Prevalence and impact of correlative science in breast cancer phase II trials.
1046 Background: Correlative science (CS) can augment clinical trial results by identifying biomarkers of response and resistance to a novel intervention. There is little data on how frequently CS is included in breast cancer phase II trials (BP2T). We evaluated recently published BP2T to identify prevalence, characteristics, and outcomes of CS. METHODS: Using search terms "breast cancer" and "phase II trials" in Pubmed, we identified BP2T of systemic therapy published between June 2005 and June 2010. Reports were abstracted by 2 independent reviewers for trial characteristics, report of CS, CS features including endpoints, material, nature of sample (clinical, optional or mandatory biopsy), adequacy of samples, biopsy safety, and CS outcomes. Descriptive statistics, Fisher's exact test, and Chi-square analysis were performed. RESULTS: 298 eligible trials enrolling a total of 18,782 patients were abstracted. 81 trials (27.2%) enrolling 5807 patients (30.9%) involved CS. Of these, 57 (70.4%) included tissue, 16 (28%) used optional research biopsy and 17 (30%) required mandatory biopsy. No trial reported biopsy safety. 55 (68%) of trials with CS specified the CS endpoint. For 44 trials (54%) the authors reported that CS was successful, while for 17 trials (21%) they specifically reported insufficient material. CS predicted response, toxicity or further characterized the tumor in 50 trials (62%). BP2T sponsored by pharmaceutical industry were more likely to involve CS (33.7% vs. 17.1%, p=0.0017). Neoadjuvant trials vs. metastatic trials and U.S.-based vs. non-U.S. trials were more likely to include CS (47% vs. 21.2%, p=0.0001; 37% vs. 21%, p=0.005). Single arm BP2T were less likely to include CS vs. randomized BP2T and phase I/II trials (22.5% vs. 42.3%, p=0.0011). There was no statistical difference between likelihood of reporting success or insufficient material for mandatory vs. optional research biopsy trials. CONCLUSIONS: A minority of recently published BP2T trials included CS. When CS is included, results frequently add to trial information. Reporting of CS endpoints, adequacy of samples, and biopsy safety is inconsistent or absent, and should be considered in all trials to inform policy debates over optimal and ethical CS design.
Duke Scholars
Published In
EISSN
Publication Date
Volume
Issue
Start / End Page
Location
Related Subject Headings
- Oncology & Carcinogenesis
- 1112 Oncology and Carcinogenesis
- 1103 Clinical Sciences
Citation
Published In
EISSN
Publication Date
Volume
Issue
Start / End Page
Location
Related Subject Headings
- Oncology & Carcinogenesis
- 1112 Oncology and Carcinogenesis
- 1103 Clinical Sciences