PubMed Health⌕ Search

Biomedical subjects

J Dempsey

Publications and source records attributed to J Dempsey.

At least 37 records · Page 2Linked to original sources
PubMed1994

How to improve a neural network for early detection of hepatic cancer.

Radiologists perform differential diagnoses of hepatic (liver) masses using ultrasonography (US), computed tomography (CT), magnetic resonance imaging (MRI), and laboratory tests, but interpretation is often difficult. In our earlier research, a backpropagation neural network was designed to diagnose five classifications of hepatic masses: metastatic carcinoma, hepatoma, cavernous hemangioma, abscess, and cirrhosis. After being trained using ultrasonographic data and laboratory tests, the network classified hepatic masses correctly in 51 of 72 cases. That accuracy of 71% is higher than the 50% scored by the average radiology resident in training but lower than the 90% scored by the typical board-certified radiologist. What do we need to do to increase that accuracy and make the network friendly enough that radiologists will use it in their diagnoses? We have reviewed the literature, discussed alternatives and developed a plan to improve the diagnostic accuracy of the networks. That plan consists of: (i) get many more patient cases and more data variables, including MRI and CT data, so the network can be more highly trained. A shortage of enough patient cases to properly train the network is the key problem; (ii) use genetic algorithms and other techniques to preprocess the data; (iii) the network should have an optical interface to read images directly; and (iv) build a user-friendly interface using the C programming language on a 486 microcomputer. Continued research along the guidelines in this study should provide a sophisticated neural network for early detection of hepatic cancer that hopefully will exceed the diagnostic abilities of most radiologists.

Carcinoma, Hepatocellular↗
PubMed1994

Sleep apnea and hypertension. A population-based study.

OBJECTIVE: To measure the independent association of sleep-disordered breathing (sleep apnea and habitual snoring) and hypertension in a healthy adult population. DESIGN: A cross-sectional study of blood pressure during wakefulness and sleep among participants with and without sleep-disordered breathing. SETTING: Community-based study. PARTICIPANTS: 147 men and women, aged 30 to 60 years, selected from Wisconsin State employees enrolled in the Wisconsin Sleep Cohort Study, an ongoing, prospective, epidemiologic study of sleep-disordered breathing. MEASUREMENTS: Sleep and medical history interview, nocturnal polysomnography, and 24-hour ambulatory blood pressure monitoring in all participants. RESULTS: Mean blood pressures were significantly higher among participants with sleep apnea (> or = 5 apneas or hypopneas per hour of sleep) compared with those without (131/80 +/- 1.7/1.1 mm Hg compared with 122/75 +/- 1.9/1.2 mm Hg during wakefulness and 113/66 +/- 1.8/1.1 mm Hg compared with 104/62 +/- 2/1.3 mm Hg during sleep, respectively; P < 0.05). The variability of the blood pressure during sleep was significantly greater in participants with sleep apnea or a history of snoring compared with those without (P < 0.05). After controlling for obesity, age, and sex, sleep apnea was significantly associated with hypertension in a dose-response fashion, with odds ratios ranging from 2.0 for 5 apneic or hypopneic episodes per hour of sleep to 5.0 for 25 apneic or hypopneic episodes. CONCLUSIONS: Our data indicate an association between hypertension and sleep apnea independent of obesity, age, and sex in a nonselected, community-based adult population.

Adult↗
PubMed1994

Determination of genotoxic and other effects in mice following short term repeated-dose and subchronic inhalation exposure to phosphine.

Phosphine is an important fumigant in the grain industry and has been reported to be genotoxic in occupationally exposed fumigators. This study reports on the effects of phosphine inhalation exposure at up to, and exceeding, occupational relevant levels in a subchronic (0.3, 1.0 and 4.5 ppm, 13 weeks) and a short term repeated-dose (5.5 ppm, 2 weeks) study in both sexes of Balb-c mice. The following end-points were examined: micronucleus induction in bone marrow, peripheral blood, spleen lymphocytes and skin keratinocytes, mutations at the hypoxanthine-guanine phosphoribosyl transferase locus in lymphocytes, and weight gain and relative organ weights (kidneys, lungs, liver, heart, brain and spleen). After subchronic exposure, there was a highly significant negative linear correlation between proportional weight gain and exposure in both sexes (multiple linear regression, r = -0.56, P < 0.0001), with female mice showing a greater effect. Females also showed an increase in relative organ weights at the highest test dose, in contrast to males where there was a slight decrease. A statistically significant increase in micronucleus frequency was seen in the bone marrow and spleen lymphocytes of both sexes, but only at the highest concentration. The short term repeated-dose study revealed a slight decrease in weight gain in both sexes, with a greater effect in females. It is concluded that phosphine is weakly genotoxic in both sexes of mice, and has an effect on weight gain. However, the weak genotoxic effect may not be biologically significant as it was seen only in the subchronic study and only at the highest test concentration of 4.5 +/- 0.8 ppm (approaching the LD50). Although such exposure conditions are unlikely to be encountered in an occupational environment, caution should continue to be exercised in the use of phosphine until more data become available.

Animals↗
PubMed1993

The occurrence of sleep-disordered breathing among middle-aged adults.

BACKGROUND: Limited data have suggested that sleep-disordered breathing, a condition of repeated episodes of apnea and hypopnea during sleep, is prevalent among adults. Data from the Wisconsin Sleep Cohort Study, a longitudinal study of the natural history of cardiopulmonary disorders of sleep, were used to estimate the prevalence of undiagnosed sleep-disordered breathing among adults and address its importance to the public health. METHODS: A random sample of 602 employed men and women 30 to 60 years old were studied by overnight polysomnography to determine the frequency of episodes of apnea and hypopnea per hour of sleep (the apnea-hypopnea score). We measured the age- and sex-specific prevalence of sleep-disordered breathing in this group using three cutoff points for the apnea-hypopnea score (> or = 5, > or = 10, and > or = 15); we used logistic regression to investigate risk factors. RESULTS: The estimated prevalence of sleep-disordered breathing, defined as an apnea-hypopnea score of 5 or higher, was 9 percent for women and 24 percent for men. We estimated that 2 percent of women and 4 percent of men in the middle-aged work force meet the minimal diagnostic criteria for the sleep apnea syndrome (an apnea-hypopnea score of 5 or higher and daytime hypersomnolence). Male sex and obesity were strongly associated with the presence of sleep-disordered breathing. Habitual snorers, both men and women, tended to have a higher prevalence of apnea-hypopnea scores of 15 or higher. CONCLUSIONS: The prevalence of undiagnosed sleep-disordered breathing is high among men and is much higher than previously suspected among women. Undiagnosed sleep-disordered breathing is associated with daytime hypersomnolence.

Adult↗
PubMed1993

Acute Page kidney phenomenon: a cause of reversible renal allograft failure.

We have reported a case of reversible acute renal transplant failure due to the Page kidney phenomenon, which followed routine diagnostic needle biopsy. Duplex Doppler ultrasonography differentiated between allograft rejection and graft failure due to perirenal and mechanical abnormalities. Renal transplant failure may result from intrinsic or extrinsic allograft compression. If this is detected and treated rapidly, renal deterioration may be reversed and the transplant salvaged.

Acute Disease↗
PubMed1992

Using an artificial neural network to diagnose hepatic masses.

Using abdominal ultrasonographic data and laboratory tests, radiologists often find differential diagnoses of hepatic masses difficult. A computerized second opinion would be especially helpful for clinicians in diagnosing liver cancer because of the difficulty of such diagnoses. A back-propagation neural network was designed to diagnose five classifications of hepatic masses: hepatoma, metastatic carcinoma, abscess, cavernous hemangioma, and cirrhosis. The network input consisted of 35 numbers per patient case that represented ultrasonographic data and laboratory tests. The network architecture had 35 elements in the input layer, two hidden layers of 35 elements each, and 5 elements in the output layer. After being trained to a learning tolerance of 1%, the network classified hepatic masses correctly in 48 of 64 cases. An accuracy of 75% is higher than the 50% scored by the average radiology resident in training but lower than the 90% scored by the typical board-certified radiologist. When sufficiently sophisticated, a neural network may significantly improve the analysis of hepatic-mass radiographs.

Diagnosis, Computer-Assisted↗
PubMed1992

Health care organization drug testing.

Health care managers are being required to respond to the growing concerns of the public about alcohol and drug use in the health care workplace. To this end, the following recommendations are offered. A drug testing policy should be developed with input from and support of employees and unions. "For cause" testing should be used because it results in more definitive results and better employee acceptance. Unless there are compelling reasons for random testing, "for cause" testing is the preferable method. All levels of employees and the medical staff should be subject to the drug-testing policy. Rehabilitation rather than punishment should be emphasized in dealing with employees with alcohol and drug problems.

Health Personnel↗
PubMed1991

Using neural networks to diagnose cancer.

While artificial brains are in the realm of science fiction, artificial neural networks (ANNs) are scientific facts. An artificial neural network is a computational structure modeled somewhat on the neural structure of the brain; both have many highly interconnected processing elements. These biologically inspired processing elements are taught by feeding examples until the results are acceptable. In the past 5 years, neural networks have become successful in providing meaningful second opinions in clinical diagnosis. In our research, a prototype artificial neural network was trained on numeral ultrasound data of 52 actual cases and then correctly identified renal cell carcinoma from renal cysts and other conditions without diagnostic errors. Our nonlinear artificial neural network was trained on software using the standard backpropagation paradigm on a 80386 microcomputer. Our ANN learned from ultrasound data in 52 cases (17 malignant, 30 cysts, and 5 other) at a Memphis hospital. The trained prototype performed without error on 47 cases which were not in the data used for training. This prototype must be validated by extending this study to more cases.

Carcinoma, Renal Cell↗
PubMed1991

Quality assurance in MRI.

"With the prospect of the federal government waiting in the wings to mandate and oversee quality assurance in radiology, managers should waste no time in establishing adequate quality assurance programs for MRI, the latest new technology," declare the authors. Determining standards for quality assurance is certainly a difficult task, considering all the stakeholders involved, say the authors. They focus on evaluating the appropriateness of the imaging test and insuring the accuracy of the diagnosis as the key elements to be monitored.

Centers for Medicare and Medicaid Services, U.S.↗
PubMed1991

Survival through productivity improvement.

There are dozens of additional case studies as well as success stories. The number of ways in which the information can be presented is limited only by the creativity of the user. In order of importance, the three principles remain 1. Convert all data to ratios. 2. Study emerging trends in the way the ratio changes over time. 3. Compare changes over time as well as static ratios to an outside source. To begin this program one needs to see what information is available at a particular hospital. If no other acceptable statistic is available use adjusted patient days. It is important that a productivity tracking system not be delayed until a system for statistical reporting is developed. It is also important to keep the system simple. Productivity monitoring is difficult because it requires a new perspective. We are used to using historical data and unique characteristics of a hospital in the budget process. Productivity gains require a detached perspective. Often only an outsider can identify some examples of overstaffing. The process is difficult, but it is not complicated. Productivity experts tell us we can produce a service faster, cheaper, or better, but not all three. Health care quality control ensures a better service. A standard cost accounting system for the health care industry may provide a less expensive service through better cost control. Productivity--the ratio of work output to resource input (health care costs)--can be increased. To make productivity improvement work the users must take the view of the entire institution rather than that of a department. Productivity measuring is necessary to ensure the survival of hospitals. This process could have saved many of the hospitals that were forced to close, and it could improve the financial health of many hospitals today.

Budgets↗