- Email: [email protected]
Ultrasonic tissue characterization in venous thrombosis
ABSTRACTS, ULTRASONIC IMAGING AND TISSUE CHARACTERIZATION SYMPOSIUM
DETECTION OF DIFFUSE LIVER DISEASE BY QUANTITATIVE ECHOGRAPHY, J.M. 3.3 Thijssen’, B.J. Oosterveld*~2, P.C. Hattman3, and G.J.E. Rosenbusch3, ‘Biophysics Laboratory of the Institute of Ophthalmology, 2presently: Institute of Radiography and 3Department of Diagnostic Radiology, University Hospital, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. Quantitative acoustic parameters and image texture parameters were used in a linear discriminant analysis. This analysis was applied to detect retrospectively the classes of diffuse liver disease against a population of normal livers. Three different sets of parameters were employed. The first set was selected by the authors, the other two were taken from the literature. The area under the ROC (or percentage correct classification) obtained with the first set ranged from 88 percent to 97 percent, depending on the disease class. It is concluded that the first order statistical parameters of the image texture (diffise scattering model), together with the slope of the attenuation coefficient are the most important parameters. As an alternative to the texture parameters, the backscattering parameters (second set of parameters) also yielded a high score. The texture analysis involving structural scattering (third set of parameters) did not produce such results. The overall conclusion is that these methods might be used for prospective diagnosis. DIFFERENTIATION OF HUMAN BREAST CANCERS BY ULTRASONIC TISSUE 3.4 CHARACTERIZATION, Robert M. Golub’, Richard E. Parsons’, Laurie Loiacono’, Bernard Sigel’, Ernest J. Feleppa2, Jeffery Justin’, Hiroshi Kimitsuki’, Issei Kodama’, Joan Sokil-Melgar’, Mary Rorke2, Frederic Lizi’, and Howard A. &en’, ‘The Medical College of Pennsylvania, Department of Surgery, Philadelphia, PA and Wverside Research Institute, New York, NY. Ultrasonic tissue characterization (UTC) parameters were investigated as a modality for diagnosis of breast cancer in vivo. Palpable breast masses (n = 86) from 72 patients were imaged prior to surgical excision with B-mode ultrasound for lesion identification and for acquisition of rf signal data. The values of slope and Y-intercept, parameters derived from the spectral analysis of backscattered echoes of four types of breast tissue (normal, fibrocystic, tibroadenoma, invasive cancer), were used as the initial set of discriminants. The parameter values from UTC of each tissue group were compared by multilinear discriminant analysis. Differences between cancer and flbroadenomas were significant @ < 0.01). These two tumors were also distinguishable from fibrocystic and normal breast tissue (p < 0.01). Fibrocystic tissue could not be well differentiated from normal breast tissue. Overall correct classification rate (for all lesions) was 88%. In this preliminary study, UTC appears to significantly improve the ability of ultrasound to diagnose breast cancer on the basis of inherent tissue properties. ULTRASONIC TISSUE CHARACTERIZATION IN VENOUS THROMBOSIS, Richard E. 3.5 Parsons’, Robert M. Golub’, Issei Kodama’, Bernard Sigel’, Ernest J. Feleppa2,Laurie Loiacono’, Jeffery Justin’, Vanlila K. Swami’, Hiroshi Kllitsukil, Mary RorkeZ, and Frederic Lixzi2, ‘The Medical College of Pennsylvania, Departments of Surgery and Pathology, Philadelphia, PA and Riverside Research Institute, New York, NY. The ability of ultrasonic tissue characterization (UTC) to detect compositional differences in venous thrombi at varying time intervals was evaluated in vivo. ‘Ihrombi were produced in 25 anesthetized 20-25 kg pigs after endothelial injury produced by application of dc current and partial ligation. Jugular veins were imaged 30 minutes after formation (day 0) and 1,2, 7, 14, and 60 days later with a color Doppler ultrasound scanner that demonstrated Row, identified the thrombi, and acquired rf data for UTC analysis. Two parameters from the normalized power spectrum of backscattered echoes were utilized: slope and intercept. Histologic examination was performed for each time period. The spectral parameter values of slope and intercept for each timed observation for thrombi and venous wall were compared by a multilinear discriminant analysis. Differences between data after 2 days were significant comparing initial to all other time intervals (p < 0.01). In addition, significant changes in the venous wall indicative of intimal hyperplasia were detected after 7 days. UTC parameters of thrombi are distinctly different for initial, 2day, 7day, lrlday, and 6Oday time periods and intimal hyperplasia. UTC appears to be capable of distinguishing age differences in thrombi, detecting changea indicative of intimal hyperplasia, and providing a means of noninvasive application in clinical evaluation of thrombi. IN-VIVO ESTIMATES OF ATTENUATION IN ATHEROSCLEROTIC PLAQUE AND 3.6 THROMBUS USING A PC-BASED DATA ACQUISITION SYSTEM, E. Yigiter’, J.P. Jones’ and
193
DETECTION OF DIFFUSE LIVER DISEASE BY QUANTITATIVE ECHOGRAPHY, J.M. 3.3 Thijssen’, B.J. Oosterveld*~2, P.C. Hattman3, and G.J.E. Rosenbusch3, ‘Biophysics Laboratory of the Institute of Ophthalmology, 2presently: Institute of Radiography and 3Department of Diagnostic Radiology, University Hospital, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands. Quantitative acoustic parameters and image texture parameters were used in a linear discriminant analysis. This analysis was applied to detect retrospectively the classes of diffuse liver disease against a population of normal livers. Three different sets of parameters were employed. The first set was selected by the authors, the other two were taken from the literature. The area under the ROC (or percentage correct classification) obtained with the first set ranged from 88 percent to 97 percent, depending on the disease class. It is concluded that the first order statistical parameters of the image texture (diffise scattering model), together with the slope of the attenuation coefficient are the most important parameters. As an alternative to the texture parameters, the backscattering parameters (second set of parameters) also yielded a high score. The texture analysis involving structural scattering (third set of parameters) did not produce such results. The overall conclusion is that these methods might be used for prospective diagnosis. DIFFERENTIATION OF HUMAN BREAST CANCERS BY ULTRASONIC TISSUE 3.4 CHARACTERIZATION, Robert M. Golub’, Richard E. Parsons’, Laurie Loiacono’, Bernard Sigel’, Ernest J. Feleppa2, Jeffery Justin’, Hiroshi Kimitsuki’, Issei Kodama’, Joan Sokil-Melgar’, Mary Rorke2, Frederic Lizi’, and Howard A. &en’, ‘The Medical College of Pennsylvania, Department of Surgery, Philadelphia, PA and Wverside Research Institute, New York, NY. Ultrasonic tissue characterization (UTC) parameters were investigated as a modality for diagnosis of breast cancer in vivo. Palpable breast masses (n = 86) from 72 patients were imaged prior to surgical excision with B-mode ultrasound for lesion identification and for acquisition of rf signal data. The values of slope and Y-intercept, parameters derived from the spectral analysis of backscattered echoes of four types of breast tissue (normal, fibrocystic, tibroadenoma, invasive cancer), were used as the initial set of discriminants. The parameter values from UTC of each tissue group were compared by multilinear discriminant analysis. Differences between cancer and flbroadenomas were significant @ < 0.01). These two tumors were also distinguishable from fibrocystic and normal breast tissue (p < 0.01). Fibrocystic tissue could not be well differentiated from normal breast tissue. Overall correct classification rate (for all lesions) was 88%. In this preliminary study, UTC appears to significantly improve the ability of ultrasound to diagnose breast cancer on the basis of inherent tissue properties. ULTRASONIC TISSUE CHARACTERIZATION IN VENOUS THROMBOSIS, Richard E. 3.5 Parsons’, Robert M. Golub’, Issei Kodama’, Bernard Sigel’, Ernest J. Feleppa2,Laurie Loiacono’, Jeffery Justin’, Vanlila K. Swami’, Hiroshi Kllitsukil, Mary RorkeZ, and Frederic Lixzi2, ‘The Medical College of Pennsylvania, Departments of Surgery and Pathology, Philadelphia, PA and Riverside Research Institute, New York, NY. The ability of ultrasonic tissue characterization (UTC) to detect compositional differences in venous thrombi at varying time intervals was evaluated in vivo. ‘Ihrombi were produced in 25 anesthetized 20-25 kg pigs after endothelial injury produced by application of dc current and partial ligation. Jugular veins were imaged 30 minutes after formation (day 0) and 1,2, 7, 14, and 60 days later with a color Doppler ultrasound scanner that demonstrated Row, identified the thrombi, and acquired rf data for UTC analysis. Two parameters from the normalized power spectrum of backscattered echoes were utilized: slope and intercept. Histologic examination was performed for each time period. The spectral parameter values of slope and intercept for each timed observation for thrombi and venous wall were compared by a multilinear discriminant analysis. Differences between data after 2 days were significant comparing initial to all other time intervals (p < 0.01). In addition, significant changes in the venous wall indicative of intimal hyperplasia were detected after 7 days. UTC parameters of thrombi are distinctly different for initial, 2day, 7day, lrlday, and 6Oday time periods and intimal hyperplasia. UTC appears to be capable of distinguishing age differences in thrombi, detecting changea indicative of intimal hyperplasia, and providing a means of noninvasive application in clinical evaluation of thrombi. IN-VIVO ESTIMATES OF ATTENUATION IN ATHEROSCLEROTIC PLAQUE AND 3.6 THROMBUS USING A PC-BASED DATA ACQUISITION SYSTEM, E. Yigiter’, J.P. Jones’ and
193