Digital Chest Images Research Articles

Analysis of image quality in digital chest imaging

An evaluation of the image quality of an amorphous silicon flat-panel detector system and a computed radiology system compared with a screen-film system was performed by means of contrast-detail phantom images. Hard and soft copy images were evaluated. Although patient dose at clinical settings was strongly decreased with the amorphous silicon system, the low-contrast visibility with this system was still significantly better than with the screen-film system. For the computed radiology system, low-contrast visibility was comparable to the screen-film system. Best results were obtained by soft copy reading at full resolution with adaptation of contrast and brightness. Changing tube voltage (102-133 kV), or additional filtration, did not significantly affect image quality. However, low-contrast visibility improved significantly with increasing exposure. It was clearly demonstrated that, in chest imaging, the amorphous silicon system has superior imaging characteristics compared to the screen-film and the computed radiology system.

Distributions of scatter-to-primary and signal-to-noise ratios per pixel in digital chest imaging

The aim of this work was to calculate distributions of scatter-to-primary ratios (epsilon(s)/epsilon(p)) and signal-to-noise ratios per pixel (SNRp) in chest images. Such distributions may provide useful information on how physical image quality (contrast, SNR) is distributed over the posterior/anterior (PA) chest image. A Monte Carlo computer program was used for the calculations, including a model of both the patient (voxel phantom) and the imaging system (X-ray tube, anti-scatter grid and image detector). The calculations were performed for three PA thicknesses 20, 24 and 28 cm. For a 24 cm patient, the epsilon(s)/epsilon(p) varies between 0.5 in the lung to 2.5 behind the spine and heart. The corresponding variation of the SNRp is a factor of 3, with the highest values in the lung. Increasing the patient thickness from 20 to 28 cm increases the epsilon(s)/epsilon(p) by a factor of 2.2 behind the spine and heart.

The influence of patient thickness and imaging system on patient dose and physical image quality in digital chest imaging

The aim of this work was to study the influence of patient thickness, tube voltage and image detector on patient dose, contrast and ideal observer signal-to-noise ratio (SNR(I)), for pathological details positioned at different regions in the image in posterior-anterior (PA) chest radiology. A Monte Carlo computational model was used to compute measures of physical image quality (contrast, SNR(I)) and patient effective dose, E. Two metastasis-like details positioned in the central right lung and right lung near the spine, respectively, were studied. The tube voltage was varied between 100 and 150 kV and the patient thickness between 20 and 28 cm. Both, a computed radiography (CR) system and a direct radiography (DR) system, were investigated. The DR system provides both lower doses and better image quality compared with the CR system. The SNR(I)2/E is approximately 2.9 times higher for the DR system compared with the CR system.

Optimization of the chest exposure condition with a contrast-detail phantom: evaluation of the flat-panel versus computed radiography systems

In this study, we evaluated the performance of a digital chest imaging system using a contrast-detail (C-D) phantom. In the initial step, 76 sample images of the C-D phantom were produced by changing the doses from 0.5, 0.75, 1.0, 1.25, 1.5, to 2.0 times the dose for a screen-film (S/F) system. The sample images were analyzed by five radiological technologists and two medical physicists, and the image quality figure (IQF) was determined. The quality of each image was examined, and appropriate doses were determined from the calculated IQF to obtain the same image quality for other digital chest imaging systems. The method of determining IQF from C-D phantom analysis was very useful for comparing image quality and determining radiographic techniques.

Irreversible JPEG Compression of Digital Chest Radiographs for Primary Interpretation: Assessment of Visually Lossless Threshold

To determine if digital chest images could be compressed in a primary interpretation context without perceived loss of fidelity (below the visually lossless threshold) at transilluminated film or cathode ray tube (CRT) display. One hundred forty-four posteroanterior radiographs were obtained with a digital chest radiography system. At both film and CRT display, an identified original image was presented side by side with a replicate, which was either an unaltered image or an image that had been Joint Photographic Experts Group (JPEG) compressed to 10:1, 20:1, or 50:1 and reconstructed. Each of the 10 readers indicated whether the replicate was "indistinguishable from the original" or "degraded" at clinical reading distance and at close inspection. The readers' ability to detect compressed images was examined for patterns; 95% CIs were used for statistical testing. With transilluminated film at clinical reading distance, readers were as likely to rate originals (48 [20%] of 240 readings) as degraded as they were to rate 20:1 replicates (106 [22%] of 480 readings) as degraded, but they frequently identified 50:1 replicates (283 [59%] of 480 readings) as degraded. At close inspection, 20:1 replicates (163 [34%] of 480 readings) were often identified as degraded, but 10:1 replicates (19 [8%] of 240 readings) were not identified as degraded more often than originals (17 [7%] of 240 readings). With CRT display, the results were nearly identical. At reading distance for primary interpretation, full-size digital chest radiographs that have been JPEG compressed to 10:1 or 20:1 and reconstructed are visually lossless at film or CRT display. Images compressed to 10:1 remain visually lossless at close inspection.

Digital chest radiography: quality assurance.

Digital chest radiography: quality assurance.

Computerized detection of pulmonary nodules in chest radiographs based on morphological features and wavelet snake model

We have developed a new computer-aided diagnosis scheme for automated detection of lung nodules in digital chest radiographs based on a combination of morphological features and the wavelet snake. In our scheme, two processes were applied in parallel to reduce the false-positive detections after initial nodule candidates were selected. One process consisted of adaptive filtering for enhancement of nodules and suppression of normal lung structures, followed by extraction of conventional morphological features. The other process consisted of a novel approach for elimination of false positives called the edge-guided wavelet snake model. In the latter process, multiscale edges of the candidate nodules were extracted to yield parts of the nodule boundaries. A wavelet snake was then used for fitting of these multiscale edges for approximation of the true boundaries of nodules. A boundary feature called the weighted overlap between the snake and the multiscale edges was calculated and used for elimination of false positives. Finally, the weighted overlap and the morphological features were combined by use of an artificial neural network for efficient reduction of false positives. Our scheme was applied to a publicly available database of digital chest images for pulmonary nodules. Receiver operating characteristic analysis was employed for evaluation of the performance of each process in the scheme. The combined features yielded a large reduction of false positives, and thus achieved a high performance in discriminating between true and false positives. These results show that our new method, in particular the false-positive reduction method based on the wavelet snake, is effective in improving the performance of a computerized scheme for detection of pulmonary nodules in chest radiographs.

Studies on computer‐aided diagnosis systems for chest radiographs and mammograms (in Japanese)

This thesis describes computer‐aided diagnosis (CAD) systems for chest radiographs and mammograms. Preprocessing and imaging processing methods for each CAD system include dynamic range compression and region segmentation technique. A new pattern recognition technique combines genetic algorithms with template matching methods to detect lung nodules. A genetic algorithm was employed to select the optimal shape of simulated nodular shadows to be compared with real lesions on digitized chest images. Detection performance was evaluated using 332 chest radiographs from the database of the Japanese Society of Radiological Technology. Our average true‐positive rate was 72.8% with an average of 11 false‐positive findings per image. A new detection method using high resolution digital images with 0.05 mm sampling is also proposed for the mammogram CAD system to detect very small microcalcifications. An automated classification method uses feature extraction based on fractal dimension analysis of masses. Using over 200 cases to evaluate the detection of mammographic masses and calcifications, the detection rate of masses and microcalcifications were 87% and 96% with 1.5 and 1.8 false‐positive findings, respectively. The classification performance on benign vs malignant lesions, the Az values that were defined by the areas under the ROC curves derived from classification schemes of masses and microcalcifications were 0.84 and 0.89. To demonstrate the practicality of these CAD systems in a computer‐network environment, we propose to use the mammogram CAD system via the Internet and WWW. A common gateway interface and server‐client approach for the CAD system via the Internet will permit display of the CAD results on ordinary computers.

Effect of dose levels on the diagnostic performance of a selenium-based digital chest system.

To evaluate the effect of dose reduction on diagnostic performance by using a digital chest imaging system in which amorphous selenium serves as the x-ray detector. Two hundred forty-seven patients were examined with the selenium system. Three sets of images were made in each patient: one set with a standard x-ray dose, one set with 55% of the standard dose, and one set with 35% of the standard dose. All 741 images were read by two radiologists. The diagnostic value of each set of images for detection of pulmonary, mediastinal, and pleural pathology was analyzed with receiver operating characteristic (ROC) methodology by using computed tomography as the reference standard. The authors also assessed the effect of sex, height, and weight of the patients on the diagnostic performance of the readers. The areas under the ROC curves for the detection of various abnormalities at 100%, 55%, and 35% of the standard dose for observer 1, respectively, were pulmonary opacities 0.82, 0.83, 0.84; interstitial disease 0.71, 0.70, 0.72; mediastinal disease 0.81, 0.80, 0.77; and pleural abnormalities 0.71, 0.72, 0.72. There were no statistically significant differences between the two observers. No statistically significant difference was found between the radiologists' performance in detecting abnormalities with standard x-ray dose images and the performance with images made with 55% and 35% of the standard dose. Sex, height, and weight had no influence on diagnostic performance.

Gradation processing algorithm of digital radiological chest image.

Gradation processing technology can enhance display contrast of interested region by using linear or nonlinear transform. Current algorithms generally calculate transform parameters depended on experimental statistical values or some simple image processing method such as histogram analysis. They cannot accord with the human subjective vision mechanism nor can they intelligently acquire required parameters. Therefore, processing results are not very accurate or reliable. We propose a gradation processing algorithm of digital radiological chest images based on region growing segmentation technology. It combines self-adapted segmentation with prior knowledge of directed segmentation, and its computation cost is low. It can perform contrast enhancement of the interested regions more accurately, and accords with the human subjective vision mechanism better.

Application of artificial neural networks for quantitative analysis of image data in chest radiographs for detection of interstitial lung disease.

The authors have developed an automated computeraided diagnostic (CAD) scheme by using artificial neural networks (ANNs) on quantitative analysis of image data. Three separate ANNs were applied for detection of interstitial disease on digitized chest images. The first ANN was trained with horizontal profiles in regions of interest (ROIs) selected from normal and abnormal chest radiographs for distinguishing between normal and abnormal patterns. For training and testing of the second ANN, the vertical output patterns obtained from the 1st ANN were used for each ROI. The output value of the second ANN was used to distinguish between normal and abnormal ROIs with interstitial infiltrates. If the ratio of the number of abnormal ROIs to the total number of all ROIs in a chest image was greater than a specified threshold level, the image was classified as abnormal. In addition, the third ANN was applied to distinguish between normal and abnormal chest images. The combination of the rule-based method and the third ANN also was applied to the classification between normal and abnormal chest images. The performance of the ANNs was evaluated by means of receiver operating characteristic (ROC) analysis. The average Az value (area under the ROC curve) for distinguishing between normal and abnormal cases was 0.976 +/- 0.012 for 100 chest radiographs that were not used in training of ANNs. The results indicate that the ANN trained with image data can learn some statistical properties associated with interstitial infiltrates in chest radiographs.

Development of an improved CAD scheme for automated detection of lung nodules in digital chest images.

Lung cancer is the leading cause of cancer deaths in men and women in the United States, with a 5-year survival rate of only about 13%. However, this survival rate can be improved to 47% if the disease is diagnosed and treated at an early stage. In this study, we developed an improved computer-aided diagnosis (CAD) scheme for the automated detection of lung nodules in digital chest images to assist radiologists, who could miss up to 30% of the actually positive cases in their daily practice. Two hundred PA chest radiographs, 100 normals and 100 abnormals, were used as the database for our study. The presence of nodules in the 100 abnormal cases was confirmed by two experienced radiologists on the basis of CT scans or radiographic follow-up. In our CAD scheme, nodule candidates were selected initially by multiple gray-level thresholding of the difference image (which corresponds to the subtraction of a signal-enhanced image and a signal-suppressed image) and then classified into six groups. A large number of false positives were eliminated by adaptive rule-based tests and an artificial neural network (ANN). The CAD scheme achieved, on average, a sensitivity of 70% with 1.7 false positives per chest image, a performance which was substantially better as compared with other studies. The CPU time for the processing of one chest image was about 20 seconds on an IBM RISC/6000 Powerstation 590. We believe that the CAD scheme with the current performance is ready for initial clinical evaluation.

Subjective Image Quality Evaluation of Image Compression Techniques

Image compression reduces the amount of data in digital images and, therefore, allows efficient storage, processing, and transmission of pictorial information. However, compression algorithms can degrade image quality by introducing artifacts, which may be unacceptable for users' tasks. This work examined the subjective effects of JPEG and wavelet compression algorithms on a series of medical images. Six digitized chest images were processed by each algorithm at various compression levels. Twelve radiologists rated the perceived image quality of the compressed images relative to the corresponding uncompressed images, as well as rated the acceptability of the compressed images for diagnostic purposes. The results indicate that subjective image quality and acceptability decreased with increasing compression levels; however, all images remained acceptable for diagnostic purposes. At high compression ratios, JPEG compressed images were judged less acceptable for diagnostic purposes than the wavelet compressed images. These results contribute to emerging system design guidelines for digital imaging workstations.

Economic aspects and cost calculations for a digital selenium-based chest imaging system.

Economic aspects and cost calculations for a digital selenium-based chest imaging system.

Computerized analysis for automated detection of lung nodules in digital chest radiographs

The objective of this study was to develop a computer‐aided diagnosis(CAD) scheme for automated detection of lung nodules in digital chest images in order to help radiologists to improve their performance in detecting lung nodules. Lung nodules in chest radiographs are the important sign of lungcancer which is the leading cause of cancer death in men and women in the United States, with the 5‐year survival rate only about 13%. Early detection and treatment are crucial for the improvement of this survival rate for patients with lungcancer. This dissertation included the following four topics: (1) accurate determination of ribcage boundary in digital chest radiographs; (2) detection of right and left hemidiaphragm edges and delineation of lung fields in digital chest radiographs; (3) development of a CAD scheme for automated detection of lung nodules in digital chest radiographs; and (4) ROC studies of effects of CAD output on radiologists’ performance for detection of lung nodules on chest radiographs. Our CAD scheme for automated detection of lung nodules in digital chest images was based on the segmentation of lung fields, the difference image technique, image feature analysis, and an artificial neural network. The performance of the CAD scheme was evaluated by the FROC methodology. The CAD scheme achieved an improved performance of 70% sensitivity with 1.7 false positives per image, on average. The ROC analysis results indicated that radiologists increased their performance significantly in detecting lung nodules by use of the CAD output as a ‘‘second opinion.’’ Radiologists did not increase their reading time when the CAD output was used. [Thesis copies are available upon request from the author.]

Effects of dose reduction on digital chest imaging using a selenium detector: a study of detecting simulated diffuse interstitial pulmonary disease.

The purpose of this study was to evaluate the effect of dose reduction on the diagnostic performance of a new digital chest imaging system in which amorphous selenium is used as the X-ray detector. Diagnostic performance was assessed for the detection of simulated diffuse interstitial pulmonary disease (DIPD). DIPD was simulated by superimposing plastic sheets that contained small radiopaque objects (birdseed) on an anthropomorphic chest phantom. We varied the number of sheets from zero to four to simulate the degree of abnormality. We made 80 images with a standard X-ray dose, 80 images with 55% of the standard dose, and 80 images with 35% of the standard dose. Six observers were asked to indicate the presence of DIPD using a five-level scale of confidence. Two hundred forty chest images were then analyzed using receiver operating characteristic (ROC) curves. The area under the ROC curve was 87.2 for all readers with standard-dose imaging (95% confidence interval [CI], 83.7-90.7), 91.7 with 55% of the standard dose (95% CI, 88.8-94.6), and 90.0 with the 35% dose (95% CI, 87.1-92.9). The area under the ROC curve for subtle DIPD (one superimposed sheet) was 75.3 for all readers with standard-dose imaging (95% CI, 67.1-83.5), 79.7 with 55% of the standard dose (95% CI, 71.9-87.5), and 70.3 with the 35% dose (95% CI, 61.7-78.9). For each dose, we observed a gradual improvement of the ROC curves with each additional sheet superimposed on the chest phantom (p < .001). We found no significant difference in diagnostic performance among images made with standard X-ray dose, those made with a 55% dose, and those made with a 35% dose (95% CI).

448. 2重IPカセッテによるディジタル胸部写真の比較

448. 2重IPカセッテによるディジタル胸部写真の比較

Dynamic Range Control Processing of Digital Chest Images

Purpose: The clinical usefulness of an advanced image-processing system called “dynamic range control processing” was investigated, with which selected parts of the dynamic range of digital chest images could be controlled.Material and Methods: A comparative study of 3 different post-processed formats of storage phosphor (SR) images was performed in 35 patients with abnormalities in the chest. The 3 formats were SR images with standard mode (SR-standard), SR images with strong edge-enhancement (SR-enhanced), and dynamic range controlled SR images (SR-controlled).Results: For lung abnormalities, there was no difference among the 3 SR image formats. For normal mediastinal structures and lung abnormalities covered by the heart or diaphragm, SR-controlled and SR-enhanced images were significantly superior to SR-standard images, while no difference was found between SR-controlled and SR-enhanced images except for the trachea and bony structures.Conclusion: Dynamic range control processing appears to be a usefu...

Digital chest imaging using a selenium detector. A simulated diffuse interstitial pulmonary disease detection study.

To assess the diagnostic performance of a new digital chest imaging system based on a selenium detector in the detection of simulated diffuse interstitial pulmonary disease (DIPD). Diffuse interstitial pulmonary disease was simulated by superimposing plastic sheets containing small radio-paque objects (birdseed) on an anthropomorphic chest phantom. the number of sheets was varied from 0 to 4 to simulate the degree of pathology. Eighty conventional radiographs and 80 images with the selenium detector were obtained under comparable conditions. Six observers were asked to indicate the presence of DIPD using a five-level scale of confidence. The total 160 chest images were analyzed using receiver operating characteristic curves. There was no significant difference in detection performance between the conventional and the digital images (95% confidence interval). None of the observers assessed the diagnostic performance of the digital chest images as significantly different from that of the conventional system. The interobserver differences were higher than the differences between digital and conventional chest images.

Comparison of hard- and soft-copy digital chest images with different matrix sizes for managing coronary care unit patients.

This study was to determine whether different digital display formats for portable chest radiographs of coronary care unit patients would provide comparable information for clinical care. In particular, we tried to ascertain whether 1024 x 1024 pixel (1K) images on a picture archiving and communication system (PACS) workstation would be comparable to 1760 x 2140 pixel (2K) images on workstations or to digital films. If comparability could be proved, we hypothesized that 1K workstations could considerably lower equipment and film costs and facilitate image transmission from point to point. Four chest radiologists read a panel of chest studies assembled from 98 coronary care unit patients, comparing 1K and 2K soft-copy images with digital hard copies. For all three image types for the 98 patients, the readers evaluated nine image parameters that the cardiologists deemed essential for clinical decision making. Two other chest radiologists reviewed each patient's three image types, historical chest images, current and prior radiologic reports, and medical record to determine the consensus, or "truth findings." With one exception (small pleural effusions), the receiver operating characteristic analysis showed no significant differences in the clinical information derived from the three image types. For clinical management in a coronary care unit, comparable information can be obtained from digital radiologic chest studies using a 1K x 1K soft-copy format, a 2K x 2K soft-copy format, or a hard copy (film). Substantial savings in cost and time are therefore possible by using soft-copy images and lower resolution (1K x 1K) workstations and, when necessary, by transmitting images on regular telephone lines.