A high-resolution image processing workstation is an important instrument for the study of computer-aided diagnosis (CAD) schemes, PACS, multi- modality image analysis, and clinically-relevant observer performance studies. The proposed instrument, which will update existing facilities, meets the research needs of a number of investigators at the University of Chicago. Drs. Doi, Metz, Giger and Levin will be able to expedite and extend their NIH-funded research using the proposed workstation. The P.I., Dr. Doi, who has extensive experience with digital imaging, will integrate basic research and clinical applications with the proposed instrument into research, teaching, and patient care at the University of Chicago.

The chest imaging project of Dr. Doi, in collaboration with Dr. MacMahon, will develop CAD schemes for detection of various abnormalities in the chest, due to interstitial disease, lung cancer, pneumothorax, and cardiomegaly. The mammography project of Dr. Giger, in collaboration with Drs. Schmidt and Nishikawa, will develop clinical, patient-oriented applications of CAD methods. These two projects aim to improve diagnostic accuracy by reducing the number of missed diagnoses and to improve the overall reproducibility of image interpretation by using the proposed instrument to generate a "second opinion".

Dr. Metz, in collaboration with Dr. MacMahon, will investigate ROC methodology for the assessment of the use of CAD in clinical environments. Automated techniques for image presentation and data collection in ROC studies will be developed.

Dr. Levin, in collaboration with Dr. Hoffmann, will develop a method for integrated 3-D display of angiographic and MR images of the brain. Software for registration and display will be developed and tested with phantoms and patient volunteers.

The angiography project of Drs. Doi and Chua, in collaboration with Dr. Hoffmann, will develop a highly accurate and reproducible method for assessing the severity of stenotic lesions and the assessment of instantaneous blood flow rates. They will use multi-frame images from cerebral and coronary angiography studies.

The facility will also serve as a resource for research training of the students and postdoctoral fellows in the Graduate Programs in Medical Physics, enabling them to learn graphics, image processing, image feature recognition, ROC analysis, and a small scale PACS. The established investigators in the Departments of Radiology, Radiation Oncology, and Cardiology would greatly benefit from accessibility of the high-resolution image processing workstation, which currently is not available.

S10RR007418

1992-05-19

1995-05-18