Integrating support vector machines and neural networks.

Support vector machines (SVMs) are a powerful technique developed in the last decade to effectively tackle classification and regression problems. In this paper we describe how support vector machines and artificial neural networks can be integrated in order to classify objects correctly. This technique has been successfully applied to the problem of determining the quality of tiles. Using an optical reader system, some features are automatically extracted, then a subset of the features is determined and the tiles are classified based on this subset.     

Spontaneous reporting of adverse drug reactions in an Italian region: Six years of analysis and observations

Pharmacovigilance started in Italy in 1965, and from 1987 reporting of ADRs has been mandatory. Doctors have to send the filled forms on suspected ADRs to the Local Health Districts which transmit biannually all the reports to the Health Department. In a Northern Italian Region (Veneto) spontaneous reporting of adverse drug reactions (ADRs) has been studied during the period 1988–1993. This Region contributes a substantial percentage of the total Italian reports. The total number of reports was 3700, most of these (54 per cent) coming from GPs. A great variability in the reporting rate among the 36 districts of Veneto Region (range: 0–8.8 per 10,000 inhabitants per year) and an important under-reporting have been evident. Underreporting is also emphasized by the fact that in 1993 the Veneto doctors who sent at least one report were only 2.7 per cent. On the whole about 35 per cent of reports concerned minor reactions caused by drugs which have a well known toxicological profile. Comparison between reports coming from Veneto and the UK in some cases show a similar safety profile (omeprazole and simvastatin), whereas in other cases (e.g. terfenadine, glafenine, fluoxetine) no correspondence can be found. This article reveals the limits of the actual spontaneous reporting in Italy and suggests some possible measures for improving it.     

Prognostic role of immunosuppressive acidic protein in advanced ovarian cancer

OBJECTIVE: Our purpose was to investigate immunosuppressive acidic protein in the prognostic characterization of advanced ovarian cancer. STUDY DESIGN: Serum levels of immunosuppressive protein were prospectively measured in 80 patients with untreated ovarian carcinoma. To evaluate the prognostic significance of immunosuppressive acidic protein levels, cutoff points were studied every 50 μg/ml between 450 and 1350 μg/ml. RESULTS: Pretreatment immunosuppressive acidic protein levels were not significantly associated with stage, histotype, grade of differentiation, postoperative residual tumor, and response to chemotherapy. The most significant association with survival was observed at a cutoff value of 1100 μg/ml (p = 0.0089). In the univariate analysis for overall survival, International Federation of Gynecology and Obstetrics stage and immunosuppressive acidic protein status were found to have a role in predicting ovarian cancer prognosis. In the multivariate analysis only immunosuppressive acidic protein status was significantly associated with survival. A statistical correlation was found between serum levels and overall survival (p = 0.0104, χ² 6.56), including immunosuppressive acidic protein as a continuous variable. CONCLUSION: Our data suggest that immunosuppressive acidic protein assay is a potentially useful tool in the prognostic characterization of advanced ovarian cancer. (Am J Obstet Gynecol 1996;175:1606-10.)     

The effects of self-generated comparison targets, BMI, and social comparison tendencies on body image appraisal

Previous studies examining the association between social comparison processes and body image dissatisfaction have yielded inconsistent findings. This study examined whether such discrepancies are due to either the use of identical comparison targets for all subjects or variability in body mass. Specifically, 216 subjects were randomly assigned to one of three experimental conditions: self-generated upward comparison group, self-generated downward comparison group, or control group. Dependent variables were measures of body image. Results indicated that increasing body mass and trait comparison tendencies were associated with increased body dissatisfaction. However, the experimental manipulation did not affect body image ratings. Results suggest that social comparison processes may operate similarly over a range of body mass index (BMI) values.     

Influence of anesthesia on the threshold of pacing-induced ischemia.

Increased myocardial oxygen demand, induced by increased heart rate, may cause myocardial ischemia in the presence of significant coronary artery disease. Alterations in anesthetic depth or technique might put at risk or protect myocardium with compromised blood flow. In 20 dogs with critical left anterior descending coronary artery (LAD) stenosis, atrial pacing rates from 100 to 160 beats/min were achieved, with end-tidal halothane 0.7% (LowH) and 1.1% (HighH), end-tidal isoflurane 1.1% (LowI) and 1.5% (HighI), as well as with continuous fentanyl plus midazolam (FM) infusion anesthesia. Despite significantly different mean arterial and coronary perfusion pressures, rate-pressure product, and left ventricular dP/dtmax, the pacing rate at which systolic shortening decreased below the lower limit of the physiologic response, indicating regional dysfunction, was the same in all investigated anesthesia conditions (LowH: 127 +/- 4 beats/min; HighH: 128 +/- 5 beats/min; LowI: 125 +/- 4 beats/min; HighI: 122 +/- 5 beats/min; FM: 124 +/- 4 beats/min [mean +/- SEM], P greater than 0.05). None of the investigated anesthesia conditions either increased ischemia tolerance or showed a detrimental effect on myocardium with compromised coronary blood flow.     

Inhibition by erythromycin of the conversion of carbamazepine to its active 10,11-epoxide metabolite.

The serum levels of carbamazepine (CBZ) and its 10,11-epoxide metabolite (CBZ-E) were determined in seven subjects after a single dose of CBZ (400 mg) in the control state and during co-administration of erythromycin (500 mg three times daily for 10 days). Erythromycin treatment was associated with a decrease in CBZ clearance and a prolongation of CBZ half-life, while CBZ-E levels were markedly reduced. These data provide evidence that erythromycin inhibits the conversion of CBZ to its epoxide metabolite.