Judy A. Bean

Quantitative texture analysis in two-dimensional echocardiography: application to the diagnosis of experimental myocardial contusion.

We postulated that the analysis of regional image texture in two-dimensional (2D) echocardiograms would be an accurate method to differentiate normal from abnormal myocardial structure. We tested this hypothesis with quantitative texture measures to study the regional, spatial distribution of echo amplitudes in 2D echocardiograms performed before and immediately after blunt left chest trauma was induced in six anesthetized dogs. After trauma the contused region of myocardium appeared brighter and exhibited an altered myocardial texture. By use of a set quantitative texture measures, we found no significant differences in pretrauma images when normal regions were compared with regions to be contused. Also, we found no difference when we compared the normal regions in each animal in pretrauma vs posttrauma images. Twelve measures, however, differentiated normal from contused regions within the posttrauma images (p values ranged from .0057 to .0001 by multivariate analysis of variance). These texture measures were capable of differentiating normal from abnormal tissue only when texture along the azimuthal (lateral) direction was calculated. We conclude that regions of myocardial contusion exhibit visibly altered local echo-amplitude patterns (altered image texture) and that these image texture alterations may be quantified with digital image analysis techniques. These findings suggest that quantitative texture calculations may be a useful approach to ultrasound tissue characterization.

A family study of anorexia nervosa and bulimia

A family study is presented that examines the association between eating disorders and affective illness. 307 relatives of 30 eating disorder patients, of 16 patients with major depression, and of 20 normal controls, were interviewed to determine lifetime histories of psychiatric illness. Using logistic regression analyses, the results supported previous findings of a familial association between eating disorders and major depression. However, there was no evidence that clinical features of eating disorder patients (presence of bulimia or major depression) can be used to predict morbid risk. In addition, the results failed to reveal a familial aggregation of either alcoholism or drug abuse in eating disorder families.

Sample size for case-control studies using Cochran's statistic.

Sample size determination for case-control studies of chronic disease are often based on the simple 2 X 2 tabular cross-classification of exposure and disease, thereby ignoring stratification which may be considered in the analysis. One consequence of this approach is that the sample size may be inadequate to attain a specified power and size when performing a statistical analysis on J 2 X 2 tables using Cochrans (1954, Biometrics 10, 417-451) statistic or the Mantel-Haenszel (1959, Journal of the National Cancer Institute 22, 719-748) statistic. A sample size formula is derived from Cochrans statistic and it is compared with the corresponding one derived when the data are treated as unstratified, and also with two other formulas proposed for stratified data analysis. The formula developed yields values slightly higher than one recently proposed by Muñoz and Rosner (1984, Biometrics 40, 995-1004), which assumes that both margins of each 2 X 2 table are fixed, while the present study considers only the case-control margin to be fixed.

Range- and azimuth-dependent variability of image texture in two-dimensional echocardiograms.

Regional two-dimensional (2D) echocardiographic amplitude patterns, or image texture, may be of diagnostic importance. Echocardiographic image texture is due in part to acoustic speckle, a complex pattern of interference of reflections from many small scatterers in tissue. The regional speckle pattern appears to be altered in several disorders associated with abnormalities in myocardial structure but also may be altered by a variety of characteristics of the scanning instrument. We hypothesized that quantitative measures of regional 2D echocardiographic image texture would vary as a function of position in range and azimuth within the field of view, even when imaging a uniform ensemble of scatterers. We tested this hypothesis by imaging a tissue-equivalent phantom with two phased-array scanners and two different methods of digitization. We analyzed the texture in several regions of interest separated in range and azimuth and found significant differences in quantitative texture measures as a function of position of the region of interest in the sector field of view (p values .006 to .0001 by multivariate analysis of variance). We found significant regional variability in texture with both scanners and both methods of digitization. We conclude that regional quantitative image texture in 2D echocardiograms varies as a function of range and azimuth, even when imaging a uniform ensemble of scatterers. This variability is related to several physical and instrument-related phenomena and precludes interpreting all regional texture alterations as indicating tissue structural abnormalities.

Estrogen use and breast cancer. Interaction with body mass.

With few exceptions, most epidemiologic studies do not show an excess relative risk of breast cancer associated with menopausal estrogen therapy. Other studies show a relationship of breast cancer to obesity, which is characterized by increased endogenous estrogen production. This study explored the possibility of an interaction between ponderosity and exogenous estrogen use in a case–control study of 113 postmenopausal breast cancer patients and pair‐matched hospital control subjects. In this series, neither menopausal estrogen use nor relative weight were significantly associated with breast cancer risk. However, among estrogen users, the relative risk of breast cancer was strikingly influenced by the ponderosity of the subjects; the relative risk was 0.41 for women whose relative weight was less than the median, compared with 1.29 for those whose relative weight exceeded the median. The mean age was also examined at diagnosis in order to explore the potential of exogenous estrogen as a tumor promotor. The mean age at breast cancer diagnosis in estrogen users, 58.1 years, was significantly lower than in nonusers, 63 years. A significant linear relationship was found between age at diagnosis and body weight among estrogen users. Estrogen‐treated women in the lowest tertile of body weight had the diagnosis of breast cancer made seven years earlier than those in the highest tertile of weight. There was no significant difference in the distribution of clinical stages at diagnosis between estrogen users and nonusers. These data suggest that relative body weight is an important modifier of the effect of exogenous estrogens on breast cancer biology.

A case-control study of breast cancer and psychotropic drug use.

The relative risk of breast cancer incidence and tumor promotion associated with psychotropic drug consumption was evaluated in 151 patients with newly diagnosed neoplasms and 151 hospital controls. No significantly altered risk of breast cancer was found in association with the use of diazepam, chlordiazepoxide, antidepressants, major tranquillizers, sedatives or hypnotics, even after adjustment for known menstrual and reproductive risk factors. No substantial evidence of tumor promotion effects was found, as measured by altered age-at-onset of disease or altered clinical stage at presentation. Psychotropic drug use was inversely related to subject ponderosity (measured by the Quetelet Index) and while this did not confound risk estimates, it may be important in exploring biologic hypotheses of psychotropic drug use and breast cancer.

Ultrasound characterization of acute myocardial ischemia by quantitative texture analysis

In this study we tested the efficacy of quantitative texture analysis in the identification of acute myocardial ischemia using an ultrasound data acquisition system that digitizes and stores echocardiographic data in polar format. In nine closed-chest dogs, data were acquired before and after coronary occlusion using a 2.4 MHz echocardiographic system. Regions of interest were analyzed at end-diastole and end-systole from the ischemic area and normal area at the same depth of field. Ultrasound data were evaluated using previously reported quantitative gray level texture measures. After occlusion, texture changes indicative of ischemia were found in systolic images. The directional component of the data analysis was important; analysis in the azimuthal direction was more accurate than in the axial direction. Six texture measures exhibited significant changes in the ischemic region from control to occlusion when analyzing data in the azimuthal direction. One false positive result occurred (significant texture change in the normal region from control to occlusion) in the azimuthal direction. Several false positive alterations in the normal regions from control to occlusion were found when the texture was evaluated in the axial direction. For accurate assessment of ischemic changes, preocclusion image data were required. We conclude that quantitative echocardiographic texture analysis using polar format data can identify subtle changes in myocardial texture such as that due to acute ischemia, using data acquired through the chest wall.

Quantitative Echocardiographic Image Texture: Normal Contraction-Related Variability

Myocardial tissue characterization using ultrasound is a growing area of investigation which attempts to evaluate the structure of the myocardium by analysis of ultrasound signals. Our laboratory has been exploring the use of texture analysis for the determination of myocardial tissue properties from two-dimensional echocardiographic images. In the present study, we tested the hypothesis that echocardiographic image texture varies with cardiac contraction in normal human subjects. In 17 subjects, we obtained long-and short-axis images at end diastole and end systole. Echo image texture was assessed using three classes of quantitative texture measures: run length, gray level difference, and busyness statistics. These statistics measure various attributes of image texture. We found significant contraction-related changes in image texture for the left ventricular posterior wall. This observation is important in that future applications of texture analysis to echocardiographic image data will require that texture be measured at a consistent point in the cardiac cycle. Moreover, it is possible that alteration in the normal variation of texture with cardiac contraction may be a sensitive indicator of abnormal myocardium.

Ultrasound Characterization of Acoustic Properties of Acute Intracardiac Thrombi: Studies in a New Experimental Model

Echocardiographic differentiation among intracavitary thrombus, cavity noise, and adjacent myocardium can be difficult. As an initial step toward quantitative thrombus characterization with ultrasound, 11 dogs were studied with an in vivo intracardiac thrombus model to delineate the acoustic properties of acute thrombi. The apical coronary arteries were ligated, and subsequently injections of 5% sodium rescinoleate and 1000 units of thrombin at the endocardium-blood interface created left ventricular mural thrombi. Echocardiographic images were obtained in long- and short-axis views with a digital acquisition system, and a statistical analysis of echo intensities was performed in regions of interest in the thrombus, surrounding ventricular cavity and adjacent myocardium. Statistical measurements used to evaluate echo intensities in each region of interest included mean gray level, standard deviation, skewness, and kurtosis. The results showed that thrombus could be distinguished from myocardium (by mean gray level and standard deviation) only in short-axis views, where regions of interest could be placed at similar depths of field. Mean gray level, standard deviation, and skewness all distinguished thrombus from intracavitary blood regardless of the region of interest placement. The phase of the cardiac cycle at which data were acquired did not alter the results. We conclude that acute intracardiac thrombi can be distinguished from surrounding blood and myocardium with ultrasound tissue characterization techniques that may have the potential for clinical application.

Clinical and biologic antecedents of the amenorrhea/hyperprolactinemia syndrome: a case-control study

Clinical, reproductive, and biologic antecedents of women with the amenorrhea/hyperprolactinemia (A/H) syndrome, most with pituitary prolactinomas, and matched healthy subjects were examined in a case-control study. Women with A/H had more frequent irregularity in early menstrual cycle patterns and excess facial hair, had less frequent catamenial symptoms, and were heavier at the time of study when compared with control subjects. The patients reported a more frequent history of thyroid disease and chest trauma. There were no significant differences between the patients and control subjects with respect to prior use of oral contraceptives, estrogens, psychotropic drugs, alcohol, or other substances known to alter serum prolactin levels, nor were there significant differences in the family history of gynecologic conditions or malignancies or the syndromes of multiple endocrine neoplasia. The implications of these findings are discussed.