Trueness, precision and stability of the LIAISON 1-84 parathyroid hormone (PTH) third-generation assay: comparison to existing intact PTH assays.

BACKGROUND: Over the past few decades, parathyroid hormone (PTH) immunoassays have progressed through successive generations resulting in increased specificity and accuracy for detecting circulating PTH. With the introduction of third-generation assays, in which the biologically active PTH(1-84) is specifically targeted, the PTH(7-84) and other fragments are not detected. The specific recognition of only PTH(1-84) whole molecule allows for more reliable standardization and calibration than with the existing assays. METHODS: Samples from patients on hemodialysis or with primary hyperparathyroidism and apparently healthy subjects were examined in different collection matrices (EDTA plasma, unspun EDTA plasma and SST) stored for 0, 24 or 72 h at room temperature to reflect the prevailing sample collection methods, shipping and processing conditions of centralized labs in the United States. Samples were analyzed by the LIAISON 1-84 PTH and N-TACT assays, and by three additional commercially available intact PTH assays. RESULTS: Defined samples, prepared using two different standards (WHO 95/646 international standard and the synthetic Bachem PTH(1-84)), show little bias with the LIAISON 1-84 PTH assay, but not with the other intact PTH assays. Furthermore, PTH is stable for up to 72 h in plasma, but less stable in serum beyond 24 h. CONCLUSIONS: The FDA-approved LIAISON 1-84 PTH assay is accurate and reliably measures the biologically active PTH molecule in plasma or serum stored at room temperature for up 72 and 24 h, respectively.

Combining clinical assessment and the Risk of Ovarian Malignancy Algorithm for the prediction of ovarian cancer.

OBJECTIVES: ACOG guidelines for the evaluation of women with a pelvic mass employ a combination of physical exam, imaging, and CA125 to guide physicians in the triage of women to gynecologic oncologists. We studied the use of ROMA with clinical assessment for cancer risk assessment in women with a pelvic mass. METHODS: This was a prospective, multicenter trial evaluating women with a pelvic mass who had an initial clinical risk assessment (ICRA) performed by a generalist. ROMA scores were calculated and sensitivity, specificity, PPV and NPV were determined for ICRA and ICRA+ROMA. RESULTS: A total of 461 women were entered into the study. There were 375 benign tumors, 48 EOC, 18 LMP tumors and 20 non-ovarian malignancies. For detection of ovarian cancer alone, ICRA had a sensitivity of 85.4%, a specificity of 84.3%, and a NPV of 97.8%. Adding ROMA to ICRA produced a significant improvement of 8.4% in sensitivity, achieving a sensitivity of 93.8% with a specificity of 67.2% and a NPV of 98.8%. Examination of all malignancies (ovarian & non-ovarian) provided a sensitivity of 89.7% for ROMA+ICRA in comparison to 77.9% for ICRA alone, a significant increase in sensitivity of 11.8%. The NPV also significantly increased from 95.5% to 97.3%. Overall, ROMA detected 13 additional malignancies missed by ICRA. CONCLUSIONS: Adjunctive use of ROMA with clinical assessment improves the stratification of women with a pelvic mass into low and high risk groups for ovarian cancer. The combination is particularly effective in ruling out malignant disease.

A Quantile-Quantile Toolbox for Reference Intervals.

BACKGROUND: Parametric statistical methods are generally better than nonparametric, but require that data follow a known, usually normal, distribution. One important application is finding reference limits and detection limits. Parametric analyses yield better estimates and measures of their uncertainty than nonparametric approaches, which rely solely on a few extreme values. Some reference data follow normal distributions; some can be transformed to normal; some are normal or transformable to normal apart from a few extreme values; and detection and quantitation limits can lead to data censoring. METHODS: A quantile-quantile (QQ) toolbox provides powerful general methodology for all these settings. RESULTS: QQ methodology leads to a family of simple methods for finding optimal power transformations, testing for normality before and after transformation, estimating reference limits, and constructing confidence intervals. CONCLUSIONS: These parametric methods have a particular appeal to clinical laboratorians because, while statistically rigorous, they do not require specialized software or statistical expertise, but can be implemented even in spreadsheets. We conclude with an exploration of reference values for amyloid beta proteins associated with Alzheimer disease.

Progranulin (GP88) tumor tissue expression is associated with increased risk of recurrence in breast cancer patients diagnosed with estrogen receptor positive invasive ductal carcinoma.

INTRODUCTION: GP88 (progranulin) has been implicated in tumorigenesis and resistance to anti-estrogen therapies for estrogen receptor positive (ER+) breast cancer. Previous pathological studies showed that GP88 is expressed in invasive ductal carcinoma (IDC), but not in normal mammary epithelial tissue, benign lesions or lobular carcinoma. Based on these results, the present study examines GP88 prognostic significance in association with recurrence and death risks for ER+ IDC patients. METHODS: Two retrospective multi-site clinical studies examined GP88 expression by immunohistochemistry (IHC) analysis of paraffin-embedded breast tumor tissue sections from ER+ IDC patients (lymph node positive and negative, stage 1 to 3) in correlation with patients' survival outcomes. The training study established a GP88 cut-off value associated with decreased disease-free (DFS) and overall (OS) survivals. The validation study verified the GP88 cut-off value and compared GP88 prognostic information with other prognostic factors, particularly tumor size, grade, disease stage and lymph node status in multivariate analysis. RESULTS: GP88 expression is associated with a statistically significant increase in recurrence risk for ER+ IDC patients. The training study established that GP88 3+ score was associated with decreased DFS (P = 0.0004) and OS (P = 0.0036). The independent validation study verified that GP88 3+ score was associated with a 5.9-fold higher hazard of disease recurrence and a 2.5-fold higher mortality hazard compared to patients with tumor GP88 < 3+. GP88 remained an independent risk predictor after considering age, ethnicity, nodal status, tumor size, tumor grade, disease stage, progesterone receptor expression and treatments. CONCLUSIONS: The survival factor GP88 is a novel prognostic biomarker, predictive of recurrence risk and increased mortality for non-metastatic ER+ IDC patients. Of importance, our data show that GP88 continues to be a prognostic factor even after five years. These results also provide evidence that GP88 provides prognostic information independent of tumor and clinical characteristics and would support prospective study to examine whether GP88 expression could help stratify patients with ER+ tumors for adjuvant therapy.

Characterization of dihydrofolate reductases from multiple strains of Plasmodium falciparum using mathematical descriptors of their inhibitors.

Two classes of graph-theoretic molecular descriptors, viz., topological indices (TIs) and atom pairs (APs), have been used to derive high-quality quantitative structure-activity relationships (QSARs) for inhibitors of dihydrofolate reductases (DHFRs) isolated from the wild and four mutant strains of Plasmodium falciparum. Of the three methods used for QSAR formulation, viz., principal-components regression (PCR), partial least squares (PLS), and ridge regression (RR), the RR method outperformed the other two. Cohen's kappa values, based on the overlap of significant and insignificant structural descriptors calculated for the QSAR development, show that DHFR from the wild strain is substantially different from the four mutant strains. The differential QSAR approach reported in this study can be used in protocols for the development of drugs to combat drug-resistant pathogens arising continuously in nature due to mutations. The pairwise kappa values in conjunction with appropriate drug targets and their corresponding set of ligands may be a useful tool in gauging the evolving mutual similarities and dissimilarities of pathogenic organisms from purely computed mathematical descriptors of the ligands.

Comparison of measurements by multiple methods or instruments.

Much work has been done on comparison of one device with another, but the problem of comparing three or more devices is less well known. Most existing work has concentrated on the possibilities of a constant relative bias between the devices, or of different linear relationships with the underlying true value. These two possibilities are placed within a hierarchy of models extending them to settings with multiplicative interaction terms. These additional terms can capture departures such as outliers, variance changing with the analyte concentration, and different measurement variances between the devices.

New HER2-negative breast cancer subtype responsive to anti-HER2 therapy identified.

PURPOSE: HER2 signaling functional activity may be important to measure in addition to HER2 protein quantification when identifying patients eligible for HER2 therapies. A HER2 Signaling Function (CELx HSF) Test for HER2-negative patients uses patient's live tumor cells on a biosensor to identify patients with abnormally high HER2-related signaling (HSFs+) likely to respond to anti-HER2 therapies. METHODS: The CELx HSF test was employed to: (1) characterize the sensitivity and specificity of the test to detect abnormal levels of HER2 signaling; (2) evaluate the inhibitory effectiveness of five different anti-HER2 therapies; (3) assess the correlation between CELx HSF test detection of abnormal HER2 signaling and response to HER2 therapy using xenograft models; and (4) confirm the prevalence of abnormal HER2 signaling amongst HER2-negative breast cancer patients (HER2-/HSFs+). RESULTS: HER2-/HSFs+ breast cancer patient samples were identified and showed sensitivity to five approved anti-HER2 therapies. Xenograft studies using both HER2+ and HER2- cell lines confirmed that CELx HER2 signaling status better predicts HER2 inhibitor efficacy than HER2 receptor status. In a study of 114 HER2-negative breast tumor patient samples, 27 (23.7%; 95% CI = 17-32%) had abnormal HER2 signaling (HSFs+). A ROC curve constructed with this dataset projects the CELx HSF Test would have greater than 90% sensitivity and specificity to detect the HER2-/HSFs+ patient population. CONCLUSIONS: The CELx HSF test is a well-characterized functional biomarker assay capable of identifying dynamic HER2-driven signaling dysfunction in tumor cells from HER2-negative breast cancer patients. This test has demonstrated efficacy of various HER2 targeted therapies in live tumor cells from the HSFs+ population and correlated the test result to HER2 drug response in mouse xenograft studies. The proportion of HER2-negative breast cancer patients found to have abnormal HER2 signaling in a 114 patient sample study, 20-25%, is significant. A clinical trial to evaluate the efficacy of anti-HER2 therapies in this patient population is warranted.

Quantitative structure-activity relationship (QSAR) modeling of human blood: air partitioning with proper statistical methods and validation.

Blood: air partition coefficient (BApc) is important in assessing toxicokinetics of chemicals. Since very few experimental data are available, quantitative structure-activity relationship (QSAR) models based on calculated molecular descriptors can be useful in estimating BApc. Since all descriptors used in the analysis are computed strictly from structure, they can be applied to any chemical, real or hypothetical. In this article, we report models for BApc estimation using four methods, including stepwise ordinary least-squares regression, which is commonly used in QSAR studies but often results in an inflated 'naïve' q2, over-representing the predictive ability of the model. The models developed using proper statistical techniques had q2 values of 0.825 and 0.841, and may be used to reliably predict BApc values for new compounds that are structurally similar to those upon which the models are based. The models developed using improper techniques had associated q2 values, as computed using naïve methods, of 0.920 and 0.934, severely overstating their actual quality.

Inferential, robust non-negative matrix factorization analysis of microarray data.

MOTIVATION: Modern methods such as microarrays, proteomics and metabolomics often produce datasets where there are many more predictor variables than observations. Research in these areas is often exploratory; even so, there is interest in statistical methods that accurately point to effects that are likely to replicate. Correlations among predictors are used to improve the statistical analysis. We exploit two ideas: non-negative matrix factorization methods that create ordered sets of predictors; and statistical testing within ordered sets which is done sequentially, removing the need for correction for multiple testing within the set. RESULTS: Simulations and theory point to increased statistical power. Computational algorithms are described in detail. The analysis and biological interpretation of a real dataset are given. In addition to the increased power, the benefit of our method is that the organized gene lists are likely to lead better understanding of the biology. AVAILABILITY: An SAS JMP executable script is available from http://www.niss.org/irMF

Determination of GP88 (progranulin) expression in breast tumor biopsies improves the risk predictive value of the Nottingham Prognostic Index.

BACKGROUND: The Nottingham Prognostic Index (NPI), which combines numerical values for nodal status, tumor size and histological grade, is used in the standard of care to provide predictive value information on post-surgery survival for patients with primary breast cancer. Attempts to improve the performance of the NPI algorithm have been carried out by testing the inclusion of other biomarker expression and morphological features such as vascular invasion. In the present study, we investigated whether expression of the autocrine growth and survival factor GP88 (progranulin), known to be overexpressed in breast cancer, would improve NPI's predictive value. METHODS: We examined by immunohistochemistry (IHC) the GP88 expression in 508 cases of estrogen receptor positive invasive ductal carcinoma with known clinical outcomes and for which NPI had been determined. GP88 IHC expression was scored by two board certified pathologists and classified into two score groups of GP88 <3+ (0, 1+, 2+) and GP88 = 3+. The correlation between GP88 scoring, NPI and disease-free (DFS) or overall survival (OS) outcomes was then examined by Kaplan-Meier analysis, Cox proportional Hazard (CPH) ratio and Pearson's X (2) test. RESULTS: Kaplan-Meier survival graphs of cases categorized by their NPI scores (<3.4, 3.4-5.4, >5.4) and GP88 expression showed that for patients within the same NPI subgroup, patients having tumors with a high GP88 expression (GP88 IHC score of 3+) had a worse DFS than patients with tumors that had a low GP88 expression (GP88 IHC score <3+). When adjusted for NPI, high GP88 score was significantly associated with recurrence with a hazard ratio of 3.30 (95 % CI 2.12 to 5.14). CONCLUSIONS: The data suggest that the determination of GP88 tumor expression at time of diagnosis for early stage breast cancer patients can provide additional survival information to that provided by NPI alone and thus may be useful for risk management of patients diagnosed with breast cancer.

1,25-Dihydroxyvitamin D to PTH(1-84) Ratios Strongly Predict Cardiovascular Death in Heart Failure.

OBJECTIVES: Vitamin D deficiency and hyperparathyroidism are common in patients with heart failure (HF). There is a growing body of evidence supporting the role of vitamin D and parathyroid hormone (PTH) in cardiac remodeling and worsening of HF. Lack of reliable automated testing of 1,25-dihydroxyvitamin D (1,25(OH)2D), the biologically active metabolite of vitamin D, has limited its contribution to the prognostic assessment of HF. Here, the association of 1,25(OH)2D and PTH(1-84) levels was evaluated for prediction of cardiovascular death in chronic HF patients. METHODS: We conducted a single center prospective cohort including 170 chronic HF patients (females n = 36; males n = 134; NYHA II-IV; mean age: 67 years; etiology: ischemic n = 119, dilated cardiomyopathy n = 51; mean LVEF: 23%). The primary outcome was cardiovascular death. RESULTS: Serum levels of 1,25(OH)2D decreased markedly with increased HF severity. Medians were 33.3 pg/mL for NYHA-II patients, 23.4 pg/mL for NYHA-III, and 14.0 pg/mL for NYHA-IV patients (p<0.001). Most patients had levels of 25(OH)D below 30ng/mL, and stratification by NYHA functional class did not show significant differences (p = 0.249). The 1,25(OH)2D to PTH(1-84) ratio and the (1,25(OH)2D)2 to PTH(1-84) ratio were found to be the most significantly related to HF severity. After a median follow-up of 4.1 years, 106 out of 170 patients reached the primary endpoint. Cox proportional hazard modeling revealed 1,25(OH)2D and the 1,25(OH)2D to PTH(1-84) ratios to be strongly predictive of outcomes. CONCLUSIONS: 1,25(OH)2D and its ratios to PTH(1-84) strongly and independently predict cardiovascular mortality in chronic HF.

Using recursive partitioning analysis to evaluate compound selection methods.

The design and analysis of a screening set for high throughput screening is complex. We examine three statistical strategies for compound selection, random, clustering, and space-filling. We examine two types of chemical descriptors, BCUTs and principal components of Dragon Constitutional descriptors. Based on the predictive power of multiple tree recursive partitioning, we reached the following tentative conclusions. Random designs appear to be as good as clustering and space-filling designs. For analysis, BCUTs appear to be better than principal components scores based upon Constitutional Descriptors. We confirm previous results that model-based selection of compounds can lead to improved screening hit rates.

QSARs for chemical mutagens from structure: ridge regression fitting and diagnostics.

QSAR models have been developed for a diverse set of mutagens using computed molecular descriptors. Such models can be used in predicting mutagenicity from structure. All common methods-regression, neural nets, k-nearest neighbors-are 'linear smoothers'-weighted averages of the activities in the calibration data with weights dependent on the descriptors. While they have been studied extensively, a vital but overlooked area is 'case diagnostics', pointers to compounds that are poorly fitted, or are unusually influential in fitting the model. This is particularly true where the measured activity is binary-present or absent. We illustrate the use of numeric and graphic diagnostics, particularly that of the FF plot, with a data set with 508 compounds and 307 structural descriptors used to predict mutagenicity.

Predicting blood:air partition coefficients using theoretical molecular descriptors.

Three regression methods, namely ridge regression (RR), partial least squares (PLS), and principal components regression (PCR), were used to develop models for the prediction of rat blood:air partition coefficient for increasingly diverse data sets. Initially, modeling was performed for a set of 13 chlorocarbons. To this set, 10 additional hydrophobic compounds were added, including aromatic and non-aromatic hydrocarbons. A set of 16 hydrophilic compounds was also modeled separately. Finally, all 39 compounds were combined into one data set for which comprehensive models were developed. A large set of diverse, theoretical molecular descriptors was calculated for use in the current study. The topostructural (TS), topochemical (TC), and geometrical or 3-dimensional (3D) indices were used hierarchically in model development. In addition, single-class models were developed using the TS, TC, and 3D descriptors. In most cases, RR outperformed PLS and PCR, and the models developed using TC indices were superior to those developed using other classes of descriptors.

Development of a two-step indirect method for modeling Ecom50.

A novel approach is developed for modeling situations in which the modeled property is an algebraically transformed version of the original experimental data. In many cases such a transformation results in a data set with a significantly smaller data range. Here we explore the effects of range-of-data on modeling statistics. We illustrate a twostep method using data on the mass spectrometry collision energy (CE) that is required to decompose 50% of precursor ions to fragments (CE50). Earlier we showed that a nonlinear center-of-mass transformation, yielding Ecom50, produces values less dependent on the specific mass spectrometric experimental conditions. For this data set the Ecom50 range is 13.5% of the CE50 range. We propose a two-step modeling method. First, the original experimental data, CE50, (larger range-of-data) is modeled by a standard modeling method (PLS). Second, the calculated dependent variable resulting from the modeling is algebraically transformed (not modeled) according to the center-of-mass transformation, providing the generally more useful data, Ecom50. As shown here, use of this two-step method for predicting Ecom50 (from previously published data) produces a standard error 21% smaller and correspondingly reduces the confidence interval for prediction. Some specific implications for prediction are given for a published data set. This work is part of the ongoing development of a system of models to assist in the development of human metabolites.

Lymphocyte-bound complement activation products as biomarkers for diagnosis of systemic lupus erythematosus.

Systemic lupus erythematosus (SLE) is frequently misdiagnosed due to the lack of definitive diagnostic tests. The purpose of this study was to determine specifically whether complement activation products (CAP) are deposited on lymphocytes of SLE patients and whether lymphocyte-bound CAP (LB-CAP) may serve as novel biomarkers for the diagnosis of SLE. We conducted a cross-sectional study of 224 patients with SLE, 179 patients with other diseases, and 114 healthy controls. LB-CAP on peripheral blood lymphocytes was measured by flow cytometry. Diagnostic utility of LB-CAP was determined by receiver operating characteristic (ROC) analysis. Significantly elevated levels of C4d and C3d were detected specifically on T and B lymphocytes (designated T-C4d, T-C3d, B-C4d, and B-C3d) of SLE patients. As diagnostic tools, T-C4d and B-C4d, respectively, were 56% sensitive/80% specific and 60% sensitive/82% specific in differentiating SLE from other diseases. Moreover, compared with measurement of anti-dsDNA, serum C3, or serum C4, measurement of T-C4d/B-C4d was significantly more sensitive in identifying SLE patients during a single clinic visit. This is the first investigation of lymphocytes bearing complement activation products in human disease. T-C4d and B-C4d have high diagnostic sensitivity and specificity for SLE and may have added value to current laboratory tests for SLE diagnosis.

Predicting allergic contact dermatitis: a hierarchical structure-activity relationship (SAR) approach to chemical classification using topological and quantum chemical descriptors.

A hierarchical classification study was carried out based on a set of 70 chemicals-35 which produce allergic contact dermatitis (ACD) and 35 which do not. This approach was implemented using a regular ridge regression computer code, followed by conversion of regression output to binary data values. The hierarchical descriptor classes used in the modeling include topostructural (TS), topochemical (TC), and quantum chemical (QC), all of which are based solely on chemical structure. The concordance, sensitivity, and specificity are reported. The model based on the TC descriptors was found to be the best, while the TS model was extremely poor.

Combining chemodescriptors and biodescriptors in quantitative structure-activity relationship modeling.

In view of the wide distribution of halocarbons in our world, their toxicity is a public health concern. Previous work has shown that various measures of toxicity can be predicted with standard molecular descriptors. In our work, biodescriptors of the effect of halocarbons on the liver were obtained by exposing hepatocytes to 14 halocarbons and a control and by producing two-dimensional electrophoresis gels to assess the expressed proteome. The resulting spot abundances provide additional biological information that might be used in toxicity prediction. QSAR models were fitted via ridge regression to predict eight dependent toxicity measures: d37, arr, EC50MTT, EC50LDH, EC20SH, LECLP, LECROS, and LECCAT. Three predictor sets were used for each-the chemodescriptors alone, the biodescriptors alone, and the combined set of both chemo- and biodescriptors. The results differed somewhat from one dependent to another, but overall it was shown that better results could be obtained by using both chemo- and biodescriptors in the model than by using either chemo- or biodescriptors alone. The library of compounds used was small and quite homogeneous, so our immediate conclusions are correspondingly limited in scope, but we believe the underlying methodologies have broad applicability at the interface of chemical and biological descriptors.

Quantitative structure-activity relationship modeling of juvenile hormone mimetic compounds for Culex pipiens larvae, with a discussion of descriptor-thinning methods.

Quantitative structure-activity relationship (QSAR) modelers often encounter the problem of multicollinearity owing to the availability of large numbers of computable molecular descriptors. Sparsity of the variables while using descriptors such as atom pairs increases the complexity. Three different predictor-thinning methods, namely, a modified Gram-Schmidt algorithm, a marginal soft thresholding algorithm, and LASSO (least absolute shrinkage and selection operator), were utilized to reduce the number of descriptors prior to developing linear models. Juvenile hormone (JH) activity of 304 compounds on Culex pipiens larvae was taken as the model data set, and predictor trimming of a large number of diverse descriptors comprising 268 global molecular descriptors (topostructural, topochemical, and geometrical), 13 quantum chemical descriptors, and 915 atom pairs (substructural counts) was applied prior to linear regression by the ridge regression method. The data set (N = 304) was split into five calibration data sets of random samples of sizes 60/110/160/210/260, and the remaining 244/194/144/94/44 compounds were used for validations. LASSO was not found to be a very effective method in handling a large set of descriptors because the number of predictors retained could not exceed the number of observations. The results indicated that the modified Gram-Schmidt algorithm could be used to trim the number of predictors in the global molecular descriptor set where collinearity of the descriptors was the major concern. On the contrary, the soft thresholding approach was found to be an effective tool in subset selection from a diverse set of descriptors having both sparsity and multicollinearity, as in the case of the combined set of atom pairs and global molecular descriptors. The final model developed after variable selection was dominated more by atom pairs, which indicated the important structural moieties that affect JH activity of the compounds. The success of the method reiterates the fact that QSAR or quantitative structure-property relationship (QSPR) models can be developed for a diverse set of compounds using properly parametrized and diverse sets of descriptors, of course, with the selection of the appropriate statistical tools.