Patient-specific meta-analysis of 12-gene colon cancer recurrence score validation studies for recurrence risk assessment after surgery with or without 5FU and oxaliplatin.

Background: Individualized estimates of the risk of recurrence in colon cancer patients are needed that reflect current medical practice and available treatment options. Methods: Three validation studies of the 12-gene colon recurrence score assay were used with pre-specified patient-specific meta-analysis (PSMA) methods to integrate the 12-gene Oncotype DX Colon Recurrence Score result (RS) with the clinical and pathology risk factors stage, T-stage, mis-match repair (MMR) status, and number of nodes examined to calculate individualized recurrence risk estimates. Baseline risk estimation used the most recent studies, so the risk estimates reflect current medical practice. The effect of fluorouracil (5FU) was estimated with a meta-analysis of two studies. The effect of oxaliplatin was estimated using one of the RS assay validation studies, in which patients were randomized to 5FU with or without oxaliplatin. Results: The RS result and each of the clinical-pathologic factors provided independent prognostic information for recurrence. Among stage II, T3, MMR-proficient patients with ≥12 nodes examined (the most common scenario), patients with RS ≤30 (approximately 48%) have estimated 5-year recurrence risk ≤10% with surgery alone. Among stage IIIA/B, T3, MMR-deficient patients with ≥12 nodes examined, patients with RS ≤19 (approximately 14%) have an estimated 5-year recurrence risk ≤10% with surgery alone. Among stage IIIA/B, T3, MMR-proficient patients with ≥12 nodes examined, those with RS ≤14 (approximately 6%) have estimated 5-year recurrence risk ≤10% with 5FU alone. Discussion: The PSMA integrates the 12-gene colon RS result with clinical and pathology factors to provide individualized recurrence risk estimates that reflect current medical practice. The risk estimates are in a range that may help inform treatment decisions for a substantial number of stage II and stage III patients.

Validating the association of adverse pathology with distant metastasis and prostate cancer mortality 20-years after radical prostatectomy.

PURPOSE: To assess the association of adverse pathology (AP), defined as high-grade (≥ Gleason Grade Group 3) and/or non-organ confined disease, with long-term oncologic outcomes after radical prostatectomy (RP). MATERIALS AND METHODS: Using a stratified cohort sampling design, we evaluated the association of AP with the risk of distant metastasis (DM) and prostate cancer-specific mortality (PCSM) up to 20 years after RP in 428 patients treated between 1987 to 2004. Cox regression of cause-specific hazards was used to estimate the absolute risk of both endpoints, with death from other causes treated as a competing risk. Additionally, subgroup analysis in patients with low and/or intermediate-risk disease, who are potentially eligible for active surveillance (AS), was performed. RESULTS: Within the cohort sample, 53% of men exhibited AP at time of RP, with median follow up of 15.5 years (IQR 14.6-16.6 years) thereafter. Adverse pathology was highly associated with DM and PCSM in the overall cohort (HR 12.30, 95% confidence interval [CI] 5.30-28.55, and HR 10.03, 95% CI 3.42-29.47, respectively, both P < 0.001). Adverse pathology was also highly associated with DM and PCSM in the low/intermediate-risk subgroup (HR 10.48, 95% CI 4.18-26.28, and 8.60, 95% CI 2.40-30.48, respectively, both P < 0.001). CONCLUSIONS: Adverse pathology at the time of RP is highly associated with future development of DM and PCSM. Accurate prediction of AP may thus be useful for individualizing risk-based surveillance and treatment strategies.

GPS Assay Association With Long-Term Cancer Outcomes: Twenty-Year Risk of Distant Metastasis and Prostate Cancer-Specific Mortality.

PURPOSE: To assess the association between the Oncotype DX Genomic Prostate Score (GPS) result and long-term oncological outcomes following radical prostatectomy (RP). METHODS: We evaluated the association of the GPS result assayed from the index lesion from RP tissue with the risk of distant metastases (DM) and prostate cancer-specific mortality (PCSM) over the 20 years following RP in a stratified cohort sample of 428 patients from 2,641 treated between 1987 and 2004. Cox regression of cause-specific hazards was used to estimate the absolute risk of both end points, with death from other causes treated as a competing risk. A correction for regression to the mean (RM) was applied since the GPS test was developed using this cohort. Exploratory analysis using presurgical parameters and the GPS test as prognostic variables was performed to assess the additional value of the GPS test on 20-year risk of DM and PCSM. Model discrimination was measured using the area under the receiver operating characteristic curve. RESULTS: The GPS test appears to be independently associated with both 20-year risk of DM and PCSM with a low false discovery rate. Per 20-unit increase in GPS, multivariable analysis with RM correction estimated hazard ratios of 2.24 (95% CI, 1.49 to 3.53) and 2.30 (95% CI, 1.45 to 4.36) for DM and PCSM, respectively. Accuracy of models including clinical risk factors alone appeared to improve when including the GPS test in assessing risk of both end points. CONCLUSION: The results suggest that the GPS test provides information on the risk for the meaningful long-term outcomes of DM and PCSM.

Development and Validation of a Tool Integrating the 21-Gene Recurrence Score and Clinical-Pathological Features to Individualize Prognosis and Prediction of Chemotherapy Benefit in Early Breast Cancer.

PURPOSE: The 21-gene recurrence score (RS) is prognostic for distant recurrence (DR) and predictive for chemotherapy benefit in early breast cancer, whereas clinical-pathological factors are only prognostic. Integration of genomic and clinical features offers the potential to guide adjuvant chemotherapy use with greater precision. METHODS: We developed a new tool (RSClin) that integrates RS with tumor grade, tumor size, and age using a patient-specific meta-analysis including 10,004 women with hormone receptor-positive, human epidermal growth factor receptor 2-negative, and node-negative breast cancer who received endocrine therapy alone in the B-14 (n = 577) and TAILORx (n = 4,854) trials or plus chemotherapy in TAILORx (n = 4,573). Cox models for RSClin were compared with RS alone and clinical-pathological features alone using likelihood ratio tests. RSClin estimates of DR used a baseline risk with TAILORx event rates to reflect current medical practice. A patient-specific estimator of absolute chemotherapy benefit was computed using individualized relative chemotherapy effect from the randomized TAILORx and B-20 trials. External validation of risk estimation was performed by comparing RSClin estimated risk and observed risk in 1,098 women in the Clalit registry. RESULTS: RSClin provides more prognostic information (likelihood ratio χ2) for DR than RS or clinical-pathological factors alone (both P < .001, likelihood ratio test). In external validation, the RSClin risk estimate was prognostic for DR risk in the Clalit registry (P < .001) and the estimated risk closely approximated the observed 10-year risk (Lin concordance 0.962). The absolute chemotherapy benefit estimate ranges from 0% to 15% as the RS ranges from 11 to 50 using RSClin in a 55-year-old woman with a 1.5-cm intermediate-grade tumor. CONCLUSION: The RSClin tool integrates clinical-pathological and genomic risk to guide adjuvant chemotherapy in node-negative breast cancer and provides more individualized information than clinical-pathological or genomic data alone.

Extensions of the absolute standardized hazard ratio and connections with measures of explained variation and variable importance.

The absolute standardized hazard ratio (ASHR) is a scale-invariant scalar measure of the strength of association of a vector of covariates with the risk of an event. It is derived from proportional hazards regression. The ASHR is useful for making comparisons among different sets of covariates. Extensions of the ASHR concept and practical considerations regarding its computation are discussed. These include a new method to conduct preliminary checks for collinearity among covariates, a partial ASHR to evaluate the association with event risk of some of the covariates conditioning on others, and the ASHR for interactions. To put the ASHR in context, its relationship to measures of explained variation and other measures of separation of risk is discussed. A new measure of the contribution of each covariate to the risk score variance is proposed. This measure, which is derived from the ASHR calculations, is interpretable as variable importance within the context of the multivariable model.

Adjusting for covariates in evaluating markers for selecting treatment, with application to guiding chemotherapy for treating estrogen-receptor-positive, node-positive breast cancer.

In many clinical contexts, biomarkers that predict treatment efficacy are highly sought after. Such treatment selection or predictive biomarkers have the potential to identify subgroups most likely to benefit from the treatment, and may therefore be used to improve clinical outcomes and reduce medical costs. A methodological challenge in evaluating these biomarkers is determining how to take into account other variables that predict clinical outcomes, or that influence the biomarker distribution, generically termed covariates. We distinguish between two questions that arise when evaluating markers in the context of covariates. First, what is the biomarker's added value for selecting treatment, over and above the covariates? Second, what is the marker's performance within covariate strata-does performance vary? We lay out statistical methodology for addressing each of these questions. We argue that the common approach of testing for the marker's statistical interaction with treatment, in the context of a multivariate model that includes the covariates as predictors, does not directly address either question. We illustrate the methodology in new analyses of the Oncotype DX Recurrence Score, a marker used to select adjuvant chemotherapy for the treatment of estrogen-receptor-positive breast cancer.

Patient-specific Meta-analysis of 2 Clinical Validation Studies to Predict Pathologic Outcomes in Prostate Cancer Using the 17-Gene Genomic Prostate Score.

OBJECTIVE: To perform patient-specific meta-analysis (MA) of two independent clinical validation studies of a 17-gene biopsy-based genomic assay as a predictor of favorable pathology at radical prostatectomy. MATERIALS AND METHODS: Patient-specific MA was performed on data from 2 studies (732 patients) using the Genomic Prostate Score (GPS; scale 0-100) together with Cancer of the Prostate Risk Assessment (CAPRA) score or National Comprehensive Cancer Network (NCCN) risk group as predictors of the likelihood of favorable pathology (LFP). Risk profile curves associating GPS with LFP by CAPRA score and NCCN risk group were generated. Decision curves and receiver operating characteristic curves were calculated using patient-specific MA risk estimates. RESULTS: Patient-specific MA-generated risk profiles ensure more precise estimates of LFP with narrower confidence intervals than either study alone. GPS added significant predictive value to each clinical classifier. A model utilizing GPS and CAPRA provided the most risk discrimination. In decision-curve analysis, greater net benefit was shown when combining GPS with each clinical classifier compared with the classifier alone. The area under the receiver operating characteristic curve improved from 0.68 to 0.73 by adding GPS to CAPRA, and 0.64 to 0.70 by adding GPS to NCCN risk group. The proportion of patients with LFP >80% increased from 11% using NCCN risk group alone to 23% using GPS with NCCN. Using GPS with CAPRA identified the highest proportion-31%-of patients with LFP >80%. CONCLUSION: Patient-specific MA provides more precise risk estimates that reflect the complete body of evidence. GPS adds predictive value to 3 widely used clinical classifiers, and identifies a larger proportion of low-risk patients than identified by clinical risk group alone.

Separate class true discovery rate degree of association sets for biomarker identification.

In 2008, Efron showed that biological features in a high-dimensional study can be divided into classes and a separate false discovery rate (FDR) analysis can be conducted in each class using information from the entire set of features to assess the FDR within each class. We apply this separate class approach to true discovery rate degree of association (TDRDA) set analysis, which is used in clinical-genomic studies to identify sets of biomarkers having strong association with clinical outcome or state while controlling the FDR. Careful choice of classes based on prior information can increase the identification power of the separate class analysis relative to the overall analysis.

Patient-specific meta-analysis for risk assessment using multivariate proportional hazards regression.

We propose a method for assessing an individual patient's risk of a future clinical event using clinical trial or cohort data and Cox proportional hazards regression, combining the information from several studies using meta-analysis techniques. The method combines patient-specific estimates of the log cumulative hazard across studies, weighting by the relative precision of the estimates, using either fixed- or random-effects meta-analysis calculations. Risk assessment can be done for any future patient using a few key summary statistics determined once and for all from each study. Generalizations of the method to logistic regression and linear models are immediate. We evaluate the methods using simulation studies and illustrate their application using real data.

Prospective calculation of identification power for individual genes in analyses controlling the false discovery rate.

Recent work on prospective power and sample size calculations for analyses of high-dimension gene expression data that control the false discovery rate (FDR) focuses on the average power over all the truly nonnull hypotheses, or equivalently, the expected proportion of nonnull hypotheses rejected. Using another characterization of power, we adapt Efron's ([2007] Ann Stat 35:1351-1377) empirical Bayes approach to post hoc power calculation to develop a method for prospective calculation of the "identification power" for individual genes. This is the probability that a gene with a given true degree of association with clinical outcome or state will be included in a set within which the FDR is controlled at a specified level. An example calculation using proportional hazards regression highlights the effects of large numbers of genes with little or no association on the identification power for individual genes with substantial association.

Gene identification using true discovery rate degree of association sets and estimates corrected for regression to the mean.

Analyses intended to identify genes with expression that is associated with some clinical outcome or state are often based on ranked p-values from tests of point null hypotheses of no association. Van de Wiel and Kim take the innovative approach of testing the interval null hypotheses that the degree of association for a gene is less than some value of interest against the alternative that it is greater. Combining this idea with the false discovery rate controlling methods of Storey, Taylor and Siegmund gives a computationally simple way to identify true discovery rate degree of association (TDRDA) sets of genes among which a specified proportion are expected to have an absolute association of a specified degree or more. This leads to a gene ranking method that uses the maximum lower bound degree of association for which each gene belongs to a TDRDA set. Estimates of each gene's actual degree of association with approximate correction for 'selection bias' due to regression to the mean (RM) can be derived using simple bivariate normal theory and Efron and Tibshirani's empirical Bayes approach. For a given data set, all possible TDRDA sets can be displayed along with the gene ranking and the RM-corrected estimates of degree of association in a concise graphical summary.

Long-term safety of a novel antianginal agent in patients with severe chronic stable angina: the Ranolazine Open Label Experience (ROLE).

OBJECTIVES: This report describes safety and tolerability data from 746 chronic angina patients treated in the ROLE (Ranolazine Open Label Experience) program. BACKGROUND: Ranolazine treats angina without depressing hemodynamic status. The long-term safety and tolerability of ranolazine have not been previously reported. METHODS: Patients with severe functional impairment from angina (mean Duke Treadmill Score [DTS] of -14.4) who completed 1 of 2 randomized treadmill trials entered the ROLE program. Ranolazine was titrated to optimal dosages between 500 and 1,000 mg twice daily. Physical examination, laboratory tests, and adverse event reporting were performed periodically. We conducted analyses to evaluate possible predictors of ranolazine intolerance, such as advanced age, diabetes, poor exercise tolerance, or history of myocardial infarctions or congestive heart failure (CHF). The ROLE program's mortality was compared against the DTS predictive model and other contemporary cohorts of high-risk CHD patients. RESULTS: Mean follow-up was 2.82 years. Two years after initial dosing, 571 patients (76.7%) remained on therapy and 72 patients (9.7%) discontinued ranolazine due to adverse events. Among 6 factors evaluated, only age > or =64 years predicted for higher withdrawal rates. Patients with a history of CHF had lower withdrawal rates. Mean QTc interval was prolonged by 2.4 ms. No treatment discontinuations occurred due to QTc prolongation, and no Torsades de Pointes was reported. Sixty-four deaths occurred during a total of 2,102 patient-years (3.0% annually) during the ROLE program. When extending observations to all patients exposed to ranolazine during the double-blind trials (n = 972) preceding the ROLE program, annual mortality was 2.8% compared with >5% as predicted by DTS. CONCLUSIONS: Long-term therapy with ranolazine seems well tolerated in high-risk CHD patients. Survival analyses suggest that symptomatic improvements attributable to ranolazine are not offset by increased mortality.