Purpose: To test whether deep learning can be used to effectively reduce artifacts in MR images of the brain. Methods: In this study, a large set of images with and without motion artifacts is needed for training. It is difficult to collect training data from clinical images because it requires a lot of effort and time. We have created motion artifact images of the brain by computer simulation. As an experimental study, we obtained original images for deep learning from 20 volunteers. These original images were used to create various images of different artifacts by computer simulation and these were used the input images for deep learning. The same method was used to create test images and these images were used to compare the structural similarity (SSIM) index and peak signal-to-noise ratio (PSNR) between the input images and output images using the three denoising methods. The network models used were U-shaped fully convolutional network (U-Net), denoising convolutional neural network (DnCNN) and wide inference network and 5 layers Residual learning and batch normalization (Win5RB). Results: U-Net was the most effective model for reducing motion artifacts. The SSIM and PSNR were 0.978 and 32.5 dB. Conclusion: This is an effective method to reduce artifacts without degrading the image quality of brain MRI images.
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.
Purpose: A sheet-like X-ray phantom on which thin Iodine is uniformly coated was developed to facilitate the handling of iodined objects used in any X-ray imaging studies. Methods: The most recommendable protocol as follows: (1) prepare undiluted 240 mg/ml Iohexol-based contrast agent and drop around 1.6 g on a horizontal surface. (2) infiltrate the agent into a membrane filter (47 mm in diameter) from the secondary side. (3) one minute later, the excess liquid components should be removed by a softy absorbent paper, and the infiltrated membrane filter should be left until substantially dried. Result: The dried membrane filter can be utilized as a sheet-like X-ray phantom on which the iodine of around 2.45 mg is almost uniformly distributed per 1 cm2. However, since the iodine density is slightly higher on the periphery part of the sheet, less than 80% area of the entire sheet is recommended to be used from a viewpoint of the spatial uniformity. Conclusion: In the practical experiments, the X-ray attenuation factor can be controlled by changing the stacking number of the sheet, and the spatial size and form can also be designed by cutting the sheet. This capability is expected to improve the efficiency of any X-ray experiments and quality control works that requires iodined materials.
Purpose: The enhancement effect ratio using ethoxybenzyl (EOB) contrast is useful in the assessment of liver fibrosis. Since the enhancement effect ratio is calculated by setting a region of interest (ROI) in the liver, the ROI setting method may affect the enhancement effect ratio. One of the methods of setting the ROI in liver fibrosis evaluation is by placing the ROI in each Quinault segment, but this method requires considerable time. Therefore, it is necessary to consider a reproducible ROI setting method in contrast to the method of placing ROIs in each Quinault segment. Method: In contrast to the method of placing one ROI in each Quinault segment, we examined the method of setting four ROIs (two in the right lobe and two in the left lobe) and two ROIs (one in the right lobe and one in the left lobe). The size of the ROI was set to 1 cm2, 4 cm2, and the maximum area that fits within each placement area. Conclusion: In the ROI setting method for CEI calculation, reproducibility can be maintained by setting the number of ROIs in four locations and by setting ROIs of 4 cm2 or more.
Purpose: In gonad protection, it is difficult to identify the position from the body surface during shielding because the position and size of the ovary vary from individual to individual, and it is not possible to evaluate whether the protective equipment is correctly placed at the position of the ovary. Therefore, the position of the ovary with respect to the pelvis was clarified, and the effectiveness of gonad protection in the front and side of the hip joint was evaluated. Methods: From the image of the pelvis taken with an MRI device, the inner and outer edges of the ovary, the upper and lower edges and the long and short axes of the pelvis, and the depth of the ovary were measured, and the position of the ovary was calculated based on the ratio of the ovary to the pelvis. A pelvic schema was created, and the position of the ovary was synthesized on the schema. In addition, the shielding rate was calculated when lead-containing rubber for the protection of the gonads was used. Results: In front of the pelvis, the ovaries were present throughout the pelvic cavity. On the anterior surface, placing the shield on the caudal side up to the line connecting the centers of the left and right femoral heads had a shielding effect of about 88%. On the lateral side, shielding the pubic upper limbs from the ischial body could reduce the exposure of the unhealthy ovaries by 99%. However, when the gonad protection was placed at the height of the line connecting the anterior superior iliac spines, the shielding rate from the left and right ovarian distribution was about 13%, so the disadvantage of using the protective equipment was greater. Conclusion: For gonad protection, the presence or absence of use should be judged by using the shielding rate according to the shape of the protective equipment as an index.
Three-dimensional (3D) images of blood vessels in the human body, which are acquired by X-ray computed tomography (CT) and cone-beam CT of Angiography devices, are widely used in medical diagnosis and treatment. Using the 3DCT images of blood vessels, we created stereo-paired color vascular anatomical charts for better understanding of vascular anatomy in clinical settings, patient explanations, and student education. Since it is difficult to distinguish branches of blood vessels that show three-dimensionally complicated running such as cerebral blood vessels, we made it easier to identify them anatomically by color-coding each branch of the blood vessel. Also, by using stereo-paired images, we can see the three-dimensional blood vessel running. In the past anatomical books and vascular anatomy atlas, there was no anatomical chart of the whole body blood vessels that could be color-coded and stereoscopically viewed. We have made it possible to identify blood vessels by the stereoscopic vision of the blood vessels using this stereo-paired color anatomical chart. In addition, this vascular anatomical chart can be additionally revised according to the needs of the clinical and educational settings to be used, and the data can be converted into an electronic file so that it can be easily used in the field of radiological examination or at home through electronic media.
Purpose: To evaluate the usefulness of single-energy metal artifact reduction (SEMAR) for target delineation in brachytherapy for cervical cancer patients with metal hip implants. Material and Methods: A series of four definitive brachytherapy sessions in the same patient was analyzed. At each brachytherapy session, the identical set of computed tomography images was subjected with or without SEMAR treatment. For both SEMAR-treated and -untreated sets, five radiation oncologists delineated the high-risk clinical target volume (HR-CTV), bladder, and rectum, for which the volume, Dice coefficient, and the dose volume parameters were compared between SEMAR-treated and -untreated datasets. Results: The bladder volume was significantly greater in the SEMAR-treated datasets compared with the SEMAR-untreated datasets. Importantly, for the bladder, Dice coefficient among five radiation oncologists was significantly higher for the SEMAR-treated datasets compared with the SEMAR-untreated datasets. These effects of SEMAR treatment were not evident for HR-CTV and the rectum. Conclusions: These data indicate that SEMAR treatment contributes to improve delineation of the bladder in brachytherapy for cervical cancer patients with metal hip implants.