2018 to December 2019, 26 patients with breast tumors measuring 2 cm or less were recruited to participate in a prospective, single-institution clinical study. The primary endpoint of this study was the evaluation of treatment efficacy for each patient. Histochemical staining with α-nicotinamide adenine dinucleotide and reduced (NADH)-diaphorase was used to determine cell viability following and efficacy of PMC. The complications and self-reported sensations from all patients during and after ablation were also assessed. The technical success of the PMC procedure was defined when the area of the NADH-diaphorase negative region fully covered the hematoxylin-eosin (H&E) staining region in the tumor. RESULTS: All patients had a complete response to ablation with no residual carcinoma on histopathological specimen. The mean energy, ablation duration, and procedure duration per tumor were 36.0 ± 4.2 kJ, 252.9 ± 30.9 S, and 104.2 ± 13.5 min, respectively. During the ablation, 14 patients underwent prolonged ablation time, and 1 patient required adjusting of the antenna position. Eleven patients had feelings of subtle heat or swelling, and 3 patients experienced slight pain. After ablation, one patient took two painkillers because of moderate pain, and no patients had postoperative oozing or other complications after PMC. Induration around the ablation area appeared in 16 patients. CONCLUSION: MR-guided PMC of small breast tumors is feasible and could be applied in clinical practice in the future. CRITICAL RELEVANCE STATEMENT: MR-guided PMC of small breast tumors is feasible and could be applied in clinical practice in the future. KEY POINTS: • MR-guided PMC of small breast tumors is feasible. • PMC was successfully performed for all patients. • All patients were satisfied with the final cosmetic result.

Asymmetric-backed multi-frequency ultrasonic transducer for conformal tumor ablation.

OBJECTIVE: Minimally invasive ultrasound ablation transducers have been widely studied. However, conventional designs are limited by the single working frequency, restricting their conformal ablation ability (i.e. ablation size and shape controllability). METHODS: New multi-frequency ultrasonic transducer design method is proposed based on the asymmetric backing layer, which divides the transducer into non-backing-layer region (i.e. front-piezoelectric region) and backing-layer region (i.e. front-piezoelectric-backing region) with multiple local thickness mode resonant frequencies. Ablation zone can be controlled by exciting the local resonance within or between the regions, and its control flexibility is further enhanced by driven under a multi-frequency modulation signal. Experiments and calculations are combined for verifying the proposal. RESULTS: The fabricated transducer with a Y-direction asymmetric backing layer shows five resonances, with two in each region and one resonance excited in both regions. Spatial ultrasound emission is demonstrated by acoustic measurements. Tissue ablation experiments verified spatial ablation zone control, and frequency modulation driving method enables the spatial transition of ablation zone from one region to the other, generating different ablation sizes and shapes. Finally, patient-specific simulations verified the effectiveness of conformal ablation. CONCLUSION: The proposed transducer enables flexible control of ablation zone. SIGNIFICANCE: This study demonstrates a new method for conformal tumor ablation.

EPolar-UNet: An edge-attending polar UNet for automatic medical image segmentation with small datasets.

BACKGROUND: Medical image segmentation is one of the most key steps in computer-aided clinical diagnosis, geometric characterization, measurement, image registration, and so forth. Convolutional neural networks especially UNet and its variants have been successfully used in many medical image segmentation tasks. However, the results are limited by the deficiency in extracting high resolution edge information because of the design of the skip connections in UNet and the need for large available datasets. PURPOSE: In this paper, we proposed an edge-attending polar UNet (EPolar-UNet), which was trained on the polar coordinate system instead of classic Cartesian coordinate system with an edge-attending construction in skip connection path. METHODS: EPolar-UNet extracted the location information from an eight-stacked hourglass network as the pole for polar transformation and extracted the boundary cues from an edge-attending UNet, which consisted of a deconvolution layer and a subtraction operation. RESULTS: We evaluated the performance of EPolar-UNet across three imaging modalities for different segmentation tasks: CVC-ClinicDB dataset for polyp, ISIC-2018 dataset for skin lesion, and our private ultrasound dataset for liver tumor segmentation. Our proposed model outperformed state-of-the-art models on all three datasets and needed only 30%-60% of training data compared with the benchmark UNet model to achieve similar performances for medical image segmentation tasks. CONCLUSIONS: We proposed an end-to-end EPolar-UNet for automatic medical image segmentation and showed good performance on small datasets, which was critical in the field of medical image segmentation.

UDRSNet: An unsupervised deformable registration module based on image structure similarity.

BACKGROUND: Image registration is a challenging problem in many clinical tasks, but deep learning has made significant progress in this area over the past few years. Real-time and robust registration has been made possible by supervised transformation estimation. However, the quality of registrations using this framework depends on the quality of ground truth labels such as displacement field. PURPOSE: To propose a simple and reliable method for registering medical images based on image structure similarity in a completely unsupervised manner. METHODS: We proposed a deep cascade unsupervised deformable registration approach to align images without reliable clinical data labels. Our basic network was composed of a displacement estimation module (ResUnet) and a deformation module (spatial transformer layers). We adopted l2 -norm to regularize the deformation field instead of the traditional l1 -norm regularization. Additionally, we utilized structural similarity (ssim) estimation during the training stage to enhance the structural consistency between the deformed images and the reference images. RESULTS: Experiments results indicated that by incorporating ssim loss, our cascaded methods not only achieved higher dice score of 0.9873, ssim score of 0.9559, normalized cross-correlation (NCC) score of 0.9950, and lower relative sum of squared difference (SSD) error of 0.0313 on CT images, but also outperformed the comparative methods on ultrasound dataset. The statistical t-test results also proved that these improvements of our method have statistical significance. CONCLUSIONS: In this study, the promising results based on diverse evaluation metrics have demonstrated that our model is simple and effective in deformable image registration (DIR). The generalization ability of the model was also verified through experiments on liver CT images and cardiac ultrasound images.

MRI non-rigid registration with tumor contraction correction for ablative margin assessment after microwave ablation of hepatocellular carcinomas.

Objective. This study aims to develop and assess a tumor contraction model, enhancing the precision of ablative margin (AM) evaluation after microwave ablation (MWA) treatment for hepatocellular carcinomas (HCCs).Approach. We utilize a probabilistic method called the coherent point drift algorithm to align pre-and post-ablation MRI images. Subsequently, a nonlinear regression method quantifies local tumor contraction induced by MWA, utilizing data from 47 HCC with viable ablated tumors in post-ablation MRI. After automatic non-rigid registration, correction for tumor contraction involves contracting the 3D contour of the warped tumor towards its center in all orientations.Main results. We evaluate the performance of our proposed method on 30 HCC patients who underwent MWA. The Dice similarity coefficient between the post-ablation liver and the warped pre-ablation livers is found to be 0.95 ± 0.01, with a mean corresponding distance between the corresponding landmarks measured at 3.25 ± 0.62 mm. Additionally, we conduct a comparative analysis of clinical outcomes assessed through MRI over a 3 month follow-up period, noting that the AM, as evaluated by our proposed method, accurately detects residual tumor after MWA.Significance. Our proposed method showcases a high level of accuracy in MRI liver registration and AM assessment following ablation treatment. It introduces a potentially approach for predicting incomplete ablations and gauging treatment success.

Time multiscale regularization for nonlinear image registration.

Regularization-based methods are commonly used for image registration. However, fixed regularizers have limitations in capturing details and describing the dynamic registration process. To address this issue, we propose a time multiscale registration framework for nonlinear image registration in this paper. Our approach replaces the fixed regularizer with a monotone decreasing sequence, and iteratively uses the residual of the previous step as the input for registration. Particularly, first, we introduce a dynamically varying regularization strategy that updates regularizers at each iteration and incorporates them with a multiscale framework. This approach guarantees an overall smooth deformation field in the initial stage of registration and fine-tunes local details as the images become more similar. We then deduce convergence analysis under certain conditions on the regularizers and parameters. Further, we introduce a TV-like regularizer to demonstrate the efficiency of our method. Finally, we compare our proposed multiscale algorithm with some existing methods on both synthetic images and pulmonary computed tomography (CT) images. The experimental results validate that our proposed algorithm outperforms the compared methods, especially in preserving details during image registration with sharp structures.

US of thyroid nodules: can AI-assisted diagnostic system compete with fine needle aspiration?

OBJECTIVES: Artificial intelligence (AI) systems can diagnose thyroid nodules with similar or better performance than radiologists. Little is known about how this performance compares with that achieved through fine needle aspiration (FNA). This study aims to compare the diagnostic yields of FNA cytopathology alone and combined with BRAFV600E mutation analysis and an AI diagnostic system. METHODS: The ultrasound images of 637 thyroid nodules were collected in three hospitals. The diagnostic efficacies of an AI diagnostic system, FNA-based cytopathology, and BRAFV600E mutation analysis were evaluated in terms of sensitivity, specificity, accuracy, and the κ coefficient with respect to the gold standard, defined by postsurgical pathology and consistent benign outcomes from two combined FNA and mutation analysis examinations performed with a half-year interval. RESULTS: The malignancy threshold for the AI system was selected according to the Youden index from a retrospective cohort of 346 nodules and then applied to a prospective cohort of 291 nodules. The combination of FNA cytopathology according to the Bethesda criteria and BRAFV600E mutation analysis showed no significant difference from the AI system in terms of accuracy for either cohort in our multicenter study. In addition, for 45 included indeterminate Bethesda category III and IV nodules, the accuracy, sensitivity, and specificity of the AI system were 84.44%, 95.45%, and 73.91%, respectively. CONCLUSIONS: The AI diagnostic system showed similar diagnostic performance to FNA cytopathology combined with BRAFV600E mutation analysis. Given its advantages in terms of operability, time efficiency, non-invasiveness, and the wide availability of ultrasonography, it provides a new alternative for thyroid nodule diagnosis. CLINICAL RELEVANCE STATEMENT: Thyroid ultrasonic artificial intelligence shows statistically equivalent performance for thyroid nodule diagnosis to FNA cytopathology combined with BRAFV600E mutation analysis. It can be widely applied in hospitals and clinics to assist radiologists in thyroid nodule screening and is expected to reduce the need for relatively invasive FNA biopsies. KEY POINTS: • In a retrospective cohort of 346 nodules, the evaluated artificial intelligence (AI) system did not significantly differ from fine needle aspiration (FNA) cytopathology alone and combined with gene mutation analysis in accuracy. • In a prospective multicenter cohort of 291 nodules, the accuracy of the AI diagnostic system was not significantly different from that of FNA cytopathology either alone or combined with gene mutation analysis. • For 45 indeterminate Bethesda category III and IV nodules, the AI system did not perform significantly differently from BRAFV600E mutation analysis.

MTANet: Multi-Task Attention Network for Automatic Medical Image Segmentation and Classification.

Medical image segmentation and classification are two of the most key steps in computer-aided clinical diagnosis. The region of interest were usually segmented in a proper manner to extract useful features for further disease classification. However, these methods are computationally complex and time-consuming. In this paper, we proposed a one-stage multi-task attention network (MTANet) which efficiently classifies objects in an image while generating a high-quality segmentation mask for each medical object. A reverse addition attention module was designed in the segmentation task to fusion areas in global map and boundary cues in high-resolution features, and an attention bottleneck module was used in the classification task for image feature and clinical feature fusion. We evaluated the performance of MTANet with CNN-based and transformer-based architectures across three imaging modalities for different tasks: CVC-ClinicDB dataset for polyp segmentation, ISIC-2018 dataset for skin lesion segmentation, and our private ultrasound dataset for liver tumor segmentation and classification. Our proposed model outperformed state-of-the-art models on all three datasets and was superior to all 25 radiologists for liver tumor diagnosis.

Densely connected convolutional networks for ultrasound image based lesion segmentation.

Delineating lesion boundaries play a central role in diagnosing thyroid and breast cancers, making related therapy plans and evaluating therapeutic effects. However, it is often time-consuming and error-prone with limited reproducibility to manually annotate low-quality ultrasound (US) images, given high speckle noises, heterogeneous appearances, ambiguous boundaries etc., especially for nodular lesions with huge intra-class variance. It is hence appreciative but challenging for accurate lesion segmentations from US images in clinical practices. In this study, we propose a new densely connected convolutional network (called MDenseNet) architecture to automatically segment nodular lesions from 2D US images, which is first pre-trained over ImageNet database (called PMDenseNet) and then retrained upon the given US image datasets. Moreover, we also designed a deep MDenseNet with pre-training strategy (PDMDenseNet) for segmentation of thyroid and breast nodules by adding a dense block to increase the depth of our MDenseNet. Extensive experiments demonstrate that the proposed MDenseNet-based method can accurately extract multiple nodular lesions, with even complex shapes, from input thyroid and breast US images. Moreover, additional experiments show that the introduced MDenseNet-based method also outperforms three state-of-the-art convolutional neural networks in terms of accuracy and reproducibility. Meanwhile, promising results in nodular lesion segmentation from thyroid and breast US images illustrate its great potential in many other clinical segmentation tasks.

Coplanarity Constrained Ultrasound Probe Calibration Based on N-Wire Phantom.

OBJECTIVE: N-wire phantom-based ultrasound probe calibration has been used widely in many freehand tracked ultrasound imaging systems. The calibration matrix is obtained by registering the coplanar point cloud in ultrasound space and non-coplanar point cloud in tracking sensor space based on the least squares method. This method is sensitive to outliers and loses the coplanar information of the fiducial points. In this article, we describe a coplanarity-constrained calibration algorithm focusing on these issues. METHODS: We verified that the out-of-plane error along the oblique wire in the N-wire phantom followed a normal distribution and used it to remove the experimental outliers and fit the plane with the Levenberg-Marquardt algorithm. Then, we projected the points to the plane along the oblique wire. Coplanarity-constrained point cloud registration was used to calculate the transformation matrix. RESULTS: Compared with the other two commonly used methods, our method had the best calibration precision and achieved 25% and 36% improvement of the mean calibration accuracy than the closed-form solution and in-plane error method respectively at depth 16. Experiments at different depths revealed that our algorithm had better performance in our setup. CONCLUSION: Our proposed coplanarity-constrained calibration algorithm achieved significant improvement in both precision and accuracy compared with existing algorithms with the same N-wire phantom. It is expected that calibration accuracy will improve when the algorithm is applied to all other N-wire phantom-based calibration procedures.

Noise-robustness test for ultrasound breast nodule neural network models as medical devices.

Background: Deep learning technology has been widely applied to medical image analysis. But due to the limitations of its own imaging principle, ultrasound image has the disadvantages of low resolution and high Speckle Noise density, which not only hinder the diagnosis of patients' conditions but also affect the extraction of ultrasound image features by computer technology. Objective: In this study, we investigate the robustness of deep convolutional neural network (CNN) for classification, segmentation, and target detection of breast ultrasound image through random Salt & Pepper Noise and Gaussian Noise. Methods: We trained and validated 9 CNN architectures in 8617 breast ultrasound images, but tested the models with noisy test set. Then, we trained and validated 9 CNN architectures with different levels of noise in these breast ultrasound images, and tested the models with noisy test set. Diseases of each breast ultrasound image in our dataset were annotated and voted by three sonographers based on their malignancy suspiciousness. we use evaluation indexes to evaluate the robustness of the neural network algorithm respectively. Results: There is a moderate to high impact (The accuracy of the model decreased by about 5%-40%) on model accuracy when Salt and Pepper Noise, Speckle Noise, or Gaussian Noise is introduced to the images respectively. Consequently, DenseNet, UNet++ and Yolov5 were selected as the most robust model based on the selected index. When any two of these three kinds of noise are introduced into the image at the same time, the accuracy of the model will be greatly affected. Conclusions: Our experimental results reveal new insights: The variation trend of accuracy with the noise level in Each network used for classification tasks and object detection tasks has some unique characteristics. This finding provides us with a method to reveal the black-box architecture of computer-aided diagnosis (CAD) systems. On the other hand, the purpose of this study is to explore the impact of adding noise directly to the image on the performance of neural networks, which is different from the existing articles on robustness in the field of medical image processing. Consequently, it provides a new way to evaluate the robustness of CAD systems in the future.

Automated Skeletal Bone Age Assessment with Two-Stage Convolutional Transformer Network Based on X-ray Images.

Human skeletal development is continuous and staged, and different stages have various morphological characteristics. Therefore, bone age assessment (BAA) can accurately reflect the individual's growth and development level and maturity. Clinical BAA is time consuming, highly subjective, and lacks consistency. Deep learning has made considerable progress in BAA in recent years by effectively extracting deep features. Most studies use neural networks to extract global information from input images. However, clinical radiologists are highly concerned about the ossification degree in some specific regions of the hand bones. This paper proposes a two-stage convolutional transformer network to improve the accuracy of BAA. Combined with object detection and transformer, the first stage mimics the bone age reading process of the pediatrician, extracts the hand bone region of interest (ROI) in real time using YOLOv5, and proposes hand bone posture alignment. In addition, the previous information encoding of biological sex is integrated into the feature map to replace the position token in the transformer. The second stage extracts features within the ROI by window attention, interacts between different ROIs by shifting the window attention to extract hidden feature information, and penalizes the evaluation results using a hybrid loss function to ensure its stability and accuracy. The proposed method is evaluated on the data from the Pediatric Bone Age Challenge organized by the Radiological Society of North America (RSNA). The experimental results show that the proposed method achieves a mean absolute error (MAE) of 6.22 and 4.585 months on the validation and testing sets, respectively, and the cumulative accuracy within 6 and 12 months reach 71% and 96%, respectively, which is comparable to the state of the art, markedly reducing the clinical workload and realizing rapid, automatic, and high-precision assessment.

Genome-Wide Identification of the NPR1-like Gene Family in Solanum tuberosum and Functional Characterization of StNPR1 in Resistance to Ralstonia solanacearum.

The NPR1 (nonexpressor of pathogenesis-related genes 1) gene is an activator of the systemic acquisition of resistance (SAR) in plants and is one of the central factors in their response to pathogenic bacterial infestation, playing an important role in plant disease resistance. Potato (Solanum tuberosum) is a crucial non-grain crop that has been extensively studied. However, the identification and analysis of the NPR1-like gene within potato have not been understood well. In this study, a total of six NPR1-like proteins were identified in potato, and phylogenetic analysis showed that the six NPR1-like proteins in Solanum tuberosum could be divided into three major groups with NPR1-related proteins from Arabidopsis thaliana and other plants. Analysis of the exon-intron patterns and protein domains of the six NPR1-like genes from potato showed that the exon-intron patterns and protein domains of the NPR1-like genes belonging to the same Arabidopsis thaliana subfamily were similar. By performing quantitative real-time PCR (qRT-PCR) analysis, we found that six NPR1-like proteins have different expression patterns in different potato tissues. In addition, the expression of three StNPR1 genes was significantly downregulated after being infected by Ralstonia solanacearum (RS), while the difference in the expression of StNPR2/3 was insignificant. We also established potato StNPR1 overexpression lines that showed a significantly increased resistance to R. solanacearum and elevated activities of chitinase, ß-1,3-glucanase, and phenylalanine deaminase. Increased peroxidase (POD), superoxide dismutase (SOD), and catalase (CAT) activities, as well as decreased hydrogen peroxide, regulated the dynamic balance of reactive oxygen species (ROS) in the StNPR1 overexpression lines. The transgenic plants activated the expression of the genes associated with the Salicylic acid (SA) defense response but suppressed the expression of the genes associated with Jasmonic acid (JA) signaling. This resulted in resistance to Ralstonia solanacearum.

Application of advanced diffusion models from diffusion weighted imaging in a large cohort study of breast lesions.

BACKGROUND: To evaluate multiple parameters in multiple b-value diffusion-weighted imaging (DWI) in characterizing breast lesions and predicting prognostic factors and molecular subtypes. METHODS: In total, 504 patients who underwent 3-T magnetic resonance imaging (MRI) with T1-weighted dynamic contrast-enhanced (DCE) sequences, T2-weighted sequences and multiple b-value (7 values, from 0 to 3000 s/mm2) DWI were recruited. The average values of 13 parameters in 6 models were calculated and recorded. The pathological diagnosis of breast lesions was based on the latest World Health Organization (WHO) classification. RESULTS: Twelve parameters exhibited statistical significance in differentiating benign and malignant lesions. alpha demonstrated the highest sensitivity (89.5%), while sigma demonstrated the highest specificity (77.7%). The stretched-exponential model (SEM) demonstrated the highest sensitivity (90.8%), while the biexponential model demonstrated the highest specificity (80.8%). The highest AUC (0.882, 95% CI, 0.852-0.912) was achieved when all 13 parameters were combined. Prognostic factors were correlated with different parameters, but the correlation was relatively weak. Among the 6 parameters with significant differences among molecular subtypes of breast cancer, the Luminal A group and Luminal B (HER2 negative) group had relatively low values, and the HER2-enriched group and TNBC group had relatively high values. CONCLUSIONS: All 13 parameters, independent or combined, provide valuable information in distinguishing malignant from benign breast lesions. These new parameters have limited meaning for predicting prognostic factors and molecular subtypes of malignant breast tumors.

Three-Dimensional Voxel-Wise Quantitative Assessment of Imaging Features in Hepatocellular Carcinoma.

Voxel-wise quantitative assessment of typical characteristics in three-dimensional (3D) multiphase computed tomography (CT) imaging, especially arterial phase hyperenhancement (APHE) and subsequent washout (WO), is crucial for the diagnosis and therapy of hepatocellular carcinoma (HCC). However, this process is still missing in practice. Radiologists often visually estimate these features, which limit the diagnostic accuracy due to subjective interpretation and qualitative assessment. Quantitative assessment is one of the solutions to this problem. However, performing voxel-wise assessment in 3D is difficult due to the misalignments between images caused by respiratory and other physiological motions. In this paper, based on the Liver Imaging Reporting and Data System (v2018), we propose a registration-based quantitative model for the 3D voxel-wise assessment of image characteristics through multiple CT imaging phases. Specifically, we selected three phases from sequential CT imaging phases, i.e., pre-contrast phase (Pre), arterial phase (AP), delayed phase (DP), and then registered Pre and DP images to the AP image to extract and assess the major imaging characteristics. An iterative reweighted local cross-correlation was applied in the proposed registration model to construct the fidelity term for comparison of intensity features across different imaging phases, which is challenging due to their distinct intensity appearance. Experiments on clinical dataset showed that the means of dice similarity coefficient of liver were 98.6% and 98.1%, those of surface distance were 0.38 and 0.54 mm, and those of Hausdorff distance were 4.34 and 6.16 mm, indicating that quantitative estimation can be accomplished with high accuracy. For the classification of APHE, the result obtained by our method was consistent with those acquired by experts. For the WO, the effectiveness of the model was verified in terms of WO volume ratio.

Classification regularized dimensionality reduction improves ultrasound thyroid nodule diagnostic accuracy and inter-observer consistency.

PROBLEM: Convolutional Neural Networks (CNNs) for medical image analysis usually only output a probability value, providing no further information about the original image or inter-relationships between different images. Dimensionality Reduction Techniques (DRTs) are used for visualization of high dimensional medical image data, but they are not intended for discriminative classification analysis. AIM: We develop an interactive phenotype distribution field visualization system for medical images to accurately reflect the pathological characteristics of lesions and their similarity to assist radiologists in diagnosis and medical research. METHODS: We propose a novel method, Classification Regularized Uniform Manifold Approximation and Projection (UMAP) referred as CReUMAP, combining the advantages of CNN and DRT, to project the extracted feature vector fused with the malignant probability predicted by a CNN to a two-dimensional space, and then apply a spatial segmentation classifier trained on 2614 ultrasound images for prediction of thyroid nodule malignancy and guidance to radiologists. RESULTS: The CReUMAP embedding correlates well with the TI-RADS categories of thyroid nodules. The parametric version that embeds external test dataset of 303 images in presence of the training data with known pathological diagnosis improves the benign and malignant nodule diagnostic accuracy (p-value = 0.016) and confidence (p-value = 1.902 × 10-6) of eight radiologists of different experience levels significantly as well as their inter-observer agreements (kappa≥0.75). CReUMAP achieve 90.8% accuracy, 92.1% sensitivity and 88.6% specificity in test set. CONCLUSION: CReUMAP embedding is well correlated with the pathological diagnosis of thyroid nodules, and helps radiologists achieve more accurate, confident and consistent diagnosis. It allows a medical center to generate its locally adapted embedding using an already-trained classification model in an updateable manner on an ever-growing local database as long as the extracted feature vectors and predicted diagnostic probabilities of the correspondent classification model can be outputted.

Clinical evaluation of malignancy diagnosis of rare thyroid carcinomas by an artificial intelligent automatic diagnosis system.

PURPOSE: To evaluate the application value of a generally trained artificial intelligence (AI) automatic diagnosis system in the malignancy diagnosis of rare thyroid carcinomas, such as follicular thyroid carcinoma, medullary thyroid carcinoma, primary thyroid lymphoma and anaplastic thyroid carcinoma and compare the diagnostic performance with radiologists of different experience levels. METHODS: We retrospectively studied 342 patients with 378 thyroid nodules that included 196 rare malignant nodules by using postoperative pathology as the gold standard, and compared the diagnostic performances of three radiologists (one junior, one mid-level, one senior) and that of AI automatic diagnosis system. RESULTS: The accuracy of the AI system in malignancy diagnosis was 0.825, which was significantly higher than that of all three radiologists and higher than the best radiologist in this study by a margin of 0.097 with P-value of 2.252 × 10-16. The mid-level radiologist and senior radiologist had higher sensitivity (0.857 and 0.959) than that of the AI system (0.847) at the cost of having much lower specificity (0.533, 0.478 versus 0.802). The junior radiologist showed relatively balanced sensitivity and specificity (0.816 and 0.549) but both were lower than that of the AI system. CONCLUSIONS: The generally trained AI automatic diagnosis system showed high accuracy in the differential diagnosis of begin nodules and rare malignancy nodules. It may assist radiologists for screening of rare malignancy nodules that even senior radiologists are not acquainted with.

Automatic volumetric diagnosis of hepatocellular carcinoma based on four-phase CT scans with minimum extra information.

Summary: We built a deep-learning based model for diagnosis of HCC with typical images from four-phase CT and MEI, demonstrating high performance and excellent efficiency. Objectives: The aim of this study was to develop a deep-learning-based model for the diagnosis of hepatocellular carcinoma. Materials and methods: This clinical retrospective study uses CT scans of liver tumors over four phases (non-enhanced phase, arterial phase, portal venous phase, and delayed phase). Tumors were diagnosed as hepatocellular carcinoma (HCC) and non-hepatocellular carcinoma (non-HCC) including cyst, hemangioma (HA), and intrahepatic cholangiocarcinoma (ICC). A total of 601 liver lesions from 479 patients (56 years ± 11 [standard deviation]; 350 men) are evaluated between 2014 and 2017 for a total of 315 HCCs and 286 non-HCCs including 64 cysts, 178 HAs, and 44 ICCs. A total of 481 liver lesions were randomly assigned to the training set, and the remaining 120 liver lesions constituted the validation set. A deep learning model using 3D convolutional neural network (CNN) and multilayer perceptron is trained based on CT scans and minimum extra information (MEI) including text input of patient age and gender as well as automatically extracted lesion location and size from image data. Fivefold cross-validations were performed using randomly split datasets. Diagnosis accuracy and efficiency of the trained model were compared with that of the radiologists using a validation set on which the model showed matched performance to the fivefold average. Student's t-test (T-test) of accuracy between the model and the two radiologists was performed. Results: The accuracy for diagnosing HCCs of the proposed model was 94.17% (113 of 120), significantly higher than those of the radiologists, being 90.83% (109 of 120, p-value = 0.018) and 83.33% (100 of 120, p-value = 0.002). The average time analyzing each lesion by our proposed model on one Graphics Processing Unit was 0.13 s, which was about 250 times faster than that of the two radiologists who needed, on average, 30 s and 37.5 s instead. Conclusion: The proposed model trained on a few hundred samples with MEI demonstrates a diagnostic accuracy significantly higher than the two radiologists with a classification runtime about 250 times faster than that of the two radiologists and therefore could be easily incorporated into the clinical workflow to dramatically reduce the workload of radiologists.