Search

Our goal was to analyze radiology report text for chest radiographs (CXRs) to identify imaging findings that have the most impact on report length and complexity. Identifying these imaging findings can highlight opportunities for designing CXR AI systems which increase radiologist efficiency. We retrospectively analyzed text from 210,025 MIMIC-CXR reports and 168,949 reports from our local institution collected from 2019 to 2022. Fifty-nine categories of imaging finding keywords were extracted from reports using natural language processing (NLP), and their impact on report length was assessed using linear regression with and without LASSO regularization. Regression was also used to assess the impact of additional factors contributing to report length, such as the signing radiologist and use of terms of perception. For modeling CXR report word counts with regression, mean coefficient of determination, R2, was 0.469 ± 0.001 for local reports and 0.354 ± 0.002 for MIMIC-CXR when considering only imaging finding keyword features. Mean R2 was significantly less at 0.067 ± 0.001 for local reports and 0.086 ± 0.002 for MIMIC-CXR, when only considering use of terms of perception. For a combined model for the local report data accounting for the signing radiologist, imaging finding keywords, and terms of perception, the mean R2 was 0.570 ± 0.002. With LASSO, highest value coefficients pertained to endotracheal tubes and pleural drains for local data and masses, nodules, and cavitary and cystic lesions for MIMIC-CXR. Natural language processing and regression analysis of radiology report textual data can highlight imaging targets for AI models which offer opportunities to bolster radiologist efficiency. [ABSTRACT FROM AUTHOR]

Accurate estimation of mortality and time to death at admission for COVID-19 patients is important and several deep learning models have been created for this task. However, there are currently no prognostic models which use end-to-end deep learning to predict time to event for admitted COVID-19 patients using chest radiographs and clinical data. We retrospectively implemented a new artificial intelligence model combining DeepSurv (a multiple-perceptron implementation of the Cox proportional hazards model) and a convolutional neural network (CNN) using 1356 COVID-19 inpatients. For comparison, we also prepared DeepSurv only with clinical data, DeepSurv only with images (CNNSurv), and Cox proportional hazards models. Clinical data and chest radiographs at admission were used to estimate patient outcome (death or discharge) and duration to the outcome. The Harrel's concordance index (c-index) of the DeepSurv with CNN model was 0.82 (0.75–0.88) and this was significantly higher than the DeepSurv only with clinical data model (c-index = 0.77 (0.69–0.84), p = 0.011), CNNSurv (c-index = 0.70 (0.63–0.79), p = 0.001), and the Cox proportional hazards model (c-index = 0.71 (0.63–0.79), p = 0.001). These results suggest that the time-to-event prognosis model became more accurate when chest radiographs and clinical data were used together. [ABSTRACT FROM AUTHOR]

Our objective is to investigate the reliability and usefulness of anatomic point–based lung zone segmentation on chest radiographs (CXRs) as a reference standard framework and to evaluate the accuracy of automated point placement. Two hundred frontal CXRs were presented to two radiologists who identified five anatomic points: two at the lung apices, one at the top of the aortic arch, and two at the costophrenic angles. Of these 1000 anatomic points, 161 (16.1%) were obscured (mostly by pleural effusions). Observer variations were investigated. Eight anatomic zones then were automatically generated from the manually placed anatomic points, and a prototype algorithm was developed using the point-based lung zone segmentation to detect cardiomegaly and levels of diaphragm and pleural effusions. A trained U-Net neural network was used to automatically place these five points within 379 CXRs of an independent database. Intra- and inter-observer variation in mean distance between corresponding anatomic points was larger for obscured points (8.7 mm and 20 mm, respectively) than for visible points (4.3 mm and 7.6 mm, respectively). The computer algorithm using the point-based lung zone segmentation could diagnostically measure the cardiothoracic ratio and diaphragm position or pleural effusion. The mean distance between corresponding points placed by the radiologist and by the neural network was 6.2 mm. The network identified 95% of the radiologist-indicated points with only 3% of network-identified points being false-positives. In conclusion, a reliable anatomic point–based lung segmentation method for CXRs has been developed with expected utility for establishing reference standards for machine learning applications. [ABSTRACT FROM AUTHOR]

COVID-19 is a highly contagious disease that can cause severe pneumonia. Patients with pneumonia undergo chest X-rays (XR) to assess infiltrates that identify the infection. However, the radiographic characteristics of COVID-19 are similar to the other acute respiratory syndromes, hindering the imaging diagnosis. In this work, we proposed identifying quantitative/radiomic biomarkers for COVID-19 to support XR assessment of acute respiratory diseases. This retrospective study used different cohorts of 227 patients diagnosed with pneumonia; 49 of them had COVID-19. Automatically segmented images were characterized by 558 quantitative features, including gray-level histogram and matrices of co-occurrence, run-length, size zone, dependence, and neighboring gray-tone difference. Higher-order features were also calculated after applying square and wavelet transforms. Mann–Whitney U test assessed the diagnostic performance of the features, and the log-rank test assessed the prognostic value to predict Kaplan–Meier curves of overall and deterioration-free survival. Statistical analysis identified 51 independently validated radiomic features associated with COVID-19. Most of them were wavelet-transformed features; the highest performance was the small dependence matrix feature of "low gray-level emphasis" (area under the curve of 0.87, sensitivity of 0.85, p < 0.001 ). Six features presented short-term prognostic value to predict overall and deterioration-free survival. The features of histogram "mean absolute deviation" and size zone matrix "non-uniformity" yielded the highest differences on Kaplan–Meier curves with a hazard ratio of 3.20 ( p < 0.05 ). The radiomic markers showed potential as quantitative measures correlated with the etiologic agent of acute infectious diseases and to stratify short-term risk of COVID-19 patients. [ABSTRACT FROM AUTHOR]

Assess the efficacy of deep convolutional neural networks (DCNNs) in detection of critical enteric feeding tube malpositions on radiographs. 5475 de-identified HIPAA compliant frontal view chest and abdominal radiographs were obtained, consisting of 174 x-rays of bronchial insertions and 5301 non-critical radiographs, including normal course, normal chest, and normal abdominal x-rays. The ground-truth classification for enteric feeding tube placement was performed by two board-certified radiologists. Untrained and pretrained deep convolutional neural network models for Inception V3, ResNet50, and DenseNet 121 were each employed. The radiographs were fed into each deep convolutional neural network, which included untrained and pretrained models. The Tensorflow framework was used for Inception V3, ResNet50, and DenseNet. Images were split into training (4745), validation (630), and test (100). Both real-time and preprocessing image augmentation strategies were performed. Receiver operating characteristic (ROC) and area under the curve (AUC) on the test data were used to assess the models. Statistical differences among the AUCs were obtained. p < 0.05 was considered statistically significant. The pretrained Inception V3, which had an AUC of 0.87 (95 CI; 0.80–0.94), performed statistically significantly better (p <.001) than the untrained Inception V3, with an AUC of 0.60 (95 CI; 0.52–0.68). The pretrained Inception V3 also had the highest AUC overall, as compared with ResNet50 and DenseNet121, with AUC values ranging from 0.82 to 0.85. Each pretrained network outperformed its untrained counterpart. (p < 0.05). Deep learning demonstrates promise in differentiating critical vs. non-critical placement with an AUC of 0.87. Pretrained networks outperformed untrained ones in all cases. DCNNs may allow for more rapid identification and communication of critical feeding tube malpositions. [ABSTRACT FROM AUTHOR]

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

It is difficult for radiologists to classify pneumoconiosis from category 0 to category 3 on chest radiographs. Therefore, we have developed a computer-aided diagnosis (CAD) system based on a three-stage artificial neural network (ANN) method for classification based on four texture features. The image database consists of 36 chest radiographs classified as category 0 to category 3. Regions of interest (ROIs) with a matrix size of 32 × 32 were selected from chest radiographs. We obtained a gray-level histogram, histogram of gray-level difference, gray-level run-length matrix (GLRLM) feature image, and gray-level co-occurrence matrix (GLCOM) feature image in each ROI. For ROI-based classification, the first ANN was trained with each texture feature. Next, the second ANN was trained with output patterns obtained from the first ANN. Finally, we obtained a case-based classification for distinguishing among four categories with the third ANN method. We determined the performance of the third ANN by receiver operating characteristic (ROC) analysis. The areas under the ROC curve (AUC) of the highest category (severe pneumoconiosis) case and the lowest category (early pneumoconiosis) case were 0.89 ± 0.09 and 0.84 ± 0.12, respectively. The three-stage ANN with four texture features showed the highest performance for classification among the four categories. Our CAD system would be useful for assisting radiologists in classification of pneumoconiosis from category 0 to category 3. [ABSTRACT FROM AUTHOR]

The objective of this study was to implement and evaluate the performance of a biplane correlation imaging (BCI) technique aimed to reduce the effect of anatomic noise and improve the detection of lung nodules in chest radiographs. Seventy-one low-dose posterior-anterior images were acquired from an anthropomorphic chest phantom with 0.28° angular separations over a range of ±10° along the vertical axis within an 11 s interval. Similar data were acquired from 19 human subjects with institutional review board approval and informed consent. The data were incorporated into a computer-aided detection (CAD) algorithm in which suspect lesions were identified by examining the geometrical correlation of the detected signals that remained relatively constant against variable anatomic backgrounds. The data were analyzed to determine the effect of angular separation, and the overall sensitivity and false-positives for lung nodule detection. The best performance was achieved for angular separations of the projection pairs greater than 5°. Within that range, the technique provided an order of magnitude decrease in the number of false-positive reports when compared with CAD analysis of single-view images. Overall, the technique yielded ~1.1 false-positive per patient with an average sensitivity of 75%. The results indicated that the incorporation of angular information can offer a reduction in the number of false-positives without a notable reduction in sensitivity. The findings suggest that the BCI technique has the potential for clinical implementation as a cost-effective technique to improve the detection of subtle lung nodules with lowered rate of false-positives. [ABSTRACT FROM AUTHOR]

It is difficult for radiologists to classify pneumoconiosis with small nodules on chest radiographs. Therefore, we have developed a computer-aided diagnosis (CAD) system based on the rule-based plus artificial neural network (ANN) method for distinction between normal and abnormal regions of interest (ROIs) selected from chest radiographs with and without pneumoconiosis. The image database consists of 11 normal and 12 abnormal chest radiographs. These abnormal cases included five silicoses, four asbestoses, and three other pneumoconioses. ROIs (matrix size, 32 × 32) were selected from normal and abnormal lungs. We obtained power spectra (PS) by Fourier transform for the frequency analysis. A rule-based method using PS values at 0.179 and 0.357 cycles per millimeter, corresponding to the spatial frequencies of nodular patterns, were employed for identification of obviously normal or obviously abnormal ROIs. Then, ANN was applied for classification of the remaining normal and abnormal ROIs, which were not classified as obviously abnormal or normal by the rule-based method. The classification performance was evaluated by the area under the receiver operating characteristic curve (Az value). The Az value was 0.972 ± 0.012 for the rule-based plus ANN method, which was larger than that of 0.961 ± 0.016 for the ANN method alone ( P ≤ 0.15) and that of 0.873 for the rule-based method alone. We have developed a rule-based plus pattern recognition technique based on the ANN for classification of pneumoconiosis on chest radiography. Our CAD system based on PS would be useful to assist radiologists in the classification of pneumoconiosis. [ABSTRACT FROM AUTHOR]

To evaluate the number of actual detections versus “accidental” detections by a computer-aided detection (CAD) system for small nodular lung cancers (≤30 mm) on chest radiographs, using two different criteria for measuring performance. A Food-and-Drug-Administration-approved CAD program (version 1.0; Riverain Medical) was applied to 34 chest radiographs with a “radiologist-missed” nodular cancer and 36 radiographs with a radiologist-mentioned nodule (a newer version 3.0 was also applied to the 36-case database). The marks applied by this CAD system consisted of 5-cm-diameter circles. A strict “nodule-in-center” criterion and a generous “nodule-in-circle” criterion were compared as methods for the calculation of CAD sensitivity. The increased sensitivities by the nodule-in-circle criterion were considered as nodules detected by chance. The number of false-positive (FP) marks was also analyzed. For the 34 radiologist-missed cancers, the nodule-in-circle criterion caused eight more cancers (24%) to be detected by chance, as compared to the nodule-in-center criterion, when using the version 1.0 results. For the 36 radiologist-mentioned nodules, the nodule-in-circle criterion caused seven more lesions (19%) to be detected by chance, as compared to the nodule-in-center criterion, when using the version 1.0 results, and three more lesions (8%) to be detected by chance when using the version 3.0 results. Version 1.0 yielded a mean of six FP marks per image, while version 3.0 yielded only three FP marks per image. The specific criteria used to define true- and false-positive CAD detections can substantially influence the apparent accuracy of a CAD system. [ABSTRACT FROM AUTHOR]

The objective of this study was to analyze image quality of chest examinations in pediatric patients using computed radiography (CR) obtained with a wide range of doses to suggest the appropriate parameters for optimal image quality. A sample of 240 chest images in four age ranges was randomly selected from the examinations performed during 2004. Images were obtained using a CR system and were evaluated independently by three radiologists. Each image was scored using criteria proposed by the European Guidelines on Quality Criteria in Pediatrics. Mean global scoring and scoring of individual criteria more sensitive to noise were used to evaluate image quality. Agfa dose level (DL) was in the range 1.20 to 2.85. It was found that there was not significant correlation ( R < 0.5) between image quality and DL for any of the age ranges for either global score or for individual criteria more related to noise. The mean value of DL was in the ranges 1.9–2.1 for the four age bands. From this study, a DL value of 1.6 is proposed for pediatric CR chest imaging. This could yield a reduction of approximately a factor of 2.5 in mean patient entrance surface doses. [ABSTRACT FROM AUTHOR]

The aim of the study is to evaluate the reliable production of temporal subtraction images in a picture archiving and communication system environment and to establish objective criteria for the evaluation of image quality. A total of 117 temporal subtraction chest images (55 in the upright position, 62 in the supine position) were obtained in five consecutive days. In all of these, we confirmed that there were no interval changes on the original images, and cases with diffuse lung disease were excluded. The temporal subtraction images were classified by three chest radiologists into five levels: 5, excellent; 4, good; 3, acceptable; 2, poor; and 1, very poor. The following were examined: (1) the yield of adequate quality of the temporal subtraction images; (2) whether the temporal subtraction images were obtained in the warping or nonwarping mode; and (3) the correlation of the overall subjective image quality with the relative shift angles, relative shift distances, and the standard deviation of gray levels in the temporal subtraction images. The percentages of acceptable temporal subtraction images were 100% and 66% in the upright and supine positions, respectively. Sixteen (26%) of the 62 supine-position images were made in nonwarping mode, whereas all upright images were made in warping mode. Significant correlations were obtained in the relative shift angle (P < 0.05), relative horizontal shift distance (P < 0.05), and standard deviation of gray levels (P < 0.0001). Temporal subtraction images with acceptable image quality were obtained in the upright position. The objective criteria may be useful for the evaluation of image quality. [ABSTRACT FROM AUTHOR]

To evaluate the usefulness of a commercially available computer-assisted diagnosis (CAD) system on operable T1 cases of lung cancer by use of digital chest radiography equipment. Fifty consecutive patients underwent surgery for primary lung cancer, and 50 normal cases were selected. All cancer cases were histopathologically confirmed T1 cases. All normal individuals were selected on the basis of chest computed tomography (CT) confirmation and were matched with cancer cases in terms of age and gender distributions. All chest radiographs were obtained with one computed radiography or two flat-panel detector systems. Eight radiologists (four chest radiologists and four residents) participated in observer tests and interpreted soft copy images by using an exclusive display system without and with CAD output. When radiologists diagnosed cases as positives, the locations of lesions were recorded on hard copies. The observers’ performance was evaluated by receiver operating characteristic analysis. The overall detectability of lung cancer cases with CAD system was 74% (37/50), and the false-positive rate was 2.28 (114/50) false positives per case for normal cases. The mean A z value increased significantly from 0.896 without CAD output to 0.923 with CAD output (P = 0.018). The main cause of the improvement in performance is attributable to changes from false negatives without CAD to true positives with CAD (19/31, 61%). Moreover, improvement in the location of the tumor was observed in 1.5 cases, on average, for radiology residents. This CAD system for digital chest radiographs is useful in assisting radiologists in the detection of early resectable lung cancer. [ABSTRACT FROM AUTHOR]

The purpose of this study was to develop and test a computerized method for the fully automated analysis of abnormal asymmetry in digital posteroanterior (PA) chest radiographs. An automated lung segmentation method was used to identify the aerated lung regions in 600 chest radiographs. Minimal a priori lung morphology information was required for this gray-level thresholding-based segmentation. Consequently, segmentation was applicable to grossly abnormal cases. The relative areas of segmented right and left lung regions in each image were compared with the corresponding area distributions of normal images to determine the presence of abnormal assymetry. Computerized diagnoses were compared with image ratings assigned by a radiologist. The ability of the automated method to distinguish normal from asymmetrically abnormal cases was evaluated by using receiver operating characteristic (ROC) analysis, which yielded an area under the ROC curve of 0.84. This automated method demonstrated promising performance in its ability to detect abnormal asymmetry in PA chest images. We believe this method could play a role in a picture archiving and communications (PACS) environment to immediately identify abnormal cases and to function as one component of a multifaceted computeraided diagnostic scheme. [ABSTRACT FROM AUTHOR]

We devised an automated classification scheme by using the rule-based method plus artificial neural networks (ANN) for distinction between normal and abnormal lungs with interstitial disease in digital chest radiographs. Four measures used in the classification scheme are determined from the texture and geometric-pattern feature analyses. The rms variation and the first moment of the power spectrum of lung patterns aredetermined as measures for the texture analysis. In addition, the total area of nodular opacities and the total length of linear opacities are determined as measures for the geometric-pattern feature analysis. In our classification scheme with these measures, we identify obviously normal and abnormal cases first by the rule-based method and then ANN is applied for the remaining difficult cases. The rulebased plus ANN method provided a sensitivity of 0.926 at the specificity of 0.900, which was considerably improved compared to performance of either the rule-based method alone or ANNs alone. [ABSTRACT FROM AUTHOR]