Auto-evaluation of skull radiograph accuracy using unsupervised anomaly detection.

BACKGROUND: Radiography plays an important role in medical care, and accurate positioning is essential for providing optimal quality images. Radiographs with insufficient diagnostic value are rejected, and retakes are required. However, determining the suitability of retaking radiographs is a qualitative evaluation. OBJECTIVE: To evaluate skull radiograph accuracy automatically using an unsupervised learning-based autoencoder (AE) and a variational autoencoder (VAE). In this study, we eliminated visual qualitative evaluation and used unsupervised learning to identify skull radiography retakes from the quantitative evaluation. METHODS: Five skull phantoms were imaged on radiographs, and 1,680 images were acquired. These images correspond to two categories: normal images captured at appropriate positions and images captured at inappropriate positions. This study verified the discriminatory ability of skull radiographs using anomaly detection methods. RESULTS: The areas under the curves for AE and VAE were 0.7060 and 0.6707, respectively, in receiver operating characteristic analysis. Our proposed method showed a higher discrimination ability than those of previous studies which had an accuracy of 52%. CONCLUSIONS: Our findings suggest that the proposed method has high classification accuracy in determining the suitability of retaking skull radiographs. Automation of optimal image consideration, whether or not to retake radiographs, contributes to improving operational efficiency in busy X-ray imaging operations.

Automated angular measurement for puncture angle using a computer-aided method in ultrasound-guided peripheral insertion.

Ultrasound guidance has become the gold standard for obtaining vascular access. Angle information, which indicates the entry angle of the needle into the vein, is required to ensure puncture success. Although various image processing-based methods, such as deep learning, have recently been applied to improve needle visibility, these methods have limitations, in that the puncture angle to the target organ is not measured. We aim to detect the target vessel and puncture needle and to derive the puncture angle by combining deep learning and conventional image processing methods such as the Hough transform. Median cubital vein US images were obtained from 20 healthy volunteers, and images of simulated blood vessels and needles were obtained during the puncture of a simulated blood vessel in four phantoms. The U-Net architecture was used to segment images of blood vessels and needles, and various image processing methods were employed to automatically measure angles. The experimental results indicated that the mean dice coefficients of median cubital veins, simulated blood vessels, and needles were 0.826, 0.931, and 0.773, respectively. The quantitative results of angular measurement showed good agreement between the expert and automatic measurements of the puncture angle with 0.847 correlations. Our findings indicate that the proposed method achieves extremely high segmentation accuracy and automated angular measurements. The proposed method reduces the variability and time required in manual angle measurements and presents the possibility where the operator can concentrate on delicate techniques related to the direction of the needle.

Quality control system for mammographic breast positioning using deep learning.

This study proposes a deep convolutional neural network (DCNN) classification for the quality control and validation of breast positioning criteria in mammography. A total of 1631 mediolateral oblique mammographic views were collected from an open database. We designed two main steps for mammographic verification: automated detection of the positioning part and classification of three scales that determine the positioning quality using DCNNs. After acquiring labeled mammograms with three scales visually evaluated based on guidelines, the first step was automatically detecting the region of interest of the subject part by image processing. The next step was classifying mammographic positioning accuracy into three scales using four representative DCNNs. The experimental results showed that the DCNN model achieved the best positioning classification accuracy of 0.7836 using VGG16 in the inframammary fold and a classification accuracy of 0.7278 using Xception in the nipple profile. Furthermore, using the softmax function, the breast positioning criteria could be evaluated quantitatively by presenting the predicted value, which is the probability of determining positioning accuracy. The proposed method can be quantitatively evaluated without the need for an individual qualitative evaluation and has the potential to improve the quality control and validation of breast positioning criteria in mammography.

[Evaluation of Radiograph Accuracy in Skull X-ray Images Using Deep Learning].

PURPOSE: Accurate positioning is essential for radiography, and it is especially important to maintain image reproducibility in follow-up observations. The decision on re-taking radiographs is entrusting to the individual radiological technologist. The evaluation is a visual and qualitative evaluation and there are individual variations in the acceptance criteria. In this study, we propose a method of image evaluation using a deep convolutional neural network (DCNN) for skull X-ray images. METHOD: The radiographs were obtained from 5 skull phantoms and were classified by simple network and VGG16. The discrimination ability of DCNN was verified by recognizing the X-ray projection angle and the retake of the radiograph. DCNN architectures were used with the different input image sizes and were evaluated by 5-fold cross-validation and leave-one-out cross-validation. RESULT: Using the 5-fold cross-validation, the classification accuracy was 99.75% for the simple network and 80.00% for the VGG16 in small input image sizes, and when the input image size was general image size, simple network and VGG16 showed 79.58% and 80.00%, respectively. CONCLUSION: The experimental results showed that the combination between the small input image size, and the shallow DCNN architecture was suitable for the four-category classification in X-ray projection angles. The classification accuracy was up to 99.75%. The proposed method has the potential to automatically recognize the slight projection angles and the re-taking images to the acceptance criteria. It is considered that our proposed method can contribute to feedback for re-taking images and to reduce radiation dose due to individual subjectivity.

Development of an Eye-Tracking Image Manipulation System for Angiography: A Comparative Study.

RATIONALE AND OBJECTIVES: Appropriate image manipulation of angiographic image display systems during interventional radiology is performed by radiological technologists and/or nurses given instructions from radiologists. However, appropriate images might not be displayed because of communication errors. Therefore, we developed a manipulation system that uses an eye tracker. The study aimed to determine if an angiographic image display system can be manipulated as well by using an eye tracker as by using a mouse. MATERIALS AND METHODS: An angiographic image display system using an eye tracker to calculate the gaze position on the screen and state of fixation was developed. Fourteen radiological technologists participated in an observer study by manipulating 10 images for each of 5 typical cases frequently performed in angiography, such as renal tumor, cerebral aneurysm, liver tumor, uterine bleeding, and hypersplenism. We measured the time from the start to the end of manipulating a series of images required when using the eye tracker and the conventional mouse. In this study, the statistical processing was done using Excel and R and R studio. RESULTS: The average time required for all observers for completing all cases was significantly shorter when using the eye tracker than when using the mouse (10.4 ± 2.1 s and 16.9 ± 2.6 s, respectively; p< 0.001 by paired t test). CONCLUSION: Radiologists were able to manipulate an angiographic image display system directly by using the newly developed eye tracker system without touching contact devices, such as a mouse or angiography console. Therefore, communication error could be avoided.

Simultaneous brain structure segmentation in magnetic resonance images using deep convolutional neural networks.

In brain magnetic resonance imaging (MRI) examinations, rapidly acquired two-dimensional (2D) T1-weighted sagittal slices are typically used to confirm brainstem atrophy and the presence of signals in the posterior pituitary gland. Image segmentation is essential for the automatic evaluation of chronological changes in the brainstem and pituitary gland. Thus, the purpose of our study was to use deep learning to automatically segment internal organs (brainstem, corpus callosum, pituitary, cerebrum, and cerebellum) in midsagittal slices of 2D T1-weighted images. Deep learning for the automatic segmentation of seven regions in the images was accomplished using two different methods: patch-based segmentation and semantic segmentation. The networks used for patch-based segmentation were AlexNet, GoogLeNet, and ResNet50, whereas semantic segmentation was accomplished using SegNet, VGG16-weighted SegNet, and U-Net. The precision and Jaccard index were calculated, and the extraction accuracy of the six convolutional network (DCNN) systems was evaluated. The highest precision (0.974) was obtained with the VGG16-weighted SegNet, and the lowest precision (0.506) was obtained with ResNet50. Based on the data, calculation times, and Jaccard indices obtained in this study, segmentation on a 2D image may be considered a viable and effective approach. We found that the optimal automatic segmentation of organs (brainstem, corpus callosum, pituitary, cerebrum, and cerebellum) on brain sagittal T1-weighted images could be achieved using SegNet with VGG16.

Evaluating medical images using deep convolutional neural networks: A simulated CT phantom image study.

BACKGROUND: Applied research on artificial intelligence, mainly in deep learning, is widely performed. If medical images can be evaluated using artificial intelligence, this could substantially improve examination efficiency. OBJECTIVE: We investigated an evaluation system for medical images with different noise characteristics using a deep convolutional neural network. METHODS: Simulated computed tomography images are the targets of the system. We used an AlexNet trained with natural images for the deep convolutional neural network and a support vector machine for classification. Synthetic computed tomography images with circular and rectangular signal bodies at different levels of contrast and added Gaussian noise were used for training and testing. RESULTS: Two transfer learning methods were tested: classification by a re-trained support vector machine using the AlexNet features, and a method that fine-tuned the deep convolutional neural network. Using the first method, all the test image noise levels could be classified correctly. The fine-tuning method achieved an accuracy rate of 92.6%. CONCLUSIONS: An image quality evaluation method using artificial intelligence will be useful for clinical images and different image quality indices in the future.

Comparison of medical image classification accuracy among three machine learning methods.

BACKGROUND: Low-quality medical images may influence the accuracy of the machine learning process. OBJECTIVE: This study was undertaken to compare accuracy of medical image classification among machine learning methods, as classification is a basic aspect of clinical image inspection. METHODS: Three types of machine learning methods were used, which include Support Vector Machine (SVM), Artificial Neural Network (ANN), and Convolution Neural Network (CNN). To investigate changes in accuracy related to image quality, we constructed a single dataset using two different file formats of DICOM (Digital Imaging and Communications in Medicine) and JPEG (Joint Photographic Experts Group). RESULTS: The JPEG format contains less color information and data capacity than the DICOM format. CNN classification was accurate for both datasets, whereas SVM and ANN accuracy decreased with the loss of data from DICOM to JPEG formats. CONCLUSIONS: CNN is more accurate than conventional machine learning methods that utilize the manual feature extraction.

Effectiveness of an e-Learning Platform for Image Interpretation Education of Medical Staff and Students.

Medical staff must be able to perform accurate initial interpretations of radiography to prevent diagnostic errors. Education in medical image interpretation is an ongoing need that is addressed by text-based and e-learning platforms. The effectiveness of these methods has been previously reported. Here, we describe the effectiveness of an e-learning platform used for medical image interpretation education. Ten third-year medical students without previous experience in chest radiography interpretation were provided with e-learning instructions. Accuracy of diagnosis using chest radiography was provided before and after e-learning education. We measured detection accuracy for two image groups: nodular shadow and ground-glass shadow. We also distributed the e-learning system to the two groups and analyzed the effectiveness of education for both types of image shadow. The mean correct answer rate after the 2-week e-learning period increased from 34.5 to 72.7%. Diagnosis of the ground glass shadow improved significantly more than that of the mass shadow. Education using the e-leaning platform is effective for interpretation of chest radiography results. E-learning is particularly effective for the interpretation of chest radiography images containing ground glass shadow.

Importance of Fractional b Value for Calculating Apparent Diffusion Coefficient in DWI.

OBJECTIVE: The purpose of this study is to examine how the fractional b value affects the calculation of apparent diffusion coefficient (ADC) using DWI. The fractional b value is the point of intersection between the fast and slow components of biexponential decay in DWI. SUBJECTS AND METHODS: Human brains were imaged using multiple b values on echo-planar DWI. The ADCs of white matter, gray matter, and thalamus were calculated using the combination of b values by two-point and multipoint methods, and the characteristics of each ADC value were compared. RESULTS: When the two selected points for calculation were smaller than the fractional b value (b = 1700 s/mm2), the ADC value was 0.0007-0.0008 mm2/s, but when the two points used for calculation were greater than the fractional b value, the ADC value was 0.0003-0.0004 mm2/s. When a range of b values was included in the fast and slow components by use of the multipoint method, the ADC value showed a statistically significant increase as the number of multiple b values increased. CONCLUSION: The ADC value fluctuated when the b values used for calculation were higher than the fractional b value. Therefore, it is important to determine the fractional b value of the target tissue.

A modified undecimated discrete wavelet transform based approach to mammographic image denoising.

In this work, the authors present an effective denoising method to attempt reducing the noise in mammographic images. The method is based on using hierarchical correlation of the coefficients of discrete stationary wavelet transforms. The features of the proposed technique include iterative use of undecimated multi-directional wavelet transforms at adjacent scales. To validate the proposed method, computer simulations were conducted, followed by its applications to clinical mammograms. Mutual information originating from information theory was used as an evaluation measure for selection of an optimal wavelet basis function. We examined the performance of the proposed method by comparing it with the conventional undecimated discrete wavelet transform (UDWT) method in terms of processing time-consuming and image quality. Our results showed that with the use of the proposed method the computation time can be reduced to approximately 1/10 of the conventional UDWT method consumed. The results of visual assessment indicated that the images processed with the proposed UDWT method showed statistically significant superior image quality over those processed with the conventional UDWT method. Our research results demonstrate the superiority and effectiveness of the proposed approach.

Forced expression of suppressor of cytokine signaling 3 in T cells protects the development of concanavalin A-induced hepatitis in mice.

T cells play central roles in liver diseases, but the regulatory mechanism by cytokine signaling is not well understood. In the present study, we explored the role of SOCS3 in T cells in concanavalin A (ConA)-induced hepatitis. Mice with T-cell-specific overexpression of SOCS3 (SOCS3-cTg) showed reduced hepatic damage and improved mice survival relative to the control, an event that was associated with decreased apoptotic signals Fas and pStat1. Expression of Th1-cytokines/chemokines was decreased in SOCS3-cTg liver with reduced expression of T-bet, a Th1-transcription factor. Flow cytometric analysis of the liver lymphocytes demonstrated that activated CD4(+) T cells, cytotoxic T cells and natural killer T cells were significantly decreased in SOCS3-cTg liver with decreased expression of perforin and granzyme B, injurious molecules for hepatocyte damage. These results suggest that forced expression of SOCS3 in T cells prevents ConA-induced liver injury by inhibiting several phases of Th1 responses.

[Evaluation of irreversible compressed images in computed radiography using physical image quality measures].

Joint photographic experts group(JPEG)and JPEG2000 are widely used as image compression algorithms in medical image database systems. Compressed images have been mainly evaluated by visual assessment on acceptable compression levels in clinical studies. However, to the best of our knowledge, little work has been done to clarify image properties based on physical analysis. In this work, investigations were made to clarify image properties based on physical analysis and to discuss the major causes of degradation related to compression ratios. The physical properties of JPEG2000-compressed and JPEG-compressed images in computed radiography(CR)were compared by measuring the characteristic curve, modulation transfer function(MTF), noise power spectrum(NPS), peak signal-to-noise ratio(PSNR), contrast-to-noise ratio(CNR), and noise equivalent quanta(NEQ). In the MTF measurement, the MTFs of JPEG at high compression ratio showed pronounced degradation at all frequencies. The NPS values of JPEG2000 tend to decrease considerably compared to that of the JPEG at all frequencies with the increase of compression ratio. Furthermore, JPEG2000 images showed higher PSNR, CNR, and NEQ values than JPEG images in the same compression ratio. In these signal-to-noise ratio measurements, good reproducibility of JPEG2000 images was achieved. Overall, JPEG2000 compressed images were far superior to JPEG compressed images. In the physical properties measured, these physical analyses are useful to comprehend physical properties for each irreversible compressed image related to compression ratios in CR.