Search

Identifying subpopulations that may be sensitive to the specific treatment is an important step toward precision medicine. On the other hand, longitudinal data with dropouts is common in medical research, and subgroup analysis for this data type is still limited. In this paper, we consider a single-index threshold linear marginal model, which can be used simultaneously to identify subgroups with differential treatment effects based on linear combination of the selected biomarkers, estimate the treatment effects in different subgroups based on regression coefficients, and test the significance of the difference in treatment effects based on treatment-subgroup interaction. The regression parameters are estimated by solving a penalized smoothed generalized estimating equation and the selection bias caused by missingness is corrected by a multiply robust weighting matrix, which allows multiple missingness models to be taken account into estimation. The proposed estimator remains consistent when any model for missingness is correctly specified. Under regularity conditions, the asymptotic normality of the estimator is established. Simulation studies confirm the desirable finite-sample performance of the proposed method. As an application, we analyze the data from a clinical trial on pancreatic cancer.

In a joint analysis of longitudinal quality of life (QoL) scores and relapse-free survival (RFS) times from a clinical trial on early breast cancer conducted by the Canadian Cancer Trials Group, we observed a complicated trajectory of QoL scores and existence of long-term survivors. Motivated by this observation, we proposed in this paper a flexible joint model for the longitudinal measurements and survival times. A partly linear mixed effect model is used to capture the complicated but smooth trajectory of longitudinal measurements and approximated by B-splines and a semiparametric mixture cure model with the B-spline baseline hazard to model survival times with a cure fraction. These two models are linked by shared random effects to explore the dependence between longitudinal measurements and survival times. A semiparametric inference procedure with an EM algorithm is proposed to estimate the parameters in the joint model. The performance of proposed procedures are evaluated by simulation studies and through the application to the analysis of data from the clinical trial which motivated this research.

Background Two germ line Fc‐γ receptor (FCGR) polymorphisms, rs1801274 [FCGR2A; His(H)131Arg(R)] and rs396991 [FCGR3A; Phe(F)158Val(V)], produce altered proteins through amino acid substitutions. We previously reported that the FCGR2A H/H genotype was associated with longer overall survival (OS) in cetuximab‐treated chemotherapy‐refractory patients with metastatic colorectal cancer. Here, we aimed to replicate and extend this finding in the Canadian Clinical Trials Group CO.20 trial. Methods After germ line DNA genotyping, polymorphic relationships with survival were assessed using log‐rank tests and hazard ratios (HR) from Cox proportional hazard models, adjusting for known prognostic factors. The dominant genetic inheritance model was used for the main analysis. Results Of 592 wild‐type KRAS patients treated with cetuximab, those with the FCGR2A H/H genotype (n = 165, 28%) had improved OS (HR: 0.66, P

In this paper, we introduce a single-index threshold Cox proportional hazard model to select and combine biomarkers to identify patients who may be sensitive to a specific treatment. A penalized smoothed partial likelihood is proposed to estimate the parameters in the model. A simple, efficient, and unified algorithm is presented to maximize this likelihood function. The estimators based on this likelihood function are shown to be consistent and asymptotically normal. Under mild conditions, the proposed estimators also achieve the oracle property. The proposed approach is evaluated through simulation analyses and application to the analysis of data from two clinical trials, one involving patients with locally advanced or metastatic pancreatic cancer and one involving patients with resectable lung cancer.

Motivated by the analysis of quality of life data from a clinical trial on early breast cancer, we propose in this paper a generalized partially linear mean-covariance regression model for longitudinal proportional data, which are bounded in a closed interval. Cholesky decomposition of the covariance matrix for within-subject responses and generalized estimation equations are used to estimate unknown parameters and the nonlinear function in the model. Simulation studies are performed to evaluate the performance of the proposed estimation procedures. Our new model is also applied to analyze the data from the cancer clinical trial that motivated this research. In comparison with available models in the literature, the proposed model does not require specific parametric assumptions on the density function of the longitudinal responses and the probability function of the boundary values and can capture dynamic changes of time or other interested variables on both mean and covariance of the correlated proportional responses. Copyright © 2017 John Wiley & Sons, Ltd.

BACKGROUND Targeting the phosphatidylinositol 3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR) pathway is of increasing interest as a therapeutic strategy in many tumors. The aim of this study was to identify molecular markers associated with mTOR inhibitor activity in women with metastatic endometrial cancer. METHODS Archival tumor samples were collected from 94 women with recurrent or metastatic endometrial cancer who participated in 3 National Cancer Insitute of Canada Clinical Trials Group phase 2 trials investigating single-agent mTOR inhibitors: IND160A and IND160B (temsirolimus) and IND192 (ridaforolimus). Analyses included mutational profiling using the OncoCarta Panel version 1.0 and immunohistochemical expression of the tumor suppressor gene PTEN (phosphatase and tensin homologue) and stathmin, a marker of PI3K activation. Associations between biomarker results and clinical outcomes were assessed. RESULTS Mutations were found in 32 of 73 analyzed tumors, PIK3CA (21 patients) was the most common mutated gene. Co-mutations were seen in 8 tumors, most frequently KRAS and PIK3CA (4 cases). PTEN loss was observed in 46 of 85 samples analyzed and increased stathmin expression was observed in 15 of 65 analyzed samples. No correlation was observed between biomarkers and response or progression. In patients taking concurrent metformin, there was a trend toward lower progression, of 11.8% versus 32.5% (P = .14). CONCLUSIONS No predictive biomarker or combination of biomarkers for mTOR inhibitor activity were identified in this study. Restriction and enrichment of study entry, especially based on archival tumor tissue, should be undertaken with caution in trials using these agents. Cancer 2014;120:603–610. © 2013 American Cancer Society.

Survival data with a sizable cure fraction are commonly encountered in cancer research. The semiparametric proportional hazards cure model has been recently used to analyze such data. As seen in the analysis of data from a breast cancer study, a variable selection approach is needed to identify important factors in predicting the cure status and risk of breast cancer recurrence. However, no specific variable selection method for the cure model is available. In this paper, we present a variable selection approach with penalized likelihood for the cure model. The estimation can be implemented easily by combining the computational methods for penalized logistic regression and the penalized Cox proportional hazards models with the expectation-maximization algorithm. We illustrate the proposed approach on data from a breast cancer study. We conducted Monte Carlo simulations to evaluate the performance of the proposed method. We used and compared different penalty functions in the simulation studies.

The joint analysis of longitudinal measurements and survival data is useful in clinical trials and other medical studies. In this paper, we consider a joint model which assumes a linear mixed model for longitudinal measurements and a promotion time cure model for survival data and links these two models through a latent variable. A semiparametric inference procedure with an EM algorithm implementation is developed for the parameters in the joint model. The proposed procedure is evaluated in a simulation study and applied to analyze the quality of life and time to recurrence data from a clinical trial on women with early breast cancer. The Canadian Journal of Statistics 40: 207–224; 2012 © 2012 Statistical Society of Canada L'analyse conjointe des mesures longitudinales et de temps de survie est pratique dans les essais cliniques et autres etudes medicales. Dans cet article, nous considerons un modele conjoint qui utilise un modele lineaire mixte (tt) pour les mesures longitudinales et un modele de guerison a temps de promotion pour les donnees de survie. Un lien entre ces deux modeles est etabli a l'aide d'une variable latente. Nous developpons une procedure d'inference semi-parametrique avec une implantation de l'algorithme EM pour les parametres du modele conjoint. A l'aide d'une etude de simulation, nous evaluons cette procedure et nous l'utilisons pour analyser des donnees sur la qualite de vie et le temps de recurrence provenant d'un essai clinique sur des femmes ayant un cancer du sein precoce. La revue canadienne de statistique 40: 207–224; 2012 © 2012 Societe statistique du Canada

Paired survival times with potential censoring are often observed from two treatment groups in clinical trials and other types of clinical studies. The ratio of marginal hazard rates may be used to quantify the treatment effect in these studies. In this paper, a recently proposed nonparametric kernel method is used to estimate the marginal hazard rate, and the method of variance estimates recovery (MOVER) is used for the construction of the confidence intervals of a time-dependent hazard ratio based on the confidence limits of a single marginal hazard rate. Two methods are proposed: one uses the delta method and another adopts the transformation method to construct confidence limits for the marginal hazard rate. Simulations are performed to evaluate the performance of the proposed methods. Real data from two clinical trials are analyzed using the proposed methods.

Combining mammalian target of rapamycin (mTOR) inhibitors and vinca alkaloids has shown therapeutic synergy in xenograft models of pediatric cancers. This phase I study assessed safety and toxicity of temsirolimus in combination with vinblastine in children. Procedure Patients ≥ 1 and ≤ 18 years with recurrent/refractory solid or CNS tumors were eligible. Vinblastine (4 mg/m2 ) and temsirolimus (15 mg/m2 ) were administered i.v. weekly, with planned dose escalation of vinblastine using a rolling six phase I design. Pharmacokinetic and pharmacodynamic data were collected. Results Seven patients with median age 12 years (range, 8-18 years) were enrolled; all were evaluable for toxicity and six for response. At dose level 1, four of six patients developed grade 3 mucositis, of which one met duration criteria for dose-limiting toxicity (DLT). Four patients required dose omissions for grade 3 or 4 hematologic toxicity, including one prolonged neutropenia DLT. A subsequent patient was enrolled on dose level -2 (temsirolimus 10 mg/m2 , vinblastine 4 mg/m2 ) with no protocol-related toxicity > grade 1, except grade 2 neutropenia. Two serious adverse events (SAE) occurred-an allergic reaction to temsirolimus (grade 2) and an intracranial hemorrhage in a CNS tumor patient (grade 3)-unlikely related to study therapy. Soluble VEGFR2 was reduced at cycle 1, day 36 in keeping with inhibition of angiogenesis. Four patients achieved prolonged stable disease for a median of 5.0 months (range, 3.1-8.3 months). Conclusion The combination of weekly temsirolimus (15 mg/m2 ) and vinblastine (4 mg/m2 ) exceeds the maximum tolerated dose in children, with frequent oral mucositis and hematologic toxicity.

BACKGROUND: National Cancer Institute of Canada Clinical Trials Group PA.3 (NCIC CTG PA.3) was a phase 3 study (n = 569) that demonstrated benefits for overall survival and progression-free survival with the addition of the epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) erlotinib to gemcitabine in patients with advanced pancreatic carcinoma (APC). Mutation status of the v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog (KRAS) and EGFR gene copy number (GCN) were evaluated as predictive markers in 26% of patients who had tumor samples available for analysis. METHODS: KRAS mutation status was evaluated by direct sequencing of exon 2, and EGFR GCN was determined by fluorescence in situ hybridization (FISH) analysis. The results were correlated with survival, which was the primary endpoint of the trial. RESULTS: KRAS analysis was successful in 117 patients, and EGFR FISH analysis was successful in 107 patients. KRAS mutations were identified in 92 patients (78.6%), and EGFR amplification or high polysomy (FISH-positive results) was identified in 50 patients (46.7%). The hazard ratio of death between gemcitabine/erlotinib and gemcitabine/placebo was 0.66 (95% confidence interval [CI], 0.28-1.57) for patients with wild-type KRAS and 1.07 (95% CI, 0.68-1.66) for patients with mutant KRAS (P value for interaction = .38), and the hazard ratio was 0.6 (95% CI, 0.34-1.07) for FISH-negative patients and 0.90 (95% CI, 0.49-1.65) for FISH-positive patients (P value for interaction = .32). CONCLUSIONS: In a molecular subset analysis of patients from NCIC CTG PA.3, EGFR GCN and KRAS mutation status were not identified as markers predictive of a survival benefit from the combination of erlotinib with gemcitabine for the first-line treatment of APC. Cancer 2010. © 2010 American Cancer Society.

In cancer clinical trials, it is often of interest in estimating the ratios of hazard rates at some specific time points during the study from two independent populations. In this paper, we consider nonparametric confidence interval procedures for the hazard ratio based on kernel estimates for the hazard rates with under-smoothing bandwidths. Two methods are used to derive the confidence intervals: one based on the asymptotic normality of the ratio of the kernel estimates for the hazard rates in two populations and another through Fieller's Theorem. The performances of the proposed confidence intervals are evaluated through Monte-Carlo simulations and applied to the analysis of data from a clinical trial on early breast cancer.

Transformation and computer intensive methods such as the jackknife and bootstrap are applied to construct accurate confidence intervals for the ratio of specific occurrence/exposure rates, which are used to compare the mortality (or survival) experience of individuals in two study populations. Monte Carlo simulations are employed to compare the performances of the proposed confidence intervals when sample sizes are small or moderate.