Nonparametric estimation of time-dependent ROC curves conditional on a continuous covariate

The receiver-operating characteristic (ROC) curve is the most widely used measure for evaluating the performance of a diagnostic biomarker when predicting a binary disease outcome. The ROC curve displays the true positive rate (or sensitivity) and the false positive rate (or 1-specificity) for different cut-off values used to classify an individual as healthy or diseased. In time-to-event studies, however, the disease status (e.g. death or alive) of an individual is not a fixed characteristic, and it varies along the study. In such cases, when evaluating the performance of the biomarker, several issues should be taken into account: first, the time-dependent nature of the disease status; and second, the presence of incomplete data (e.g. censored data typically present in survival studies). Accordingly, to assess the discrimination power of continuous biomarkers for time-dependent disease outcomes, time-dependent extensions of true positive rate, false positive rate, and ROC curve have been recently proposed. In this work, we present new nonparametric estimators of the cumulative/dynamic time-dependent ROC curve that allow accounting for the possible modifying effect of current or past covariate measures on the discriminatory power of the biomarker. The proposed estimators can accommodate right-censored data, as well as covariate-dependent censoring. The behavior of the estimators proposed in this study will be explored through simulations and illustrated using data from a cohort of patients who suffered from acute coronary syndrome.

Applicability of Two International Risk Scores in Cardiac Surgery in a Reference Center in Brazil

Background:The applicability of international risk scores in heart surgery (HS) is not well defined in centers outside of North America and Europe.Objective:To evaluate the capacity of the Parsonnet Bernstein 2000 (BP) and EuroSCORE (ES) in predicting in-hospital mortality (IHM) in patients undergoing HS at a reference hospital in Brazil and to identify risk predictors (RP).Methods:Retrospective cohort study of 1,065 patients, with 60.3% patients underwent coronary artery bypass grafting (CABG), 32.7%, valve surgery and 7.0%, CABG combined with valve surgery. Additive and logistic scores models, the area under the ROC (Receiver Operating Characteristic) curve (AUC) and the standardized mortality ratio (SMR) were calculated. Multivariate logistic regression was performed to identify the RP.Results:Overall mortality was 7.8%. The baseline characteristics of the patients were significantly different in relation to BP and ES. AUCs of the logistic and additive BP were 0.72 (95% CI, from 0.66 to 0.78 p = 0.74), and of ES they were 0.73 (95% CI; 0.67 to 0.79 p = 0.80). The calculation of the SMR in BP was 1.59 (95% CI; 1.27 to 1.99) and in ES, 1.43 (95% CI; 1.14 to 1.79). Seven RP of IHM were identified: age, serum creatinine > 2.26 mg/dL, active endocarditis, systolic pulmonary arterial pressure > 60 mmHg, one or more previous HS, CABG combined with valve surgery and diabetes mellitus.Conclusion:Local scores, based on the real situation of local populations, must be developed for better assessment of risk in cardiac surgery.

Association between LDL-C, Non HDL-C, and Apolipoprotein B Levels with Coronary Plaque Regression

Background:Previous reports have inferred a linear relationship between LDL-C and changes in coronary plaque volume (CPV) measured by intravascular ultrasound. However, these publications included a small number of studies and did not explore other lipid markers.Objective:To assess the association between changes in lipid markers and regression of CPV using published data.Methods:We collected data from the control, placebo and intervention arms in studies that compared the effect of lipidlowering treatments on CPV, and from the placebo and control arms in studies that tested drugs that did not affect lipids. Baseline and final measurements of plaque volume, expressed in mm3, were extracted and the percentage changes after the interventions were calculated. Performing three linear regression analyses, we assessed the relationship between percentage and absolute changes in lipid markers and percentage variations in CPV.Results:Twenty-seven studies were selected. Correlations between percentage changes in LDL-C, non-HDL-C, and apolipoprotein B (ApoB) and percentage changes in CPV were moderate (r = 0.48, r = 0.47, and r = 0.44, respectively). Correlations between absolute differences in LDL-C, non‑HDL-C, and ApoB with percentage differences in CPV were stronger (r = 0.57, r = 0.52, and r = 0.79). The linear regression model showed a statistically significant association between a reduction in lipid markers and regression of plaque volume.Conclusion:A significant association between changes in different atherogenic particles and regression of CPV was observed. The absolute reduction in ApoB showed the strongest correlation with coronary plaque regression.

Incidence and Predictors of Cardiovascular Complications and Death after Vascular Surgery

AbstractBackground:Patients undergoing arterial vascular surgery are considered at increased risk for post-operative complications.Objective:To assess the incidence and predictors of complications and death, as well as the performance of two models of risk stratification, in vascular surgery.Methods:This study determined the incidence of cardiovascular complications and deaths within 30 days from surgery in adults. Univariate comparison and logistic regression assessed the risk factors associated with the outcomes, and the receiver operating characteristic (ROC) curve assessed the discriminatory capacity of the revised cardiac risk index (RCRI) and vascular study group of New England cardiac risk index (VSG-CRI).Results:141 patients (mean age, 66 years; 65% men) underwent the following surgeries: carotid (15); lower limbs (65); abdominal aorta (56); and others (5). Cardiovascular complications and death occurred within 30 days in 28 (19.9%) and 20 (14.2%) patients, respectively. The risk predictors were: age, obesity, stroke, poor functional capacity, altered scintigraphy, surgery of the aorta, and troponin change. The scores RCRI and VSG-CRI had area under the curve of 0.635 and 0.639 for early cardiovascular complications, and 0.562 and 0.610 for death in 30 days.Conclusion:In this small and selected group of patients undergoing arterial vascular surgery, the incidence of adverse events was elevated. The risk assessment indices RCRI and VSG-CRI did not perform well for complications within 30 days.

Echocardiographic Predictors of Worse Outcome After Cardiac Resynchronization Therapy

Abstract Background: Cardiac resynchronization therapy (CRT) is the recommended treatment by leading global guidelines. However, 30%-40% of selected patients are non-responders. Objective: To develop an echocardiographic model to predict cardiac death or transplantation (Tx) 1 year after CRT. Method: Observational, prospective study, with the inclusion of 116 patients, aged 64.89 ± 11.18 years, 69.8% male, 68,1% in NYHA FC III and 31,9% in FC IV, 71.55% with left bundle-branch block, and median ejection fraction (EF) of 29%. Evaluations were made in the pre‑implantation period and 6-12 months after that, and correlated with cardiac mortality/Tx at the end of follow-up. Cox and logistic regression analyses were performed with ROC and Kaplan-Meier curves. The model was internally validated by bootstrapping. Results: There were 29 (25%) deaths/Tx during follow-up of 34.09 ± 17.9 months. Cardiac mortality/Tx was 16.3%. In the multivariate Cox model, EF < 30%, grade III/IV diastolic dysfunction and grade III mitral regurgitation at 6‑12 months were independently related to increased cardiac mortality or Tx, with hazard ratios of 3.1, 4.63 and 7.11, respectively. The area under the ROC curve was 0.78. Conclusion: EF lower than 30%, severe diastolic dysfunction and severe mitral regurgitation indicate poor prognosis 1 year after CRT. The combination of two of those variables indicate the need for other treatment options.

Association between Functional Variables and Heart Failure after Myocardial Infarction in Rats

Background: Heart failure prediction after acute myocardial infarction may have important clinical implications. Objective: To analyze the functional echocardiographic variables associated with heart failure in an infarction model in rats. Methods: The animals were divided into two groups: control and infarction. Subsequently, the infarcted animals were divided into groups: with and without heart failure. The predictive values were assessed by logistic regression. The cutoff values predictive of heart failure were determined using ROC curves. Results: Six months after surgery, 88 infarcted animals and 43 control animals were included in the study. Myocardial infarction increased left cavity diameters and the mass and wall thickness of the left ventricle. Additionally, myocardial infarction resulted in systolic and diastolic dysfunction, characterized by lower area variation fraction values, posterior wall shortening velocity, E-wave deceleration time, associated with higher values of E / A ratio and isovolumic relaxation time adjusted by heart rate. Among the infarcted animals, 54 (61%) developed heart failure. Rats with heart failure have higher left cavity mass index and diameter, associated with worsening of functional variables. The area variation fraction, the E/A ratio, E-wave deceleration time and isovolumic relaxation time adjusted by heart rate were functional variables predictors of heart failure. The cutoff values of functional variables associated with heart failure were: area variation fraction < 31.18%; E / A > 3.077; E-wave deceleration time < 42.11 and isovolumic relaxation time adjusted by heart rate < 69.08. Conclusion: In rats followed for 6 months after myocardial infarction, the area variation fraction, E/A ratio, E-wave deceleration time and isovolumic relaxation time adjusted by heart rate are predictors of heart failure onset.

Very Long-Term Prognostic Role of Admission BNP in Non-ST Segment Elevation Acute Coronary Syndrome

Abstract Background: BNP has been extensively evaluated to determine short- and intermediate-term prognosis in patients with acute coronary syndrome, but its role in long-term mortality is not known. Objective: To determine the very long-term prognostic role of B-type natriuretic peptide (BNP) for all-cause mortality in patients with non-ST segment elevation acute coronary syndrome (NSTEACS). Methods: A cohort of 224 consecutive patients with NSTEACS, prospectively seen in the Emergency Department, had BNP measured on arrival to establish prognosis, and underwent a median 9.34-year follow-up for all-cause mortality. Results: Unstable angina was diagnosed in 52.2%, and non-ST segment elevation myocardial infarction, in 47.8%. Median admission BNP was 81.9 pg/mL (IQ range = 22.2; 225) and mortality rate was correlated with increasing BNP quartiles: 14.3; 16.1; 48.2; and 73.2% (p < 0.0001). ROC curve disclosed 100 pg/mL as the best BNP cut-off value for mortality prediction (area under the curve = 0.789, 95% CI= 0.723-0.854), being a strong predictor of late mortality: BNP < 100 = 17.3% vs. BNP ≥ 100 = 65.0%, RR = 3.76 (95% CI = 2.49-5.63, p < 0.001). On logistic regression analysis, age >72 years (OR = 3.79, 95% CI = 1.62-8.86, p = 0.002), BNP ≥ 100 pg/mL (OR = 6.24, 95% CI = 2.95-13.23, p < 0.001) and estimated glomerular filtration rate (OR = 0.98, 95% CI = 0.97-0.99, p = 0.049) were independent late-mortality predictors. Conclusions: BNP measured at hospital admission in patients with NSTEACS is a strong, independent predictor of very long-term all-cause mortality. This study allows raising the hypothesis that BNP should be measured in all patients with NSTEACS at the index event for long-term risk stratification.

Prognostic Value of Pulmonary Vascular Resistance by Magnetic Resonance in Systolic Heart Failure

Abstract Background: Pulmonary hypertension is associated with poor prognosis in heart failure. However, non-invasive diagnosis is still challenging in clinical practice. Objective: We sought to assess the prognostic utility of non-invasive estimation of pulmonary vascular resistances (PVR) by cardiovascular magnetic resonance to predict adverse cardiovascular outcomes in heart failure with reduced ejection fraction (HFrEF). Methods: Prospective registry of patients with left ventricular ejection fraction (LVEF) < 40% and recently admitted for decompensated heart failure during three years. PVRwere calculated based on right ventricular ejection fraction and average velocity of the pulmonary artery estimated during cardiac magnetic resonance. Readmission for heart failure and all-cause mortality were considered as adverse events at follow-up. Results: 105 patients (average LVEF 26.0 ±7.7%, ischemic etiology 43%) were included. Patients with adverse events at long-term follow-up had higher values of PVR (6.93 ± 1.9 vs. 4.6 ± 1.7estimated Wood Units (eWu), p < 0.001). In multivariate Cox regression analysis, PVR ≥ 5 eWu(cutoff value according to ROC curve) was independently associated with increased risk of adverse events at 9 months follow-up (HR2.98; 95% CI 1.12-7.88; p < 0.03). Conclusions: In patients with HFrEF, the presence of PVR ≥ 5.0 Wu is associated with significantly worse clinical outcome at follow-up. Non-invasive estimation of PVR by cardiac magnetic resonance might be useful for risk stratification in HFrEF, irrespective of etiology, presence of late gadolinium enhancement or LVEF.

Optimal designs for mixed effects poisson regression models

Magdeburg, Univ., Fak. für Mathematik, Diss., 2010

Optimal designs for the prediction in hierarchical random coefficient regression models

Magdeburg, Univ., Fak. für Mathematik, Diss., 2015

T Cells Bearing a Chimeric Antigen Receptor against Prostate-Specific Membrane Antigen Mediate Vascular Disruption and Result in Tumor Regression.

Aberrant blood vessels enable tumor growth, provide a barrier to immune infiltration, and serve as a source of protumorigenic signals. Targeting tumor blood vessels for destruction, or tumor vascular disruption therapy, can therefore provide significant therapeutic benefit. Here, we describe the ability of chimeric antigen receptor (CAR)-bearing T cells to recognize human prostate-specific membrane antigen (hPSMA) on endothelial targets in vitro as well as in vivo. CAR T cells were generated using the anti-PSMA scFv, J591, and the intracellular signaling domains: CD3ζ, CD28, and/or CD137/4-1BB. We found that all anti-hPSMA CAR T cells recognized and eliminated PSMA(+) endothelial targets in vitro, regardless of the signaling domain. T cells bearing the third-generation anti-hPSMA CAR, P28BBζ, were able to recognize and kill primary human endothelial cells isolated from gynecologic cancers. In addition, the P28BBζ CAR T cells mediated regression of hPSMA-expressing vascular neoplasms in mice. Finally, in murine models of ovarian cancers populated by murine vessels expressing hPSMA, the P28BBζ CAR T cells were able to ablate PSMA(+) vessels, cause secondary depletion of tumor cells, and reduce tumor burden. Taken together, these results provide a strong rationale for the use of CAR T cells as agents of tumor vascular disruption, specifically those targeting PSMA. Cancer Immunol Res; 3(1); 68-84. ©2014 AACR.

Giant warts in a kidney transplant patient: regression with sirolimus.

Purpose: Recent reports have suggested that intraabdominal postoperative infection is associated with higher rates of overall and local recurrence and cancer-specific mortality. However, the mechanisms responsible for this association are unknown. We hypothesized that the greater inflammatory response in patients with postoperative intraabdominal infection is associated to an increase in local and systemic angiogenesis. Methods: We designed a prospective cohorts study with matched controls. Patients with postoperative intra-abdominal infection (abscess and/or anastomotic leakage) (group 1; n=17) after elective colorectal cancer resection operated on for cure were compared to patients with an uncomplicated postoperative course (group 2; n=17). IL-6 and VEGF levels were determined by ELISA in serum and peritoneal fluid at baseline, 48 hours and postoperative day 4 or at the time the peritoneal infection occurred. Results: No differences were observed in age, gender, preoperative CEA, tumor stage and location and type of procedure performed. Although there were no differences in serum IL-6 levels at 48 hours, this pro-inflammatory cytokine was higher in group 1 on postoperative day 4 (group 1: 21533 + 27900 vs. group 2: 1130 + 3563 pg/ml; p < 0.001). Serum VEGF levels were higher in group 1 on postoperative day 4 (group 1: 1212 + 1025 vs. group 2: 408 + 407 pg/ml; p < 0.01). Peritoneal fluid VEGF levels were also higher in group 1 at 48 hours (group 1: 4857 + 4384 vs. group 2: 630 + 461 pg/ml; p < 0.001) and postoperative day 4 (group 1: 32807 + 98486 vs. group 2: 1002 + 1229 pg/ml; p < 0.001). A positive correlation between serum IL-6 and VEGF serum levels was observed on postoperative day 4 (r=0.7; p<0.01). Conclusions: These results suggest that not only the inflammatory response but also the angiogenic pathways are stimulated in patients with intra-abdominal infection after surgery for colorectal cancer. The implications of this finding on long-term follow-up need to be evaluated.

Long-term outcome after cognitive and behavioral regression in nonlesional epilepsy with continuous spike-waves during slow-wave sleep.

PURPOSE: To present the long-term follow-up of 10 adolescents and young adults with documented cognitive and behavioral regression as children due to nonlesional focal, mainly frontal, epilepsy with continuous spike-waves during slow wave sleep (CSWS). METHODS: Past medical and electroencephalography (EEG) data were reviewed and neuropsychological tests exploring main cognitive functions were administered. KEY FINDINGS: After a mean duration of follow-up of 15.6 years (range, 8-23 years), none of the 10 patients had recovered fully, but four regained borderline to normal intelligence and were almost independent. Patients with prolonged global intellectual regression had the worst outcome, whereas those with more specific and short-lived deficits recovered best. The marked behavioral disorders resolved in all but one patient. Executive functions were neither severely nor homogenously affected. Three patients with a frontal syndrome during the active phase (AP) disclosed only mild residual executive and social cognition deficits. The main cognitive gains occurred shortly after the AP, but qualitative improvements continued to occur. Long-term outcome correlated best with duration of CSWS. SIGNIFICANCE: Our findings emphasize that cognitive recovery after cessation of CSWS depends on the severity and duration of the initial regression. None of our patients had major executive and social cognition deficits with preserved intelligence, as reported in adults with early destructive lesions of the frontal lobes. Early recognition of epilepsy with CSWS and rapid introduction of effective therapy are crucial for a best possible outcome.

Innovation sources and productivity in Catalonian firms: a quantile regression analysis

This paper explores the effects of two main sources of innovation -intramural and external R&D- on the productivity level in a sample of 3,267 Catalonian firms. The data set used is based on the official innovation survey of Catalonia which was a part of the Spanish sample of CIS4, covering the years 2002-2004. We compare empirical results by applying usual OLS and quantile regression techniques both in manufacturing and services industries. In quantile regression, results suggest different patterns at both innovation sources as we move across conditional quantiles. The elasticity of intramural R&D activities on productivity decreased when we move up the high productivity levels both in manufacturing and services sectors, while the effects of external R&D rise in high-technology industries but are more ambiguous in low-technology and knowledge-intensive services. JEL codes: O300, C100, O140. Keywords: Innovation sources, R&D, Productivity, Quantile regression

A priori ratemaking using bivariate poisson regression models

In automobile insurance, it is useful to achieve a priori ratemaking by resorting to gene- ralized linear models, and here the Poisson regression model constitutes the most widely accepted basis. However, insurance companies distinguish between claims with or without bodily injuries, or claims with full or partial liability of the insured driver. This paper exa- mines an a priori ratemaking procedure when including two di®erent types of claim. When assuming independence between claim types, the premium can be obtained by summing the premiums for each type of guarantee and is dependent on the rating factors chosen. If the independence assumption is relaxed, then it is unclear as to how the tari® system might be a®ected. In order to answer this question, bivariate Poisson regression models, suitable for paired count data exhibiting correlation, are introduced. It is shown that the usual independence assumption is unrealistic here. These models are applied to an automobile insurance claims database containing 80,994 contracts belonging to a Spanish insurance company. Finally, the consequences for pure and loaded premiums when the independence assumption is relaxed by using a bivariate Poisson regression model are analysed.