Journal of X-Ray Science and Technology - Volume 30, issue 5

Authors: Zhuang, Yan | Rahman, Md Fashiar | Wen, Yuxin | Pokojovy, Michael | McCaffrey, Peter | Vo, Alexander | Walser, Eric | Moen, Scott | Xu, Honglun | Tseng, Tzu-Liang (Bill)

Article Type: Research Article

Abstract: BACKGROUND: With the emergence of continuously mutating variants of coronavirus, it is urgent to develop a deep learning model for automatic COVID-19 diagnosis at early stages from chest X-ray images. Since laboratory testing is time-consuming and requires trained laboratory personal, diagnosis using chest X-ray (CXR) is a befitting option. OBJECTIVE: In this study, we proposed an interpretable multi-task system for automatic lung detection and COVID-19 screening in chest X-rays to find an alternate method of testing which are reliable, fast and easily accessible, and able to generate interpretable predictions that are strongly correlated with radiological findings. …METHODS: The proposed system consists of image preprocessing and an unsupervised machine learning (UML) algorithm for lung region detection, as well as a truncated CNN model based on deep transfer learning (DTL) to classify chest X-rays into three classes of COVID-19, pneumonia, and normal. The Grad-CAM technique was applied to create class-specific heatmap images in order to establish trust in the medical AI system. RESULTS: Experiments were performed with 15,884 frontal CXR images to show that the proposed system achieves an accuracy of 91.94% in a test dataset with 2,680 images including a sensitivity of 94.48% on COVID-19 cases, a specificity of 88.46% on normal cases, and a precision of 88.01% on pneumonia cases. Our system also produced state-of-the-art outcomes with a sensitivity of 97.40% on public test data and 88.23% on a previously unseen clinical data (1,000 cases) for binary classification of COVID-19-positive and COVID-19-negative films. CONCLUSION: Our automatic computerized evaluation for grading lung infections exhibited sensitivity comparable to that of radiologist interpretation in clinical applicability. Therefore, the proposed solution can be used as one element of patient evaluation along with gold-standard clinical and laboratory testing. Show more

Keywords: COVID-19, chest X-ray, lung detection, transfer learning, multi-task system, clinical applicability

DOI: 10.3233/XST-221151

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 847-862, 2022

Authors: Davis, Graham R.

Article Type: Research Article

Abstract: BACKGROUND: Beam-hardening in tomography with polychromatic X-ray sources results from the nonlinear relationship between the amount of substance in the X-ray beam and attenuation. Simple linearisation curves can be derived with the use of an appropriate step wedge, however, this does not yield good results when different materials are present whose relationships between X-ray attenuation and energy are very different. OBJECTIVE: To develop a more accurate method of beam-hardening correction for two-phase samples, particularly immersed or embedded biological hard tissue. METHODS: Use of a two-dimensional step wedge is proposed in this study. This is not created …physically but is derived from published X-ray attenuation coefficients in conjunction with a modelled X-ray spectrum, optimised from X-ray attenuation measurements of a calibration carousel. To test this method, a hydroxyapatite disk was scanned twice; first dry, and then immersed in 70% ethanol solution (commonly used to preserve biological specimens). RESULTS: With simple linearisation the immersed disk reconstruction exhibited considerable residual beam hardening, with edges appearing approximately 10% more attenuating. With 2-dimensional correction, the attenuation coefficient showed only around 0.5% deviation from the dry case. CONCLUSION: Two-dimensional beam-hardening correction yielded accurate results and does not require segmentation of the two phases individually. Show more

Keywords: Beam hardening correction, microtomography, micro-CT, mineral concentration, calibration

DOI: 10.3233/XST-221178

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 863-874, 2022

Authors: Cui, Xueying | Guo, Yingting | Zhang, Xiong | Shangguan, Hong | Liu, Bin | Wang, Anhong

Article Type: Research Article

Abstract: BACKGROUND AND OBJECTIVE: Since low-dose computed tomography (LDCT) images typically have higher noise that may affect accuracy of disease diagnosis, the objective of this study is to develop and evaluate a new artifact-assisted feature fusion attention (AAFFA) network to extract and reduce image artifact and noise in LDCT images. METHODS: In AAFFA network, a feature fusion attention block is constructed for local multi-scale artifact feature extraction and progressive fusion from coarse to fine. A multi-level fusion architecture based on skip connection and attention modules is also introduced for artifact feature extraction. Specifically, long-range skip connections are used to …enhance and fuse artifact features with different depth levels. Then, the fused shallower features enter channel attention for better extraction of artifact features, and the fused deeper features are sent into pixel attention for focusing on the artifact pixel information. Besides, an artifact channel is designed to provide rich artifact features and guide the extraction of noise and artifact features. The AAPM LDCT Challenge dataset is used to train and test the network. The performance is evaluated by using both visual observation and quantitative metrics including peak signal-noise-ratio (PSNR), structural similarity index (SSIM) and visual information fidelity (VIF). RESULTS: Using AAFFA network improves the averaged PSNR/SSIM/VIF values of AAPM LDCT images from 43.4961, 0.9595, 0.3926 to 48.2513, 0.9859, 0.4589, respectively. CONCLUSIONS: The proposed AAFFA network is able to effectively reduce noise and artifacts while preserving object edges. Assessment of visual quality and quantitative index demonstrates the significant improvement compared with other image denoising methods. Show more

Keywords: Low-dose computed tomography (LDCT), artifact removal, image denoising, deep learning, multi-scale feature fusion, attention mechanism

DOI: 10.3233/XST-221149

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 875-889, 2022

Authors: Guo, Peiyuan | Wang, Zhentian | Wu, Chengpeng | Zhu, Xiaohua | Zhang, Li

Article Type: Research Article

Abstract: BACKGROUND: X-ray grating interferometry normally requires multiple steps and exposures, causing a prolonged imaging time. There is motivation to use fewer steps to reduce scanning time and complexity, while keeping fidelity of the retrieved signals. OBJECTIVE: We propose an iterative signal retrieval method, extracting attenuation, dark field contrast (DFC), and differential phase contrast (DPC) signals from two X-ray exposures. METHODS: Two shots were captured at G2 grating positions with difference of 1/4 grating period. The algorithm consists of two stages. At the first stage, amplitude of sample phase stepping curve retrieved by virtual phase stepping (VPS) …method, visibility and local phase of background phase stepping curve are used to limit the results to the proximity of the ground truth. After the second stage, three high-quality parameters, amplitude, visibility, and local phase, are retrieved through finetuning, and three signals are calculated. Simulated and real-sample experiments were conducted to validate this method. RESULTS: We used standard phase stepping result as benchmark and calculated structural similarity (SSIM) and peak signal-to-noise ratio (PSNR) between benchmark and parameters retrieved by our dual-shot method and virtual phase stepping (VPS) method. For both simulated and real-sample experiments, the SSIM and PSNR value of dual-shot method are higher than those of VPS method. For real-sample method, we also conducted a three-step PS, and the SSIM and PSNR value of dual-shot method are slightly lower than those of three-step PS. CONCLUSION: Using our dual-shot method demonstrates higher performance than other single-shot method in retrieving high-quality signals, and it also reduces radiation dose and time. Show more

Keywords: Grating interferometry, signal retrieval, medical imaging

DOI: 10.3233/XST-221162

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 891-901, 2022

Authors: Zhang, Zhi-Bin | Zou, Yong-Ning | Huang, Ye-Ling | LI, Qi

Article Type: Research Article

Abstract: Industrial computed tomography (CT) crack segmentation is a key technology in industrial CT image processing. Unfortunately, the interference of artifact and noise in CT image often bring great trouble to the crack segmentation. In order to improve the segmentation accuracy of cracks in CT images, we propose to develop and test a new crack segmentation algorithm based on linear feature enhancement by analyzing the features of cracks in CT images. Firstly, the total variational model is used to denoise the input image. Next, a Frangi multiscale filter is used to extract linear structures in the image, and then the extracted …linear structures are used to enhance the contrast of the image. Finally, the cracks in the image are detected and segmented by Otsu algorithm. By comparing with the manual segmentation results, the average intersection-over-union (IOU) reaches 86.10% and the average F1 score reaches 92.44%, which verifies the effectiveness and correctness of the algorithm developed in this study. Overall, experiments demonstrate that the new algorithm improves the accuracy of crack segmentation and it is effective applying to industry CT images. Show more

Keywords: Image segmentation, industrial CT image, hessian matrix, total variational model

DOI: 10.3233/XST-221171

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 903-917, 2022

Authors: Bourkache, Noureddine | Laghrouche, Mourad | Lahdir, Mourad | Sidhom, Sahbi

Article Type: Research Article

Abstract: BACKGROUND: Medical diagnostic support systems are important tools in the field of radiology. However, the precision obtained, during the exploitation of high homogeneity image datasets, needs to be improved. OBJECTIVE: To develop a new learning system dedicated to public health practitioners. This study presents an upgraded version dedicated to radiology experts for better clinical decision-making when diagnosing and treating the patient (CAD approach). METHODS: Our system is a hybrid approach based on a matching of semantic and visual attributes of images. It is a combination of two complementary subsystems to form the intermodal system. The first …one named α based on semantic attributes. Indexing and image retrieval based on specific keywords. The second system named β based on low-level attributes. Vectors characterizing the digital content of the image (color, texture and shape) represent images. Our image database consists of 930 X-ray images including 320 mammograms acquired from the mini-MIAS database of mammograms and 610 X-rays acquired from the Public Hospital Establishment (EPH-Rouiba Algeria). The combination of two subsystems gives rise to the intermodal system: α-subsystem offers an overall result (based on visual descriptors), then β-subsystem (low level descriptors) refines the result and increases relevance. RESULTS: Our system can perform a specific image search (in a database of images with very high homogeneity) with an accuracy of around 90% for a recall of 25%. The average (overall) accuracy of the system exceeds 70%. CONCLUSION: The results obtained are very encouraging, and demonstrate efficiency of our approach, particularly for the intermodal system. Show more

Keywords: Image indexing, medical image, data base indexing, information retrieval, content base retrieval, query by image, cancer diagnosis

DOI: 10.3233/XST-221180

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 919-939, 2022

Authors: Liao, Fengxiang | Huang, Zizhen | Xu, Rong | Luo, Zhehuang | Qi, Wanling | Fan, Bing | Yu, Juhong

Article Type: Research Article

Abstract: OBJECTIVE: To investigate 18 F-FDG PET/CT findings of tuberculous lymphadenitis and analyze the causes of misdiagnosis. METHOD: Between 2013 and 2021, a retrospective review was conducted on 22 patients at Jiangxi Provincial People’s Hospital Affiliated with Nanchang University who had lymph node tuberculosis confirmed by histology or clinical investigation. Subjective judgment and quantitative analysis were adopted. RESULTS: Out of 22 patients, 14 are male and 8 are female. The average age was 55.5 years (55.5±12.4). The most common site of lymph node tuberculosis (LNTB) is the mediastinum (41.5%), followed by the neck (24.4%) and the abdominal …cavity (21.9%). Half of the patients have more than one site affected. More than half of LNTB patients (54.5%) are concurrent with other types of TB, especially PTB. Among the 41 biggest affected lymph nodes, the average maximum diameter, minimum diameter, SUVmax and the lesion SUVmax/SUVmean liver ratio are 22.04±8.39, 16.93±6.75, 9.72±5.04 and 6.72±3.60, respectively. There is a poor correlation coefficient of 0.236 between the FDG uptake and the size of the biggest affected lymph node. Patients who are concurrent with no other TB have the significantly higher FDG uptake than patients who are concurrent with other TB (12.42 vs 8.02) (p  = 0.005). Among these cases, 6 cases (27.3%) are accurately diagnosed with LNTB, all of which have pulmonary tuberculosis as a complication. However, 16 cases (72.7%) are misdiagnosed as lymphoma (50%), sarcoidosis (13.6%), and lymph node metastasis (9%). CONCLUSIONS: This study demonstrates that 18 F-FDG PET/CT is very useful in detecting LNTB because tuberculous granulomas show significant levels of glucose uptake. It proves to be an effective method for revealing lesion extent and discovering additional lesions that morphological imaging is missed. However, 18 F-FDG PET/CT is not able to reliably distinguish LNTB from lymphoma, sarcoidosis, and metastatic lymph nodes. Nonetheless, 18 F-FDG PET/CT allows for the selection of the most optimal biopsy location, and thus has potential to detect early treatment response and distinguish between active and inactive lesions. Show more

Keywords: 18F-FDG, PET/CT, tuberculous lymphadenitis, misdiagnosis, finding

DOI: 10.3233/XST-221195

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 941-951, 2022

Authors: Chen, Chih-I | Lu, Nan-Han | Huang, Yung-Hui | Liu, Kuo-Ying | Hsu, Shih-Yen | Matsushima, Akari | Wang, Yi-Ming | Chen, Tai-Been

Article Type: Research Article

Abstract: BACKGROUND: Dividing liver organs or lesions depicting on computed tomography (CT) images could be applied to help tumor staging and treatment. However, most existing image segmentation technologies use manual or semi-automatic analysis, making the analysis process costly and time-consuming. OBJECTIVE: This research aims to develop and apply a deep learning network architecture to segment liver tumors automatically after fine tuning parameters. METHODS AND MATERIALS: The medical imaging is obtained from the International Symposium on Biomedical Imaging (ISBI), which includes 3D abdominal CT scans of 131 patients diagnosed with liver tumors. From these CT scans, there are …7,190 2D CT images along with the labeled binary images. The labeled binary images are regarded as gold standard for evaluation of the segmented results by FCN (Fully Convolutional Network). The backbones of FCN are extracted from Xception, InceptionresNetv2, MobileNetv2, ResNet18, ResNet50 in this study. Meanwhile, the parameters including optimizers (SGDM and ADAM), size of epoch, and size of batch are investigated. CT images are randomly divided into training and testing sets using a ratio of 9:1. Several evaluation indices including Global Accuracy, Mean Accuracy, Mean IoU (Intersection over Union), Weighted IoU and Mean BF Score are applied to evaluate tumor segmentation results in the testing images. RESULTS: The Global Accuracy, Mean Accuracy, Mean IoU, Weighted IoU, and Mean BF Scores are 0.999, 0.969, 0.954, 0.998, 0.962 using ResNet50 in FCN with optimizer SGDM, batch size 12, and epoch 9. It is important to fine tuning the parameters in FCN model. Top 20 FNC models enable to achieve higher tumor segmentation accuracy with Mean IoU over 0.900. The occurred frequency of InceptionresNetv2, MobileNetv2, ResNet18, ResNet50, and Xception are 9, 6, 3, 5, and 2 times. Therefore, the InceptionresNetv2 has higher performance than others. CONCLUSIONS: This study develop and test an automated liver tumor segmentation model based on FCN. Study results demonstrate that many deep learning models including InceptionresNetv2, MobileNetv2, ResNet18, ResNet50, and Xception have high potential to segment liver tumors from CT images with accuracy exceeding 90%. However, it is still difficult to accurately segment tiny and small size tumors by FCN models. Show more

Keywords: Abdominal computerized tomography (CT), fully convolutional network, liver tumor segmentation

DOI: 10.3233/XST-221194

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 953-966, 2022

Authors: Zhou, Sibo | Qiu, Yuxuan | Han, Lin | Liao, Guoliang | Zhuang, Yan | Ma, Buyun | Luo, Yan | Lin, Jiangli | Chen, Ke

Article Type: Research Article

Abstract: BACKGROUND: The intelligent diagnosis of thyroid nodules in ultrasound image is an important research issue. Automatically locating the region of interest (ROI) of thyroid nodules and providing pre-diagnosis results can help doctors to diagnose faster and more accurate. OBJECTIVES: This study aims to propose a model, which can detect multiple nodules stably and accurately in order to avoid missed detection and misjudgment. In addition, the detection speed of the model needs to be fast for real-time diagnosis in ultrasound images. METHODS: Based on the object detection technology, we propose an accurate, robust and high-speed network with …multiscale fusion strategy called Efficient-YOLO, which can realize the localization and recognition of nodules at the same time. Finally, multiple metrics are used to measure the diagnostic ability of the model. RESULTS: Experimental results conducted on 3,562 ultrasound images show that our new model greatly increases the accuracy and speed of the detection compared with the baseline model. The best mAP is 92.64%, and the fastest detection speed is 45.1 frames per second. CONCLUSIONS: This study proposed an effective method to diagnosis thyroid nodules automatically, which can meet the real-time requirements, indicating that its effectiveness and feasibility for future clinical application. Show more

Keywords: Thyroid nodules, object detection, generalization, real-time detection, YOLO

DOI: 10.3233/XST-221206

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 967-981, 2022

Authors: Priyanka, | Kadavigere, Rajagopal | Sukumar, Suresh

Article Type: Research Article

Abstract: BACKGROUND: Pediatric population is more sensitive to the effects of radiation than adults. Establishing diagnostic reference level (DRL) is an efficient dose optimization technique implemented by many countries for reducing radiation dose during Computed Tomography (CT) examinations. OBJECTIVES: To estimate radiation dose and establish a new local diagnostic reference level for CT head examination in the pediatric population. MATERIALS AND METHODS: We prospectively recruited 143 pediatric patients referred for CT head examination with age ranging from 0–5 years old. All patients had undergone CT head examination using the standard pediatric head protocol. Volumetric CT dose index …(CTDIvol) and dose length product (DLP) were recorded. The effective dose was first calculated. Then, 75th percentile of dose indices was calculated to establish DRLs. RESULTS: DRLs in terms of CTDIvol and DLP are 23.84 mGy, 555.99 mGy.cm for patients <1 years old and 28.65 mGy, 794.99 mGy.cm for patients from 1–5 years old, respectively. Mean effective doses for <1 years old patients and 1–5 years old patients are 2.91 mSv and 2.78 mSv respectively. CONCLUSION: The study concludes that DRL in terms of CTDIvol is lower but DRL in terms of DLP and the effective dose is higher compared to a few other studies which necessitate the need for dose optimization. Show more

Keywords: Diagnostic reference levels, pediatric computed tomography, volumetric computed tomography dose index, dose length product, effective dose estimation

DOI: 10.3233/XST-221172

Citation: Journal of X-Ray Science and Technology, vol. 30, no. 5, pp. 983-991, 2022