Use of ROC curve analysis for prediction gives fallacious results: Use predictivity-based indices.

ABSTRACT: The area under the ROC curve is frequently used for assessing the predictive efficacy of a model, and the Youden index is commonly used to provide the optimal cut-off. Both are misleading tools for predictions. A ROC curve is drawn for the sensitivity of a quantitative test against its (1 - specificity) at different values of the test. Both sensitivity and specificity are retrospective in nature as these are indicators of correct classification of already known conditions. They are not indicators of future events and are not valid for predictions. Predictivity intimately depends on the prevalence which may be ignored by sensitivity and specificity. We explain this fallacy in detail and illustrate with several examples that the actual predictivity could differ greatly from the ROC curve-based predictivity reported by many authors. The predictive efficacy of a test or a model is best assessed by the percentage correctly predicted in a prospective framework. We propose predictivity-based ROC curves as tools for providing predictivities at varying prevalence in different populations. For optimal cut-off for prediction, in place of the Youden index, we propose a P-index where the sum of positive and negative predictivities is maximum after subtracting 1. To conclude, for correctly assessing adequacy of a prediction models, predictivity-based ROC curves should be used instead of the usual sensitivity-specificity-based ROC curves and the P-index should replace the Youden index.

The minimum effective dose (ED90) of prophylactic oxytocin infusion during cesarean delivery in patients with and without obesity: an up-down sequential allocation dose-response study.

BACKGROUND: Obesity is associated with greater oxytocin requirement during labor induction or augmentation. There are scant data exploring the intra-operative requirement during cesarean delivery in patients with obesity, and none comparing it with those without obesity. We evaluated the minimum effective dose (ED90) of an oxytocin infusion to achieve adequate uterine tone during cesarean delivery in patients with and without obesity. METHODS: Patients (body mass index ≥30 kg/m2 represented patients with obesity) undergoing cesarean delivery using subarachnoid block were included. This prospective dual-arm dose-finding study used a 9:1 biased sequential allocation design. Oxytocin infusion was initiated at 13 IU/h at cord clamping in the first patient of each group. Uterine tone was graded as satisfactory or unsatisfactory by the obstetrician four minutes after initiation of the infusion. The dose of oxytocin infusion for subsequent patients was determined according to the response of the previous patient in the group. Oxytocin-associated side effects were evaluated. Dose-response data for the groups was evaluated using log-logistic function and ED90 estimates derived from fitted equations using the delta method. RESULTS: The ED90 of oxytocin was significantly higher for patients with obesity (n = 40) compared with those without obesity (n = 40) [25.7 IU/h, 95% CI 18.6 to 32.9) vs. 16.6 IU/h, 95% CI 14.9 to 18.3)]; relative ratio 1.55 [95% CI 1.09 to 2.01] (P = 0.019). CONCLUSIONS: Patients with obesity require a higher intra-operative oxytocin infusion dose rate to achieve a satisfactorily contracted uterus after fetal delivery when compared with patients without obesity.

Calculation of effective dose of phenylephrine bolus for treatment of post-spinal hypotension in pre-eclamptic patients undergoing caesarean section - a non-randomised controlled trial.

BACKGROUND: Patients with pre-eclampsia require smaller vasopressor doses compared with those with normotension for management of post-spinal hypotension during caesarean section. However, the literature has little evidence as to the phenylephrine dose required for patients with pre-eclampsia. METHODS: Fifty patients, with either pre-eclampsia or normotension, and developing post-spinal hypotension during caesarean section under spinal anaesthesia, were studied. Women in both groups did not receive prophylactic vasopressors. The first patient in each group received phenylephrine 50 µg to treat the first episode of hypotension, defined as fall of systolic blood pressure ≥20% from baseline or an absolute value <100 mmHg. If hypotension was corrected within one minute it was considered a 'success'. The doses for the subsequent patients were determined by responses to all previous patients, according to a variation of Narayana's rule for the up-down sequential allocation method. RESULTS: The 95% effective dose (ED95) and 50% effective dose (ED50) of phenylephrine was 41.7 µg (95% CI 33.8 to 49.6 µg) and 29.1 µg (95% CI 26.0 to 32.2 µg) respectively in the pre-eclampsia group, and 64.9 µg (95% CI 54.1 to 75.7 µg) and 47.3 µg (95% CI 39.7 to 54.9 µg) respectively in the normotensive group. The proportionate reduction in phenylephrine dose ranged from 33% (95% CI 18 to 44%) to 40% (95% CI 19 to 52%). CONCLUSION: Patients with pre-eclampsia may need a 33% to 40% reduction in the first phenylephrine bolus dose, compared with patients with normotension, for the treatment of the first episode of post-spinal hypotension.

Neonatal outcomes following phenylephrine or norepinephrine for treatment of spinal anaesthesia-induced hypotension at emergency caesarean section in women with fetal compromise: a randomised controlled study.

BACKGROUND: Norepinephrine is as effective as phenylephrine for management of spinal anaesthesia-induced hypotension. Most of the studies comparing these vasopressors have been conducted in healthy pregnant women undergoing elective caesarean section. In the current study, we tested the null hypothesis that there is no difference in neonatal outcome when phenylephrine or norepinephrine is used to treat spinal anaesthesia-induced hypotension in women undergoing emergency caesarean section for fetal compromise. METHODS: Patients undergoing caesarean section for fetal compromise who developed spinal anaesthesia-induced hypotension were randomised to receive phenylephrine 100 µg or norepinephrine 8 µg for treatment of each hypotensive episode, defined as systolic blood pressure <100 mmHg. Umbilical cord arterial and venous blood samples were obtained for blood gas analysis. The primary outcome measure was umbilical artery pH. RESULTS: One hundred patients (50 in each group) were studied. There was no significant difference in umbilical artery pH between the two groups (mean difference 0.001; 95% CI -0.032 to 0.034). The number of hypotensive episodes, vasopressor boluses required, the incidence of bradycardia, heart rate and blood pressure trends following vasopressor administration, and the incidence of nausea/vomiting were not significantly different between groups. CONCLUSION: Phenylephrine 100 µg and norepinephrine 8 µg were not significantly different in terms of neonatal outcome when administered as intravenous boluses for treatment of spinal anaesthesia-induced hypotension in parturients undergoing emergency caesarean sections for fetal compromise.

A randomised controlled trial of phenylephrine and noradrenaline boluses for treatment of postspinal hypotension during elective caesarean section.

Phenylephrine is currently the vasopressor of choice during elective caesarean section, but it can cause reflex bradycardia. Noradrenaline, a potent α-agonist and weak ß-agonist, may be associated with a lower incidence of bradycardia. However, comparative information is limited. This double-blind randomised controlled trial compared the effects of 100 µg phenylephrine and 5 µg noradrenaline administered as boluses for the treatment of postspinal hypotension during elective caesarean section in women with an uncomplicated singleton pregnancy. Hypotension was defined as a decrease of ≥ 20% from baseline systolic arterial pressure, or an absolute value < 100 mmHg. Ninety women were included in the study. The primary outcome was the incidence of maternal bradycardia < 60 beats.min-1 . There was no difference in the incidence of bradycardia (37.8% with phenylephrine vs. 22.2% with noradrenaline; p = 0.167), number of hypotensive episodes, number of boluses required to treat the first hypotensive episode or reactive hypertension. The total number of boluses used was higher in the phenylephrine group (p = 0.01). Maternal heart rate at 1 min after vasopressor administration was non-significantly lower using phenylephrine vs. noradrenaline (p = 0.034, considering p < 0.01 as statistically significant). The umbilical artery pH was higher using phenylephrine than with noradrenaline (p = 0.034). In conclusion, both vasopressors reversed postspinal hypotension without a statistically significant difference in maternal bradycardia. However, in view of the lower umbilical artery pH when using noradrenaline, further research is warranted to study its placental transfer and fetal metabolic effects.

Comparison of the potency of phenylephrine and norepinephrine bolus doses used to treat post-spinal hypotension during elective caesarean section.

BACKGROUND: Phenylephrine, although considered the vasopressor of choice, can cause reflex bradycardia and a fall in cardiac output. Norepinephrine, due to its direct positive chronotropic and reflex negative chronotropic actions, is expected to overcome this problem. However, limited information about its effective dose for management of post-spinal hypotension, and its potency compared to phenylephrine, is available. METHODS: One hundred consecutive patients who developed post-spinal hypotension were treated with a predetermined dose of either phenylephrine or norepinephrine. Correction of hypotension after one minute was considered 'success'. The starting dose for the first patient and testing interval (the incremental or decremental dosing) were 100 µg and 10 µg in the phenylephrine group, and 6 µg and 0.5 µg in the norepinephrine group. Doses for subsequent patients were determined by the responses of previous patients according to the Narayana rule for up-down sequential allocation. ED95 and ED50 of phenylephrine and norepinephrine boluses and their potency ratio were calculated. RESULTS: Using Probit analysis, ED95 and ED50 values were 43.1 µg (95% CI 39.5 to 65.0 µg) and 33.2 µg (95% CI 5.1 to 37.0 µg) for phenylephrine, and 3.7 µg (95% CI 3.5 to 4.7 µg) and 3.2 µg (95% CI 1.8 to 3.4 µg) for norepinephrine. The relative potency ratio of norepinephrine and phenylephrine was 11.3 (95% CI 8.1 to 16.9). CONCLUSION: Based on the results of this study, norepinephrine is about 11 times more potent than phenylephrine. When used as bolus doses for treatment of hypotension, 100 µg phenylephrine should be approximately equivalent to 9 µg norepinephrine.

Environment-friendly route for stable aqueous dispersion of reduced graphene oxide for heat transfer application.

A simultaneous functionalization and reduction route to prepare stable dispersion of reduced graphene oxide from graphene oxide has been described. Diethanol amine has been introduced for the first time as an environment friendly reducing agent in a simple reflux reaction. Diethanol amine acts as a reducing agent and helps to enhance the stability of dispersion, making hydrogen bonding by virtue of two functional groups. The prepared dispersion of 0.025 mg/mL concentration is stable for months together and has a zeta potential value -45 V at room temperature. UV-Vis study shows peak at 264 nm that is signatory for reduced graphene oxide. TEM images confirm spread thin sheets of graphene of few hundred nanometer lateral dimension. Thermal diffusivity studies suggest nearly 60% enhancement for the dispersion in comparison to base fluid, water. This suggests graphene dispersion is promising for heat transfer applications.

Primary cutaneous large B-cell lymphoma shows activation of nuclear factor kappa B and low incidence of Epstein-Barr virus.

Primary cutaneous lymphomas are defined as lymphomas, which are present in the skin without evidence of extracutaneous disease at the time of diagnosis. Primary cutaneous large B-cell lymphoma (PCLBCL) is a subtype of primary cutaneous B-cell lymphoma with a female predominance, occurring in elderly patients and known to have unfavorable prognosis. We evaluated 10 cases of PCLBCL in immunocompetent patients between 2005 and 2008. A panel of immunoperoxidase stains; CD3, CD10, CD20, BCL2, BCL6, and MUM1 were performed on all cases. Nuclear factor kappa B (NF-kappaB) pathway activation was evaluated using an immunostain for P65. The presence of Epstein-Barr virus (EBV) was assessed using Epstein-Barr virus encoded RNA (EBER) in situ hybridization probe. All cases were CD20 positive and CD3 negative. CD10, BCL6, BCL2, and MUM1 were positive in 4/10 (40%), 6/10 (60%), 7/10 (70%), and 7/10 (70%) cases, respectively. NF-kappaB activation was detected in 7/10 (70%) cases. One (10%) case was positive for EBV by in situ hybridization. Interestingly, the EBV positive case was also positive for MUM1 and negative for CD10, indicating an activated immunophenotype. In conclusion, majority of PCLBCL shows activation of NF-kappaB pathway with a low incidence of EBV.

Biodiesel surrogates: achieving performance demands.

Synthesis of surrogate molecules is particularly useful for generating in sight of structural-activity relationships, understanding processes and improving the performance. In order to improve upon the physico-chemical properties of biodiesel, methyl, ethyl, isopropyl and n-butyl esters of beta-branched fatty acid have been synthesized, initiating from beta-branched alcohols. Beta-branched alcohols upon oxidation gave corresponding acids, which were converted to their esters. The synthesized esters have substantially better oxidative stability, exhibited by Rancimat oxidation induction period of more than 24 h. The cloud point of synthesized esters is < -36 degrees C, pour point is < -42 degrees C and CFPP is < -21 degrees C, which is substantially better than fatty acid methyl esters. Besides achieving the objective of better oxidative stability and improved low temperature properties, the synthesized surrogate esters have viscosity in the range of 4.2-4.6 cSt at 40 degrees C, meeting the international diesel and biodiesel standards. The cetane number of synthesized esters is 62-69, which is much better than diesel and biodiesel. The blends of the synthesized esters in diesel at 5% and 10% meet Indian standards of diesel.

Estimation of trace impurities in reactor-grade uranium using ICP-AES.

Estimation of impurities in reactor grade uranium is important from the point of view of neutron economy. For chemical separation, ion exchange and solvent extraction techniques have been employed although the latter is generally preferred. Amongst various extractants TBP (tri-n-butyl phosphate), TBP-TOPO (tri-n-octyl phosphine oxide), or TOPO only (in CCl(4), xylene, dodecane) is most often used. New reagents like Cyanex-923 (mixture of 4 tri-alkyl phosphine oxides)/TEHP (tri-ethylhexyl phosphoric acid) are also being used. This communication reports chemical separation of uranium by precipitation using 1,2-diaminocyclohexane NNN'N'-tetra acetic acid (CyDTA)/ammonium hydroxide in presence of 1,10-phenanthroline and estimation of impurities in the filtrate by ICP-AES. Quantitative separation of U, a high spectral interferent in plasma and recovery of impurities have been achieved. Recovery of Cd has been improved by using 1,10-phenanthroline. The method is accurate and precise, offering a relative standard deviation ranging from less than 4% (3.8% for Eu at the 10mug g(-1) level) to 12.9% (for Ce at the 2.5 mug g(-1) level) for all the elements studied.

Determination of tungsten in niobium-tantalum, vanadium and molybdenum bearing geological samples using derivative spectrophotometry and ICP-AES.

Two different procedures, one using derivative spectrophotometry and another using inductively coupled plasma atomic emission spectrometry (ICP-AES) have been developed for the determination of tungsten in niobate-tantalates, tin slag samples, ores, concentrates and vanadium and molybdenum bearing geological materials. In the first method involving derivative spectrophotometry, 0.05-0.5 g of the sample is fused with sodium hydroxide, the tungsten is extracted by leaching the melt with distilled water and estimated as thiocyanate using a second derivative spectrophotometric method in the presence of interferents, i.e. Nb, Mo and V, without separating them. Mixtures of tungsten with V, Nb and Mo are used for standardizing the various parameters like zero-crossing wavelength, wavelength range, etc. Tolerance limits for V, Nb and Mo have also been evaluated. In the second method involving ICP-AES, 0.05-0.5 g of sample is fused with KHSO(4) to a clear melt and dissolved in ammonium oxalate solution. Ammonium hydroxide precipitation is then carried out to separate Nb and Ta as hydroxides and the filtrate is boiled with nitric acid to destroy the oxalates before aspiration into the plasma for measurement of tungsten values by ICP-AES using the 207.911 nm emission line. Both methods have been applied to niobate-tantalate and tin slag samples and the results obtained are reported in this paper. The values obtained by both methods are in good agreement with each other. The proposed methods have also been applied to the determination of tungsten in two Canadian Certified Reference Standards (CT-1 and MP-2) and the values obtained are in good agreement with the certified values and the R.S.D.% in case of the ICP-AES method varied from 1-2% at >1000 mug g(-1) level to 9.4% at the 20 mug g(-1) level whereas the R.S.D.% in case of the derivative method varied from 1 to 7.8%.

Myosin isoforms in uterine smooth muscle during pregnancy in rat.

Effects of uterine stretching and physiological hypertrophy on myosin isozyme were investigated in rat during pregnancy. Both nonpregnant and pregnant rat uteri express a single myosin band on native gels. Analysis of native myosin under denaturing conditions revealed two myosin heavy chains (MHCs) with molecular mass of 204 and 200 kDa respectively. Filamin, a 240 kDa protein co-electrophoreses with myosin on native gels. No correlation is found between regulatory myosin light chain phosphorylation and pattern of myosin isozymes or the MHC. The results suggest that uterine stretching and physiological hypertrophy during pregnancy do not induce any changes in uterine myosin isozyme.

Effect of vitamin C on androgen independent prostate cancer cells (PC3 and Mat-Ly-Lu) in vitro: involvement of reactive oxygen species-effect on cell number, viability and DNA synthesis.

Studies have described the protective role of vitamin C (ascorbic acid) in certain types of cancer. In this study, we report the effects of vitamin C treatment of two androgen independent prostate cancer cell lines from human (PC3) and rat (Mat-Ly-Lu or MLL) sources. In vitro treatment of PC3 and MLL with sodium ascorbate acid (0-10 mM) resulted in a decrease in cell viability and thymidine incorporation into DNA. These effects of vit. C were dose and time dependent. Ascorbate induced these changes through the production of hydrogen peroxide since addition of catalase (100-300 units/ml), an enzyme that degrades hydrogen peroxide, inhibited the effects of ascorbate on these cell lines. In contrast, superoxide dismutase, an enzyme that dismutates superoxide and generates hydrogen peroxide did not prevent ascorbate-induced changes emphasizing the involvement of reactive oxygen species (ROS) in cellular damage. That singlet oxygen scavengers such as sodium azide and hydroquinone, hydroxyl radical scavengers such as D-mannitol and DL-alpha-tocopherol did not counteract the effects of ascorbate on thymidine incorporation suggests that these free radicals are not involved in cellular damage. In conclusion, these results suggest that vitamin C inhibits tumor growth by virtue of producing reactive oxygen species. These results suggest that ascorbate is a potent anticancer agent for prostate cancer cells.

Changes in myosin heavy chain (MHC) expression during pregnancy in rat uterus.

Expression of myosin heavy chain (MHC) was investigated in rat uterus during pregnancy. Two MHC isoforms, SM MHCI (204 kDa) and SM MHCII (200 kDa), were resolved following the analysis of Guba Straub extract of non-pregnant and pregnant uterus using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Whereas SM MHCI showed an increase from 69.26 +/- 3.26% in nonpregnant uterus to 81.40 +/- 5.36% in pregnant uterus, the SM MHCII exhibited a corresponding decline from 30.73 +/- 3.29 to 18.59 +/- 5.36%. The native myosin separated under non-dissociating conditions and subsequently analyzed in denaturing SDS-PAGE also demonstrated two MHC isoforms with identical electrophoretic mobilities. A SDS-PAGE analysis of native myosin from pregnant rats extracted at room temperature and in the absence of proteolytic inhibitors revealed a characteristically increased proteolysis of MHC into two peptide products of 153 and 140 kDa during pregnancy. Such a proteolysis of MHC, but in very low proportions, was noticed in nonpregnant uterus too.

Role of substance P in several models of bladder inflammation.

Substance P (SP) is a peptide found in the sensory nervous system which has multiple biologic effects including stimulation of muscle contraction, pain nociception, immune cell functions, plasma extravasation and a constellation of inflammatory effects. Here we investigate the role of SP in several animals models of bladder inflammation. Using the female Lewis rat, inflammation was induced using either xylene, lipopolysaccharide (LPS) or polyinosinic-polycytidylic acid (polyIC). Inflammation occurred rapidly (4 h) and was maintained in each model for at least 7 days. Each of these protocols decreased the bladder content of immunoreactive SP by approximately 50%, suggesting enhanced release. There was no change in the urinary frequency of these animals over 3 weeks, suggesting that urinary frequency changes are not mediated by acute inflammation. We also found that the SP receptor (NK1) antagonist, (-)CP96345, was unable to block the inflammation produced by polyIC, suggesting that SP is not an obligatory mediator of immune cell stimulation in this model.

Antiproliferative effects of c-myc antisense oligonucleotide in prostate cancer cells: a novel therapy in prostate cancer.

OBJECTIVES: To explore the possibility of using antisense oligonucleotide therapy for prostate cancer, we investigated the effect of c-myc-antisense-oligonucleotide (c-myc-As-ODN) in human prostate cancer cell lines such as LNCaP, PC3, and DU145. METHODS: LNCaP, PC3, and DU145 cells were incubated in the presence of c-myc-As-ODN. Dose (0 to 10 microM) and time dependent (1 to 6 days) effects on proliferation and viability were examined by [3H]thymidine incorporation and MTT assay, respectively. Flow cytometry analysis was carried out to analyze cell cycle status by determining the DNA content in LNCaP cells. Control cultures received either c-myc-sense-ODN or scrambled (nonsense) nucleotides. RESULTS: Time- and dose-dependent decreases in DNA synthesis and cell viability were noted for all three prostate cancer cell lines after c-myc-As-ODN treatment. Further studies using LNCaP cells indicated that these changes were accompanied by an increase in the percentage of cells with less than 2N DNA content after c-myc-As-ODN treatment. The results suggest that c-myc-As-ODN induces cell death. Comparison of a c-myc-As-ODN-treated group with a group subjected to isoleucine deprivation revealed that thymidine incorporation was almost the same in c-myc-As-ODN-treated LNCaP cells and in LNCaP cells at early S phase. CONCLUSIONS: These results suggest that c-myc-As-ODN inhibits prostate cancer cell growth and proliferation mainly by decreasing cell viability.

17 beta-estradiol induced prostatitis in the rat is an autoimmune disease.

PURPOSE: Studies suggest that alteration in steroid hormone levels may be one of the factors causing nonbacterial prostatitis (NBP) in rats. We hypothesized that hormonally induced prostatitis in the rat may be an autoimmune disease. Studies were carried out to prove this hypothesis. MATERIALS AND METHODS: We injected 17 beta-estradiol (E2; 250 micrograms/kg. subcutaneously) or vehicle into 1-year-old male rats for 30 days, and isolated and cultured the splenocytes in the presence of con-A (Experiment 1). Approximately 10(7) splenocytes were adoptively transferred into young naive syngeneic rats. To find out whether or not the inflammation is mediated by T-lymphocytes, a pure population of T-lymphocytes from E2-treated 3-month-old rats was injected into young naive syngeneic rats (Experiment 2). To rule out the possibility that the inflammation was due to con-A itself, we cross-linked some T-cells with anti-CD3 antibody before adoptive transfer (Experiment 2). RESULTS: The recipients of splenocytes from E2-treated animals presented evidence of inflammation in terms of lymphocytic infiltration and presence of degranulated mast cells. Furthermore, we observed in these animals an increase in histamine-releasing peptide (HRP) levels, an indication of plasma extravasation. The T-cells stimulated by anti-CD3 antibody produced a similar degree of inflammation, thereby ruling out the possibility that the inflammation was due to con-A. The results also indicated that the immune response to antigen(s) is not dependent on the age of the animal but is dependent on a population of CD3+ T-cells. CONCLUSION: Our results demonstrate that hormonal imbalance and autoimmunity in male rats produce the symptoms of nonbacterial prostatitis.