Patient Identification Based on Deep Metric Learning for Preventing Human Errors in Follow-up X-Ray Examinations.

Biological fingerprints extracted from clinical images can be used for patient identity verification to determine misfiled clinical images in picture archiving and communication systems. However, such methods have not been incorporated into clinical use, and their performance can degrade with variability in the clinical images. Deep learning can be used to improve the performance of these methods. A novel method is proposed to automatically identify individuals among examined patients using posteroanterior (PA) and anteroposterior (AP) chest X-ray images. The proposed method uses deep metric learning based on a deep convolutional neural network (DCNN) to overcome the extreme classification requirements for patient validation and identification. It was trained on the NIH chest X-ray dataset (ChestX-ray8) in three steps: preprocessing, DCNN feature extraction with an EfficientNetV2-S backbone, and classification with deep metric learning. The proposed method was evaluated using two public datasets and two clinical chest X-ray image datasets containing data from patients undergoing screening and hospital care. A 1280-dimensional feature extractor pretrained for 300 epochs performed the best with an area under the receiver operating characteristic curve of 0.9894, an equal error rate of 0.0269, and a top-1 accuracy of 0.839 on the PadChest dataset containing both PA and AP view positions. The findings of this study provide considerable insights into the development of automated patient identification to reduce the possibility of medical malpractice due to human errors.

[Assessment Tool of Image Quality for Medical Liquid-crystal Displays Using Commercially Available Digital Cameras].

We developed a simplified tool for measuring image quality of medical liquid-crystal displays (LCDs) using a commercially available color digital camera. This tool implemented as a plug-in software for ImageJ (open-source image processing program) was designed to compute modulation transfer functions (MTFs) and Wiener spectra (WS) of monochrome and color LCDs from LCD photographed images captured by a camera. The intensities of the red (R), green (G), and blue (B) signals of the unprocessed image data depend on the spectral sensitivity of the image sensor used in the camera. In order to evaluate image quality based on LCD luminance, the plug-in software calibrates the RGB signals from the camera using measured luminance of the LCD and converts them into grayscale signals that correspond to the luminance of the LCD. The MTFs and WS are determined based on the line response from a one-pixel line image and the one-dimensional noise profiles acquired by scanning the uniform image using numerically synthesized slit, respectively. With this plug-in software for ImageJ, we are able to readily compute MTFs and WS of both monochrome and color LCDs from unprocessed image data of cameras. Our simplified tool is helpful to evaluate and understand the physical performance of LCDs for a large number of display users in hospitals and medical centers.

Detectability of BI-RADS category 3 or higher breast lesions and reading time on mammography: comparison between 5-MP and 8-MP LCD monitors.

Background Five-megapixel (MP) displays are recommended as soft copy devices for digital mammogram. An 8-MP liquid crystal display (LCD) (two 4-MP displays within one display) might offer the advantage of being able to view biplane mammography more easily than the dual planes of 5-MP LCDs. Purpose To compare detectability of Breast Imaging Reporting and Data System (BI-RADS) category 3 or higher lesions and reading time on mammography between 5- MP and 8-MP LCDs. Material and Methods The mammograms of 240 breasts of 120 patients including 60 breasts with BI-RADS category 3 or higher lesions and 180 breasts with normal or category 2 lesions were enrolled. All bilateral mammograms were displayed on bifacial 5-MP LCDs or an 8-MP LCD (two 4-MP displays within one display). Six radiologists assessed 240 breasts on each display. The observations were analyzed using receiver operating characteristic (ROC) analysis. A jack-knife method was used for statistical analysis. We employed a paired t-test to determine whether any significant differences existed in the reading time between two different displays. A P value < 0.05 was considered significant. Results The mean areas under the ROC curve obtained using 5-MP and 8-MP LCDs were 0.925 and 0.915, respectively, and there was no significant difference ( P = 0.46). There was also no significant difference in the reading time between two types of displays (57.8 min. vs. 51.5 min, P = 0.39). Conclusion The detectability of BI-RADS category 3 or higher lesions and reading time using an 8-MP LCD were comparable to those using a 5-MP LCD.

[Construction and usefulness of the dose management system using DICOM radiation dose structured report (Dose SR) in angiography].

The International Commission on Radiological Protection recommends diagnostic reference levels (DRL) in each radiological examination for justification and optimization of patients' dose in medicine. The aim of our study was to propose the dose management system by utilizing dose information in diagnostic X-ray radiation dose structured report (Dose SR) in The Digital Imaging and Communications in Medicine to optimize radiation dose in institutions. Our dose management system is able to organize dose information obtained from various angiography systems and CTs. It is possible to provide this information to operators for justification and optimization of patient dose. Our system would be useful for the estimation of organ dose and could be used for the determination of local DRL (LDRL) for each radiological practice. In addition, the optimization became possible to compare LDRL with national DRL.

Chromaticity and correlated color temperature of the white point in medical liquid-crystal display.

PURPOSE: The color characteristics of medical liquid-crystal displays (LCDs) have become one of the main interests in quality assurance and quality control of medical display devices. It is necessary to evaluate variations of the color characteristics in medical LCDs in order to provide consistent color characteristics for image reading. In this study, we characterize the color characteristics in grayscale images displayed on medical LCDs. METHODS: Eight medical color LCDs and eight medical monochrome LCDs were used in this study with operating times ranging from 800 to 25 000 h. Default settings of correlated color temperature (CCT) of white points for all color LCDs were set at 7500 K. We measured chromaticity in the Commission Internationale de l'Éclairage (CIE) u'v' color space and CCT of the white point at 18 luminance levels. A colorimeter and grayscale test patterns with various luminance levels were used for the measurements. First, we examined differences in chromaticity and CCT at 18 luminance levels for each LCD. Second, we compared chromaticities of the different LCDs. Then, chromaticity and CCT of LCDs were measured at different ambient light conditions (30 and 560 lux) and compared to those measured in a dark room. Finally, the relationship between operating times and CCT was evaluated. RESULTS: The chromaticities of monochrome LCDs changed toward the blue and CCT increased as luminance decreased. In color LCDs, the chromaticity and CCT of the white point except at the minimum luminance were nearly constant to those measured at the maximum luminance. We observed differences in chromaticity between color and monochrome LCDs because of the nonadjustable white point of monochrome LCDs. When the ambient light increased, chromaticity varied, and CCT decreased on every LCD. Ambient light had a much greater impact on chromaticity and CCT at lower luminance levels. The CCT of LCDs decreased in longer operating times. The grayscale of LCDs with longer operating times was measured to be more yellowish due to the degradation of cold cathode fluorescent lamps. CONCLUSIONS: Our results clearly indicate that the white point in grayscale images displayed on medical LCDs changes depending on luminance levels, color or monochrome LCDs, ambient light conditions, and operating times. These data provide useful information to understand color characteristics of LCD for image reading, especially for quality assurance and quality control of LCDs.

Biological fingerprint for patient verification using trunk scout views at various scan ranges in computed tomography.

Immediate verification of whether a patient being examined is correct is desirable, even if the scan ranges change during different examinations for the same patient. This study proposes an advanced biological fingerprint technique for the rapid and reliable verification of various scan ranges in computed tomography (CT) scans of the torso of the same patient. The method comprises the following steps: geometric correction of different scans, local feature extraction, mismatch elimination, and similarity evaluation. The geometric magnification correction was aligned at the scanner table height in the first two steps, and the local maxima were calculated as the local features. In the third step, local features from the follow-up scout image are matched to those in the corresponding baseline scout image via template matching and outlier elimination via a robust estimator. We evaluated the correspondence rate based on the inlier ratio between corresponding scout images. The ratio of inliers between the baseline and follow-up scout images was assessed as the similarity score. The clinical dataset, including chest, abdomen-pelvis, and chest-abdomen-pelvis scans, included 600 patients (372 men, 68 ± 12 years) who underwent two routine torso CT examinations. The highest area under the receiver operating characteristic curve (AUC) was 0.996, which was sufficient for patient verification. Moreover, the verification results were comparable to the conventional method, which uses scout images in the same scan range. Patient identity verification was achieved before the main scan, even in follow-up torso CT, under different scan ranges.

Potential for personal identification using the volume of the mastoid air cells extracted from postmortem computed tomographic images.

This study revealed the usefulness of volumetric analysis of mastoid air cells (MACs) extracted from postmortem computed tomography (PMCT) images in characterizing individuals. To characterize deceased persons, the MACs volumes of 61 Japanese PMCT images were measured after thresholding in Hounsfield units and based on the number of voxels on the right and left sides and the voxel size for each person. The volume differences between the right and left MACs and sex were examined. Although there were no obvious volume differences between males and females, the order of sizes on the two sides varied for each person. Moreover, deceased persons could be roughly classified using the total volume of MACs. Deceased persons with similar total volumes could be distinguished further by comparing the ratio of volumes in bilateral MACs. Although the identification process is dependent on samples and different sizes of bilateral MACs, our pilot study indicated that 81.9% (50/61) of deceased persons could be distinguished. In conclusion, volumetric analysis of MACs measured using PMCT imaging has the potential to identify individuals and reduce the number of candidates.

A novel algorithm for comprehensive quality assessment of clinical magnetic resonance images based on natural scene statistics in spatial domain.

BACKGROUND: A magnetic resonance imaging (MRI)-specific objective image quality assessment (IQA) algorithm, the quality evaluation using multidirectional filters for MRI (QEMDIM), was previously reported. QEMDIM requires a set of reference images to calculate the quality score (SQ) for an assessed image. SQ may be affected by the quality of the reference set owing to the calculation procedure. PURPOSE: To propose a modified version of the IQA algorithm and compare the IQA performance of the original and modified algorithms. ASSESSMENT: Brain axial T1- and T2-weighted spin-echo images of varying quality levels (noise and blurring) were acquired from seven healthy men. Subjective IQA (paired comparisons) was conducted on the images, and subjective quality scores were obtained. With reference sets of various quality levels, QEMDIM and modified IQA were applied to the same images that underwent the subjective IQA. The correlation of each SQ and modified score (Smod) with the subjective scores was evaluated for content-related subsets of assessed images and for each reference set. The effect of the reference-set quality on the distribution of the correlation coefficients (CCs) was statistically evaluated for SQ and Smod using a one-way analysis of variance test with a significance level of 0.05. We also evaluated the variation in Smod for images with almost the same qualities using the standard deviation (SD). RESULTS: The CCs of SQ varied significantly with the quality of the reference set, whereas that of Smod did not. The SD of Smod for almost-same-quality images was less than that corresponding to the confidence interval of the subjective scores. CONCLUSION: Our modified algorithm was superior to QEMDIM in terms of IQA performance in clinical practice, especially in terms of accuracy, robustness, and reproducibility.

Influence of monitor display resolution and displayed image size on the spatial resolution of ultra-high-resolution CT images: a phantom study.

To determine the optimal display conditions for ultra-high-resolution computed tomography (UHRCT) images in clinical practice, this study investigated the effects of liquid-crystal display (LCD) resolution and displayed image size on the spatial resolution of phantom images acquired using a UHRCT system. A phantom designed to evaluate the high-contrast resolution was scanned. The scan data were reconstructed into four types of UHRCT image series consisting of the following possible combinations: two types of reconstruction kernels on the filtered back-projection method (for the lung and mediastinum) and two types of matrix sizes (10242 and 20482). These images were displayed under eight types of display conditions: three image sizes displayed on a 2-megapixel (MP) and 3-MP color LCD and two image sizes on an 8-MP color LCD. A total of 32 samples (four image series × eight display conditions) were evaluated by eight observers for high-contrast resolution. The high-contrast resolution of the displayed UHRCT images was significantly affected by the displayed image size, although the largest (full-screen) displayed image size did not necessarily show the maximum high-contrast resolution. When the images were displayed in the full-screen size, LCD resolution affected the high-contrast resolution of only the 20482-matrix-size images reconstructed using the lung kernel. In conclusion, the spatial resolution of UHRCT images may be affected by LCD resolution and displayed image size. To optimize the clinical display conditions for UHRCT images, it is necessary to adopt an LCD with an adequate resolution for each viewing situation.

[New solutions for automated image recognition and identification: challenges to radiologic technology and forensic pathology].

This is a review on biological fingerprint for radiologic technology and forensic pathology by JSRT and JSMP (https://www.jsmp.org/en).

New solutions for automated image recognition and identification: challenges to radiologic technology and forensic pathology.

This paper outlines the history of biometrics for personal identification, the current status of the initial biological fingerprint techniques for digital chest radiography, and patient verification during medical imaging, such as computed tomography and magnetic resonance imaging. Automated image recognition and identification developed for clinical images without metadata could also be applied to the identification of victims in mass disasters or other unidentified individuals. The development of methods that are adaptive to a wide range of recent imaging modalities in the fields of radiologic technology, patient safety, forensic pathology, and forensic odontology is still in its early stages. However, its importance in practice will continue to increase in the future.

Medical display application for degraded image sharpness restoration based on the modulation transfer function: initial assessment for a five-megapixel mammography display monitor.

An image-display application for medical liquid-crystal display (LCD) monitors called the sharpness recovery (SR) function has been developed to compensate for image sharpness as a function of deficiencies in the modulation transfer function (MTF) of a monitor. We investigated the effects of the SR function for a five-megapixel (MP) mammography LCD monitor on the resolution and noise properties of the displayed images by measuring the MTF and overall noise power spectrum (NPS), respectively. Furthermore, the effectiveness of the SR function for the 5-MP monitor in displaying subtle microcalcifications on digital mammograms was verified using a two-alternative-forced-choice sensitivity measurement as an initial application for medical image interpretation. Four radiologists compared the visibility of 45 regions of interest with a malignant microcalcification cluster shown on SR-processed and unprocessed mammograms. SR processing improved the MTF of the displayed images by approximately 40% at the Nyquist frequency of the 5-MP monitor, whereas it slightly increased the overall NPS values. All observers indicated that the fraction of cases considered to have better visibility of microcalcifications with the SR processing was significantly greater than that without the processing (averaging 82%, with the 95% confidence interval ranging from 70 to 93%). The SR processing for the 5-MP monitor yielded a significant improvement in the resolution properties of the displayed images, with a certain increase in the image noise. The SR function has the potential to improve the observer performance of radiologists, particularly when reading subtle microcalcifications reproduced on 5-MP monitors.

Proper Management of the Clinical Exposure Index Based on Body Thickness Using Dose Optimization Tools in Digital Chest Radiography: A Phantom Study.

In radiography, the exposure index (EI), as per the International Electrotechnical Commission standard, depends on the incident beam quality and exposure dose to the digital radiography system. Today automatic exposure control (AEC) systems are commonly employed to obtain the optimal image quality. An AEC system can maintain a constant incident exposure dose on the image receptor regardless of the patient thickness. In this study, we investigated the relationship between body thickness, entrance surface dose (ESD), EI, and the exposure indicator (S value) with the aim of using EI as the dose optimization tool in digital chest radiography (posterior-anterior and lateral projection). The exposure condition from the Korean national survey for determining diagnostic reference levels and two digital radiography systems (photostimulable phosphor plate and indirect flat panel detector) were used. As a result, ESD increased as the phantom became thicker with constant exposure indicator, which indicates similar settings to an AEC system, but the EI indicated comparatively constant values without following the tendency of ESD. Therefore, body thickness should be considered under the AEC system for introducing EI as the dose optimization tool in digital chest radiography.

[Objective evaluation of visual fatigue for reading of radiographs displayed on medical-grade liquid-crystal displays].

The aim of this study was to evaluate visual fatigue objectively by measuring accommodation time and critical fusion frequency (CFF) before and after reading posteroanterior chest radiographs displayed on medical-grade liquid-crystal displays (LCDs) under different monitor conditions. A color LCD (500, 170 cd/m²) and a monochrome LCD (500 cd/m²) were used in this study. Six observers independently kept reading the radiographs for two hours to understand various lung nodules in the "Fatigue Session". Objective visual fatigue was measured by using the accommodation device and the CFF meter before and after the Fatigue Session. The ambient lighting of the laboratory was set at 35 lux. Both the accommodation time and the CFF between before and after the Fatigue Session indicated statistically significant differences (p<0.05). Our results on accommodation time and CFF before and after reading the radiographs on medical-grade LCDs indicated that visual fatigue could be evaluated objectively.

[Application of the grayscale standard display function to general purpose liquid-crystal display monitors for clinical use].

Interpretations of medical images have been shifting to soft-copy readings with liquid-crystal display (LCD) monitors. The display function of the medical-grade LCD monitor for soft-copy readings is recommended to calibrate the grayscale standard display function (GSDF) in accordance with the guidelines of Japan and other countries. In this study, the luminance and display function of five models of eight general purpose LCD monitors were measured to gain an understanding of their characteristics. Moreover, the display function (gamma 2.2 or gamma 1.8) of general purpose LCD monitors was converted to GSDF through the use of a look-up table, and the detectability of a simulated lung nodule in the chest x-ray image was examined. As a result, the maximum luminance, contrast ratio, and luminance uniformity of general purpose LCD monitors, except for one model of two LCD monitors, met the management grade 1 standard in the guideline JESRA X-0093-2005. In addition, the detectability of simulated lung nodule in the mediastinal space was obviously improved by converting the display function of a general purpose LCD monitor into GSDF.

[Evaluation of image quality using the normalized-rank approach for primary class liquid-crystal display (LCD) monitors with different colors and resolution].

The purposes of this study were to evaluate the image quality of five types of liquid-crystal display (LCD) monitors by utilizing the normalized-rank approach and to investigate the effect of LCD monitor specifications, such as display colors, luminance, and resolution, on the evaluators' ranking. The LCD monitors used in this study were 2, 3 and 5 mega-pixel monochrome LCD monitors, and 2 and 3 mega-pixel color LCD monitors (Eizo Nanao Corporation). All LCD monitors were calibrated to the grayscale standard display function (GSDF) with different maximum luminance (recommended luminance) settings. Also, four kinds of radiographs were used for observer study based on the normalized-rank approach: three adult chest radiographs, three pediatric chest radiographs, three ankle joint radiographs, and four double-contrasted upper gastrointestinal radiographs. Ten radiological technologists participated in the observer study. Monochrome LCD monitors exhibited superior ranking with statistically significant differences (p<0.05) compared to color LCD monitors in all kinds of radiographs. The major difference between monochrome and color monitors was luminance. Therefore, it is considered that the luminance of LCD monitors affects observers' evaluations based on image quality. Moreover, in the case of radiographs that include high frequency image components, the monitor resolution also affects the evaluation. In clinical practice, it is necessary to optimize the luminance and choose appropriate LCD monitors for diagnostic images.

[Correlation between basic imaging properties and subjective evaluations of two digital radiographic X-ray systems based on direct-conversion flat panel detector].

The purpose of this study was to examine the correlation between the basic imaging properties of two digital radiographic X-ray systems with a direct conversion flat-panel detector and their image qualities, which were evaluated by the observer in hard copy and soft copy studies. The subjective image quality was evaluated and compared in terms of the low-contrast detectability and image sharpness in the two digital radiographic X-ray systems. We applied the radiographs of a contrast detail phantom to the evaluation of low-contrast detectability and analyzed the contrast detail diagrams. Finally, low-contrast detectability was evaluated by the image quality figure (IQF) calculated from the contrast detail diagrams. Also, the subjective image sharpness of human dry bones of two systems was examined and evaluated by the normalized-rank method. The results indicated that System A tended to provide superior subjective image quality compared to System B in both observer studies. We also found high correlations between IQFs and basic imaging properties, such as the noise power spectrum (NPS) and the noise equivalent quantum (NEQ). In conclusion, the low-contrast detectability of the two digital radiographic X-ray systems with a direct conversion flat-panel detector corresponded to the NPS and the NEQ in both outputs (soft copy and hard copy). On the other hand, the subjective image sharpness of human dry bones was affected by their noise properties.

A simple method for the automatic classification of body parts and detection of implanted metal using postmortem computed tomography scout view.

Information on medical devices embedded in the body is important in the identification of an unidentified body. Computed tomography (CT) is a powerful imaging modality; however, metallic artifacts deteriorate the image quality because of the reconstruction method. On the contrary, CT scout view is less affected by metallic artifacts compared to CT. It is a simple method to classify the body into three rough parts for postmortem CT (PMCT) scout view, and an algorithm used to detect the location of the implanted metal has been developed for personal identification in forensic pathology. Of the test images, 97% were correctly classified into the three body parts. The true-positive rate for detection of the implanted metal in the scout view was 96.5%. Therefore, our simple methods are applicable in PMCT scout views and would be particularly useful for forensic pathology.

Correction to: A simple method for the automatic classification of body parts and detection of implanted metal using postmortem computed tomography scout view.

In the original publication of the article, the authors affiliations were incorrectly published. The corrected affiliations are given in this correction.

USE OF CLINICAL EXPOSURE INDEX AND DEVIATION INDEX BASED ON NATIONAL DIAGNOSTIC REFERENCE LEVEL AS DOSE-OPTIMIZATION TOOLS FOR GENERAL RADIOGRAPHY IN KOREA.

The International Electrotechnical Commission introduced the concepts of exposure index (EI), target exposure index (EIT) and deviation index (DI) to manage and optimize patient dose in real time. In this study, we have proposed an appropriate method for setting the EIT based on the Korean national diagnostic reference levels (DRLs). Furthermore, we evaluated the use of clinical EI, EIT and DI as tools for patient dose optimization in clinical environments by observing the changes in DI with those in EIT. According to the Korean national exposure conditions, we conducted experiments on three representative radiographic examinations (chest posterior-anterior, lateral and abdomen anterior-posterior) of clinical environments. As the exposure conditions and DRLs varied, the clinical EI, EIT and DI also varied. These results reveal that the clinical EI, EIT and DI can be used as tools for optimizing the patient dose if EIT is periodically and properly updated.