Sample size and threshold estimation for clinical trials with predictive biomarkers.

With the increasing availability of newly discovered biomarkers personalized drug development is becoming more commonplace. Unless evidence of the dependence of clinical benefit on biomarker classification is a priori unequivocal, personalized drug development needs to jointly investigate treatments and biomarkers in clinical trials. Motivated by the development of contemporary cancer treatments, we propose targeting three main questions sequentially in order to determine (1) whether a drug is efficacious, (2) whether a biomarker can personalize treatment, and (3) how to define personalization. For time-to-event data satisfying the Cox proportional hazards model, we show that (1) and (2) may not directly involve the variance of an interaction term but of a contrast with smaller variance. An asymptotically exact covariance matrix for the parameter vector in the CPH model is derived to construct sample size formulae and an inference approach for thresholds of continuous biomarkers. The covariance matrix also reveals strategies for greater efficiency in trial design, for example, when the biomarker is binary or does not modulate the effect of treatment in the control arm. We motivate our approach by studying the outcome of a contemporary cancer study.

A randomized phase II trial of vismodegib versus placebo with FOLFOX or FOLFIRI and bevacizumab in patients with previously untreated metastatic colorectal cancer.

PURPOSE: Vismodegib, a Hedgehog pathway inhibitor, has preclinical activity in colorectal cancer (CRC) models. This trial assessed the efficacy, safety, and pharmacokinetics of adding vismodegib to first-line treatment for metastatic CRC (mCRC). EXPERIMENTAL DESIGN: Patients were randomized to receive vismodegib (150 mg/day orally) or placebo, in combination with FOLFOX or FOLFIRI chemotherapy plus bevacizumab (5 mg/kg) every 2 weeks until disease progression or intolerable toxicity. The primary endpoint was progression-free survival (PFS). Key secondary objectives included evaluation of predictive biomarkers and pharmacokinetic drug interactions. RESULTS: A total of 199 patients with mCRC were treated on protocol (124 FOLFOX, 75 FOLFIRI). The median PFS hazard ratio (HR) for vismodegib treatment compared with placebo was 1.25 (90% CI: 0.89-1.76; P = 0.28). The overall response rates for placebo-treated and vismodegib-treated patients were 51% (90% CI: 43-60) and 46% (90% CI: 37-55), respectively. No vismodegib-associated benefit was observed in combination with either FOLFOX or FOLFIRI. Increased tumor tissue Hedgehog expression did not predict clinical benefit. Grade 3 to 5 adverse events reported for more than 5% of patients that occurred more frequently in the vismodegib-treated group were fatigue, nausea, asthenia, mucositis, peripheral sensory neuropathy, weight loss, decreased appetite, and dehydration. Vismodegib did not alter the pharmacokinetics of FOLFOX, FOLFIRI, or bevacizumab. CONCLUSIONS: Vismodegib does not add to the efficacy of standard therapy for mCRC. Compared with placebo, treatment intensity was lower for all regimen components in vismodegib-treated patients, suggesting that combined toxicity may have contributed to lack of efficacy.

Efficacy and safety of vismodegib in advanced basal-cell carcinoma.

BACKGROUND: Alterations in hedgehog signaling are implicated in the pathogenesis of basal-cell carcinoma. Although most basal-cell carcinomas are treated surgically, no effective therapy exists for locally advanced or metastatic basal-cell carcinoma. A phase 1 study of vismodegib (GDC-0449), a first-in-class, small-molecule inhibitor of the hedgehog pathway, showed a 58% response rate among patients with advanced basal-cell carcinoma. METHODS: In this multicenter, international, two-cohort, nonrandomized study, we enrolled patients with metastatic basal-cell carcinoma and those with locally advanced basal-cell carcinoma who had inoperable disease or for whom surgery was inappropriate (because of multiple recurrences and a low likelihood of surgical cure, or substantial anticipated disfigurement). All patients received 150 mg of oral vismodegib daily. The primary end point was the independently assessed objective response rate; the primary hypotheses were that the response rate would be greater than 20% for patients with locally advanced basal-cell carcinoma and greater than 10% for those with metastatic basal-cell carcinoma. RESULTS: In 33 patients with metastatic basal-cell carcinoma, the independently assessed response rate was 30% (95% confidence interval [CI], 16 to 48; P=0.001). In 63 patients with locally advanced basal-cell carcinoma, the independently assessed response rate was 43% (95% CI, 31 to 56; P<0.001), with complete responses in 13 patients (21%). The median duration of response was 7.6 months in both cohorts. Adverse events occurring in more than 30% of patients were muscle spasms, alopecia, dysgeusia (taste disturbance), weight loss, and fatigue. Serious adverse events were reported in 25% of patients; seven deaths due to adverse events were noted. CONCLUSIONS: Vismodegib is associated with tumor responses in patients with locally advanced or metastatic basal-cell carcinoma. (Funded by Genentech; Erivance BCC ClinicalTrials.gov number, NCT00833417.).

Inhibition of the hedgehog pathway in advanced basal-cell carcinoma.

BACKGROUND: Mutations in hedgehog pathway genes, primarily genes encoding patched homologue 1 (PTCH1) and smoothened homologue (SMO), occur in basal-cell carcinoma. In a phase 1 clinical trial, we assessed the safety and pharmacokinetics of GDC-0449, a small-molecule inhibitor of SMO, and responses of metastatic or locally advanced basal-cell carcinoma to the drug. METHODS: We selected 33 patients with metastatic or locally advanced basal-cell carcinoma to receive oral GDC-0449 at one of three doses; 17 patients received 150 mg per day, 15 patients received 270 mg per day, and 1 patient received 540 mg per day. We assessed tumor responses with the use of Response Evaluation Criteria in Solid Tumors (RECIST), physical examination, or both. Molecular aspects of the tumors were examined. RESULTS: The median duration of the study treatment was 9.8 months. Of the 33 patients, 18 had an objective response to GDC-0449, according to assessment on imaging (7 patients), physical examination (10 patients), or both (1 patient). Of the patients who had a response, 2 had a complete response and 16 had a partial response. The other 15 patients had either stable disease (11 patients) or progressive disease (4 patients). Eight grade 3 adverse events that were deemed to be possibly related to the study drug were reported in six patients, including four with fatigue, two with hyponatremia, one with muscle spasm, and one with atrial fibrillation. One grade 4 event, asymptomatic hyponatremia, was judged to be unrelated to GDC-0449. One patient withdrew from the study because of adverse events. We found evidence of hedgehog signaling in tumors that responded to the treatment. CONCLUSIONS: GDC-0449, an orally active small molecule that targets the hedgehog pathway, appears to have antitumor activity in locally advanced or metastatic basal-cell carcinoma. (ClinicalTrials.gov number, NCT00607724.)