Digital chest radiography: an update on modern technology, dose containment and control of image quality

Robust Stochastic Neural Ensemble Learning With Noisy Labels for Thoracic Disease Classification

Chest radiography is the most common radiology examination for thoracic disease diagnosis, such as pneumonia. A tremendous number of chest X-rays prompt data-driven deep learning models in constructing computer-aided diagnosis systems for thoracic diseases. However, in realistic radiology practice, a deep learning-based model often suffers from performance degradation when trained on data with noisy labels possibly caused by different types of annotation biases. To this end, we present a novel stochastic neural ensemble learning (SNEL) framework for robust thoracic disease diagnosis using chest X-rays. The core idea of our method is to learn from noisy labels by constructing model ensembles and designing noise-robust loss functions. Specifically, we propose a fast neural ensemble method that collects parameters simultaneously across model instances and along optimization trajectories. Moreover, we propose a loss function that both optimizes a robust measure and characterizes a diversity measure of ensembles. We evaluated our proposed SNEL method on three publicly available hospital-scale chest X-ray datasets. The experimental results indicate that our method outperforms competing methods and demonstrate the effectiveness and robustness of our method in learning from noisy labels. Our code is available at https://github.com/hywang01/SNEL.

Performance of artificial intelligence-based software for the automatic detection of lung lesions on chest radiographs of patients with suspected lung cancer

PURPOSE

This study aimed to evaluate the performance of the commercially available artificial intelligence-based software CXR-AID for the automatic detection of pulmonary nodules on the chest radiographs of patients suspected of having lung cancer.

MATERIALS AND METHODS

This retrospective study included 399 patients with clinically suspected lung cancer who underwent CT and chest radiography within 1 month between June 2020 and May 2022. The candidate areas on chest radiographs identified by CXR-AID were categorized into target (properly detected areas) and non-target (improperly detected areas) areas. The non-target areas were further divided into non-target normal areas (false positives for normal structures) and non-target abnormal areas. The visibility score, characteristics and location of the nodules, presence of overlapping structures, and background lung score and presence of pulmonary disease were manually evaluated and compared between the nodules detected or undetected by CXR-AID. The probability indices calculated by CXR-AID were compared between the target and non-target areas.

RESULTS

Among the 450 nodules detected in 399 patients, 331 nodules detected in 313 patients were visible on chest radiographs during manual evaluation. CXR-AID detected 264 of these 331 nodules with a sensitivity of 0.80. The detection sensitivity increased significantly with the visibility score. No significant correlation was observed between the background lung score and sensitivity. The non-target area per image was 0.85, and the probability index of the non-target area was lower than that of the target area. The non-target normal area per image was 0.24. Larger and more solid nodules exhibited higher sensitivities, while nodules with overlapping structures demonstrated lower detection sensitivities.

CONCLUSION

The nodule detection sensitivity of CXR-AID on chest radiographs was 0.80, and the non-target and non-target normal areas per image were 0.85 and 0.24, respectively. Larger, solid nodules without overlapping structures were detected more readily by CXR-AID.

Pulmonary abnormality screening on chest x-rays from different machine specifications: a generalized AI-based image manipulation pipeline

BACKGROUND

Chest x-ray is commonly used for pulmonary abnormality screening. However, since the image characteristics of x-rays highly depend on the machine specifications, an artificial intelligence (AI) model developed for specific equipment usually fails when clinically applied to various machines. To overcome this problem, we propose an image manipulation pipeline.

METHODS

A total of 15,010 chest x-rays from systems with different generators/detectors were retrospectively collected from five institutions from May 2020 to February 2021. We developed an AI model to classify pulmonary abnormalities using x-rays from a single system. Then, we externally tested its performance on chest x-rays from various machine specifications. We compared the area under the receiver operating characteristics curve (AUC) of AI models developed using conventional image processing pipelines (histogram equalization [HE], contrast-limited histogram equalization [CLAHE], and unsharp masking [UM] with common data augmentations) with that of the proposed manipulation pipeline (XM-pipeline).

RESULTS

The XM-pipeline model showed the highest performance for all the datasets of different machine specifications, such as chest x-rays acquired from a computed radiography system (n = 356, AUC 0.944 for XM-pipeline versus 0.917 for HE, 0.705 for CLAHE, 0.544 for UM, p [Formula: see text] 0.001, for all) and from a mobile x-ray generator (n = 204, AUC 0.949 for XM-pipeline versus 0.933 for HE, p = 0.042, 0.932 for CLAHE (p = 0.009), 0.925 for UM (p = 0.001).

CONCLUSIONS

Applying the XM-pipeline to AI training increased the diagnostic performance of the AI model on the chest x-rays of different machine configurations.

RELEVANCE STATEMENT

The proposed training pipeline would successfully promote a wide application of the AI model for abnormality screening when chest x-rays are acquired using various x-ray machines.

KEY POINTS

• AI models developed using x-rays of a specific machine suffer from generalization. • We proposed a new image processing pipeline to address the generalization problem. • AI models were tested using multicenter external x-ray datasets of various machines. • AI with our pipeline achieved the highest diagnostic performance than conventional methods.

Seven general radiography x-ray detectors with pixel sizes ranging from 175 to 76μm: technical evaluation with the focus on orthopaedic imaging

Aim. Flat panel detectors with small pixel sizes general can potentially improve imaging performance in radiography applications requiring fine detail resolution. This study evaluated the imaging performance of seven detectors, covering a wide range of pixel sizes, in the frame of orthopaedic applications.Material and methods. Pixel sizes ranged from 175 (detector A175) to 76μm (detector G76). Modulation transfer function (MTF) and detective quantum efficiency (DQE) were measured using International Electrotechnical Commission (IEC) RQA3 beam quality. Threshold contrast (CT) and a detectability index (d') were measured at three air kerma/image levels. Rabbit shoulder images acquired at 60 kV, over five air kerma levels, were evaluated in a visual grading study for anatomical sharpness, image noise and overall diagnostic image quality by four radiologists. The detectors were compared to detector E124.Results. The 10% point of the MTF ranged from 3.21 to 4.80 mm-1, in going from detector A175to detector G76. DQE(0.5 mm-1) measured at 2.38μGy/image was 0.50 ± 0.05 for six detectors, but was higher for F100at 0.62. High frequency DQE was superior for the smaller pixel detectors, howeverCTfor 0.25 mm discs correlated best with DQE(0.5 mm-1). Correlation betweenCTand the detectability model was good (R2= 0.964).CTfor 0.25 mm diameter discs was significantly higher for D150and F100compared to E124. The visual grading data revealed higher image quality ratings for detectors D125and F100compared to E124. An increase in air kerma was associated with improved perceived sharpness and overall quality score, independent of detector. Detectors B150, D125, F100and G76, performed well in specific tests, however only F100consistently outperformed the reference detector.Conclusion. Pixel size alone was not a reliable predictor of small detail detectability or even perceived sharpness in a visual grading analysis study.

Deep learning-based automatic detection for pulmonary nodules on chest radiographs: The relationship with background lung condition, nodule characteristics, and location

PURPOSE

Computer-aided diagnosis (CAD), which assists in the interpretation of chest radiographs, is becoming common. However, few studies have evaluated the benefits and pitfalls of CAD in the real world. This study aimed to evaluate the independent performance of commercially available deep learning-based automatic detection (DLAD) software, EIRL Chest X-ray Lung Nodule, in a cohort that included patients with background pulmonary abnormalities often encountered in clinical situations.

METHODS

Patients with clinically suspected lung cancer for whom chest radiography was performed within a month before or after CT scan between June 2020 and May 2022 in our institution were enrolled. The reference standard was created using a bounding box annotated by two radiologists with reference to the CT. The visibility score, characteristics, location of the pulmonary nodules, presence of overlapping structures or pulmonary disease, and background lung score were manually determined.

RESULTS

We included 388 patients. The DLAD software detected 222 of the 322 nodules visible on manual evaluation, with a sensitivity of 0.689 and a false-positive rate of 0.168. The detectability of the DLAD software was significantly lower for small and subsolid and nodules with overlapping structures. The visibility score and sensitivity of detection by the DLAD software were positively correlated. The relationship between the background lung score and detection by the DLAD software was unclear.

CONCLUSION

The standalone performance of DLAD in detecting pulmonary nodules exhibited a sensitivity of 0.689 and a false-positive rate of 0.168. Understanding the characteristics of DLAD is crucial when interpreting chest radiographs with the assistance of the DLAD.

Effect of Contrast Level and Image Format on a Deep Learning Algorithm for the Detection of Pneumothorax with Chest Radiography

Under the black-box nature in the deep learning model, it is uncertain how the change in contrast level and format affects the performance. We aimed to investigate the effect of contrast level and image format on the effectiveness of deep learning for diagnosing pneumothorax on chest radiographs. We collected 3316 images (1016 pneumothorax and 2300 normal images), and all images were set to the standard contrast level (100%) and stored in the Digital Imaging and Communication in Medicine and Joint Photographic Experts Group (JPEG) formats. Data were randomly separated into 80% of training and 20% of test sets, and the contrast of images in the test set was changed to 5 levels (50%, 75%, 100%, 125%, and 150%). We trained the model to detect pneumothorax using ResNet-50 with 100% level images and tested with 5-level images in the two formats. While comparing the overall performance between each contrast level in the two formats, the area under the receiver-operating characteristic curve (AUC) was significantly different (all p < 0.001) except between 125 and 150% in JPEG format (p = 0.382). When comparing the two formats at same contrast levels, AUC was significantly different (all p < 0.001) except 50% and 100% (p = 0.079 and p = 0.082, respectively). The contrast level and format of medical images could influence the performance of the deep learning model. It is required to train with various contrast levels and formats of image, and further image processing for improvement and maintenance of the performance.

The effect of breast shielding outside the field of view on breast entrance surface dose in axial X-ray examinations: a phantom study

PURPOSE

The purpose of this study was to evaluate the effect of outside-field-of-view (FOV) lead shielding on the entrance surface dose (ESD) of the breast on an anthropomorphic X-ray phantom for a variety of axial skeleton X-ray examinations.

METHODS

Using an anthropomorphic phantom and radiation dosimeter, the ESD of the breast was measured with and without outside-FOV shielding in anterior-posterior (AP) abdomen, AP cervical spine, occipitomental 30° (OM30) facial bones, AP lumbar spine, and lateral lumbar spine radiography. The effect of several exposure parameters, including a low milliampere-seconds technique, grid use, automatic exposure control use, wraparound lead (WAL) use, trolley use, and X-ray table use, on the ESD of the breast with and without outside-FOV shielding was investigated. The mean ESD (μSv) and standard deviation for each radiographic protocol were calculated. A one-tailed Student's t-test was carried out to evaluate whether ESD to the breast was reduced with the use of outside-FOV shielding.

RESULTS

A total of 920 breast ESD measurements were recorded across the different protocol parameters. The largest decrease in mean ESD of the breast with outside-FOV shielding was 0.002 μSv (P = 0.084), recorded in the AP abdomen on the table with a grid, OM30 on the table with a grid, OM30 standard protocol on the trolley, and OM30 on the trolley with WAL protocols. This decrease was found to be statistically non-significant.

CONCLUSION

This study found no significant decrease in the ESD of the breast with the use of outside-FOV shielding for the AP abdomen, AP cervical spine, OM30 facial bones, AP lumbar spine, or lateral lumbar spine radiography across a range of protocols.

Exposure index in digital radiography: initial results of awareness and knowledge from Nigerian digital radiography practices

INTRODUCTION

Exposure Index (EI) is incorporated into Digital Radiography (DR) systems to indicate incorrect exposure to enable matching exposure to the desired speed class of operation. However, knowledge of the utilization of EI by radiographers in a low-income country has not been investigated.

METHODS

A pre-tested questionnaire designed using Google forms, with open and close-ended questions was shared online with radiographers working with DR systems in public and private health facilities in some cities located in southern Nigeria. The 32-item questionnaire had two parts: Part A focused on socio-demographic characteristics and Part B focused on the respondents' awareness and knowledge of EI in DR systems. A 5-point Likert scale with 5 test items was used to assess the respondents' knowledge of EI. Statistical analyses were conducted using the Statistical Package for Social Sciences (SPSS) version 21.0. The probability value of p < 0.05 was considered statistically significant.

RESULTS

About 8.3% of the respondents had good knowledge of EI in DR systems in spite of the awareness level of 24.7%. The absence of the EI concept in DR curriculum for undergraduates, the lack of EI software in DR systems, and equipment training by the vendor engineers were reasons for the low level of knowledge of EI in DR systems.

CONCLUSION

There is low awareness and knowledge of EI by radiographers in this study, which suggests the need to maximize the benefits of EI concepts by ensuring its integration into clinical radiography practice and curriculum for undergraduates program, to improve knowledge, awareness, and practice in DR.

Evaluation of the clinical utility of temporal subtraction using bone suppression processing in digital chest radiography

Rationale and objectives

To evaluate the usefulness of temporal subtraction using the bone suppression method in digital chest radiography for the detection of pulmonary lesions.

Materials and methods

The images of 31 patients with pulmonary lesions and 19 normal cases were included in the study. Conventional and bone suppression temporal subtraction were performed in the 50 cases selected and used for an observer performance study. Five radiologists participated in the study, and the differences between using conventional and bone suppression temporal subtraction were assessed using jackknife free-response receiver operating characteristic analysis.

Results

The average figure-of-merit values for all radiologists increased significantly using the bone suppression method, from 0.619 (conventional) to 0.696 (p = 0.032). The average sensitivity for detecting pulmonary lesions improved from 67.9% to 75.4%, and the average number of false-positive per case decreased from 0.336 to 0.252 using bone suppression temporal subtraction.

Conclusion

Bone suppression temporal subtraction processing can assist with the detection of subtle pulmonary lesions in digital chest radiographs.

Commissioning and performance evaluation of commercially available mobile imager for image guided total body irradiation

BACKGROUND

The setup of lung shield (LS) in total body irradiation (TBI) with the computed radiography (CR) system is a time-consuming task and has not been quantitatively evaluated. The TBI mobile imager (TBI-MI) can solve this problem through real-time monitoring. Therefore, this study aimed to perform commissioning and performance evaluation of TBI-MI to promote its use in clinical practice.

METHODS

The source-axis distance in TBI treatment, TBI-MI (CNERGY TBI, Cablon Medical B.V.), and the LS position were set to 400, 450, and 358 cm, respectively. The evaluation items were as follows: accuracy of image scaling and measured displacement error of LS, image quality (linearity, signal-to-noise ratio, and modulation transfer function) using an EPID QC phantom, optimal thresholding to detect intra-fractional motion in the alert function, and the scatter radiation dose from TBI-MI.

RESULTS

The accuracy of image scaling and the difference in measured displacement of the LS was <4 mm in any displacements and directions. The image quality of TBI imager was slightly inferior to the CR image but was visually acceptable in clinical practice. The signal-to-noise ratio was improved at high dose rate. The optimal thresholding value to detect a 10-mm body displacement was determined to be approximately 5.0%. The maximum fraction of scattering radiation to irradiated dose was 1.7% at patient surface.

CONCLUSION

MI-TBI can quantitatively evaluate LS displacement with acceptable image quality. Furthermore, real-time monitoring with alert function to detect intrafraction patient displacement can contribute to safe TBI treatment.

Comparison of inspiratory and expiratory airway volumes and luminal areas among standing, sitting, and supine positions using upright and conventional CT

Upright computed tomography (CT) provides physiologically relevant images of daily life postures (sitting and standing). The volume of the human airway in sitting or standing positions remains unclear, and no clinical study to date has compared the inspiratory and expiratory airway volumes and luminal areas among standing, sitting, and supine positions. In this prospective study, 100 asymptomatic volunteers underwent both upright (sitting and standing positions) and conventional (supine position) CT during inspiration and expiration breath-holds and the pulmonary function test (PFT) within 2 h of CT. We compared the inspiratory/expiratory airway volumes and luminal areas on CT among the three positions and evaluated the correlation between airway volumes in each position on CT and PFT measurements. The inspiratory and expiratory airway volumes were significantly higher in the sitting and standing positions than in the supine position (inspiratory, 4.6% and 2.5% increase, respectively; expiratory, 14.9% and 13.4% increase, respectively; all P < 0.001). The inspiratory and expiratory luminal areas of the trachea, bilateral main bronchi, and average third-generation airway were significantly higher in the sitting and standing positions than in the supine position (inspiratory, 4.2‒10.3% increases, all P < 0.001; expiratory, 6.4‒12.8% increases, all P < 0.0001). These results could provide important clues regarding the pathogenesis of orthopnea. Spearman's correlation coefficients between the inspiratory airway volume on CT and forced vital capacity and forced expiratory volume in 1 s on PFT were numerically higher in the standing position than in the supine position (0.673 vs. 0.659 and 0.669 vs. 0.643, respectively); however, no statistically significant differences were found. Thus, the airway volumes on upright and conventional supine CT were moderately correlated with the PFT measurements.

Performance of a Chest Radiography AI Algorithm for Detection of Missed or Mislabeled Findings: A Multicenter Study

Purpose: We assessed whether a CXR AI algorithm was able to detect missed or mislabeled chest radiograph (CXR) findings in radiology reports. Methods: We queried a multi-institutional radiology reports search database of 13 million reports to identify all CXR reports with addendums from 1999–2021. Of the 3469 CXR reports with an addendum, a thoracic radiologist excluded reports where addenda were created for typographic errors, wrong report template, missing sections, or uninterpreted signoffs. The remaining reports contained addenda (279 patients) with errors related to side-discrepancies or missed findings such as pulmonary nodules, consolidation, pleural effusions, pneumothorax, and rib fractures. All CXRs were processed with an AI algorithm. Descriptive statistics were performed to determine the sensitivity, specificity, and accuracy of the AI in detecting missed or mislabeled findings. Results: The AI had high sensitivity (96%), specificity (100%), and accuracy (96%) for detecting all missed and mislabeled CXR findings. The corresponding finding-specific statistics for the AI were nodules (96%, 100%, 96%), pneumothorax (84%, 100%, 85%), pleural effusion (100%, 17%, 67%), consolidation (98%, 100%, 98%), and rib fractures (87%, 100%, 94%). Conclusions: The CXR AI could accurately detect mislabeled and missed findings. Clinical Relevance: The CXR AI can reduce the frequency of errors in detection and side-labeling of radiographic findings.

Blunt thoracic trauma: role of chest radiography and comparison with CT - findings and literature review

In the setting of acute trauma where identification of critical injuries is time-sensitive, a portable chest radiograph is broadly accepted as an initial diagnostic test for identifying benign and life-threatening pathologies and guiding further imaging and interventions. This article describes chest radiographic findings associated with various injuries resulting from blunt chest trauma and compares the efficacy of the chest radiograph in these settings with computed tomography (CT). Common chest radiographic findings in blunt thoracic injuries will be reviewed to improve radiologic identification, expedite management, and improve trauma morbidity and mortality. This article discusses demographic information, mechanism of specific injuries, common imaging findings, imaging pearls, and pitfalls and exhibits several classic imaging findings in blunt chest trauma. Thoracic structures commonly injured in blunt trauma that will be discussed in this article include vasculature structures (aortic trauma), the heart (cardiac contusion, pericardial effusion), the esophagus (esophageal perforation), pleural space and airways (pneumothorax, hemothorax, bronchial injury), lungs (pulmonary contusion), the diaphragm (diaphragmatic rupture), and the chest wall (flail chest). Chest radiography plays an important role in the initial evaluation of blunt chest trauma. While CT imaging has a higher sensitivity than chest radiography, it remains a valuable tool due to its ability to provide rapid diagnostic information in time-sensitive trauma situations and is ubiquitously available in the trauma bay. Familiarity with the gamut of injuries that may occur as well as identification of the associated chest radiograph findings can aid in timely diagnoses and prompt management in the setting of acute blunt chest trauma.

Optimization of Image Quality and Organ Absorbed Dose for Pediatric Chest X-Ray Examination: In-House Developed Chest Phantom Study

Purpose

This study aimed to identify proper exposure techniques to maintain optimal diagnostic image quality with minimum radiation dose for anteroposterior chest X-ray projection in pediatric patients.

Methods

Briefly, an in-house developed pediatric chest phantom was constructed. Next, nanodot OSLDs were used for organ absorbed dose measurement and placed in the lung area, and the phantom was exposed to various exposure techniques (ranging from 50 to 70 kVp with 1.6, 2, and 2.5 mAs). After that, the phantom was used to assess image quality parameters, including SNR and CNR. Two radiologists assessed the subjective image quality using a visual grading analysis (VGA) technique. Finally, the figure of merit (FOM) was analyzed.

Results

The developed phantom was constructed successfully and could be useful for dose measurement and image quality assessment. The absorbed dose varied from 0.009 to 0.031 mGy for the range of exposure techniques used. SNR and CNR showed a gradually increasing trend, while kVp and mAs values were increased. The highest kVp (70 kVp) produced the highest SNR and CNR, exhibiting a significant difference compared with 50 and 60 kVp (P < 0.05). The overall VGA score was 3.2 ± 0.3, and the low kVp technique demonstrated better image quality compared with the reference image.

Conclusion

The optimized exposure technique was identified as 60 kV and 2.5 mAs, indicating the highest FOM score. This work revealed practicable techniques that could be implemented into clinical practice for performing pediatric chest radiography.

Clinical utility of postprocessed low-dose radiographs in skeletal imaging

OBJECTIVES

Radiography remains the mainstay of diagnostic and follow-up imaging. In view of the risks and the increasing use of ionizing radiation, dose reduction is a key issue for research and development. The introduction of digital radiography and the associated access to image postprocessing have opened up new opportunities to minimize the radiation dosage. These advances are contingent upon quality controls to ensure adequate image detail and maintenance of diagnostic confidence. The purpose of this study was to investigate the clinical applicability of postprocessed low-dose images in skeletal radiography.

METHODS

In our study setting, the median radiation dose for full dose X-rays was 9.61 dGy*cm2 for pelvis, 1.20 dGy*cm2 for shoulder and 18.64 dGy*cm2 for lumbar spine exams. Based on these values, we obtained 200 radiographs for each anatomic region in four consecutive steps, gradually reducing the dose to 84%, 71%, 60% and 50% of the baseline using an automatic exposure control (AEC). 549 patients were enrolled for a total of 600 images. All X-rays were postprocessed with a spatial noise reduction algorithm. Two radiologists assessed the diagnostic value of the radiographs by rating the visualization of anatomical landmarks and image elements on a five-point Likert scale. A mean-sum score was calculated by averaging the two reader's total scores. Given the non-parametric distribution, we used the Mann-Whitney U test to evaluate the scores.

RESULTS

Median dosage at full dose accounted for 38.4%, 48 and 53.2% of the German reference dose area product for shoulder, pelvis and lumbar spine, respectively. The applied radiation was incrementally reduced to 21.5%, 18.4% and 18.7% of the respective reference value for shoulder, pelvis and lumbar spine. Throughout the study, we observed an estimable tendency of superior quality at higher dosage in overall image quality. Statistically significant differences in image quality were restricted to the 50% dose groups in shoulder and lumbar spine images. Regardless of the applied dosage, 598 out of 600 images were of sufficient diagnostic value.

CONCLUSION

In digital radiography image postprocessing allows for extensive reduction of radiation dosage. Despite a trend of superior image detail at higher dose levels, overall quality and, more importantly, diagnostic utility of low-dose images was not significantly affected. Therefore, our results not only confirm the clinical utility of postprocessed low-dose radiographs, but also suggest a widespread deployment of this advanced technology to ensure further dose limitations in clinical practice.

ADVANCES IN KNOWLEDGE

The diagnostic image quality of postprocessed skeletal radiographs is not significantly impaired even after extensive dose reduction by up to 20% of the reference value.

The diagnostic value of grey-scale inversion technique in chest radiography

Purpose

We investigated whether the additional use of grey-scale inversion technique improves the interpretation of eight chest abnormalities, in terms of diagnostic performance and interobserver variability.

Material and methods

A total of 507 patients who underwent a chest computed tomography (CT) examination and a chest radiography (CXR) within 24 h were enrolled. CT was the standard of reference. Images were retrospectively reviewed for the presence of atelectasis, consolidation, interstitial abnormality, nodule, mass, pleural effusion, pneumothorax and rib fractures. Four CXR reading settings, involving 3 readers were organized: only standard; only inverted; standard followed by inverted; and inverted followed by standard. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and accuracy, assessed with the area under the curve (AUC), and their 95% confidence interval were calculated for each reader and setting. Interobserver agreement was tested by Cohen’s K test with quadratic weights (K_w) and its 95%CI.

Results

CXR sensitivity % for any finding was 35.1 (95% CI: 33 to 37) for setting 1, 35.9 (95% CI: 33 to 37), for setting 2, 32.59 (95% CI: 30 to 34) for setting 3, and 35.56 (95% CI: 33 to 37) for setting 4; specificity % 93.78 (95% CI: 91 to 95), 93.92 (95% CI: 91 to 95), 94.43 (95% CI: 92 to 96), 93.86 (95% CI: 91 to 95); PPV % 56.22 (95% CI: 54.2 to 58.2), 56.49 (95% CI: 54.5 to 58.5), 57.15 (95% CI: 55 to 59), 56.75 (95% CI: 54 to 58); NPV % 85.66 (95% CI: 83 to 87), 85.74 (95% CI: 83 to 87), 85.29 (95% CI: 83 to 87), 85.73 (95% CI: 83 to 87); AUC values 0.64 (95% CI: 0.62 to 0.66), 0.65 (95% CI: 0.63 to 0.67), 0.64 (95% CI: 0.62 to 0.66), 0.65 (95% CI: 0.63 to 0.67); K_w values 0.42 (95% CI: 0.4 to 0.44), 0.40 (95% CI: 0.38 to 0.42), 0.42 (95% CI: 0.4 to 0.44), 0.41 (95% CI: 0.39 to 0.43) for settings 1, 2, 3 and 4, respectively.

Conclusions

No significant advantages were observed in the use of grey-scale inversion technique neither over standard display mode nor in combination at the detection of eight chest abnormalities.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11547-022-01453-0.

Assessment of the effect of a comprehensive chest radiograph deep learning model on radiologist reports and patient outcomes: a real-world observational study

OBJECTIVES

Artificial intelligence (AI) algorithms have been developed to detect imaging features on chest X-ray (CXR) with a comprehensive AI model capable of detecting 124 CXR findings being recently developed. The aim of this study was to evaluate the real-world usefulness of the model as a diagnostic assistance device for radiologists.

DESIGN

This prospective real-world multicentre study involved a group of radiologists using the model in their daily reporting workflow to report consecutive CXRs and recording their feedback on level of agreement with the model findings and whether this significantly affected their reporting.

SETTING

The study took place at radiology clinics and hospitals within a large radiology network in Australia between November and December 2020.

PARTICIPANTS

Eleven consultant diagnostic radiologists of varying levels of experience participated in this study.

PRIMARY AND SECONDARY OUTCOME MEASURES

Proportion of CXR cases where use of the AI model led to significant material changes to the radiologist report, to patient management, or to imaging recommendations. Additionally, level of agreement between radiologists and the model findings, and radiologist attitudes towards the model were assessed.

RESULTS

Of 2972 cases reviewed with the model, 92 cases (3.1%) had significant report changes, 43 cases (1.4%) had changed patient management and 29 cases (1.0%) had further imaging recommendations. In terms of agreement with the model, 2569 cases showed complete agreement (86.5%). 390 (13%) cases had one or more findings rejected by the radiologist. There were 16 findings across 13 cases (0.5%) deemed to be missed by the model. Nine out of 10 radiologists felt their accuracy was improved with the model and were more positive towards AI poststudy.

CONCLUSIONS

Use of an AI model in a real-world reporting environment significantly improved radiologist reporting and showed good agreement with radiologists, highlighting the potential for AI diagnostic support to improve clinical practice.

Diagnostic Performance of Dual-Energy Subtraction Radiography for the Detection of Pulmonary Emphysema: An Intra-Individual Comparison

PURPOSE/OBJECTIVES

To compare the diagnostic performance of dual-energy subtraction (DE) and conventional radiography (CR) for detecting pulmonary emphysema using computed tomography (CT) as a reference standard.

METHODS AND MATERIALS

Sixty-six patients (24 female, median age 73) were retrospectively included after obtaining lateral and posteroanterior chest X-rays with a dual-shot DE technique and chest CT within ±3 months. Two experienced radiologists first evaluated the standard CR images and, second, the bone-/soft tissue weighted DE images for the presence (yes/no), degree (1-4), and quadrant-based distribution of emphysema. CT was used as a reference standard. Inter-reader agreement was calculated. Sensitivity and specificity for the correct detection and localization of emphysema was calculated. Further degree of emphysema on CR and DE was correlated with results from CT. A p-value < 0.05 was considered as statistically significant.

RESULTS

The mean interreader agreement was substantial for CR and moderate for DE (kCR = 0.611 vs. kDE = 0.433; respectively). Sensitivity, as well as specificity for the detection of emphysema, was comparable between CR and DE (sensitivityCR 96% and specificityCR 75% vs. sensitivityDE 91% and specificityDE 83%; p = 0.157). Similarly, there was no significant difference in the sensitivity or specificity for emphysema localization between CR and DE (sensitivityCR 50% and specificityCR 100% vs. sensitivityDE 57% and specificityDE 100%; p = 0.157). There was a slightly better correlation with CT of emphysema grading in DE compared to CR (rDE = 0.75 vs. rCR = 0.68; p = 0.108); these differences were not statistically significant, however.

CONCLUSION

Diagnostic accuracy for the detection, quantification, and localization of emphysema between CR and DE is comparable. Interreader agreement, however, is better with CR compared to DE.

Chest radiographs and machine learning - Past, present and future

Despite its simple acquisition technique, the chest X-ray remains the most common first-line imaging tool for chest assessment globally. Recent evidence for image analysis using modern machine learning points to possible improvements in both the efficiency and the accuracy of chest X-ray interpretation. While promising, these machine learning algorithms have not provided comprehensive assessment of findings in an image and do not account for clinical history or other relevant clinical information. However, the rapid evolution in technology and evidence base for its use suggests that the next generation of comprehensive, well-tested machine learning algorithms will be a revolution akin to early advances in X-ray technology. Current use cases, strengths, limitations and applications of chest X-ray machine learning systems are discussed.

Performance and educational training of radiographers in lung nodule or mass detection: Retrospective comparison with different deep learning algorithms

The aim of this investigation was to compare the diagnostic performance of radiographers and deep learning algorithms in pulmonary nodule/mass detection on chest radiograph.A test set of 100 chest radiographs containing 53 cases with no pathology (normal) and 47 abnormal cases (pulmonary nodules/masses) independently interpreted by 6 trained radiographers and deep learning algorithems in a random order. The diagnostic performances of both deep learning algorithms and trained radiographers for pulmonary nodules/masses detection were compared.QUIBIM Chest X-ray Classifier, a deep learning through mass algorithm that performs superiorly to practicing radiographers in the detection of pulmonary nodules/masses (AUCMass: 0.916 vs AUCTrained radiographer: 0.778, P < .001). In addition, heat-map algorithm could automatically detect and localize pulmonary nodules/masses in chest radiographs with high specificity.In conclusion, the deep-learning based computer-aided diagnosis system through 4 algorithms could potentially assist trained radiographers by increasing the confidence and access to chest radiograph interpretation in the age of digital age with the growing demand of medical imaging usage and radiologist burnout.

A Survey on Exposure Parameters Variation due to Aging in Radiology Devices

The inevitable use of medical imaging examinations and lack of a suitable alternative lead to the need to control and minimize the amount of radiation from such artificial medical sources. To assess the relation between exposure parameters and lifetime of radiology devices, quality control tests were carried out on 13 radiology devices in 11 general hospitals. In this study, a barracuda dosimeter, SE-43137 Sweden, was calibrated to measure and record the quantities of kVp, mAs and exposure parameters. In all the devices using applying the minimum and maximum values of kVp, the minimum and maximum values of the internal resistances were calculated. The lowest mR/mA for the device C was observed at a flow rate of 200 mA (equal to 2,425), while the highest value was for the device A (2) at a current intensity of 200 mA (equal to 14.625). By increasing the age of the device, the output of the device is reduced. Therefore, to compensate for this decrease in the output, higher exposure conditions are usually applied to the device, which can greatly increase the damage to the device.

[Relationship between Radiation Quality and Image Quality in Digital Chest Radiography: Validation Study Using Human Soft Tissue-equivalent Phantom]

PURPOSE

To evaluate image quality for chest radiography at different radiation qualities, using phantoms with scatter fractions similar to those of lungs.

METHODS

Two base phantoms with 10 and 4 cm thicknesses, respectively, made of a soft tissue-equivalent material, were used to mimic the X-ray attenuation of the human lung. Two plates with soft tissue- and bone-equivalent materials, respectively, were placed on the base phantom as contrast objects. The image data were obtained with the same entrance surface dose in each radiation quality. Six radiation qualities generated using 120 and 90 kV, and additional copper filters with thicknesses 0, 0.1, and 0.2 mm were selected. The signal-difference-to-noise ratio (SdNR) and a contrast ratio of the soft tissue to the bone were measured for the six radiation qualities.

RESULTS

The thicker the additional filter, the better the SdNR at both tube voltages. The SdNR values were not significantly different between 120 and 90 kV for the same filter thickness. The contrast ratio was higher at 120 than at 90 kV by approximately 8%.

CONCLUSIONS

Because of the advantage of the contrast ratio and the highest SdNR, the radiation quality with 120 kV and 0.2-mm copper filtration was the best. It was indicated that the conventional tube voltage of 120 kV remains to be better than the lower tube voltage of 90 kV.

Prescription practice and clinical utility of chest radiographs in a pediatric intensive care unit: a prospective observational study

Background

Chest radiograph (CXR) prescribing pattern and practice vary widely among pediatric intensive care units (PICU). ‘On demand’ approach is increasingly recommended as against daily ‘routine’ CXRs; however, the real-world practice is largely unknown.

Methods

This was a prospective observational study performed in children younger than 12 years admitted to PICU of a tertiary care teaching hospital in India. Data were collected on all consecutive CXRs performed between December 2016 and April 2017. The primary outcome was to assess the factors that were associated with higher chest radiograph prescriptions in PICU. Secondary outcomes were to study the indications, association with mechanical ventilation, image quality and avoidable radiation exposure.

Results

Of 303 children admitted during the study period, 159 underwent a total of 524 CXRs in PICU. Median (IQR) age of the study cohort was 2 (0.6–5) years. More than two thirds [n = 115, 72.3%] were mechanically ventilated. Most CXRs (n = 449, 85.7%) were performed on mechanically ventilated patients, amounting to a median (IQR) of 3 (2–5) radiographs per ventilated patient. With increasing duration of ventilation, the number of CXRs proportionately increased in the first two weeks of mechanical ventilation. In non-ventilated children, about two thirds (68%) underwent only one CXR. Majority of the prescriptions were on demand (n = 461, 88%). Most common indications were peri-procedure prescriptions (37%) followed by evaluation for respiratory disease status (24%). About 40% CXRs resulted in interventions; adjustment in ventilator settings (13.5%) was the most frequent intervention. In 26% (n = 138) of radiographs, image quality required improvement. One or more additional body part exposure other than chest and upper abdomen were noted 336 (64%) images. Children with > 3 CXR had higher PRISM III score, more often mechanically ventilated, had higher number of indwelling devices [mean (SD) 2.6 (1.2) vs. 1.7 (1.0)] and stayed longer in PICU [median (IQR) 11(7.5–18.5) vs. 6 (3–9)].

Conclusion

On demand prescription was the prevalent practice in our PICU. Most non-ventilated children underwent only one CXR while duration of PICU stay and the number of devices determined the number of CXRs in mechanically ventilated children. Quality improvement strategies should concentrate on the process of acquisition of images and limiting the radiation exposure to unwanted body parts.

ESTIMATION OF ENTRANCE SURFACE AIR KERMA IN DIGITAL RADIOGRAPHIC EXAMINATIONS

PURPOSE

Contribution of radiation doses from medical X-ray examination to collective dose is significant. Unusually, high doses may increase the risk of stochastic effects of radiations. Therefore, radiation dose assessment was performed in 241 digital X-ray examinations in the study and was compared with published dose reference levels (DRLs).

METHODS

Entrance surface air kerma (ESAK) was calculated in chest PA, cervical AP/Lat, abdomen AP, lumbar AP/Lat and pelvis AP digital radiographic examinations (119 male and 122 female) following the International Atomic Energy Agency recommended protocol. Initially, 270 digital examinations were selected, reject analysis was performed and final 241 examinations were enrolled in the study for dose calculations. The exposure parameters and X-ray tube output were used for dose calculations. Effective doses were estimated with the help of conversion coefficients from ICRP 103.

RESULTS

Median ESAK (mGy) and associated effective doses obtained were cervical spine AP (1.30 mGy, 0.045 mSv), cervical spine Lat (0.25 mGy, 0.005 mSv), chest PA (0.11 mGy, 0.014 mSv), abdomen AP (0.90 mGy, 0.118 mSv), lumbar spine AP (1.52 mGy, 0.177 mSv), lumbar spine Lat (7.76 mGy, 0.209 mSv) and pelvis AP (0.82 mGy, 0.081 mSv). Results were compared with the studies of UK, Oman, India and Canada.

CONCLUSION

The calculated ESAK and effective dose values were less than or close to previously published literature except for cervical spine AP and lumbar spine Lat. The results reinforce the need for radiation protection optimization, improving examination techniques and appropriate use of automatic exposure control in digital radiography. ESAK values reported in this study could further contribute to establishing local DRLs, regional DRLs and national DRLs.

Comprehensive Computer-Aided Decision Support Framework to Diagnose Tuberculosis From Chest X-Ray Images: Data Mining Study

Background

Tuberculosis (TB) is one of the most infectious diseases that can be fatal. Its early diagnosis and treatment can significantly reduce the mortality rate. In the literature, several computer-aided diagnosis (CAD) tools have been proposed for the efficient diagnosis of TB from chest radiograph (CXR) images. However, the majority of previous studies adopted conventional handcrafted feature-based algorithms. In addition, some recent CAD tools utilized the strength of deep learning methods to further enhance diagnostic performance. Nevertheless, all these existing methods can only classify a given CXR image into binary class (either TB positive or TB negative) without providing further descriptive information.

Objective

The main objective of this study is to propose a comprehensive CAD framework for the effective diagnosis of TB by providing visual as well as descriptive information from the previous patients’ database.

Methods

To accomplish our objective, first we propose a fusion-based deep classification network for the CAD decision that exhibits promising performance over the various state-of-the-art methods. Furthermore, a multilevel similarity measure algorithm is devised based on multiscale information fusion to retrieve the best-matched cases from the previous database.

Results

The performance of the framework was evaluated based on 2 well-known CXR data sets made available by the US National Library of Medicine and the National Institutes of Health. Our classification model exhibited the best diagnostic performance (0.929, 0.937, 0.921, 0.928, and 0.965 for F1 score, average precision, average recall, accuracy, and area under the curve, respectively) and outperforms the performance of various state-of-the-art methods.

Conclusions

This paper presents a comprehensive CAD framework to diagnose TB from CXR images by retrieving the relevant cases and their clinical observations from the previous patients’ database. These retrieval results assist the radiologist in making an effective diagnostic decision related to the current medical condition of a patient. Moreover, the retrieval results can facilitate the radiologists in subjectively validating the CAD decision.

Social Distancing with Portable Chest Radiographs During the COVID-19 Pandemic: Assessment of Radiograph Technique and Image Quality Obtained at 6 Feet and Through Glass

PURPOSE

To develop a technique that allows portable chest radiography to be performed through the glass door of a patient's room in the emergency department.

MATERIALS AND METHODS

A retrospective review of 100 radiographs (50 [mean age 59.4 ± 17.3, range 22-87; 30 women] performed with the modified technique in April 2020, randomized with 50 [mean age 59 ± 21.6, range 19-100; 31 men] using the standard technique was completed by three thoracic radiologists to assess image quality. Radiation exposure estimates to patient and staff were calculated. A survey was created and sent to 32 x-ray technologists to assess their perceptions of the modified technique. Unpaired Ttests were used for numerical data. A P value < .05 was considered statistically significant.

RESULTS

The entrance dose for a 50th percentile patient was the same between techniques, measuring 169 µGy. The measured technologist exposure from the modified technique assuming a 50th percentile patient and standing 6 feet to the side of the glass was 0.055 µGy, which was lower than standard technique technologist exposure of 0.088 µGy. Of the 100 portable chest radiographs evaluated by three reviewers, two reviewers rated all images as having diagnostic quality, while the other reviewer believed two of the standard images and one of the modified technique images were non-diagnostic. A total of 81% (26 of 32) of eligible technologists completed the survey. Results showed acceptance of the modified technique with the majority feeling safer and confirming conservation of PPE. Most technologists did not feel the modified technique was more difficult to perform.

CONCLUSIONS

The studies acquired with the new technique remained diagnostic, patient radiation doses remained similar, and technologist dose exposure were decreased with modified positioning. Perceptions of the new modified technique by frontline staff were overwhelmingly positive.

Triple attention learning for classification of 14 thoracic diseases using chest radiography

Chest X-ray is the most common radiology examinations for the diagnosis of thoracic diseases. However, due to the complexity of pathological abnormalities and lack of detailed annotation of those abnormalities, computer-aided diagnosis (CAD) of thoracic diseases remains challenging. In this paper, we propose the triple-attention learning (A ³ Net) model for this CAD task. This model uses the pre-trained DenseNet-121 as the backbone network for feature extraction, and integrates three attention modules in a unified framework for channel-wise, element-wise, and scale-wise attention learning. Specifically, the channel-wise attention prompts the deep model to emphasize the discriminative channels of feature maps; the element-wise attention enables the deep model to focus on the regions of pathological abnormalities; the scale-wise attention facilitates the deep model to recalibrate the feature maps at different scales. The proposed model has been evaluated on 112,120images in the ChestX-ray14 dataset with the official patient-level data split. Compared to state-of-the-art deep learning models, our model achieves the highest per-class AUC in classifying 13 out of 14 thoracic diseases and the highest average per-class AUC of 0.826 over 14 thoracic diseases.

Comparison of inspiratory and expiratory lung and lobe volumes among supine, standing, and sitting positions using conventional and upright CT

Currently, no clinical studies have compared the inspiratory and expiratory volumes of unilateral lung or of each lobe among supine, standing, and sitting positions. In this prospective study, 100 asymptomatic volunteers underwent both low-radiation-dose conventional (supine position, with arms raised) and upright computed tomography (CT) (standing and sitting positions, with arms down) during inspiration and expiration breath-holds and pulmonary function test (PFT) on the same day. We compared the inspiratory/expiratory lung/lobe volumes on CT in the three positions. The inspiratory and expiratory bilateral upper and lower lobe and lung volumes were significantly higher in the standing/sitting positions than in the supine position (5.3–14.7% increases, all P < 0.001). However, the inspiratory right middle lobe volume remained similar in the three positions (all P > 0.15); the expiratory right middle lobe volume was significantly lower in the standing/sitting positions (16.3/14.1% decrease) than in the supine position (both P < 0.0001). The Pearson’s correlation coefficients (r) used to compare the total lung volumes on inspiratory CT in the supine/standing/sitting positions and the total lung capacity on PFT were 0.83/0.93/0.95, respectively. The r values comparing the total lung volumes on expiratory CT in the supine/standing/sitting positions and the functional residual capacity on PFT were 0.83/0.85/0.82, respectively. The r values comparing the total lung volume changes from expiration to inspiration on CT in the supine/standing/sitting positions and the inspiratory capacity on PFT were 0.53/0.62/0.65, respectively. The study results could impact preoperative CT volumetry of the lung in lung cancer patients (before lobectomy) for the prediction of postoperative residual pulmonary function, and could be used as the basis for elucidating undetermined pathological mechanisms. Furthermore, in addition to morphological evaluation of the chest, inspiratory and expiratory upright CT may be used as an alternative tool to predict lung volumes such as total lung capacity, functional residual capacity, and inspiratory capacity in situation in which PFT cannot be performed such as during an infectious disease pandemic, with relatively more accurate predictability compared with conventional supine CT.

ESTIMATES OF PATIENT DOSES AND KERMA-AREA PRODUCT MONITORING IN DIGITAL RADIOGRAPHY

The application of the kerma-area product (PKA) meter is increased rapidly in dosimetry. This study presents measurements of PKA in adherence to the International Atomic Energy Agency protocol for 300 adult patients in digital radiographic procedures. Effective doses (ED) were calculated from PKA measurements and conversion coefficients (E-103/PKA) obtained from the International Commission on radiological protection 103. In skull posteroanterior (PA), skull lateral (LAT), cervical spine anteroposterior (AP), cervical spine LAT, chest PA, abdomen AP, lumbar spine AP, pelvis AP and lumbar spine LAT, the third-quartile PKA values were found to be 0.2, 0.28, 0.33, 0.19, 0.26, 0.95, 0.93, 0.96 and 3.15 Gycm2, and estimated mean EDs were 0.005, 0.008, 0.056, 0.021, 0.037, 0.146, 0.165, 0.097 and 0.258 mSv, respectively. The third-quartile PKA values were suggested as local diagnostic reference levels (LDRLs). Results were compared with the diagnostic reference levels (DRLs) of the UK, the European Commission, previously published LDRLs in Greece and China by Metaxas et al. and Zhang and Chu, respectively. The PKA (third-quartile) value for cervical spine AP was 120% higher than UK 2010 DRLs, lumbar spine LAT was 123% higher than LDRLs given by Metaxas et al. and chest PA was 160% higher than UK 2010 DRLs and 225% higher than Metaxas et al. provided LDRLs. The PKA results were lower than the UK, and two studies in Greece by Metaxas et al. except for chest PA, cervical spine AP and lumbar spine LAT showed the need for further optimization. The LDRLs reported in this study may further contribute to establishing future national DRLs.

Radiologists' Knowledge and Attitudes towards CT Radiation Dose and Exposure in Saudi Arabia-A Survey Study

Computed tomography (CT) is a key imaging technique in diagnostic radiology, providing highly sensitive and specific information. While its use has increased dramatically in recent years, the quantity and associated risks of radiation from CT scans present major challenges, particularly in paediatrics. The fundamental principles of radiation protection require that radiation quantities be as low as reasonably achievable and CT use must be justified, particularly for paediatric patients. CT radiation knowledge is a key factor in optimising and minimising radiation risk. The objective of this study was to analyse knowledge level, expertise, and competency regarding CT radiation dose and its hazards in paediatrics among radiologists in Saudi Arabian hospitals. A self-reported, multiple-choice questionnaire assessed the attitudes and opinions of radiologists involved in imaging studies using ionising radiation. Among the total respondents, 65% ± 13.5% had a good comprehension of the dangers of carcinogenicity to the patient resulting from CT scans, with 80% presuming that cancer risks were elevated. However, only 48.5%, 56.5%, and 65% of the respondents were aware of specific radiation risks in head, chest, and abdominal paediatric examinations, respectively. Regular, frequent, and specific training courses are suggested to improve the fundamental knowledge of CT radiation among radiologists and other physicians.

Calculating the target exposure index using a deep convolutional neural network and a rule base

PURPOSE

The objective of this study is to determine the quality of chest X-ray images using a deep convolutional neural network (DCNN) and a rule base without performing any visual assessment. A method is proposed for determining the minimum diagnosable exposure index (EI) and the target exposure index (EIt).

METHODS

The proposed method involves transfer learning to assess the lung fields, mediastinum, and spine using GoogLeNet, which is a type of DCNN that has been trained using conventional images. Three detectors were created, and the image quality of local regions was rated. Subsequently, the results were used to determine the overall quality of chest X-ray images using a rule-based technique that was in turn based on expert assessment. The minimum EI required for diagnosis was calculated based on the distribution of the EI values, which were classified as either suitable or non-suitable and then used to ascertain the EIt.

RESULTS

The accuracy rate using the DCNN and the rule base was 81%. The minimum EI required for diagnosis was 230, and the EIt was 288.

CONCLUSION

The results indicated that the proposed method using the DCNN and the rule base could discriminate different image qualities without any visual assessment; moreover, it could determine both the minimum EI required for diagnosis and the EIt.

Radiation dose reduction and improvement of image quality in digital chest radiography by new spatial noise reduction algorithm

Purpose

To evaluate the image quality of low-dose chest digital radiographic images obtained with a new spatial noise reduction algorithm, compared to a conventional de-noising technique.

Materials and methods

In 69 patients, the dose reduction protocol was divided into A, B, and C test groups– 60% (n = 22), 50% (n = 23), and 40% (n = 24) of the baseline dose. In each patient, baseline dose radiographs were obtained with conventional image processing while low-dose images were acquired with new image processing. A set of baseline and low-dose radiographic images per patient was evaluated and scored on a 5-point scale over seven anatomical landmarks (radiolucency of unobscured lung, pulmonary vascularity, trachea, edge of rib, heart border, intervertebral disc space, and pulmonary vessels in the retrocardiac area) and three representative abnormal findings (nodule, consolidation, and interstitial marking) by two thoracic radiologists. A comparison of paired baseline and low-dose images was statistically analyzed using a non-inferiority test based on the paired t-test or the Wilcoxon signed-rank test.

Results

In A, B, and C test groups, the mean dose reduction rate of the baseline radiation dose was 63.4%, 53.9%, and 47.8%, respectively. In all test groups, the upper limit of the 95% confidence interval was less than the non-inferiority margin of 0.5 every seven anatomical landmarks and three representative abnormal findings, which suggested that the image quality of the low-dose image was not inferior to that of the baseline dose image even if the maximum average dose reduction rate was reduced to 47.8% of the baseline dose.

Conclusion

In our study, an image processing technique integrating a new noise reduction algorithm achieved dose reductions of approximately half without compromising image quality for abnormal lung findings and anatomical landmarks seen on chest radiographs. This feature-preserving, noise reduction algorithm adopted in the proposed engine enables a lower radiation dose boundary for the sake of patient’s and radiography technologist’s radiation safety in routine clinical practice, in compliance with regulatory guidelines.

A multi institutional comparison of imaging dose and technique protocols for neonatal chest radiography

INTRODUCTION

The focus on paediatric radiation dose reduction supports reevaluation of paediatric imaging protocols. This is particularly important in the neonates where chest radiographs are frequently requested to assess respiratory illness and line placement. This study aims to assess the impact of neonatal chest radiographic protocols on patient dose in four hospitals in different countries.

METHODS

Exposure parameters, collimation, focus to skin distance (FSD) and radiation dose from 200 neonatal chest radiographs were registered prospectively. Inclusion criteria consisted of both premature and full-term neonates weighing between 1000 and 5000 g. Only data from the examinations meeting diagnostic criteria and approved for the clinical use were included. Radiation dose was assessed using dose area product (DAP).

RESULTS

The lowest DAP value (4.58 mGy cm²) was recorded in the Norwegian hospital, employing a high kVp, low mAs protocol using a DR system. The Canadian hospital recorded the highest DAP (9.48), using lower kVp and higher mAs with a CR system, including the addition of a lateral projection. The difference in the mean DAP, weight, field of view (FOV) and kVp between the hospitals is statistically significant (p < 0.001).

CONCLUSION

Use of non-standardised imaging protocols in neonatal chest radiography results in differences in patient dose across hospitals included in the study. Using higher kVp, lower mAs and reducing the number of lateral projections to clinically relevant indications result in a lower DAP measured in the infant sample studied. Further studies to examine image quality based on exposure factors and added filtration are recommended.

IMPLICATIONS FOR PRACTICE

Reevaluation of paediatric imaging protocols presents an opportunity to reduce patient dose in a population with increased sensitivity to ionising radiation.

Evaluation of radiographers' knowledge and attitudes of image quality optimisation in paediatric digital radiography in Saudi Arabia and Australia: a survey-based study

INTRODUCTION

Digital radiography (DR) systems enable radiographers to reduce the radiation dose to patients while maintaining optimised image quality. However, concerns still exist about paediatric patients who may be exposed to an increased level of radiation dose which is not needed for clinical practice. The purpose of this study was to evaluate the knowledge, awareness and attitudes, in terms of image quality optimisation of radiographers undertaking paediatric DR in Australia and Saudi Arabia.

METHODS

A survey-based study was devised and distributed to radiographers from Australia and Saudi Arabia. Questions focused on Australian and Saudi Arabian radiographers' knowledge and attitude of paediatric DR examinations.

RESULTS

There were 376 participants who responded to the survey from both countries. A major finding showed that most participants lack knowledge in the area of paediatric DR examinations. Most participants from Australia had received no formal training in paediatric digital radiography (79%), whereas nearly half of the participants from Saudi Arabia received no training (45%). Approximately three out of four radiographers from both countries believed that when using DR they did not need to change the way they collimate the beam as DR images can be cropped using post-processing methods.

CONCLUSION

The finding of this study demonstrates that radiographers from both countries should improve their understanding and clinical use of DR in paediatric imaging. More education and training for both students and clinicians is needed to enhance radiographer performance in digital radiography and improve their clinical practices.

A theoretical study of digital silicon photomultiplier utilization in diffuse optical imaging systems

Digital silicon photomultiplier (dSiPM) is introduced for diffuse optical imaging (DOI) applications instead of conventional photomultiplier tubes and avalanche photodiodes (APDs) as a state-of-the-art detector. According to the low-level light regime in DOI applications, high sensitivity and high dynamic range (DR) image sensors are needed for DOI systems. dSiPM is proposed as a developing detector which can detect low-level lights. Also, an accurate equation is obtained for calculating the DR of dSiPMs. Different dSiPMs and the corresponding benefits are studied for DOI applications. Furthermore, a 120 dB DR dSiPM is chosen for use in DOI systems. It is shown that dSiPMs can be utilized in DOI configurations such as time domain (TD), frequency domain (FD) and continuous wave (CW) systems. Ultimately, by utilizing dSiPM in DOI systems, the DOI method can be used for thoracic imaging due to the high DR and signal-to-noise ratio (SNR) of the detector.

Inorganic, Organic, and Perovskite Halides with Nanotechnology for High–Light Yield X- and γ-ray Scintillators

Trends in scintillators that are used in many applications, such as medical imaging, security, oil-logging, high energy physics and non-destructive inspections are reviewed. First, we address traditional inorganic and organic scintillators with respect of limitation in the scintillation light yields and lifetimes. The combination of high–light yield and fast response can be found in Ce 3 + , Pr 3 + and Nd 3 + lanthanide-doped scintillators while the maximum light yield conversion of 100,000 photons/MeV can be found in Eu 3 + doped SrI 2 . However, the fabrication of those lanthanide-doped scintillators is inefficient and expensive as it requires high-temperature furnaces. A self-grown single crystal using solution processes is already introduced in perovskite photovoltaic technology and it can be the key for low-cost scintillators. A novel class of materials in scintillation includes lead halide perovskites. These materials were explored decades ago due to the large X-ray absorption cross section. However, lately lead halide perovskites have become a focus of interest due to recently reported very high photoluminescence quantum yield and light yield conversion at low temperatures. In principle, 150,000–300,000 photons/MeV light yields can be proportional to the small energy bandgap of these materials, which is below 2 eV. Finally, we discuss the extraction efficiency improvements through the fabrication of the nanostructure in scintillators, which can be implemented in perovskite materials. The recent technology involving quantum dots and nanocrystals may also improve light conversion in perovskite scintillators.

A novel method for comparing radiation dose and image quality, between and within different x-ray units in a series of hospitals

OBJECTIVES

To develop a novel method for comparing radiation dose and image quality (IQ) to evaluate adult chest x-ray (CXR) imaging among several hospitals.

METHODS

CDRAD 2.0 phantom was used to acquire images in eight hospitals (17 digital x-ray units) using local adult CXR protocols. IQ was represented by image quality figure inverse (IQF_inv), measured using CDRAD analyser software. Signal to noise ratio, contrast to noise ratio and conspicuity index were calculated as additional measures of IQ. Incident air kerma (IAK) was measured using a solid-state dosimeter for each acquisition. Figure of merit (FOM) was calculated to provide a single estimation of IQ and radiation dose.

RESULTS

IQ, radiation dose and FOM varied considerably between hospitals and x-ray units. For IQF_inv, the mean (range) between and within the hospitals were 1.42 (0.83-2.18) and 1.87 (1.52-2.18), respectively. For IAK, the mean (range) between and within the hospitals were 93.56 (17.26-239.15) μGy and 180.85 (122.58-239.15) μGy, respectively. For FOM, the mean (range) between and within hospitals were 0.05 (0.01-0.14) and 0.03 (0.02-0.05), respectively.

CONCLUSIONS

The suggested method for comparing IQ and dose using FOM concept along with the new proposed FOM, is a valid, reliable and effective approach for monitoring and comparing IQ and dose between and within hospitals. It is also can be beneficial for the optimisation of x-ray units in clinical practice. Further optimisation for the hospitals/x-ray units with low FOM are required to minimise radiation dose without degrading IQ.

Chest CT Scan at Radiation Dose of a Posteroanterior and Lateral Chest Radiograph Series: A Proof of Principle in Lymphangioleiomyomatosis

BACKGROUND

Given the rising utilization of medical imaging and the risks of radiation, there is increased interest in reducing radiation exposure. The objective of this study was to evaluate, as a proof of principle, CT scans performed at radiation doses equivalent to that of a posteroanterior and lateral chest radiograph series in the cystic lung disease lymphangioleiomyomatosis (LAM).

METHODS

From November 2016 to May 2018, 105 consecutive subjects with LAM received chest CT scans at standard and ultra-low radiation doses. Standard and ultra-low-dose images, respectively, were reconstructed with routine iterative and newer model-based iterative reconstruction. LAM severity can be quantified as cyst score (percentage of lung occupied by cysts), an ideal benchmark for validating CT scans performed at a reduced dose compared with a standard dose. Cyst scores were quantified using semi-automated software and evaluated by linear correlation and Bland-Altman analysis.

RESULTS

Overall, ultra-low-dose CT scans represented a 96% dose reduction, with a median dose equivalent to 1 vs 22 posteroanterior and lateral chest radiograph series (0.14 mSv; 5th-95th percentile, 0.10-0.20 vs standard dose 3.4 mSv; 5th-95th percentile, 1.5-7.4; P < .0001). The mean difference in cyst scores between ultra-low- and standard-dose CT scans was 1.1% ± 2.0%, with a relative difference in cyst score of 11%. Linear correlation coefficient was excellent at 0.97 (P < .0001).

CONCLUSIONS

In LAM chest CT scan at substantial radiation reduction to doses equivalent to that of a posteroanterior and lateral chest radiograph series provides cyst score quantification similar to that of standard-dose CT scan.

TRIAL REGISTRY

ClinicalTrials.gov; Nos.: NCT00001465 and NCT00001532; URL: www.clinicaltrials.gov.

Automated Chest X-Ray Screening: Can Lung Region Symmetry Help Detect Pulmonary Abnormalities?

Our primary motivator is the need for screening HIV+ populations in resource-constrained regions for exposure to Tuberculosis, using posteroanterior chest radiographs (CXRs). The proposed method is motivated by the observation that radiological examinations routinely conduct bilateral comparisons of the lung field. In addition, the abnormal CXRs tend to exhibit changes in the lung shape, size, and content (textures), and in overall, reflection symmetry between them. We analyze the lung region symmetry using multi-scale shape features, and edge plus texture features. Shape features exploit local and global representation of the lung regions, while edge and texture features take internal content, including spatial arrangements of the structures. For classification, we have performed voting-based combination of three different classifiers: Bayesian network, multilayer perception neural networks, and random forest. We have used three CXR benchmark collections made available by the U.S. National Library of Medicine and the National Institute of Tuberculosis and Respiratory Diseases, India, and have achieved a maximum abnormality detection accuracy (ACC) of 91.00% and area under the ROC curve (AUC) of 0.96. The proposed method outperforms the previously reported methods by more than 5% in ACC and 3% in AUC.

Visual grading analysis of digital neonatal chest phantom X-ray images: Impact of detector type, dose and image processing on image quality

OBJECTIVES

To evaluate the impact of digital detector, dose level and post-processing on neonatal chest phantom X-ray image quality (IQ).

METHODS

A neonatal phantom was imaged using four different detectors: a CR powder phosphor (PIP), a CR needle phosphor (NIP) and two wireless CsI DR detectors (DXD and DRX). Five different dose levels were studied for each detector and two post-processing algorithms evaluated for each vendor. Three paediatric radiologists scored the images using European quality criteria plus additional questions on vascular lines, noise and disease simulation. Visual grading characteristics and ordinal regression statistics were used to evaluate the effect of detector type, post-processing and dose on VGA score (VGAS).

RESULTS

No significant differences were found between the NIP, DXD and CRX detectors (p>0.05) whereas the PIP detector had significantly lower VGAS (p< 0.0001). Processing did not influence VGAS (p=0.819). Increasing dose resulted in significantly higher VGAS (p<0.0001). Visual grading analysis (VGA) identified a detector air kerma/image (DAK/image) of ~2.4 μGy as an ideal working point for NIP, DXD and DRX detectors.

CONCLUSIONS

VGAS tracked IQ differences between detectors and dose levels but not image post-processing changes. VGA showed a DAK/image value above which perceived IQ did not improve, potentially useful for commissioning.

KEY POINTS

• A VGA study detects IQ differences between detectors and dose levels. • The NIP detector matched the VGAS of the CsI DR detectors. • VGA data are useful in setting initial detector air kerma level. • Differences in NNPS were consistent with changes in VGAS.

Digital Radiography in Total Hip Arthroplasty: Technique and Radiographic Results

BACKGROUND

Obtaining the ideal acetabular cup position in total hip arthroplasty remains a challenge. Advancements in digital radiography and image analysis software allow the assessment of the cup position during the surgical procedure. This study describes a validated technique for evaluating cup position during total hip arthroplasty using digital radiography.

METHODS

Three hundred and sixty-nine consecutive patients undergoing total hip arthroplasty were prospectively enrolled. Preoperative supine anteroposterior pelvic radiographs were made. Intraoperative anteroposterior pelvic radiographs were made with the patient in the lateral decubitus position. Radiographic beam angle adjustments and operative table adjustments were made to approximate rotation and tilt of the preoperative radiograph. The target for cup position was 30° to 50° abduction and 15° to 35° anteversion. Intraoperative radiographic measurements were calculated and final cup position was determined after strict impingement and range-of-motion testing. Postoperative anteroposterior pelvic radiographs were made. Two independent observers remeasured all abduction and anteversion angles.

RESULTS

Of the cups, 97.8% were placed within 30° to 50° of abduction, with a mean angle (and standard deviation) of 39.5° ± 4.6°. The 2.2% of cups placed outside the target zone were placed so purposefully on the basis of intraoperative range-of-motion testing and patient factors, and 97.6% of cups were placed between 15° and 35° of anteversion, with a mean angle of 26.6° ± 4.7°. Twenty-eight percent of cups were repositioned on the basis of intraoperative measurements. Subluxation during range-of-motion testing occurred in 3% of hips despite acceptable measurements, necessitating cup repositioning. There was 1 early anterior dislocation.

CONCLUSIONS

Placing the acetabular component within a target range is a critical component to minimizing dislocation and polyethylene wear in total hip arthroplasty. Using digital radiography, we positioned the acetabular component in our desired target zone in 97.8% of cases and outside the target zone, purposefully, in 2.2% of cases. When used in conjunction with strict impingement testing, digital radiography allows for predictable cup placement in total hip arthroplasty.

Comparison and optimization of imaging techniques in suspected physical abuse paediatric radiography

OBJECTIVE

This study compares protocols in clinical use for paediatric suspected physical abuse (SPA) radiographic imaging across four National Health Service Trusts in the UK. The patient dose delivered from a SPA skeletal survey was compared between two sites using different imaging technology. Additionally, the technique in use for the abdomen anteroposterior (AP) radiographic projection was optimized at one of the participant sites.

METHODS

Retrospective data collection was performed to compare SPA protocols. Exposure details and patient dose data for SPA skeletal surveys were collected and compared. SPA skeletal surveys were performed on two anthropomorphic paediatric phantoms using two digital imaging systems. Effective dose (ED) was calculated using a dose calculator software (PCXMC v. 2.0, STUK, Helsinki, Finland) and used as a quantification of the radiation risk. For the optimization study, abdomen AP radiographs of the phantoms were acquired over a range of tube potentials (40-117 kV) for constant ED on a digital radiography (DR) system. The contrast-to-noise ratio (CNR) between "bone" and "soft tissue" in the images was measured and used as an indicator of image quality.

RESULTS

This study showed that there is a variation in the protocols and a range of techniques in use for SPA imaging across the four participant sites. The skeletal surveys undertaken on the newborn phantom at two sites resulted in an ED of 57 ± 3 µSv and 90 ± 4 µSv, on the DR unit and digital radiography/fluoroscopy (dRF) unit, respectively. Measurements of the abdomen AP projection achieved an improved CNR (4%) at a lower tube potential (55 kV) without increasing ED, compared with the current clinical setting (64 kV). Advances in knowledge: This study showed that an improved CNR can be achieved for newborn and 1-year-old abdomen AP radiographs using 0.1 mm copper filtration and a reduced kV (55 kV) without increasing ED.

Digital imaging and radiographic practise in diagnostic radiography: An overview of current knowledge and practice in Europe

INTRODUCTION

Recent research has identified the issue of 'dose creep' in diagnostic radiography and claims it is due to the introduction of CR and DR technology. More recently radiographers have reported that they do not regularly manipulate exposure factors for different sized patients and rely on pre-set exposures. The aim of the study was to identify any variation in knowledge and radiographic practice across Europe when imaging the chest, abdomen and pelvis using digital imaging.

METHODS

A random selection of 50% of educational institutes (n = 17) which were affiliated members of the European Federation of Radiographer Societies (EFRS) were contacted via their contact details supplied on the EFRS website. Each of these institutes identified appropriate radiographic staff in their clinical network to complete an online survey via SurveyMonkey. Data was collected on exposures used for 3 common x-ray examinations using CR/DR, range of equipment in use, staff educational training and awareness of DRL. Descriptive statistics were performed with the aid of Excel and SPSS version 21.

RESULTS

A response rate of 70% was achieved from the affiliated educational members of EFRS and a rate of 55% from the individual hospitals in 12 countries across Europe. Variation was identified in practice when imaging the chest, abdomen and pelvis using both CR and DR digital systems. There is wide variation in radiographer training/education across countries.

CONCLUSION

There is a need for standardisation of education and training including protocols and exposure parameters to ensure that there is continued adherence to the ALARA principle.

Foreign Circular Element Detection in Chest X-Rays for Effective Automated Pulmonary Abnormality Screening

In automated chest X-ray screening (to detect pulmonary abnormality: Tuberculosis (TB), for instance), the presence of foreign element such as buttons and medical devices hinders its performance. In this paper, using digital chest radiographs, the authors present a new technique to detect circular foreign element, within the lung regions. They first compute edge map by using several different edge detection algorithms, which is followed by morphological operations for potential candidate selection. These candidates are then confirmed by using circular Hough transform (CHT). In their test, the authors have achieved precision, recall, and F1 score of 96%, 90%, and 92%, respectively with lung segmentation. Compared to state-of-the-art work, their technique excels performance in terms of both detection accuracy and computational time.

Supine to upright lung mechanics: do changes in lung shape influence lung tissue deformation?

In this study we analyze lung shape change between the upright and supine postures and the effect of this shape change on the deformation of lung tissue under gravity. We use supine computed tomography images along with upright tomosynthesis images obtained on the same day to show that there is significant diaphragmatic movement between postures. Using a continuum model of lung tissue deformation under gravity we show that the shape changes due to this diaphragmatic movement could result in different lung tissue expansion patterns between supine and upright lungs. This is an essential consideration when interpreting imaging data acquired in different postures or translating data acquired in supine imaging to upright function.

Substrate and Passivation Techniques for Flexible Amorphous Silicon-Based X-ray Detectors

Flexible active matrix display technology has been adapted to create new flexible photo-sensing electronic devices, including flexible X-ray detectors. Monolithic integration of amorphous silicon (a-Si) PIN photodiodes on a flexible substrate poses significant challenges associated with the intrinsic film stress of amorphous silicon. This paper examines how altering device structuring and diode passivation layers can greatly improve the electrical performance and the mechanical reliability of the device, thereby eliminating one of the major weaknesses of a-Si PIN diodes in comparison to alternative photodetector technology, such as organic bulk heterojunction photodiodes and amorphous selenium. A dark current of 0.5 pA/mm² and photodiode quantum efficiency of 74% are possible with a pixelated diode structure with a silicon nitride/SU-8 bilayer passivation structure on a 20 µm-thick polyimide substrate.