[Questionnaire Survey of Radiation Protection Education in Hospitals].

PURPOSE: The present study aimed to investigate the current situation of radiation protection education for designated radiation workers in hospitals. METHODS: A web-based questionnaire survey was conducted at 1,883 hospitals nationwide with 200 or more beds. RESULTS: Responses from 186 hospitals were included in the analysis. Seven hospitals (6.7%) regulated by the Act on the Regulation of Radioisotopes and six hospitals (7.4%) regulated by only the Ordinance on Prevention of Ionizing Radiation Hazards did not implement radiation protection education. In approximately 6% of the hospitals, designated radiation workers-including physicians, nurses, and radiological technologist-did not attend the education program. The education program attendance rate of physicians was lower than that of nurses. In more than 90% of the hospitals, the frequency of the periodical education program was once every year and lecture time spanned one or less than one hour. The topics of lecture in more than 90% of the hospitals were health effects of radiation and methods of radiation protection for occupational exposure. The radiological technologist was the instructor of the education program in approximately 70% of the hospitals. CONCLUSION: The implementation of radiation protection for designated radiation workers varied from hospital to hospital, and some hospitals did not comply with laws and regulations. Effective and efficient radiation protection education models should be implemented in hospitals.

Improved detection of cholesterol gallstones using quasi-material decomposition images generated from single-energy computed tomography images via deep learning.

In this study, we developed a method for generating quasi-material decomposition (quasi-MD) images from single-energy computed tomography (SECT) images using a deep convolutional neural network (DCNN). Our aim was to improve the detection of cholesterol gallstones and to determine the clinical utility of quasi-MD images. Four thousand pairs of virtual monochromatic images (70 keV) and MD images (fat/water) of the same section, obtained via dual-energy computed tomography (DECT), were used to train the DCNN. The trained DCNN can automatically generate quasi-MD images from the SECT images. Additional SECT images were obtained from 70 patients (40 with and 30 without cholesterol gallstones) to generate quasi-MD images for testing. The presence of gallstones in this dataset was confirmed by ultrasonography. We conducted a receiver operating characteristic (ROC) observer study with three radiologists to validate the clinical utility of the quasi-MD images for detecting cholesterol gallstones. The mean area under the ROC curve for the detection of cholesterol gallstones improved from 0.867 to 0.921 (p = 0.001) when quasi-MD images were added to SECT images. The clinical utility of quasi-MD imaging for detecting cholesterol gallstones was showed. This study demonstrated that the lesion detection capability of images obtained from SECT can be improved using a DCNN trained with DECT images obtained using high-end computed tomography systems.

Development of an individual display optimization system based on deep convolutional neural network transition learning for somatostatin receptor scintigraphy.

Somatostatin receptor scintigraphy (SRS) is an essential examination for the diagnosis of neuroendocrine tumors (NETs). This study developed a method to individually optimize the display of whole-body SRS images using a deep convolutional neural network (DCNN) reconstructed by transfer learning of a DCNN constructed using Gallium-67 (67Ga) images. The initial DCNN was constructed using U-Net to optimize the display of 67Ga images (493 cases/986 images), and a DCNN with transposed weight coefficients was reconstructed for the optimization of whole-body SRS images (133 cases/266 images). A DCNN was constructed for each observer using reference display conditions estimated in advance. Furthermore, to eliminate information loss in the original image, a grayscale linear process is performed based on the DCNN output image to obtain the final linearly corrected DCNN (LcDCNN) image. To verify the usefulness of the proposed method, an observer study using a paired-comparison method was conducted on the original, reference, and LcDCNN images of 15 cases with 30 images. The paired comparison method showed that in most cases (29/30), the LcDCNN images were significantly superior to the original images in terms of display conditions. When comparing the LcDCNN and reference images, the number of LcDCNN and reference images that were superior to each other in the display condition was 17 and 13, respectively, and in both cases, 6 of these images showed statistically significant differences. The optimized SRS images obtained using the proposed method, while reflecting the observer's preference, were superior to the conventional manually adjusted images.

Task-based assessment of resolution properties of CT images with a new index using deep convolutional neural network.

In this study, we propose a method for obtaining a new index to evaluate the resolution properties of computed tomography (CT) images in a task-based manner. This method applies a deep convolutional neural network (DCNN) machine learning system trained on CT images with known modulation transfer function (MTF) values to output an index representing the resolution properties of the input CT image [i.e., the resolution property index (RPI)]. Sample CT images were obtained for training and testing of the DCNN by scanning the American Radiological Society phantom. Subsequently, the images were reconstructed using a filtered back projection algorithm with different reconstruction kernels. The circular edge method was used to measure the MTF values, which were used as teacher information for the DCNN. The resolution properties of the sample CT images used to train the DCNN were created by intentionally varying the field of view (FOV). Four FOV settings were considered. The results of adapting this method to the filtered back projection (FBP) and hybrid iterative reconstruction (h-IR) images indicated highly correlated values with the MTF10% in both cases. Furthermore, we demonstrated that the RPIs could be estimated in the same manner under the same imaging conditions and reconstruction kernels, even for other CT systems, where the DCNN was trained on CT systems produced by the same manufacturer. In conclusion, the RPI, which is a new index that represents the resolution property using the proposed method, can be used to evaluate the resolution of a CT system in a task-based manner.

Utilization of upper and lower limits of exposure index in clinical digital radiography.

In many digital X-ray imaging systems, although air kerma on a surface of each detector is used, a standardized dose index called an exposure index (EI) has been proposed by the IEC, which is expected to be utilized for dose management. In clinical practices, EI is effectively utilized using a deviation index (DI), which is a deviation between a target EI (EIT) set for each imaging region and an EIT of the acquired image. However, an important issue in clinical uses of EI is a suppression of excessive doses. It is difficult to achieve a reliable reduction in exposure doses by indicating DI. In this study, physical image characteristics of detectors, visual detectability by charts, and observer experiments using a chest phantom were examined to determine upper (DImax) and lower (DImin) limits of the EIT and DI to achieve a reliable dose reduction in chest examinations. As the result, the tolerance ranges indicated by DImax and DImin, which were set based on the results of physical and visual evaluations, proved to be almost consistent with the distribution of EI values in 735 clinical images taken with a photo-timer control in real clinical practices.

Deep learning versus the human visual system for detecting motion blur in radiography.

Purpose: The necessity of image retakes is initially determined on a preview monitor equipped with an operating system; therefore, some image blurring is only noticed later, on a high-resolution monitor. The purpose of this study is to investigate blur detection performance on radiographs via a deep learning approach compared with human observers. Approach: A total of 99 radiographs (blurry 57, nonblurry 42) were independently observed and rated by six observers using preview and diagnostic liquid crystal displays (LCDs). The deep convolution neural network (DCNN) was trained and tested using ninefold cross-validation. The average areas under the ROC curves (AUCs) were calculated for each observer with LCDs and by stand-alone DCNN for each test session and then statistically tested using a 95% confidence interval. Results: The average AUCs were 0.955 for stand-alone DCNN and 0.827 and 0.947 for human observers using preview and diagnostic LCDs, respectively. The DCNN revealed a high performance for image motion blur on digital radiographs (sensitivity 94.8%, specificity 96.8%, and accuracy 95.6%), along with the capability to detect a slight motion blur that was overlooked by human observers with a preview LCD. There were no cases of motion blur overlooked by the stand-alone DCNN, of which some were incorrectly recognized as nonblurry by human observers. Conclusions: The deep learning-based approach was capable of distinguishing slight motion blur that was unnoticeable on a preview LCD, and thus, is expected to aid the human visual system for detecting blurred images in the initial review of digital radiographs.

[Radiomics for Estimating Recurrence Risk of Patients with Lung Cancer by Using Survival Analysis].

PURPOSE: Because of the promotion of cancer screening, the number of patients with lung cancer detected at the early stage has increased. However, it was reported that 30-40% of the lung cancer patients at stage I relapsed. If the recurrence risk can be accurately predicted, it is possible to give medical care for improving the prognosis of lung cancer patients. The purpose of this study was to develop a method for the prediction of recurrence risk of patients with lung cancer by using survival analysis of radiomics approach. METHOD: A public database was used in this study. Fifty patients (25 recurrences and 25 censored cases) classified as stage I or II were selected and their pretreatment computed tomography (CT) images were obtained. First, we selected one slice containing the largest tumor area and manually segmented the tumor regions. We subsequently calculated 367 radiomic features such as tumor size, shape, CT values, and texture. Radiomic features were selected by using least absolute shrinkage and selection (Lasso). Cox regression model and random survival forest (RSF) with the selected radiomic features were used for estimating the recurrence functions of fifty patients. RESULT: The experimental result showed that average area under the curve (AUC) values of Cox regression model and RSF for the prediction accuracy were 0.81 and 0.93, respectively. CONCLUSION: Since our scheme can predict recurrence risk of patients with lung cancer by using non-invasive image examinations, it would be useful for the selection of treatment and the follow-up after the treatment.

[Application of Convolutional Neural Network for Evaluating CT Dose Using Image Noise Classification: A Phantom Study].

PURPOSE: It is well known that there is a trade-off relationship between image noise and exposure dose in X-ray computed tomography (CT) examination. Therefore, CT dose level was evaluated by using the CT image noise property. Although noise power spectrum (NPS) is a common measure for evaluating CT image noise property, it is difficult to evaluate noise performance directly on clinical CT images, because NPS requires CT image samples with uniform exposure area for the evaluation. In this study, various noise levels of CT phantom images were classified for estimating dose levels of CT images using convolutional neural network (CNN). METHOD: CT image samples of water phantom were obtained with a combination of mAs value (50, 100, 200 mAs) and X-ray tube voltage (80, 100, 120 kV). The CNN was trained and tested for classifying various noise levels of CT image samples by keeping 1) a constant kV and 2) a constant mAs. In addition, CT dose levels (CT dose index: CTDI) for all exposure conditions were estimated by using regression approach of the CNN. RESULT: Classification accuracies for various noise levels were very high (more than 99.9%). The CNN-estimated dose level of CT images was highly correlated (r=0.998) with the actual CTDI. CONCLUSION: CT image noise level classification using CNN can be useful for the estimation of CT radiation dose.

[Investigation of Clinical Utility of Radiological Technologist's Reading Report as a Second Opinion for Medical Doctor Reading of Digital Mammogram].

PURPOSE: We investigated the clinical utility of a radiological technologist's (RT)'s reports (RRs) as a second opinion by the free-response receiver operating characteristic (FROC) observer study that compared the performance of medical doctors' (MDs') reading of digital mammogram with and without consulting the RR. METHOD: One hundred women (39 malignant, 61 benign or normal) who underwent diagnostic mammography were selected from among 1674 routine clinical images classified by the degree of difficulty and categories for inclusion in the FROC study. The first FROC study performed by three RTs (RT 1-3) was conducted to collect the data for RR utilized in the second FROC study. The second FROC study was performed by five MDs, and the statistical significance of MDs' performances with and without reference to the RR was investigated by figure of merit (FOM). RESULT: The FOM values of three RTs obtained in the first FROC study were 0.529, 0.576, and 0.539, respectively. In the second FROC study, RT 2 had the highest FOM, RT 1 the lowest false positives/case, and RT 3 the highest sensitivity. The average FOM values in the second FROC study for the five MDs with/without reference to the RR were as follows: RT 2's RR was 0.534/0.588 (p=0.003), RT 1's RR was 0.500/0.545 (p=0.099), and RT 3's RR was 0.569/0.592 (p=0.324). CONCLUSION: We concluded that the MDs' performance of reading mammogram was statistically improved by consulting the RR when the RT's reading skill was high.