Akiko Kano

Digital image subtraction of temporally sequential chest images for detection of interval change

An automated digital image subtraction technique for temporally sequential chest images has been developed in order to aid radiologists in the detection of interval changes. A number of small regions of interest (ROIs) are selected automatically in the lung areas of two temporally sequential chest images. A local matching, based on a cross-correlation method, is performed on each pair of corresponding ROIs in order to determine shift values for the coordinates of two images. A proper warping of x,y coordinates is obtained by fitting two-dimensional polynomials to the distributions of shift values. One of the images is warped and then subtracted from the other. Forty six pairs of chest images (42 with interval changes and 4 without interval change) were processed using this method. The subtraction images were able to enhance various important interval changes, such as differences in the size of tumor masses, changes in heart size, and changes in pulmonary infiltrates or pleural effusions. Approximately 70% of the pairs showed reasonably good registration.

IMAGE FEATURE ANALYSIS OF FALSE-POSITIVE DIAGNOSES PRODUCED BY AUTOMATED DETECTION OF LUNG NODULES

RATIONALE AND OBJECTIVES To reduce the number of false-negative diagnoses by radiologists, the authors are developing a computer-aided diagnosis scheme for detection of lung nodules in digital chest images. In this study, the authors attempted to reduce the number of false-positive diagnoses obtained with a previous computer scheme by incorporating additional knowledge from experienced chest radiologists into the computer scheme. METHODS The authors applied their previous computer scheme, using less-strict criteria, to 60 clinical chest radiographs; this yielded 735 candidate nodules (23 true nodules and 712 false-positive diagnoses). These candidates were analyzed using region-growing, trend-correction, and edge-gradient techniques to determine measures by which to quantify image features of candidate nodules. RESULTS The 712 false-positive diagnoses represented various anatomic structures that were located throughout the chest image. From this analysis, we were able to decrease the number of false-positive errors from an average of 12 to approximately 5 per image without eliminating any true nodules. CONCLUSION Our results show that incorporating knowledge from experienced chest radiologists into the computer algorithm will play an important role in the development of computerized schemes for the detection of pulmonary nodules.

Computer-aided diagnosis in chest radiography. Preliminary experience.

RATIONALE AND OBJECTIVES.Computer-aided diagnosis (CAD) schemes for chest radiography are being developed with which to alert radiologists to possible lesions, and thus potentially improve diagnostic accuracy. However, CAD schemes have not been tested on a large number of clinical cases. The authors identify design parameters that would be required for development of an intelligent workstation. METHODS.Computer-aided diagnosis programs were applied for the automated detection of lung nodules, cardiomegaly, and interstitial infiltrates to 310 consecutive chest radiographs, and were analyzed for potential usefulness and limitations. Computer-aided diagnosis output was evaluated by radiologists and physicists for accuracy and technical problems, respectively. RESULTS.Approximately 70% of the results were judged to be potentially acceptable; however, the number of false-positive findings was relatively high. Technical problems included failure to detect subtle abnormalities and the occurrence of false-positive detections caused by normal anatomical structures. CONCLUSION.Computer-aided diagnosis has the potential to be a valuable aid to radiologists in clinical practice, if certain technical problems can be overcome and if optimal operating points can be defined for clinical use.

Development of a digital duplication system for portable chest radiographs

To provide high-quality duplicate chest images for the intensive care units, we have developed a digital duplication system in which film digitization is performed in conjunction with nonlinear density correction, contrast adjustment, and unsharp mask filtering. This system provides consistent image densities over a wide exposure range and enhancement of structures in the mediastinum and upper abdominal areas, improving visibility of catheters and tubes. The image quality is often superior to that of the original radiograph and is more consistent from day to day. Repeat rates for portable chest radiographs have been reduced by more than a factor of two since implementation of digitization in December 1991, and the number of repeat examinations caused by exposure errors have been substantially reduced.

A study on computer-aided diagnosis based on temporal subtraction of sequential chest radiographs (in Japanese)

An automated digital image subtraction technique for use with pairs of temporally sequential chest radiographs has been developed to aid radiologists in the detection of interval changes. Automated image registration based on nonlinear geometric warping is performed prior to subtraction in order to deal with complicated radiographic misregistration. Processing includes global matching, to achieve rough registration between the entire lung fields in the two images, and local matching, based on a cross-correlation method, to determine local shift values for a number of small regions. A proper warping of x,y -coordinates is determined by fitting two-dimensional polynomials to the distributions of the shift values. One image is warped and then subtracted from the other. The resultant subtraction images were able to enhance the conspicuity of various types of interval changes. Improved global matching based on a weighted template matching method achieved robust registration even with photofluorographs taken in chest mass screening programs, which had previously presented us with a relatively large number of poor-registration images. The new method was applied to 129 pairs of chest mass screening images, and offered registration accuracy as good as manual global matching. An observer test using 114 cases including 57 lungcancer cases presented better sensitivity and specificity on average compared to conventional comparison readings. In addition, newly developed image processing that eliminates the rib edge artifacts in subtraction images was applied to 26 images having pathological interval changes; results showed the potential for application to automated schemes for the detection of interval change patterns. With its capacity to improve the diagnostic accuracy of chest radiographs, the chest temporal subtraction technique promises to become an important element of computer-aided diagnosis(CAD) systems.

High-Quality Portable Chest Images Using Enhanced Film-Digitization and Computed Radiography

In order to provide high-quality duplicate portable chest images for the intensive care units, we have implemented two digital radiography systems, an enhanced film-digitization system and a storage phosphor computed radiography system. Films are digitized into 2048x2430x10 bit matrices using a laser film scanner. The densities in the films are corrected to that of properly exposed radiographs based on automated histogram analysis, and non-linear unsharp mask filtering is applied to enhance mediastinal and upper abdominal detail. The storage phosphor computed radiography images are obtained in the standard manner, except that the processing parameters have been modified to produce images similar in terms of density and edge enhancement to those obtained using the film duplication system. Since implementation of digital radiography in our hospital, the repeat rate has decreased by over fifty percent. We have found that the image quality of the digital images is often visually superior to that of the original radiograph as well as consistent from day to day.