Search PubMed⌕ Search

Biomedical subjects

M R Conaway

Publications and source records attributed to M R Conaway.

30 records · Page 2Linked to original sources
PubMed1995

Combinations of multiple serum markers are superior to individual assays for discriminating malignant from benign pelvic masses.

To determine whether measurement of the levels of multiple tumor markers in the preoperative serum of women presenting with a pelvic mass distinguished benign from malignant disease better than the assay of CA 125 alone, sera from 429 patients, 192 of whom had malignant histology, were assayed for 8 different markers: CA 125, macrophage colony-stimulating factor, OVX1, lipid-associated sialic acid (LASA), CA15-3, CA72-4, CA19-9, and CA54/61. The sensitivity and specificity of CA 125 alone (> 35 U/ml) was 78.1 and 76.8%, respectively. A panel consisting of CA 125, OVX1, LASA, CA15-3, and CA72-4 had a sensitivity of 83.3% and specificity of 84.0% when two or more markers were elevated. Using the concentrations of these five markers, logistic regression analysis had a sensitivity of 85.4% and a specificity of 83.1%. Considering the values of markers in different sequences, classification and regression tree analysis substantially improved the sensitivity to 90.6% and the specificity to 93.2%. When applied in clinical practice this approach could improve the management of women presenting with a pelvic mass and may also have application in screening for ovarian cancer.

Biomarkers, Tumor↗
PubMed1995

Bivariate sequential designs for phase II trials.

In this paper we propose methods for designing group sequential phase II trials with two dependent binary endpoints. The emphasis is on the derivation of stopping rules for phase II trials which require the enrollment of a small number of patients. The methods are based on enumerating the exact distribution for the binary endpoints. We illustrate the methods with a recent study which required the use of group sequential design to monitor antitumor activity and toxicity.

Antineoplastic Agents↗
PubMed1994

Overexpression of the protein tyrosine phosphatase PTP1B in human breast cancer: association with p185c-erbB-2 protein expression.

BACKGROUND: The p185c-erbB-2 growth factor receptor protein tyrosine kinase (PTK) is overexpressed in one third of human breast cancer patients and indicates a poor prognosis in these patients. Protein tyrosine phosphatases (PTPs) may balance PTK activity as part of normal growth-regulation pathways. PTP1B is an intracellular PTP that is involved in linkage between signal transduction pathways and may interface with inappropriate PTK activity in transformed cells. PURPOSE: The aim of this study was to determine if PTP1B is overexpressed in human mammary tumors and to determine if such overexpression is associated with the overexpression of the p185c-erbB-2 receptor PTK. METHODS: Our samples were frozen sections from 29 human mammary tumors (19 pure infiltrating, two pure intraductal, and eight combined intraductal and infiltrating) and nine sections from normal breast tissue. The sections were immunohistochemically stained for PTP1B and p185c-erbB-2, and the results were analyzed statistically for association between overexpression of the two proteins. Northern blot analysis was used to assess if PTP1B overexpression was coincident with increased transcription of the PTP1B gene. RESULTS: Overexpression of the PTP1B protein was observed in 72.4% of the tumor sections compared with normal epithelium, with maximal expression occurring in 37.9% of the tumors. All of the tumor subtypes, including the pure intraductal lesions, overexpressed PTP1B. Statistical analyses demonstrated a significant association between PTP1B overexpression and breast cancer (P < .038) and between the overexpression of PTP1B and the overexpression of p185c-erbB-2 (P < .006). PTP1B messenger RNA steady-state transcription was consistently increased in tumors versus normal tissues. CONCLUSIONS: PTP1B overexpression is a common phenotypic manifestation in human breast cancers and is associated with over-expression of p185c-erbB-2. Steady-state PTP1B transcription is increased in tumor tissue. IMPLICATIONS: Further studies are needed to determine if PTP1B overexpression balances or augments PTK activity. PTP1B overexpression might also be evaluated as a clinical prognostic factor in human breast cancers.

Blotting, Northern↗
PubMed1994

Overexpression of the tyrosine phosphatase PTP1B is associated with human ovarian carcinomas.

OBJECTIVE: Our purpose was to determine whether protein tyrosine phosphatase 1B is overexpressed in ovarian cancers, possibly altering the balance of intracellular tyrosine phosphorylation. STUDY DESIGN: The expression of protein tyrosine phosphatase 1B was assayed in frozen sections from 54 human ovarian carcinomas and seven normal ovaries by immunochemical staining with monoclonal antibody AE4-2J, which is specific for protein tyrosine phosphatase 1B. The expression of protein tyrosine phosphatase 1B-specific messenger ribonucleic acid in tumors was determined by Northern analysis. The results were analyzed statistically by means of Fisher's exact test. RESULTS: Minimal staining was observed in normal ovarian epithelium. In contrast, 43 of 54 (79.6%) tumors displayed increased protein tyrosine phosphatase 1B expression, which is statistically associated with malignancy. Overexpression was associated with the expression of the p185c-erbB-2, p170EGFR, and p165mCSFR growth factor receptor protein tyrosine kinases. Protein tyrosine phosphatase 1B messenger ribonucleic acid expression was inconsistently increased in tumor cells. CONCLUSION: Increased expression of protein tyrosine phosphatase 1B in ovarian cancers that also express protein tyrosine kinases suggests that protein tyrosine phosphatase 1B may play a role in the growth regulation of ovarian cancers.

Female↗
PubMed1994

Causal nonresponse models for repeated categorical measurements.

This paper uses causal models for nonresponse (Fay, 1986, Journal of the American Statistical Association 81, 354-365) to extend the conditional likelihood procedure for repeated categorical outcome variables to allow for nonrandomly missing data. The extension based on causal models is similar to the log-linear approach presented by Conaway (1992, Journal of the American Statistical Association, 87, 817-824), but has the advantage that the parameters are directly interpretable in terms of the distribution of the outcome variables. As with log-linear model approach, all of the computations can be done with standard statistical software. The methods are first described in terms of a simple example with three binary responses with no covariates and then are applied to a more complicated example. A simulation study evaluates the properties of the estimates based on the proposed method.

Adolescent↗
PubMed1994

Evaluation of prognostic factors and staging in gestational trophoblastic tumor.

OBJECTIVE: To evaluate factors influencing survival and compare current classification systems in women treated for malignant gestational trophoblastic tumor. METHODS: A consecutive series of 454 women treated between 1968-1992 was reviewed retrospectively to identify potential clinical prognostic factors using univariate analysis of life tables. All patients were evaluated using clinical classification, World Health Organization, and recently modified International Federation of Gynecology and Obstetrics (FIGO) staging systems, applied retrospectively. Multivariate Cox regression analysis was used to model potential independent prognostic factors within subsets of the patient population. RESULTS: Factors identified by univariate analysis as potential prognostic influences included age, duration of disease, type of antecedent pregnancy, clinicopathologic diagnosis, site of metastases, number of metastatic sites and foci, tumor size, and prior therapy. The pre-therapy hCG level was not significantly associated with survival (P < .04). Multivariate Cox modeling consistently identified prior therapy, type of antecedent pregnancy, number of metastatic sites, and duration of disease as independent prognostic factors. Clinicopathologic diagnosis and hCG level were of borderline significance only in some models of the total patient population. All classification systems were able to identify low- and high-risk subsets of patients with approximately equal efficiency. The addition of FIGO substages enhanced discrimination between prognostic groups in patients with stage III disease. CONCLUSIONS: Existing systems for the classification of malignant gestational trophoblastic tumor are based in part on factors that are not independently prognostic, such as hCG level or tumor size. These systems discriminate between low- and high-risk patients with approximately equal efficiency. The clinical classification system is currently preferred for determining initial therapy in women with malignant gestational trophoblastic tumors.

Chorionic Gonadotropin↗
PubMed1993

Impact of patient-controlled compression on the mammography experience.

The authors tested the hypothesis that giving women control over the compression portion of the mammography examination results in a less painful experience, greater overall patient satisfaction, and a radiographic image as good as that produced by means of technologist-controlled compression. One hundred nine women undergoing screening mammography at a hospital-based outpatient clinic were studied. Each underwent two-view, screen-film mammography performed in routine fashion except that, by random assignment, one breast was compressed by the technologist and the other breast, by the patient. Patient-controlled compression was significantly (P = .003) less painful than technologist-controlled compression. Overall patient satisfaction (96% [105 of 109]) and willingness to repeat the experience were extremely high. The majority of images (93.5% [202 of 216]) were rated as having good to excellent compression. With minimal patient education, self-compression resulted in an image at least as good as that produced with technologist-applied compression. Further study of this technique is warranted.

Adult↗
PubMed1992

Pre-natal blood lead levels and learning difficulties in children: an analysis of non-randomly missing categorical data.

This paper presents an analysis of categorical variables subject to non-response. We incorporate the incomplete data into the analysis by modelling the distribution of the variables of interest and the non-response mechanism. We discuss issues of model selection and interpretation and the effect of discarding incomplete observations. In addition, we describe how to perform all of the computations with standard statistical software. We discuss the problem of incomplete categorical data within the context of a study of the effect of lead exposure on learning difficulties in children. In this study, many of the children are not observed on some of the variables of interest. It is particularly important in this study to incorporate the incomplete data, since there is evidence that non-response is related to the variables of interest. We reach different conclusions when we incorporate the incomplete data into the analysis than we reach when we discard the incomplete data. We also examine the sensitivity of our conclusions to the choice of a model for the non-response mechanism.

Algorithms↗
PubMedDate not supplied

A comparison of developmental versus functional assessment in the rehabilitation of young children.

BACKGROUND: The purpose of this study was to explore the differences in and potential uses of information derived from developmental vs. functional assessment during the acute rehabilitation of very young children with acquired brain injury. Both methods of assessment are typically used during hospitalization in order to assist in developing individualized goals and outcome measures. With the trend of shortened hospital stays, effective assessment for determining optimal treatment goals and outcomes becomes increasingly important. The results from a developmental and a functional assessment obtained on 23 inpatient children below 6 years of age who had experienced either an acquired brain injury or encephalitis were compared. The data was collected through a retrospective chart review spanning 4 years. METHODS AND OUTCOME MEASURES: Each child received a cognitive and a language test using either the Early Learning Accomplishment Profile (E-LAP) or the Learning Accomplishment Profile Diagnostic (LAP-D) for the developmental assessment measure. The Functional Independence Measure for Children (WeeFIM) was used as a functional assessment. Summary statistics and frequencies were calculated for variables including age and diagnosis. Partial Pearson correlations and 95% confidence intervals were calculated between the functional and developmental assessments, adjusting for the amount of time between administrations of the two exams. Pearson correlations were computed between length of hospital stay and performance on the developmental and functional quotients. RESULTS: Moderate, statistically significant Pearson partial correlations were found between the E-LAP/LAP-D cognitive quotient and the WeeFIM cognitive quotient (r = 0.42, 95% CI (0, 0.72)), the E-LAP/LAP-D language quotient and the WeeFIM cognitive quotient (r = 0.55, 95% CI (0.17, 0.79)) and the E-LAP/LAP-D cognitive quotient and the WeeFIM total quotient (r = 0.50, 95% CI (0.10, 0.76)). An inverse correlation was found between the length of stay and the E-LAP/ LAP-D cognitive quotient (r = -0.68, 95% CI (-0.86, -0.34)) as well as the E-LAP/LAP-D language quotient (r = -0.61, 95% CI (-0.83, -0.23)). CONCLUSIONS: The moderate but limited correlations between developmental and functional assessments may be attributed to differences in the two forms of assessment including the test items, their administration and scoring. While both forms of assessment were thought to be useful for developing individualized treatment goals and measuring outcomes, there were advantages and disadvantages to each.

Activities of Daily Living↗
PubMedDate not supplied

Cancer screening practices among women in a community health center population.

BACKGROUND: Cancer takes a disproportionate toll on disadvantaged Americans. Poverty and low education are risk factors for underuse of cancer screening. METHODS: In this report, we discuss predictors of adherence to cancer screening (mammography, clinical breast exam [CBE], and Pap tests) among 926 women who receive care at a community health center that serves a predominantly low-income and minority population. We examine predictors for each of the tests and for a composite measure of overall cancer screening test compliance. In studying multiple screening behaviors we not only investigate factors associated with each individual behavior, but we also identify consistently effective factors across several behaviors. RESULTS: The analysis indicates consistent effects of age, education, and insurance status on cancer screening. In addition, decisional balance, a measure of a person's beliefs about the pros and cons of complying with the screening test, is associated strongly with adherence. We have extended earlier findings about the positive relationship between decisional balance and mammography to include decisional balance and Pap tests, as well. This finding suggests that behavioral interventions that target decisional balance can effectively promote adherence to cancer screening tests.

Adolescent↗
PubMedDate not supplied

Serum androgens: associations with prostate cancer risk and hair patterning.

Cancer of the prostate is the leading cancer among American men, yet few risk factors have been established. Hair growth and development are influenced by androgens, and it has long been suspected that prostate cancer also is responsive to these hormones. A blinded, case-control study was undertaken to determine if hair patterning is associated with risk of prostate cancer, as well as specific hormonal profiles. The study accrued 315 male subjects who were stratified with regard to age, race, and case-control status (159 prostate cancer cases/156 controls). Hair-patterning classification and serum levels of total and free testosterone (T), sex hormone binding globulin, and dihydrotestosterone (DHT) were performed. Data indicate that hair patterning did not differ between prostate cancer cases and controls; however, significant hormonal differences were detected between the two groups. Free T was greater among cases than in controls (16.4 +/- 6.1 vs. 14.9 +/- 4.8 pg/ml, P = 0.02). Conversely, DHT-related ratios were greater among controls (P = 0.03 for DHT/T and P = 0.01 for DHT/free T). Several strong associations also were found between hormone levels and hair patterning. Men with vertex and frontal baldness had higher levels of free T (16.5 +/- 5.5 and 16.2 +/- 8.0 pg/ml, respectively) when compared to men with either little or no hair loss (14.8 +/- 4.7 pg/ml) (P = 0.01). Data suggest that increased levels of free T may be a risk factor for prostatic carcinoma. In addition, although no differences in hair patterning were detected between cases and controls within this older population, further research (i.e., prospective trials or case-control studies among younger men) may be necessary to determine if hair patterning serves as a viable biomarker for this disease, especially given the strong association between free T levels and baldness.

Aged↗