A two-step deep learning method for 3DCT-2DUS kidney registration during breathing.

This work proposed KidneyRegNet, a novel deep registration pipeline for 3D CT and 2D U/S kidney scans of free breathing, which comprises a feature network, and a 3D-2D CNN-based registration network. The feature network has handcrafted texture feature layers to reduce the semantic gap. The registration network is an encoder-decoder structure with loss of feature-image-motion (FIM), which enables hierarchical regression at decoder layers and avoids multiple network concatenation. It was first pretrained with a retrospective dataset cum training data generation strategy and then adapted to specific patient data under unsupervised one-cycle transfer learning in onsite applications. The experiment was performed on 132 U/S sequences, 39 multiple-phase CT and 210 public single-phase CT images, and 25 pairs of CT and U/S sequences. This resulted in a mean contour distance (MCD) of 0.94 mm between kidneys on CT and U/S images and MCD of 1.15 mm on CT and reference CT images. Datasets with small transformations resulted in MCDs of 0.82 and 1.02 mm, respectively. Large transformations resulted in MCDs of 1.10 and 1.28 mm, respectively. This work addressed difficulties in 3DCT-2DUS kidney registration during free breathing via novel network structures and training strategies.

End to End Unsupervised Rigid Medical Image Registration by Using Convolutional Neural Networks.

In this paper, we focus on the issue of rigid medical image registration using deep learning. Under ultrasound, the moving of some organs, e.g., liver and kidney, can be modeled as rigid motion. Therefore, when the ultrasound probe keeps stationary, the registration between frames can be modeled as rigid registration. We propose an unsupervised method with Convolutional Neural Networks. The network estimates from the input image pair the transform parameters first then the moving image is wrapped using the parameters. The loss is calculated between the registered image and the fixed image. Experiments on ultrasound data of kidney and liver verified that the method is capable of achieve higher accuracy compared with traditional methods and is much faster.

Coronary Artery Segmentation by Deep Learning Neural Networks on Computed Tomographic Coronary Angiographic Images.

Coronary artery lumen delineation, to localize and grade stenosis, is an important but tedious and challenging task for coronary heart disease evaluation. Deep learning has recently been successful applied to many applications, including medical imaging. However for small imaged objects such as coronary arteries and their segmentation, it remains a challenge. This paper investigates coronary artery lumen segmentation using 3D U-net convolutional neural networks, and tests its utility with multiple datasets on two settings. We adapted the computed tomography coronary angiography (CTCA) volumes into small patches for the networks and tuned the kernels, layers and the batch size for machine learning. Our experiment involves additional efforts to select and test various data transform, so as to reduce the problem of overfitting. Compared with traditional normalization of data, we showed that subject-specific normalization of dataset was superior to patch based normalization. The results also showed that the proposed deep learning approach outperformed other methods, evaluated by the Dice coefficients.

Advanced analyses of computed tomography coronary angiography can help discriminate ischemic lesions.

BACKGROUND: Computed tomography coronary angiography (CTCA) image analysis enables plaque characterization and non-invasive fractional flow reserve (FFR) calculation. We analyzed various parameters derived from CTCA images and evaluated their associations with ischemia. METHODS: 49 (61 lesions) patients underwent CTCA and invasive FFR. Lesions with diameter stenosis (DS) ≥ 50% were considered obstructive. CTCA image processing incorporating analytical and numerical methods were used to quantify anatomical parameters of lesion length (LL) and minimum lumen area (MLA); plaque characteristic parameters of plaque volume, low attenuation plaque (LAP) volume, dense calcium volume (DCV), normalized plaque volume (NP Vol), plaque burden, eccentricity index and napkin-ring (NR) sign; and hemodynamic parameters of resistance index, stenosis flow reserve (SFR) and FFRB. Ischemia was defined as FFR ≤ 0.8. RESULTS: Plaque burden and plaque volume were inversely related to FFR. Multivariable logistic regression analysis identified the best anatomical, plaque and hemodynamic predictors, respectively, as DS (≥50% vs <50%; OR: 8.0; 95% CI: 1.6-39.4), normalized plaque volume (NP Vol) (≥4.3 vs <4.3; OR: 3.9; 95% CI: 1.1-14.0) and NR Sign (0 vs 1; OR: 13.6; 95% CI: 1.3-146.1), and FFRB (≤0.8 vs >0.8; OR: 44.4; 95% CI: 8.8-224.8). AUC increased from 0.70 with DS as the sole predictor to 0.81 after adding NP Vol and NR Sign; further addition of FFRB increased AUC to 0.93. CONCLUSION: Normalized plaque volume, napkin-ring derived from plaque analysis, and FFRB from numerical simulations on CTCA images substantially improved discrimination of ischemic lesions, compared to assessment by DS alone.

Three-dimensional spatiotemporal features for fast content-based retrieval of focal liver lesions.

Content-based image retrieval systems for 3-D medical datasets still largely rely on 2-D image-based features extracted from a few representative slices of the image stack. Most 2 -D features that are currently used in the literature not only model a 3-D tumor incompletely but are also highly expensive in terms of computation time, especially for high-resolution datasets. Radiologist-specified semantic labels are sometimes used along with image-based 2-D features to improve the retrieval performance. Since radiological labels show large interuser variability, are often unstructured, and require user interaction, their use as lesion characterizing features is highly subjective, tedious, and slow. In this paper, we propose a 3-D image-based spatiotemporal feature extraction framework for fast content-based retrieval of focal liver lesions. All the features are computer generated and are extracted from four-phase abdominal CT images. Retrieval performance and query processing times for the proposed framework is evaluated on a database of 44 hepatic lesions comprising of five pathological types. Bull's eye percentage score above 85% is achieved for three out of the five lesion pathologies and for 98% of query lesions, at least one same type of lesion is ranked among the top two retrieved results. Experiments show that the proposed system's query processing is more than 20 times faster than other already published systems that use 2-D features. With fast computation time and high retrieval accuracy, the proposed system has the potential to be used as an assistant to radiologists for routine hepatic tumor diagnosis.

Content-based image retrieval of multiphase CT images for focal liver lesion characterization.

PURPOSE: Characterization of focal liver lesions with various imaging modalities can be very challenging in the clinical practice and is experience-dependent. The authors' aim is to develop an automatic method to facilitate the characterization of focal liver lesions (FLLs) using multiphase computed tomography (CT) images by radiologists. METHODS: A multiphase-image retrieval system is proposed to retrieve a preconstructed database of FLLs with confirmed diagnoses, which can assist radiologists' decision-making in FLL characterization. It first localizes the FLL on multiphase CT scans using a hybrid generative-discriminative FLL detection method and a nonrigid B-spline registration method. Then, it extracts the multiphase density and texture features to numerically represent the FLL. Next, it compares the query FLL with the model FLLs in the database in terms of the feature and measures their similarities using the L1-norm based similarity scores. The model FLLs are ranked by similarities and the top results are finally provided to the users for their evidence studies. RESULTS: The system was tested on a database of 69 four-phase contrast-enhanced CT scans, consisting of six classes of liver lesions, and evaluated in terms of the precision-recall curve and the Bull's Eye Percentage Score (BEP). It obtained a BEP score of 78%. Compared with any single-phase based representation, the multiphase-based representation increased the BEP scores of the system, from 63%-65% to 78%. In a pilot study, two radiologists performed characterization of FLLs without and with the knowledge of the top five retrieved results. The results were evaluated in terms of the diagnostic accuracy, the receiver operating characteristic (ROC) curve and the mean diagnostic confidence. One radiologist's accuracy improved from 75% to 92%, the area under ROC curves (AUC) from 0.85 to 0.95 (p = 0.081), and the mean diagnostic confidence from 4.6 to 7.3 (p = 0.039). The second radiologist's accuracy did not change, at 75%, with AUC increasing from 0.72 to 0.75 (p = 0.709), and the mean confidence from 4.5 to 4.9 (p = 0.607). CONCLUSIONS: Multiphase CT images can be used in content-based image retrieval for FLL's categorization and result in good performance in comparison with single-phase CT images. The proposed method has the potential to improve the radiologists' diagnostic accuracy and confidence by providing visually similar lesions with confirmed diagnoses for their interpretation of clinical studies.

Computer-aided focal liver lesion detection.

PURPOSE: Our aim is to develop an automatic method which can detect diverse focal liver lesions (FLLs) in 3D CT volumes. METHOD: A hybrid generative-discriminative framework is proposed. It first uses a generative model to describe non-lesion components and then identifies all candidate FLLs within a 3D liver volume by eliminating non-lesion components. It subsequently uses a discriminative approach to suppress false positives with the advantage of tumoroid, a novel measurement combining three shape features spherical symmetry, compactness and size. RESULTS: This method was tested on 71 abdominal CT datasets (5,854 slices from 61 patients, with 261 FLLs covering six pathological types) and evaluated using the free-response receiver operating characteristic (FROC) curves. Overall, it achieved a true positive rate of 90 % with one false positive per liver. It degenerated gently with the decrease in lesion sizes to 30 ml. It achieved a true-positive rate of 36 % when tested on the lesions less than 4 ml. The average computing time of the lesion detection is 4 min and 28 s per CT volume on a PC with 2.67 GHz CPU and 4.0 GB RAM. CONCLUSIONS: The proposed method is comparable to the radiologists' visual investigation in terms of efficiency. The tool has great potential to reduce radiologists' burden in going through thousands of images routinely.

Modeling n-furcated liver vessels from a 3-D segmented volume using hole-making and subdivision methods.

It is difficult to build an accurate and smooth liver vessel model due to the tiny size, noise, and n-furcations of vessels. To overcome these problems, we propose an n-furcation vessel tree modeling method. In this method, given a segmented volume and a point indicating the root of the vessels, centerlines and cross-sectional contours of the vessels are extracted and organized as a tree first. Then, the tree is broken up into separate branches in descending order of length, and polygonal meshes of all the branches are separately constructed from the cross-sectional contours. Finally, all the meshes are combined sequentially using our hole-making approach. Holes are made on a coarse mesh, and a final fine mesh is generated using a subdivision method. The hole-making approach with the subdivision method provides good efficiency in mesh construction as well as great flexibilities in mesh editing. Experiments show that our method can automatically construct smooth mesh models for n-furcated vessels with mean absolute error of 0.92 voxel and mean relative error of 0.17. It is promising to be used in diagnosis, analysis, and surgery simulation of liver diseases, and is able to model tubular structures with tree topology.

Segmentation of liver vasculature from contrast enhanced CT images using context-based voting.

A novel vessel context-based voting is proposed for automatic liver vasculature segmentation in CT images. It is able to conduct full vessel segmentation and recognition of multiple vasculatures effectively. The vessel context describes context information of a voxel related to vessel properties, such as intensity, saliency, direction and connectivity. Voxels are grouped to liver vasculatures hierarchically based on vessel context. They are first grouped locally into vessel branches with the advantage of a vessel junction measurement, and then grouped globally into vasculatures, which is implemented using a multiple feature point voting mechanism. The proposed method has been evaluated on 10 clinical CT datasets. Segmentation of third-order vessel trees from CT images (0.76 × 0.76 × 2.0mm) of the portal venous phase takes less than 3 min on a PC with 2.0 GHz dual core processor and the average segmentation accuracy is up to 98%.

Part-based object retrieval in cluttered environment.

A novel local structural approach, which is a sequel to our previous work, is proposed in this paper for object retrieval in a cluttered and occluded environment without identifying the outlines of an object. It works by first extracting consistent and structurally unique local neighborhood from inputs or models and then voting on the optimal matches employing dynamic programming and a novel hypercube-based indexing structure. The proposed concepts have been tested on a database with thousands of images and compared with the six nearest-neighbors shape description with superior results.

[X-ray diffraction quantitative analysis with polynomial-fit method for polycrystalline materials].

The polynomial-fit method for XRD quantitative analysis of polycrystalline materials is presented in this paper, which combines a mathematic function model with computer technology. Based on the construction of diffraction peak mathematic function model, the XRD atlases from experiments were analyzed by means of polynomial whole pattern fitting to the spectral lines using computer software, then the integral intensities of every peak and weight percentages of each phase could be obtained accurately. This paper mainly includes three parts: 1. According to the fact that the diffraction atlases of mixture are weighted superposition of the powder diffraction atlases of each component and the weight factor of each component is relevant to its volume fraction or weight percentage, the theory of polynomial whole pattern fitting is constructed and the weight factor and weight percentage can be worked out. 2. Describe the method and procedure of data acquisition and analysis. 3. Discuss the result of quantitative analysis. The polynomial-fit method not only simplifies the process of data handling, but also increases the accuracy of analytical result.