Liver CT Images Research Articles

Edge Detection을 이용한 간 혈관 추출

Hepatic vessel tree is the key structure for hepatic disease diagnosis and liver surgery planning. Especially, it is used to evaluate the donors` and recipients` liver for the LDLT(Living Donors Liver Transplantation) and estimate the volumes of left and right hepatic lobes for securing their life in the LDLT. In this study, we propose a method to apply canny edge detection that is not affected by noise to the liver images for automatic segmentation of hepatic vessels tree in contrast abdominal MDCT image. Using histograms and average pixel values of the various liver CT images, optimized parameters of the Canny algorithm are determined. It is more time-efficient to use the common parameters than to change parameters manually according to CT images. Candidates of hepatic vessels are extracted by threshold filtering around the detected the vessel edge. Finally, using a system which detects the true-negatives and the false-positives in horizontal and vertical direction, the true-negatives are added in candidate of hepatic vessels and the false-positives are removed. As a result of the process, the various hepatic vessel trees of patients are accurately reconstructed in 3D.

Evaluation of Hepatic Cystic Echinococcosis’ CT image in Xinjiang Uygur Autonomous Region based on Kolmogorov Complexity Model

Designing and developing computer-assisted image processing techniques to help doctors improve their diagnosis has received considerable interests over the past years. In this paper, we used the kolmogorov complexity model to analyze the CT images of the healthy liver and multiple daughter hydatid cysts. Before the complexity characteristic calculating, the image preprocessing methods had been used for image standardization. From the kolmogorov complexity model, complexity characteristic were calculated in order to quantify the complexity, between healthy liver and multiple daughter hydatid cysts. Then we use statistical method to analyze the complexity characteristic of those two types of images. Our preliminary results show that the complexity characteristic has statistically significant (p<0.05) to analyze these two types CT images, between the healthy liver and the multiple daughter hydatid cysts. Furthermore, the result leads us to the conclusion that the kolmogorov complexity model could use for analyze the hydatid disease and will also extend the analysis the other lesions of liver.

Adaptive statistical iterative reconstruction versus filtered back projection in the same patient: 64 channel liver CT image quality and patient radiation dose

To compare routine dose liver CT reconstructed with filtered back projection (FBP) versus low dose images reconstructed with FBP and adaptive statistical iterative reconstruction (ASIR). In this retrospective study, patients had a routine dose protocol reconstructed with FBP, and again within 17 months (median 6.1 months), had a low dose protocol reconstructed twice, with FBP and ASIR. These reconstructions were compared for noise, image quality, and radiation dose. Nineteen patients were included. (12 male, mean age 58). Noise was significantly lower in low dose images reconstructed with ASIR compared to routine dose images reconstructed with FBP (liver: p < .05, aorta: p < 0.001). Low dose FBP images were scored significantly lower for subjective image quality than low dose ASIR (2.1 ± 0.5, 3.2 ± 0.8, p < 0.001). There was no difference in subjective image quality scores between routine dose FBP images and low dose ASIR images (3.6 ± 0.5, 3.2 ± 0.8, NS).Radiation dose was 41% less for the low dose protocol (4.4 ± 2.4 mSv versus 7.5 ± 5.5 mSv, p < 0.05). Our initial results suggest low dose CT images reconstructed with ASIR may have lower measured noise, similar image quality, yet significantly less radiation dose compared with higher dose images reconstructed with FBP.

A Comprehensive Descriptor of Shape: Method and Application to Content-Based Retrieval of Similar Appearing Lesions in Medical Images

We have developed a method to quantify the shape of liver lesions in CT images and to evaluate its performance for retrieval of images with similarly-shaped lesions. We employed a machine learning method to combine several shape descriptors and defined similarity measures for a pair of shapes as a weighted combination of distances calculated based on each feature. We created a dataset of 144 simulated shapes and established several reference standards for similarity and computed the optimal weights so that the retrieval result agrees best with the reference standard. Then we evaluated our method on a clinical database consisting of 79 portal-venous-phase CT liver images, where we derived a reference standard of similarity from radiologists' visual evaluation. Normalized Discounted Cumulative Gain (NDCG) was calculated to compare this ordering with the expected ordering based on the reference standard. For the simulated lesions, the mean NDCG values ranged from 91% to 100%, indicating that our methods for combining features were very accurate in representing true similarity. For the clinical images, the mean NDCG values were still around 90%, suggesting a strong correlation between the computed similarity and the independent similarity reference derived the radiologists.

Accuracy of Multislice Liver CT and MRI for Preoperative Assessment of Colorectal Liver Metastases after Neoadjuvant Chemotherapy

Introduction: To determine the best imaging modality for preoperative detection, characterization and measurement of colorectal liver metastases (CRLM) after neoadjuvant chemotherapy (NAC). Methods: A total of 79 lesions in 15 patients with CRLM were included. Following NAC, all patients received multislice liver CT (MSCT) and magnetic resonance imaging (MRI) that were scored by two observers for lesion number, type, diameter (mm) and segmental location. Intraoperative findings, histopathology and follow-up imaging were used as reference standard for surgically treated patients; non-surgical candidates underwent follow-up imaging. Results: Lesion detection rate was similar for MSCT and MRI (76 and 80%, respectively, p = 0.648). Lesion characterization was significantly superior (p = 0.021) at MRI (89%, ĸ 0.747, p = 0.001) compared to MSCT (77%, ĸ 0.235, p = 0.005). Interobserver variability for diameter measurement was not significant at MRI (p = 0.909 [95% CI –1.245 to 1.395]), but significant at MSCT (p = 0.028 [95% CI –3.349 to –2.007]). Differences in diameter measurement were independent of observer (p = 0.131), and no statistical effect from imaging modality on diameter measurement was observed (p = 0.095). Conclusion: MRI is superior to MSCT in preoperative characterization and measurement of CRLM after NAC. Lesion detection rates for both modalities are comparable.

Hepatocellular nodules in liver cirrhosis: state of the art CT evaluation (perfusion CT/volume helical shuttle scan/dual-energy CT, etc.)

The purpose of this article is to explain the role of advanced liver CT imaging, including perfusion CT, dual-energy CT, and volume helical shuttle (VHS) scanning, with regard to its clinical applications. Perfusion CT is a promising method for calculating hepatic blood flow and portal blood flow, including microcirculation, using a color-encoded display of parameters obtained from the liver time-density curve, with iodine contrast agent. Tumor angiogenesis and assessment of the response to antiangiogenesis treatment (e.g., Sorafenib) can be analyzed by perfusion CT of the liver. VHS scan has very high temporal resolution due to the reciprocating movement employed during scanning, enabling the acquisition of 24 scans of the whole liver in the arterial dominant phase during a 40-s breath hold, and a reduction in radiation dose. Dual-energy CT enables differentiation of materials and tissues based on their CT density values, using two different energy spectra. This method includes a low tube voltage CT technique that increases the contrast enhancement of vascular structures while simultaneously reducing radiation dose. Images obtained at the preferred settings of low tube voltage and high tube current, with dose reduction in the hepatic arterial phase, are useful for detecting hypervascular hepatocellular carcinoma.

CT Color Mapping of the Arterial Enhancement Fraction of VX2 Carcinoma Implanted in Rabbit Liver: Comparison With Perfusion CT

The purpose of this study was to compare the arterial enhancement fraction (AEF) calculated at multiphasic liver CT with the hepatic perfusion index (HPI) measured with cine mode perfusion CT. Cine mode perfusion CT was performed after VX2 tumor implantation in the livers of 10 rabbits. HPI and its color map were obtained with a computer application. With raw data from cine mode perfusion CT, images were extracted in the unenhanced, arterial, and portal venous phases to simulate multiphasic liver CT. On the basis of simulated multiphasic CT images, the AEF color map was obtained with prototype software. HPI and AEF were compared for the same regions of interest in the liver parenchyma, whole liver tumor, and viable tumor portion. In the liver parenchyma, the mean HPI was 23.3% ± 2.6% (SD) and the AEF 24.4% ± 2.8%; in whole liver tumor, 73.4% ± 9.5% and 78.4% ± 10.5%; and in the viable tumor portion, 78.0% ± 7.7% and 78.3% ± 7.5%. The differences were not statistically significant (p > 0.05, Wilcoxon's signed rank test). Measurement agreement between the two parameters was moderate (Bland-Altman 95% limits of agreement, -14.9% and 19.2%), but there was a strong positive correlation between AEF and HPI (within-subject r = 0.91, p < 0.001). Functional maps of HPI and AEF correlated with the histologic findings. AEF calculated from simulated multiphasic liver CT images correlates strongly with HPI obtained at cine mode perfusion CT. Further study of the AEF is warranted to explore its value in providing hepatic perfusion information without additional radiation exposure.

デジタルファントムと臨床画像を用いた肝臓領域における低コントラスト分解能の評価

The high performance of multi detector-row computed tomography (MDCT), enables a flow dynamic study of the liver that can be carried out within 10 seconds per one phase. In addition, it has been possible to obtain a stable image quality by utilizing the auto exposure control system. However, no clear standard of minimal CT value differences and minimal tumor sizes for detecting liver tumors has ever been shown. In order to help in their standardization, we developed novel software to make an arbitrary digital phantom. The limit of space occupied lesion (SOL) detection in each image quality could be determined by a detection study using digital phantoms superimposed on liver CT images. As a result, the detection rate of simulated tumors was significantly different, in comparisons of different image qualities for the late arterial phase and in comparison between the late arterial and equilibrium phases for the same image quality. Our method could help standardize the scan conditions for performing a liver CT.

Study on the methods of image enhancement for liver CT images

Image enhancement methods for liver CT images are studied in this paper. The liver region is first segmented from the whole CT image by simply using the characteristics of the gray distribution of the liver. The segmented liver CT image is processed by direct gray stretching, logarithmic transformation and linear stretching with histogram fitting. In addition, the method of selective histogram equalization is proposed to enhance the segmented liver CT image. It is proven from the experimental results that by this two-step method, the visual effect of the segmented image can be effectively improved and the focus is obviously highlighted.

Liver segmentation by intensity analysis and anatomical information in multi-slice CT images

Quantitative assessment and essentially segmentation of liver and its tumours are of clinical importance in various procedures such as diagnosis, treatment planning, and monitoring. Moreover, segmentation of liver is the basis of further processing such as visualization, liver resection planning, and liver shape analysis. In this paper, we propose an algorithm to estimate an initial liver boundary. The proposed method consists of four steps as follows: first, we compute statistical parameters of liver's intensity range, associated with a large cross-section of liver CT image, utilizing expectation maximization (EM) algorithm. Second, by automatic extraction of ribs and segmentation of the heart, we define a ROI to confine the liver region for the next operations. Third, we propose a double thresholding approach to divide the liver intensity range into two overlapping ranges. In this case, based on a decision table, we label an object as a liver candidate or disregard it from the rest of the procedures. Finally, we employ an anatomical based rule to finalize a candidate as a liver tissue. In this case, we propose a color-map transformation scheme to convert the liver gray images into color images. In this way, we attempt to visually differentiate the liver from its surrounding tissues. We have evaluated the techniques in the presence of 14 randomly selected local datasets as well as all datasets from the MICCAI 2007 Grand Challenge workshop database. For the local datasets, the average overlap error and average volume difference were of values of 15.3 and 2.8%, respectively. In the case of the MICCAI datasets, the above values were estimated as 20.3 and -4.5%, respectively. The results reveal that the proposed technique is feasible to perform consistent initial liver borders. The boundary might be then employed in an 'Active Contour' algorithm to finalize the liver mask.

A case of primary hepatic epithelioid hemangioendothelioma with spontaneous rupture

Primary hepatic epithelioid hemangioendothelioma is a rare neoplasm of endothelial origin. The clinical manifestations are nonspecific, ranging from complete absence of symptoms to hepatic failure and death. Spontaneous rupture of a hepatic epithelioid hemangioendothelioma is an extremely rare presentation. We present a case of primary hepatic epithelioid hemangioendothelioma in a 65-year-old male patient with alcoholic liver cirrhosis. He was hospitalized due to epigastric pain and multiple liver masses on abdominal ultrasound. Dynamic liver CT imaging revealed multiple peripheral nodular enhanced mass lesions with delayed centripetal enhancement, and the adjacent collection of high-attenuation fluid along the liver capsule. Abdominal tapping revealed blood in the peritoneal cavity. Primary hepatic epithelioid hemangioendothelioma with spontaneous rupture was finally diagnosed based on a histopathologic examination revealing positive immunohistochemical staining for CD34.

Orthogonal moments based texture analysis of CT liver images

Orthogonal moments such as Zernike moments and Legendre moments have been proven to have superior feature representation capability and low information redundancy. The number of orthogonal moments to be used as features or numerical attributes to perform any application is minimal due to the orthogonal nature. However, the information possessed by each moment order needs to be analysed to identify the appropriate moment orders for the undertaken task. In this work, a statistical significance test has been performed to select the best moment orders to discriminate normal and abnormal tissues in liver images. The experimental results reveal the efficacy of the proposed features.

Use of three‐dimensional (3D) optical flow method in mapping 3D anatomic structure and tumor contours across four‐dimensional computed tomography data

A three‐dimensional (3D) optical flow program that includes a multi‐resolution feature has been developed and applied to 3D anatomic structure and gross tumor volume (GTV) contour mapping for four‐dimensional computed tomography (4D CT) data. The present study includes contour mapping for actual CT data sets from 3 patients and also for a thoracic phantom in which the displacement for each voxel was known. Of the CT data sets for the actual patients, one set was used to map lung and GTV contours over all respiration phases, and the other two were studied using only the end inspiration and end expiration phases, in which the displacements between phases were the largest. Including the residual motion in the 4D CT data and motion from table shaking, the optical flow calculation agrees with the known displacement to within 1 mm. Excluding errors not introduced by the optical flow algorithm, agreement for a displacement magnitude of 24 mm can be within 0.1 mm. The mapped contours in 4D CT images of lungs, liver, esophagus, GTV, and other structures for actual patients were acceptable to clinicians. The 3D optical flow program is a good tool for contour mapping of anatomic structure and tumor volume across 4D CT scans.PACS numbers: 87.55.D‐, 87.59.bd

Advanced fuzzy cellular neural network: Application to CT liver images

To achieve better boundary integrities and recall accuracies for segmented liver images, use of the advanced fuzzy cellular neural network (AFCNN), as a variant of the fuzzy cellular neural network (FCNN), is proposed to effectively segment CT liver images. In order to better utilize relevant contour and gray information from liver images, we have improved the FCNN [Wang S, Wang M. A new algorithm NDA based on fuzzy cellular neural networks for white blood cell detection. IEEE Trans Inform Technol Biomed, in press], which proved to be very effective for the segmentation of microscopic white blood cell images, to create the novel neural network, AFCNN. Its convergent property and global stability are proved. Based on the FCNN-based NDA algorithm [Wang S, Wang M. A new algorithm NDA based on fuzzy cellular neural networks for white blood cell detection. IEEE Trans Inform Technol Biomed, in press], we developed the AFCNN-based NDA algorithm, which we used to segment 5 CT liver images. For comparison, we also segmented the same 5 CT liver images using the FCNN-based NDA algorithm. : AFCNN has distinct advantages over FCNN in both boundary integrity and recall accuracy. In particular, the performance index Binary_rate is generally much higher for AFCNN than for FCNN when applied to CT liver images.

SU‐FF‐J‐82: Improving Soft Tissue Contrast in 4D CT Images of Liver Cancer Patients Using Deformable Image Registration Method

Purpose: To investigate image quality improvement in 4DCT images of liver cancer patients by using deformable image registration. The low soft‐tissue contrast in liver CT images is always a challenge for accurate target delineation. Method and Materials: Ten patients with liver cancers were selected in this study. These patients received 4DCT scans for radiotherapy treatment planning using 120kVp and 150mA on a GE PET/CT system. The 4DCT images were retrospectively sorted and binned into 10 equispaced phases. The end‐expiration phase was chosen to be a reference phase, and the images from the other 9 phases were co‐registered to the reference phase using an intensity‐based, automatic deformable image registration algorithm. Then the 10 matched 4DCT images were averaged to give a single, high quality reference‐phase CT image for tumor target delineation. The image quality enhancement was quantified relative to the original CT by calculating the signal‐to‐noise ratio (SNR) inside the liver region. The incremental improvement in image quality was also studied by combining fewer 4DCT data sets. Results: The image contrast in the soft tissue region is noticeably improved. SNRs inside the liver increased for all patients by a factor of at least 2.3 (average at 3.0). The improvement in image quality is not linearly proportional to the number of images averaged. Averaging 6 CTs can achieve 85% of the SNR enhancement obtained by averaging all 10 CTs of different phases. Conclusion: We developed an effective method to improve soft tissue contrast in the liver by co‐registering and combining multiple CTs within the 4DCT data set using a deformable image registration method. The resultant, high‐quality, single‐phase CT could be used for better delineation of tumor target volume and critical avoidance structures. The deformable image registration method can also map these contoured structures back to each individual phases for motion‐compensated 4DCT planning.

Liver CT image processing: A short introduction of the technical elements

In this paper, we describe the technical aspects of image analysis for liver diagnosis and treatment, including the state-of-the-art of liver image analysis and its applications. After discussion on modalities for liver image analysis, various technical elements for liver image analysis such as registration, segmentation, modeling, and computer-assisted detection are covered with examples performed with clinical data sets. Perspective in the imaging technologies is also reviewed and discussed.

Computed-aided diagnosis system for detection of liver tumor in CT liver image

AIM: To evaluate prospectively the influence of pulmonary nodule characteristics on detection performances of a computer-aided diagnosis (CAD) tool and experienced chest radiologists using multislice CT (MSCT).MATERIALS AND METHODS: MSCT scans of 20 consecutive patients were evaluated by a CAD system and two independent chest radiologists for presence of pulmonary nodules. Nodule size, position, margin, matrix characteristics, vascular and pleural attachments and reader confidence were recorded and data compared with an independent standard of reference. Statistical analysis for predictors influencing nodule detection or reader performance included chi-squared, retrograde stepwise conditional logistic regression with odds ratios and nodule detection proportion estimates (DPE), and ROC analysis.RESULTS: For 135 nodules, detection rates for CAD and readers were 76.3, 52.6 and 52.6%, respectively; false-positive rates were 0.55, 0.25 and 0.15 per examination, respectively. In consensus with CAD the reader detection rate increased to 93.3%, and the false-positive rate dropped to 0.1/scan. DPEs for nodules ≤5 mm were significantly higher for ICAD than for the readers (p<0.05). Absence of vascular attachment was the only significant predictor of nodule detection by CAD (p=0.0006–0.008). There were no predictors of nodule detection for reader consensus with CAD. In contrast, vascular attachment predicted nodule detection by the readers (p=0.0001–0.003). Reader sensitivity was higher for nodules with vascular attachment than for unattached nodules (sensitivities 0.768 and 0.369; 95% confidence intervals=0.651–0.861 and 0.253–0.498, respectively).CONCLUSION: CAD increases nodule detection rates, decreases false-positive rates and compensates for deficient reader performance in detection of smallest lesions and of nodules without vascular attachment.

Automated liver segmentation using multislice CT images

AbstractWith practical applications of multislice CT, high‐speed imaging has drastically become possible in recent years as compared with the past and imaging the liver twice during the time a patient holds his breath once has become possible in liver CT imaging using contrast media. Thus, three‐dimensional images of two time phases differing in the circulatory state of a contrast medium are obtainable without positional discrepancies. The authors propose an automated liver area extracting (segmenting) method using three‐dimensional CT images of two time phases of a contrast medium for extracting the shape of the liver. This method first constructs a two‐dimensional characteristic (feature) space having the CT values of time phases on each axis from the three‐dimensional images of a two time phase portion imaged immediately after and 10‐plus seconds after the infusion of a contrast medium. The method extracts areas showing changes in the contrast medium corresponding to the liver in this space and generates images with the liver region emphasized using these. The method then determines the liver region by performing opening and closing procedures on the liver area‐emphasized images generated. Finally, it determines the liver region by a three‐dimensional component connecting procedure using geometric shape characteristics of the blood vessels, tumors, and so on. The efficacy of using images of two time phases simultaneously has been verified by applying the proposed method to nine cases. © 2003 Wiley Periodicals, Inc. Syst Comp Jpn, 34(9): 71–82, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.10210

A high-performance computing service over the Internet for nonrigid image registration

This paper presents a novel high-performance computing (HPC) system for compute-intensive medical applications. Our system provides surgeons a framework for utilizing remote HPC resources with a secure execution environment by a public key cryptography and a high-speed data transmission by a real-time lossless data compression algorithm. Our implementation on a cluster of 64 off-the-shelf PCs with 128 processors has successfully demonstrated HPC benefits by reducing the turn-around time for the nonrigid registration of liver CT images of 512×512×159 voxels from about 17 h on a sequential system to about 12 min.

Liver tissue classification by bounded-depth threshold circuits

We present a new stochastic learning algorithm and first results of computational experiments on fragments of liver CT images. The algorithm is designed to compute a depth-three threshold circuit, where the first layer is calculated by an extension of the Perceptron algorithm by a special type of simulated annealing. The fragments of CT images are of size 119×119 with eight bit grey levels. From 348 positive (focal liver tumours) and 348 negative examples a number of hypotheses of the type w1x1+⋯+wnxn≥ϑ were calculated for n=14161. The threshold functions at levels two and three were determined by computational experiments. The circuit was tested on various sets of 50+50 additional positive and negative examples. For depth-three circuits, we obtained a correct classification of about 97%. The input to the algorithm is derived from the DICOM standard representation of CT images. The simulated annealing procedure employs a logarithmic cooling schedule c(k)=Γ/ln(k+2), where Γ is a parameter that depends on the underlying configuration space. In our experiments, the parameter Γ is chosen according to estimations of the maximum escape depth from local minima of the associated energy landscape.