Evlin L. Kinney

Prediction of survival in progressive systemic sclerosis by multivariate analysis of clinical features

The records of 64 patients with progressive systemic sclerosis (PSS) were reviewed. More than 100 items from the patients history, physical examination, and laboratory tests were encoded. A two-tailed Gehans test was used to evaluate the effect of each variable on patient survival. Variables significant at the 0.05 level were then analyzed simultaneously, using the Cox regression model, in order to determine which clinical features are most predictive of survival in PSS. By Cox regression, the presence of an S3 gallop and advanced age at onset of PSS related negatively to survival. The relative risk of death for a patient with an S3 gallop was 5.44 (ratio of risk) times that for a patient without an S3 gallop, and the risk of death for a 70-year-old patient at disease onset was 20 times that of a 20-year-old patient at disease onset. In addition, when patient selection was taken into account, corticosteroid administration had a significant effect, and seemed to actually shorten survival. Although one of the factors, age at onset, is not modifiable, the presence of either variable should alert the clinician to the patient being at high risk for short survival. The application of this technique should also prove helpful in analyzing the efficacy of various proposed therapies.

Expert system detection of drug interactions: Results in consecutive inpatients

Although drug interactions (DI) are a common cause of morbidity, their large number precludes remembering them. To address this problem, we constructed a microcomputer-based expert system and assessed its efficacy in 90 consecutive inpatients. It was found that, without the expert system, a knowledge of the patients medication list did not affect the frequency of occurrence of DI. Also, without the expert system, no DI were predicted, clinically, whereas the expert system predicted 27 DI of which, in retrospect, 10 actually occurred. Unsuspected DI were most likely if: a drug was not within the specialty of the clinician, DI host factors were present, or the DI involved a commonly prescribed drug pair. Although none of the drug interactions were life-threatening, in two cases, the DI was the cause for admission. Since the offending medications could usually be adjusted in dose, drug interactions were easily corrected once clinicians were made aware of them.

Classification of left ventricular thrombi by their history of systemic embolization using pattern recognition of two-dimensional echocardiograms

Although one can diagnose left ventricular (LV) thrombi by two-dimensional echocardiography (2DE), the factors associated with peripheral embolization, given a 2DE with LV thrombi, have not been well delineated. Therefore we looked at 2DE and clinical variables that included texture features in the 2DE of 38 patients whose 2DE had LV thrombi and questioned these patients to see if clinical embolization had occurred in the 8.9 +/- 6.1 month (+/- SD) average follow-up period. Eight patients, four with acute myocardial infarction (AMI) and four with dilated LV and decreased LV systolic wall motion, had clinically apparent leg or brain emboli, whereas the remaining patients did not. Emboli occurred within a week of obtaining the 2DE in question. The variables considered were the age of the patient, the type of heart disease present, warfarin administration, exercise tolerance, standard M-mode measurements, LV dyssynergy by 2DE, clot size and mobility, and gray scale statistics which include run length, Sobel edge points followed by 50% gradient thresholding, gray level second-order statistics, offset 1 and gray level difference statistics, offset 1. The values of the variables were then entered into an expert system (Expert Ease) in order to achieve classification of patients into emboli versus no emboli groups, while using a minimal number of variables. The only variables that were needed included run length, long runs emphasis, gray level difference statistics (entropy, contrast, mean, and angular second moment), gray level second-order statistics (contrast), and warfarin status. When probability statistics were applied to this schema, its accuracy was predicted to be at least 96%.(ABSTRACT TRUNCATED AT 250 WORDS)

Cardiac size and motion during water immersion: implications for volume homeostasis.

Prior work has shown that head-out water immersion (WI) produces a prompt central hypervolemia, a natriuresis, and a diuresis. To assess if cardiac chamber enlargement modulates these effects, we measured cardiac size, shape, wall motion, and cardiac displacement via serial underwater two-dimensional echocardiography. Six normal volunteers underwent 2.5 hours of WI to the neck, seated, at 34.5 degrees C. Recovery was 30 minutes. The size of both atria increased significantly but transiently during the study (p less than 0.0001 for left atrium, and p = 0.0020 for right atrium). Both atria returned to baseline size during WI. Moreover, for left atrium there was a small overshoot in recovery. Neither left ventricular nor right ventricular dimensions nor ejection fraction changed significantly. Also, no shape changes were detected, although WI was associated with upward and lateral displacement of the acoustic windows. Correlation coefficients (r) for left atrial, or right atrial size vs urinary excretion of sodium or urine volume size ranged from 0.05 to 0.36. These results, in sum, suggest that strong compensatory mechanisms are counteracting the effect of WI on distended cardiac receptors, and that cardiac receptor activation alone does not constitute the afferent limb of the reflex mediating the renal effects of head-out WI.

Comparison of the ID3 algorithm versus discriminant analysis for performing feature selection

Having obtained disappointing results in a small medical data set despite the fact that our data seemed to be well suited for induction via ID3, we decided to compare the performance of ID3 to discriminant analysis. Performance was gauged by the percentage of correct classification in a second, independent data set. Examples were obtained from a cardiology project on the accuracy of auscultation. There were 107 examples in the first data set and 67 cases in the second. We found that ID3 and discriminant analysis performed equally poorly, with ID3 classifying only 60% of the second set correctly and discriminant analysis classifying 66% of the second set correctly. Also, the ID3 probability statistic for estimating the accuracy of ID3 for classifying further cases was markedly optimistic compared to our actual second data set results. Moreover, with an increase in sample size, ID3 seemed to break down, producing a large, complex decision tree of dubious generality, whereas discriminant analysis, with a larger sample size, used more independent variables but maintained its first set accuracy. These data suggest that there is a need for more sophisticated algorithms than ID3, even at the risk of giving up some computational efficiency.

An expert system on the diagnosis of ascites

We constructed an expert system on the diagnosis of ascites, using a combination of case reports and unpublished patient data. Rule production was by induction from examples, and the program operated on an algorithm which was a modification of Quinlans ID3. The result was a small, but formally complete expert system. When tested against a new data set of patients, our expert system predicted the clinical diagnosis 82% of the time.

Case Report: Transient Myocardial Uptake of Tc-99m Methylene Diphosphonate in a Patient with Amyloidosis

A patient with autopsy-proven myocardial amyloidosis had been observed for 10 years with Tc-99m methylene diphosphonate bone scans. The bone scans manifested transient myocardial uptake. No cause other than the myocardial amyloid could be found to explain the fluctuating scan findings.

Should echocardiography be used to screen dizzy patients

The value of echocardiography in the evaluation of dizziness was studied in 151 consecutive patients. Twenty-two patients, on further review, were found to have related complaints, seizures and syncope mainly, but not dizziness. Twenty-four patients had vertigo, 67 had nonvertiginous dizziness, and in 38 patients, there was insufficient information to categorize the dizziness as being vertigo or nonvertiginous dizziness. When the 22 nondizzy patients were com pared with the 129 dizzy patients, the only statistically significant differences were that there were more blacks in the nondizzy group and that the aortic root dimension was, on average, 0.4 cm larger in the nondizzy group. These find ings, although statistically significant, appeared to be clinically insignificant. There was, however, a high prevalence of valvular heart disease in both the nondizzy and two of the three dizzy subgroups. In most cases, the valvular abnormality had not been suspected clinically. But in no case was significant information added by echocardiography that helped in patient management. Thus, in the dizzy patient, echocardiography should be reserved for specific cardiac indications and not used as a routine screening test.