Bayesian Methods in Human Drug and Biological Products Development in CDER and CBER.

The Center for Drug Evaluation and Research (CDER) and the Center for Biologics Evaluation and Research (CBER) of the U.S. Food and Drug Administration (FDA) have been leaders in protecting and promoting the U.S. public health by helping to ensure that safe and effective drugs and biological products are available in the United States for those who need them. The null hypothesis significance testing approach, along with other considerations, is typically used to demonstrate the effectiveness of a drug or biological product. The Bayesian framework presents an alternative approach to demonstrate the effectiveness of a treatment. This article discusses the Bayesian framework for drug and biological product development, highlights key settings in which Bayesian approaches may be appropriate, and provides recent examples of the use of Bayesian approaches within CDER and CBER.

First FDA approval of dual anti-HER2 regimen: pertuzumab in combination with trastuzumab and docetaxel for HER2-positive metastatic breast cancer.

On June 8, 2012, the U.S. Food and Drug Administration (FDA) approved pertuzumab (Perjeta, Genentech) for use in combination with trastuzumab (Herceptin, Genentech) and docetaxel for the treatment of patients with HER2-positive metastatic breast cancer (MBC) who have not received prior anti-HER2 therapy or chemotherapy for metastatic disease. Approval was based on the results of a randomized, double-blind, placebo-controlled trial conducted in 808 patients with HER2-positive MBC. Patients were randomized (1:1) to receive pertuzumab (n = 402) or placebo (n = 406) in combination with trastuzumab and docetaxel. The primary endpoint was progression-free survival (PFS) and a key secondary endpoint was overall survival (OS). A statistically significant improvement in PFS (difference in medians of 6.1 months) was observed in patients receiving pertuzumab [HR, 0.62; 95% confidence interval (CI), 0.51-0.75; P < 0.0001]. A planned interim analysis suggested an improvement in OS (HR, 0.64; 95% CI, 0.47-0.88; P = 0.0053) but the HR and P value did not cross the stopping boundary. Common adverse reactions (>30%) observed in patients on the pertuzumab arm included diarrhea, alopecia, neutropenia, nausea, fatigue, rash, and peripheral neuropathy. No additive cardiac toxicity was observed. Significant manufacturing issues were identified during the review. On the basis of substantial evidence of efficacy for pertuzumab in MBC and the compelling public health need, FDA did not delay availability to patients pending final resolution of all manufacturing concerns. Therefore, FDA approved pertuzumab but limited its approval to lots not affected by manufacturing problems. The applicant agreed to multiple manufacturing and testing postmarketing commitments under third-party oversight to resolve manufacturing issues.

Vandetanib for the treatment of symptomatic or progressive medullary thyroid cancer in patients with unresectable locally advanced or metastatic disease: U.S. Food and Drug Administration drug approval summary.

On April 6, 2011, the U.S. Food and Drug Administration approved vandetanib (Caprelsa tablets; AstraZeneca Pharmaceuticals LP) for the treatment of symptomatic or progressive medullary thyroid cancer in patients with unresectable, locally advanced, or metastatic disease. Vandetanib is the first drug approved for this indication, and this article focuses on the basis of approval. Approval was based on the results of a double-blind trial conducted in patients with medullary thyroid carcinoma. Patients were randomized 2:1 to vandetanib, 300 mg/d orally (n = 231), or to placebo (n = 100). The primary objective was demonstration of improvement in progression-free survival (PFS) with vandetanib compared with placebo. Other endpoints included evaluation of overall survival and objective response rate. The PFS analysis showed a marked improvement for patients randomized to vandetanib (hazard ratio = 0.35; 95% confidence interval, 0.24-0.53; P < 0.0001). The objective response rate for the vandetanib arm was 44% compared with 1% for the placebo arm. The most common grade 3 and 4 toxicities (>5%) were diarrhea and/or colitis, hypertension and hypertensive crisis, fatigue, hypocalcemia, rash, and corrected QT interval (QTc) prolongation. This approval was based on a statistically significant and clinically meaningful improvement in PFS. Given the toxicity profile, which includes prolongation of the QT interval and sudden death, only prescribers and pharmacies certified through the vandetanib Risk Evaluation Mitigation Strategy Program are able to prescribe and dispense vandetanib. Treatment-related risks should be taken into account when considering the use of vandetanib in patients with indolent, asymptomatic, or slowly progressing disease.

Approval summary: erlotinib maintenance therapy of advanced/metastatic non-small cell lung cancer (NSCLC).

On April 16, 2010, the U. S. Food and Drug Administration (FDA) approved erlotinib tablets (Tarceva®; OSI Pharmaceuticals, Inc., Melville, NY) for maintenance treatment of patients with stage IIIB/IV non-small cell lung cancer (NSCLC) whose disease had not progressed after four cycles of platinum-based first-line chemotherapy. In total, 889 patients received either erlotinib (150 mg) or placebo once daily. Progression-free survival (PFS), in all patients and in patients with epidermal growth factor receptor (EGFR)(+) tumors by immunohistochemistry (IHC), was the primary efficacy endpoint. Overall survival (OS) was a secondary sponsor endpoint but was the primary regulatory endpoint. Median PFS times were 2.8 months and 2.6 months in the erlotinib and placebo arms, respectively (hazard ratio [HR], 0.71; 95% confidence interval [CI], 0.62-0.82; p < .001). Median OS times were 12.0 months and 11.0 months, favoring erlotinib (HR, 0.81; 95% CI, 0.70-0.95). The PFS and OS HRs in patients with EGFR(+) tumors by IHC were 0.69 (95% CI, 0.58-0.82) and 0.77 (95% CI, 0.64-0.93), respectively. The PFS and OS HRs in patients with EGFR(-) tumors by IHC were 0.77 (95% CI, 0.51-1.14) and 0.91 (95% CI, 0.59-1.38), respectively. Following disease progression, 57% of placebo-treated patients received additional chemotherapy, compared with 47% of erlotinib-treated patients. Fourteen percent of placebo-treated patients received erlotinib or gefitinib, 31% received docetaxel, and 14% received pemetrexed. In total, 59% of placebo-treated patients who received treatment received FDA approved second-line NSCLC drugs. The most common adverse reactions in patients receiving erlotinib were rash and diarrhea.

Sorafenib for the treatment of unresectable hepatocellular carcinoma.

PURPOSE: To describe the U.S. Food and Drug Administration (FDA) review and approval of sorafenib (Nexavar; Bayer Pharmaceuticals Corp., Montville, NJ, and Onyx Pharmaceuticals Corp., Emeryville, CA), an oral kinase inhibitor, for the treatment of patients with unresectable hepatocellular carcinoma (HCC). EXPERIMENTAL DESIGN: The FDA independently analyzed an international, double-blind, placebo-controlled trial comparing the effect of best supportive care plus sorafenib or matching placebo on overall survival. Eligible patients had unresectable, biopsy-proven HCC and had not received prior systemic therapy. RESULTS: Among the 602 randomized patients (placebo, 303; sorafenib, 299), baseline characteristics were well balanced, and 97% were Child-Pugh score A. HCC was "advanced" in 70% overall, as defined by extrahepatic metastases or by tumor radiographically visible in venous structures outside the liver. Underlying liver diseases included hepatitis B (18%), hepatitis C (28%), and alcohol-related (26%). The trial was stopped following a prespecified second interim analysis showing a statistically significant survival advantage for sorafenib [median, 10.7 vs 7.9 months; hazard ratio, 0.69 (95% confidence interval, (0.55, 0.87)), p = 0.00058]. Adverse events in sorafenib-treated patients included diarrhea in 55% (grade 3, 10%), hand-foot syndrome in 21% (grade 3, 8%), rash in 19% (grade 3, 1%), and cardiac ischemia or infarction in 2.7% (versus 1.3% for placebo). On sorafenib, treatment-emergent hypertension occurred in 9% of patients (placebo, 4%) and was grade 3 in 4% (placebo, 1%); elevated serum lipase occurred in 40% (placebo, 37%); hypophosphatemia occurred in 35% (placebo, 11%). CONCLUSIONS: Sorafenib is the first systemic therapy to demonstrate a survival benefit in a randomized trial for unresectable HCC and has received FDA approval for this indication.

Composite model of time-varying appearance and disappearance of neurohormone pulse signals in blood.

Blood-borne neurohormonal signals reflect the intermittent burst-like release of peptides and steroids from neurons, glands and target tissues. Hormones control basic physiological processes, such as growth, metabolism, reproduction and stress-related adaptations. Secreted molecules undergo combined diffusion, advection and irreversible elimination from the circulation. Quantification of these interdependent processes by a structurally relevant model embodying discrete event times, continuous rates of secretion and elimination, and stochastic variations poses a formidable challenge. In an experimental setting, one observes only the hormone concentrations, which comprise a time-varying composite of secretion and elimination. The number, shape and location of underlying bursts (pulses) and attendant secretion and kinetic parameters are unobserved. The ability to estimate the properties of these processes from the observed data is fundamental to an understanding of regulated hormonal dynamics. The present formulation allows objective simultaneous appraisal of discrete (pulse times) and continuous (secretion/elimination) properties of neuroglandular activity in the presence of random variability. A probability distribution is constructed for the structural parameters (secretion/elimination, pulsing), and an algorithm is developed by which one can, based upon observed hormone concentration data, make probabilistic statements about the underlying structure: pulse frequency per day, total basal (constitutive) and pulsatile secretion per day, and half-lives of elimination. The algorithm consists of the following steps: first, explicit construction of a family of sequentially decreasing putative pulse-time sets for a given neurohormone concentration time series; and then, recursive iteration between the following two: (a) for a given pulse-time set, generate a sample from the probability distribution of unknown underlying hormone secretion and elimination rates; and (b) determine whether or not a probability-based transition from one pulse-time set to another is merited (i.e., add/remove a pulse-time or stay the same). We apply this procedure illustratively to joint estimation of pulse times, secretion rates and elimination kinetics of selected pituitary hormones (ACTH, LH and GH).