Factors Associated With Mailed Fecal Immunochemical Test Completion in an Integrated Academic-Community Healthcare System.

INTRODUCTION: Mailed fecal immunochemical test (FIT) outreach is an effective strategy to increase colorectal cancer (CRC) screening. The aim of this study was to determine the patient-level, clinic-level, and geographic-level factors associated with CRC screening completion in a mailed FIT outreach program. METHODS: This retrospective cohort study was conducted in the integrated healthcare system of University of Washington Medicine and included patients aged 50-75 years, who were due for CRC screening, and had a primary care encounter in the past 3 years. Eligible patients received mailed outreach that included a letter with information about CRC screening, FIT kit, and a prepaid return envelope. CRC screening and factors associated with completion were obtained from electronic health records and the CRC screening program database. RESULTS: Of the 9,719 patients who received mailed outreach, 29.6% completed FIT mailed outreach. The median FIT return time was 27 days (interquartile range 14-54). On multivariate analysis, patients with a higher area deprivation index, insured through Medicaid, living without a partner, and whose last primary care visit was >12 months ago were less likely to complete a FIT compared with their counterparts. Over a 12-month period, overall CRC screening across the health system increased by 2 percentage points (68%-70%). DISCUSSION: Mailed FIT outreach in an integrated academic-community practice was feasible, with 32% of invited patients completing CRC screening by FIT or colonoscopy, on par with published literature. Patient and geographic-level factors were associated with CRC screening completion. These data will inform additional interventions aimed to increase CRC screening participation in this population.

Combining biomarkers by maximizing the true positive rate for a fixed false positive rate.

Biomarkers abound in many areas of clinical research, and often investigators are interested in combining them for diagnosis, prognosis, or screening. In many applications, the true positive rate (TPR) for a biomarker combination at a prespecified, clinically acceptable false positive rate (FPR) is the most relevant measure of predictive capacity. We propose a distribution-free method for constructing biomarker combinations by maximizing the TPR while constraining the FPR. Theoretical results demonstrate desirable properties of biomarker combinations produced by the new method. In simulations, the biomarker combination provided by our method demonstrated improved operating characteristics in a variety of scenarios when compared with alternative methods for constructing biomarker combinations. Thus, use of our method could lead to the development of better biomarker combinations, increasing the likelihood of clinical adoption.

Case-Only Analysis of Gene-Environment Interactions Using Polygenic Risk Scores.

Investigations of gene (G)-environment (E) interactions have led to limited findings to date, possibly due to weak effects of individual genetic variants. Polygenic risk scores (PRS), which capture the genetic susceptibility associated with a set of variants, can be a powerful tool for detecting global patterns of interaction. Motivated by the case-only method for evaluating interactions with a single variant, we propose a case-only method for the analysis of interactions with a PRS in case-control studies. Assuming the PRS and E are independent, we show how a linear regression of the PRS on E in a sample of cases can be used to efficiently estimate the interaction parameter. Furthermore, if an estimate of the mean of the PRS in the underlying population is available, the proposed method can estimate the PRS main effect. Extensions allow for PRS-E dependence due to associations between variants in the PRS and E. Simulation studies indicate the proposed method offers appreciable gains in efficiency over logistic regression and can recover much of the efficiency of a cohort study. We applied the proposed method to investigate interactions between a PRS and epidemiologic factors on breast cancer risk in the UK Biobank (United Kingdom, recruited 2006-2010).

Biomarker combinations for diagnosis and prognosis in multicenter studies: Principles and methods.

Many investigators are interested in combining biomarkers to predict a binary outcome or detect underlying disease. This endeavor is complicated by the fact that many biomarker studies involve data from multiple centers. Depending upon the relationship between center, the biomarkers, and the target of prediction, care must be taken when constructing and evaluating combinations of biomarkers. We introduce a taxonomy to describe the role of center and consider how a biomarker combination should be constructed and evaluated. We show that ignoring center, which is frequently done by clinical researchers, is often not appropriate. The limited statistical literature proposes using random intercept logistic regression models, an approach that we demonstrate is generally inadequate and may be misleading. We instead propose using fixed intercept logistic regression, which appropriately accounts for center without relying on untenable assumptions. After constructing the biomarker combination, we recommend using performance measures that account for the multicenter nature of the data, namely the center-adjusted area under the receiver operating characteristic curve. We apply these methods to data from a multicenter study of acute kidney injury after cardiac surgery. Appropriately accounting for center, both in construction and evaluation, may increase the likelihood of identifying clinically useful biomarker combinations.

Development of biomarker combinations for postoperative acute kidney injury via Bayesian model selection in a multicenter cohort study.

BACKGROUND: Acute kidney injury (AKI) is a frequent complication of cardiac surgery. We sought prognostic combinations of postoperative biomarkers measured within 6 h of surgery, potentially in combination with cardiopulmonary bypass time (to account for the degree of insult to the kidney). We used data from a large cohort of patients and adapted methods for developing biomarker combinations to account for the multicenter design of the study. METHODS: The primary endpoint was sustained mild AKI, defined as an increase of 50% or more in serum creatinine over preoperative levels lasting at least 2 days during the hospital stay. Severe AKI (secondary endpoint) was defined as a serum creatinine increase of 100% or more or dialysis during hospitalization. Data were from a cohort of 1219 adults undergoing cardiac surgery at 6 medical centers; among these, 117 developed sustained mild AKI and 60 developed severe AKI. We considered cardiopulmonary bypass time and 22 biomarkers as candidate predictors. We adapted Bayesian model averaging methods to develop center-adjusted combinations for sustained mild AKI by (1) maximizing the posterior model probability and (2) retaining predictors with posterior variable probabilities above 0.5. We used resampling-based methods to avoid optimistic bias in evaluating the biomarker combinations. RESULTS: The maximum posterior model probability combination included plasma N-terminal-pro-B-type natriuretic peptide, plasma heart-type fatty acid binding protein, and change in serum creatinine from before to 0-6 h after surgery; the median probability combination additionally included plasma interleukin-6. The center-adjusted, optimism-corrected AUCs for these combinations were 0.80 (95% CI: 0.78, 0.87) and 0.81 (0.78, 0.87), respectively, for predicting sustained mild AKI, and 0.81 (0.76, 0.90) and 0.83 (0.76, 0.90), respectively, for predicting severe AKI. For these data, the Bayesian model averaging methods yielded combinations with prognostic capacity comparable to that achieved by standard frequentist methods but with more parsimonious models. CONCLUSIONS: Pending external validation, the identified combinations could be used to identify individuals at high risk of AKI immediately after cardiac surgery and could facilitate clinical trials of renoprotective agents.

Using ordinal outcomes to construct and select biomarker combinations for single-level prediction.

BACKGROUND: Biomarker studies may involve an ordinal outcome, such as no, mild, or severe disease. There is often interest in predicting one particular level of the outcome due to its clinical significance. METHODS: A simple approach to constructing biomarker combinations in this context involves dichotomizing the outcome and using a binary logistic regression model. We assessed whether more sophisticated methods offer advantages over this simple approach. It is often necessary to select among several candidate biomarker combinations. One strategy involves selecting a combination based on its ability to predict the outcome level of interest. We propose an algorithm that leverages the ordinal outcome to inform combination selection. We apply this algorithm to data from a study of acute kidney injury after cardiac surgery, where kidney injury may be absent, mild, or severe. RESULTS: Using more sophisticated modeling approaches to construct combinations provided gains over the simple binary logistic regression approach in specific settings. In the examples considered, the proposed algorithm for combination selection tended to reduce the impact of bias due to selection and to provide combinations with improved performance. CONCLUSIONS: Methods that utilize the ordinal nature of the outcome in the construction and/or selection of biomarker combinations have the potential to yield better combinations.

Urinalysis findings and urinary kidney injury biomarker concentrations.

INTRODUCTION: Urinary biomarkers of kidney injury are presumed to reflect renal tubular damage. However, their concentrations may be influenced by other factors, such as hematuria or pyuria. We sought to examine what non-injury related urinalysis factors are associated with urinary biomarker levels. METHODS: We examined 714 adults who underwent cardiac surgery in the TRIBE-AKI cohort that did not experience post-operative clinical AKI (patients with serum creatinine change of ≥ 20% were excluded). We examined the association between urinalysis findings and the pre- and first post-operative urinary concentrations of 4 urinary biomarkers: neutrophil gelatinase-associated lipocalin (NGAL), interleukin-18 (IL-18), kidney injury molecule-1 (KIM-1), and liver fatty acid binding protein (L-FABP). RESULTS: The presence of leukocyte esterase and nitrites on urinalysis was associated with increased urinary NGAL (R2 0.16, p < 0.001 and R2 0.07, p < 0.001, respectively) in pre-operative samples. Hematuria was associated with increased levels of all 4 biomarkers, with a much stronger association seen in post-operative samples (R2 between 0.02 and 0.21). Dipstick proteinuria concentrations correlated with levels of all 4 urinary biomarkers in pre-operative and post-operative samples (R2 between 0.113 and 0.194 in pre-operative and between 0.122 and 0.322 in post-operative samples). Adjusting the AUC of post-operative AKI for dipstick proteinuria lowered the AUC for all 4 biomarkers at the pre-operative time point and for 2 of the 4 biomarkers at the post-operative time point. CONCLUSIONS: Several factors available through urine dipstick testing are associated with increased urinary biomarker concentrations that are independent of clinical kidney injury. Future studies should explore the impact of these factors on the prognostic and diagnostic performance of these AKI biomarkers.

Pre-diagnostic circulating sex hormone levels and risk of prostate cancer by ERG tumour protein expression.

BACKGROUND: Experimental studies have shown androgen receptor stimulation to facilitate formation of the TMPRSS2:ERG gene fusion in prostate cell lines. No study has tested whether higher pre-diagnostic circulating sex hormone levels in men increase risk of developing TMPRSS2:ERG-positive prostate cancer specifically. METHODS: We conducted a nested case-control study of 200 prostate cancer cases and 1057 controls from the Physicians' Health Study and Health Professionals Follow-up Study. We examined associations between pre-diagnostic circulating levels of total testosterone, free testosterone, DHT, androstanediol glucuronide, estradiol, and SHBG and risk of prostate cancer by TMPRSS2:ERG status. TMPRSS2:ERG was estimated by ERG immunohistochemistry. We used multivariable unconditional polytomous logistic regression to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for risk of ERG-positive (n=94) and, separately, ERG-negative (n=106) disease. RESULTS: Free testosterone was significantly associated with the risk of ERG-positive prostate cancer (OR: 1.37, 95% CI: 1.05-1.77), but not ERG-negative prostate cancer (OR: 1.09, 95% CI: 0.86-1.38) (Pdiff=0.17). None of the remaining hormones evaluated showed clear differential associations with ERG-positive vs ERG-negative disease. CONCLUSIONS: These findings provide some suggestive evidence that higher pre-diagnostic free testosterone levels are associated with an increased risk of developing TMPRSS2:ERG-positive prostate cancer.

Methodological issues in current practice may lead to bias in the development of biomarker combinations for predicting acute kidney injury.

Individual biomarkers of renal injury are only modestly predictive of acute kidney injury (AKI). Using multiple biomarkers has the potential to improve predictive capacity. In this systematic review, statistical methods of articles developing biomarker combinations to predict AKI were assessed. We identified and described three potential sources of bias (resubstitution bias, model selection bias, and bias due to center differences) that may compromise the development of biomarker combinations. Fifteen studies reported developing kidney injury biomarker combinations for the prediction of AKI after cardiac surgery (8 articles), in the intensive care unit (4 articles), or other settings (3 articles). All studies were susceptible to at least one source of bias and did not account for or acknowledge the bias. Inadequate reporting often hindered our assessment of the articles. We then evaluated, when possible (7 articles), the performance of published biomarker combinations in the TRIBE-AKI cardiac surgery cohort. Predictive performance was markedly attenuated in six out of seven cases. Thus, deficiencies in analysis and reporting are avoidable, and care should be taken to provide accurate estimates of risk prediction model performance. Hence, rigorous design, analysis, and reporting of biomarker combination studies are essential to realizing the promise of biomarkers in clinical practice.

Validity of a Neurological Scoring System for Canine X-Linked Myotubular Myopathy.

A simple clinical neurological test was developed to evaluate response to gene therapy in a preclinical canine model of X-linked myotubular myopathy (XLMTM). This devastating congenital myopathy is caused by mutation in the myotubularin (MTM1) gene. Clinical signs include muscle weakness, early respiratory failure, and ventilator dependence. A spontaneously occurring canine model has a similar clinical picture and histological abnormalities on muscle biopsy compared with patients. We developed a neuromuscular assessment score, graded on a scale from 10 (normal) to 1 (unable to maintain sternal recumbency). We hypothesize that this neurological assessment score correlates with genotype and established measures of disease severity and is reliable when performed by an independent observer. At 17 weeks of age, there was strong correlation between neurological assessment scores and established methods of severity testing. The neurological severity score correctly differentiated between XLMTM and wild-type dogs with good interobserver reliability, on the basis of strong agreement between neurological scores assigned by independent observers. Together, these data indicate that the neurological scoring system developed for this canine congenital neuromuscular disorder is reliable and valid. This scoring system may be helpful in evaluating response to therapy in preclinical testing in this disease model, such as response to gene therapy.

Androgen receptor CAG repeat polymorphism and risk of TMPRSS2:ERG-positive prostate cancer.

BACKGROUND: The androgen receptor (AR) is an essential gene in prostate cancer pathogenesis and progression. Genetic variation in AR exists, including a polymorphic CAG repeat sequence that is inversely associated with transcriptional activity. Experimental data suggest that heightened AR activity facilitates formation of TMPRSS2:ERG, a gene fusion present in approximately 50% of tumors of patients with prostate cancer. METHODS: We undertook a nested case-control study to investigate the hypothesis that shorter CAG repeat length would be associated with prostate cancer risk defined by TMPRSS2:ERG status. The study included 291 men with prostate cancer (147 ERG-positive) and 1,221 cancer-free controls. ORs and 95% confidence intervals (CI) were calculated using logistic regression. RESULTS: Median CAG repeat length (interquartile range) among controls was 22 (20-24). Men with shorter CAG repeats had an increased risk of ERG-positive (OR, 1.07 per 1 repeat decrease; 95% CI, 1.00-1.14), but not ERG-negative prostate cancer (OR, 0.99 per 1 repeat decrease; 95% CI, 0.93-1.05). CONCLUSIONS: These data suggest that shorter CAG repeats are specifically associated with development of TMPRSS2:ERG-positive prostate cancer. IMPACT: Our results provide supportive evidence that androgen signaling underlies the development of prostate tumors that harbor TMPRSS2:ERG. Moreover, these results suggest that TMPRSS2:ERG may represent a unique molecular subtype of prostate cancer with an etiology distinct from TMPRSS2:ERG-negative disease.

Modification of the association between obesity and lethal prostate cancer by TMPRSS2:ERG.

BACKGROUND: TMPRSS2:ERG is a hormonally regulated gene fusion present in about half of prostate tumors. We investigated whether obesity, which deregulates several hormonal pathways, interacts with TMPRSS2:ERG to impact prostate cancer outcomes. METHODS: The study included 1243 participants in the prospective Physicians' Health Study and Health Professionals Follow-Up Study diagnosed with prostate cancer between 1982 and 2005. ERG overexpression (a TMPRSS2:ERG marker) was assessed by immunohistochemistry of tumor tissue from radical prostatectomy or transurethral resection of the prostate. Body mass index (BMI) and waist circumference, measured on average 1.3 years and 5.3 years before diagnosis, respectively, were available from questionnaires. Data on BMI at baseline was also available. We used Cox regression to calculate hazard ratios and 95% confidence intervals (CIs). All statistical tests were two-sided. RESULTS: During a mean follow-up of 12.8 years, 119 men developed lethal disease (distant metastases or prostate cancer death). Among men with ERG-positive tumors, the multivariable hazard ratio for lethal prostate cancer was 1.48 (95% CI = 0.98 to 2.23) per 5-unit increase in BMI before diagnosis, 2.51 (95% CI = 1.26 to 4.99) per 8-inch increase in waist circumference before diagnosis, and 2.22 (95% CI = 1.35 to 3.63) per 5-unit increase in BMI at baseline. The corresponding hazard ratios among men with ERG-negative tumors were 1.10 (95% CI = 0.76 to1.59; P interaction = .24), 1.14 (95% CI = 0.62 to 2.10; P interaction = .09), and 0.78 (95% CI = 0.52 to 1.19; P interaction = .001). CONCLUSIONS: These results suggest that obesity is linked with poorer prostate cancer prognosis primarily in men with tumors harboring the gene fusion TMPRSS2:ERG.

A single nucleotide polymorphism in inflammatory gene RNASEL predicts outcome after radiation therapy for localized prostate cancer.

PURPOSE: To study associations between single nucleotide polymorphisms (SNP) in Ribonuclease L (RNASEL), a gene implicated in inflammation and prostate cancer risk, and outcomes after radiation therapy. EXPERIMENTAL DESIGN: We followed participants in the prospective US Health Professionals Follow-Up Study treated with radiation therapy for early-stage prostate cancer. Three SNPs were genotyped based on previously determined functional and biological significance. We used multivariable Cox proportional hazards models to assess per-allele associations with the primary outcome defined as time to a composite endpoint including development of lethal prostate cancer or biochemical recurrence. RESULTS: We followed 434 patients treated with radiation therapy for a median of 9 years. On multivariate analysis, the rs12757998 variant allele was associated with significantly decreased risk of the composite endpoint [HR: 0.65; 95% confidence interval (CI), 0.45-0.94%; P = 0.02] driven by decreased biochemical recurrence (HR: 0.60; 95% CI, 0.40-0.89%; P = 0.01) and men treated with external beam (HR: 0.58; 95% CI, 0.36-0.93%; P = 0.02). In contrast, in 516 men from the same cohort treated with radical prostatectomy, we found no significant impact of this variant on outcome. Furthermore, the rs12757998 variant allele significantly modified the association between androgen deprivation therapy and outcomes after radiation therapy (p-interaction = 0.02). CONCLUSION: We show an association between RNASEL SNP rs12757998 and outcome after radiation therapy for prostate cancer. This SNP is associated with increased circulating C-reactive protein and interleukin-6, suggesting a potential role for inflammation in the response to radiation. If validated, genetic predictors of outcome may help inform prostate cancer management.

CD4 helper T cells, CD8 cytotoxic T cells, and FOXP3(+) regulatory T cells with respect to lethal prostate cancer.

Prostate cancer represents a major contributor to cancer mortality, but the majority of men with prostate cancer will die of other causes. Thus, a challenge is identifying potentially lethal disease at diagnosis. Conflicting results have been reported when investigating the relationship between infiltration of lymphocytes and survival in prostate cancer. One of the mechanisms suggested is the recruitment of regulatory T cells (T(regs)), a subpopulation of T cells that have a role in promoting tumor growth. T(regs) counteract tumor rejection through suppressive functions on the anti-immune response but their prognostic significance is still unknown. We report here the results of a conducted case-control study nested in a cohort of men treated with transurethral resection of the prostate and diagnosed incidentally with prostate cancer. Cases are men who died of prostate cancer (n=261) and controls are men who survived >10 years after their diagnosis (n=474). Infiltration of both T(helper) and T(cytotoxic) cells was frequently observed and the majority of the T(regs) were CD4(+). T(helper) or T(cytotoxic) cells were not associated with lethal prostate cancer. However, we found a nearly twofold increased risk of lethal prostate cancer when comparing the highest with the lowest quartile of CD4(+) T(regs) cells (95% confidence interval: 1.3-2.9). Our conclusion is that men with greater numbers of CD4(+) T(regs) in their prostate tumor environment have an increased risk of dying of prostate cancer. Identification of CD4(+) T(regs) in tumor tissue may predict clinically relevant disease at time of diagnosis independently of other clinical factors.

Associations of carotid artery intima-media thickness (IMT) with risk factors and prevalent cardiovascular disease: comparison of mean common carotid artery IMT with maximum internal carotid artery IMT.

OBJECTIVE: The goal of this study was to compare internal carotid artery (ICA) intima-media thickness (IMT) with common carotid artery (CCA) IMT as global markers of cardiovascular disease (CVD). METHODS: Cross-sectional measurements of the mean CCA IMT and maximum ICA IMT were made on ultrasound images acquired from the Framingham Offspring cohort (n = 3316; mean age, 58 years; 52.7% women). Linear regression models were used to study the associations of the Framingham risk factors with CCA and ICA IMT. Multivariate logistic regression models and receiver operating characteristic (ROC) curve analysis were used to compare the associations of prevalent CVD with CCA and ICA IMT and determine sensitivity and specificity. RESULTS: The association between age and the mean CCA IMT corresponded to an increase of 0.007 mm/y; the increase was 0.037 mm/y for the ICA IMT. Framingham risk factors accounted for 28.6% and 27.5% of the variability in the CCA and ICA IMT, respectively. Age and gender contributed 23.5% to the variability of the CCA IMT and 22.5% to that of the ICA IMT, with the next most important factor being systolic blood pressure (1.9%) for the CCA IMT and smoking (1.6%) for the ICA IMT. The CCA IMT and ICA IMT were statistically significant predictors of prevalent CVD, with the ICA IMT having a larger area under the ROC curve (0.756 versus 0.695). CONCLUSIONS: Associations of risk factors with CCA and ICA IMT are slightly different, and both are independently associated with prevalent CVD. Their value for predicting incident cardiovascular events needs to be compared in outcome studies.

Apolipoprotein E/C1 locus variants modify renal cell carcinoma risk.

Lipid peroxidation is considered a unifying mechanistic pathway through which known risk factors induce renal cell carcinoma (RCC). We hypothesized that genes selected a priori for their role in lipid peroxidation would modify cancer risk. We genotyped 635 single nucleotide polymorphisms (SNP) in 38candidate genes in 777 Caucasian RCC cases and 1,035 controls enrolled in a large European case-control study. Top candidate SNPs were confirmed among 718 Caucasian cases and 615 controls in a second study in the United States. Two of the three SNPs (rs8106822 and rs405509) that replicated in the U.S. study were within a regulatory region of the APOE promoter. The OR for rs8106822 A>G variant was 1.22(AG) and 1.41(GG) (P(trend) = 0.01) in the European study, 1.05(AG) and 1.51(GG) (P(trend) = 0.03) in the U.S. study, and 1.15(AG) and 1.44(GG) (P(trend) = 0.001) among 1,485 cases and 1,639 controls combined. The rs405509 G>T variant was associated with risk in the European (OR, 0.87(TG); OR, 0.71(TT); P(trend) = 0.02), the U.S. (OR, 0.68(TG); OR, 0.71(TT); P(trend) = 0.02), and both studies combined (OR(TG), 0.79; OR(TT), 0.71; P(trend) = 0.001), as was the G-G haplotype (r(2) = 0.64; P= 4.7 x 10(-4)). This association is biologically plausible as SNP rs405509 was shown to modify protein binding and transcriptional activity of the APOE protein in vitro and is in linkage disequilibrium with key known variants defining the e2, e3, and e4 alleles that modify risk of atherosclerosis, Alzheimer's disease risk, and progression to AIDS. In two large case-control studies, our findings further define a functional region of interest at the APOE locus that increases RCC susceptibility.