Some Thoughts on the Use of Predictive Probability in Clinical Trials.

In this short note, we express our viewpoint regarding declaring study success based on Bayesian predictive probability of study success.

Circulating tumor cells: application as a biomarker for molecular characterization and predictor of survival in an all-comer solid tumor phase I clinical study.

PURPOSE: Clinical development of cancer drugs has a low success rate. Prognostic and predictive biomarkers using minimally invasive approaches hold promise for increasing the probability of success by enabling disease characterization, patient selection and early detection of drug treatment effect. Enumeration and molecular characterization of circulating tumor cells (CTC) may address some of these needs, and thus were evaluated for utility in a Phase I solid tumor clinical study. EXPERIMENTAL DESIGN: Blood samples for CTC analysis were obtained from 24 cancer patients in a multi-center all-comer Phase I study of MEDI-575, a novel anti-PDGFRα antibody. Samples were taken at screening and analyzed for enumeration of CTC using the CellSearch(®) platform and for molecular characterization using a novel quantitative RT-PCR assay. RESULTS: Fifty-nine percent of the patients showed at least 1 CTC per 7.5 ml of blood at baseline. Progression-free survival (PFS) and overall survival (OS) of patients with 0 CTCs at baseline were longer than PFS and Os for patients with 1-3 and >3 CTCs (8.8 versus 1.4 and 1.3 months PFS, P = 0.02; 9.0 vs 7.4 and 3.5 months OS, P = 0.20, respectively). Patients with 0 CTC showed a greater percentage of stable disease than the other 2 groups with 1-3 and >3 CTCs (57% vs 29% and 0%). The multimarker qRT-PCR method detected CTC in 40% of the patients, and 80% of these patients were positive for pre-selected drug target genes. CONCLUSION: CTC enumeration of patients in an all-comer study is feasible and may allow for patient stratification for PFS and Os to evaluate the clinical response of investigational agents. Gene expression profiling of isolated CTC may provide a means for molecular characterization of selected tumor targets.

Phase 1, open-label study of MEDI-547 in patients with relapsed or refractory solid tumors.

BACKGROUND: Targeting the cell-surface receptor EphA2, which is highly expressed in some solid tumors, is a novel approach for cancer therapy. We aimed to evaluate the safety profile, maximum tolerated dose (MTD), pharmacokinetics, and antitumor activity of MEDI-547, an antibody drug conjugate composed of the cytotoxic drug auristatin (toxin) linked to a human anti-EphA2 monoclonal antibody (1C1), in patients with solid tumors relapsed/refractory to standard therapy. METHODS: In this phase 1, open-label study with planned dose-escalation and dose-expansion cohorts, patients received a 1-h intravenous infusion of MEDI-547 (0.08 mg/kg) every 3 weeks. RESULTS: Six patients received 0.08 mg/kg; all discontinued treatment. Dose escalation was not pursued. The study was stopped before cohort 2 enrollment due to treatment-related bleeding and coagulation events (hemorrhage-related, n = 3; epistaxis, n = 2). Therefore, lower doses were not explored and an MTD could not be selected. The most frequently reported treatment-related adverse events (AEs) were increased liver enzymes, decreased hemoglobin, decreased appetite, and epistaxis. Three patients (50%) experienced treatment-related serious AEs, including conjunctival hemorrhage, pain (led to study drug discontinuation), liver disorder, and hemorrhage. Best response included progressive disease (n = 5; 83.3%) and stable disease (n = 1; 16.7%). Minimal or no dissociation of toxin from 1C1 conjugate occurred in the blood. Serum MEDI-547 concentrations decreased rapidly, ~70% by 3 days post-dose. No accumulation of MEDI-547 was observed at 0.08 mg/kg upon administration of a second dose 3 weeks following dose 1. CONCLUSIONS: The safety profile of MEDI-547 does not support further clinical investigation in patients with advanced solid tumors.

Pharmacodynamic (phase 0) study using etaracizumab in advanced melanoma.

AlphaVbeta3 (alphavbeta3) is an important molecule for tumor-induced angiogenesis and is upregulated in metastatic melanoma (MM). We proposed to study the mechanism of action of etaracizumab, a monoclonal antibody targeting alphavbeta3, in MM. Patients with MM and biopsiable tumor were treated with etaracizumab in 3 dose cohorts starting from 8 mg/kg. Tumor saturation by etaracizumab using LM609 immunohistochemical staining of tumor sections was the primary endpoint. Subsequent dose cohorts were defined based on the tumor saturation by etaracizumab. Secondary end points were analysis of clinical benefit and changes from baseline of several tumor and peripheral blood biomarkers. Eighteen patients were enrolled at 3 dose levels. Etaracizumab showed better melanoma cell saturation at the 8mg/kg and 1 mg/kg dose compared with the 4 mg/kg dose and better vascular endothelial cell saturation at 8 mg/kg compared with lower dose groups. Etaracizumab demonstrated an acceptable safety profile. The optimal biologic dose out of those selected for investigation was 8 mg/kg. Patients treated at the highest dose may have had better clinical benefit secondary to suppression of the activated immediate downstream effector of alphavbeta3 signaling, FAK, in melanoma cells, but this alone did not ultimately affect melanoma cell proliferation or apoptosis. No apparent antiangiogenic or immunomodulatory effects of etaracizumab were noted.

A randomized phase 2 study of etaracizumab, a monoclonal antibody against integrin alpha(v)beta(3), + or - dacarbazine in patients with stage IV metastatic melanoma.

BACKGROUND: The alpha (v) beta (3) (alpha(v)beta(3)) integrin is involved in intracellular signaling regulating cell proliferation, migration, and differentiation and is important for tumor-induced angiogenesis. METHODS: This phase 2, randomized, open-label, 2-arm study was designed to capture safety data and evaluate the antitumor efficacy of etaracizumab (Abegrin), an IgG1 humanized monoclonal antibody against the alpha(v)beta(3) integrin, in patients with previously untreated metastatic melanoma. The objective was to evaluate whether etaracizumab + or - dacarbazine had sufficient clinical activity to warrant further study in a phase 3 clinical trial. RESULTS: One hundred twelve patients were randomized to receive etaracizumab alone (N = 57) or etaracizumab + dacarbazine (N = 55). Safety of etaracizumab + or - dacarbazine was acceptable with infusion-related, gastrointestinal, and metabolic reactions being the most common adverse events (AEs). The majority of AEs were grade 1 or 2 in severity in both study arms; most events were not considered serious, except for cardiovascular (myocardial infarction, atrial fibrillation) and thromboembolic events, which occurred in 3 and 5 patients, respectively. None of the patients in the etaracizumab-alone study arm and 12.7% of patients in the etaracizumab + dacarbazine study arm achieved an objective response. The median duration of objective response in the etaracizumab + dacarbazine study arm was 4.2 months. Stable disease rate, time to progression (TTP), and progression-free survival (PFS) appeared to be similar between the 2 treatment arms. Stable disease occurred in 45.6% of patients in the etaracizumab-alone study arm and 40.0% of patients in the etaracizumab + dacarbazine study arm. Median TTP and median PFS were both 1.8 months in the etaracizumab-alone study arm and 2.5 and 2.6 months in the etaracizumab + dacarbazine study arm, respectively. Median overall survival was 12.6 months in the etaracizumab-alone study arm and 9.4 months in the etaracizumab + dacarbazine study arm. CONCLUSIONS: The survival results in both treatment arms of this study were considered unlikely to result in clinically meaningful improvement over dacarbazine alone.

Optimal Bayesian design for patient selection in a clinical study.

Bayesian experimental design for a clinical trial involves specifying a utility function that models the purpose of the trial, in this case the selection of patients for a diagnostic test. The best sample of patients is selected by maximizing expected utility. This optimization task poses difficulties due to a high-dimensional discrete design space and, also, to an expected utility formula of high complexity. A simulation-based optimal design method is feasible in this case. In addition, two deterministic algorithms that perform a systematic search over the design space are developed to address the computational issues.

Adjusting for verification bias in diagnostic test evaluation: a Bayesian approach.

Obtaining accurate estimates of the performance of a diagnostic test for some population of patients might be difficult when the sample of subjects used for this purpose is not representative for the whole population. Thus, in the motivating example of this paper a test is evaluated by comparing its results with those given by a gold standard procedure, which yields the disease status verification. However, this procedure is invasive and has a non-negligible risk of serious complications. Moreover, subjects are selected to undergo the gold standard based on some risk factors and the results of the test under study. The test performance estimates based on the selected sample of subjects are biased. This problem was presented in previous studies under the name of verification bias. The current paper introduces a Bayesian method to adjust for this bias, which can be regarded as a missing data problem. In addition, it addresses the case of non-ignorable verification bias. The proposed Bayesian estimation approach provides test performance estimates that are consistent with the results obtained using likelihood-based approach. In addition, the paper studies how valuable the statistical findings are from the perspective of clinical decision making.

HLA-specific antibodies are risk factors for lymphocytic bronchiolitis and chronic lung allograft dysfunction.

Bronchiolitis obliterans syndrome (BOS) represents a major limitation in lung transplantation. While acute rejection is widely considered the most important risk factor for BOS, the impact of HLA-specific antibodies is less understood. Of 51 lung recipients who were prospectively tested during a 4.2 +/- 1.6-year period, 14 patients developed HLA-specific antibodies. A multi-factorial analysis was performed to correlate the prevalence of BOS with HLA antibodies, persistent-recurrent acute rejection (ACR-PR), lymphocytic bronchiolitis, and HLA-A, -B, and -DR mismatches. HLA-specific antibodies were associated with ACR-PR (10/14 vs. 11/37 with no antibodies, p < 0.05), lymphocytic bronchiolitis (8/14 vs. 10/37, p < 0.05), and BOS (10/14, vs. 9/37, p < 0.005). Other risk factors for BOS were: lymphocytic bronchiolitis (13/18 vs. 6/33 with no lymphocytic bronchiolitis, p < 0.0001), ACR-PR (12/21 vs. 7/30 with no ACR-PR, p < 0.05), and the number of HLA-DR mismatches (1.7 +/- 0.48 in BOS vs. 1.2 +/- 0.63 without BOS, p < 0.05). The presence of antibodies exhibited a cumulative effect on BOS when it was associated with either lymphocytic bronchiolitis or ACR-PR. The complex relationship between the development of HLA antibodies and acute and chronic lung allograft rejection determines the importance of post-transplant screening for HLA-specific antibodies as a prognostic element for lung allograft outcome.