An objective response (OR) is defined as a patient having a best overall response of either complete response (CR) or partial response (PR). A patient has best overall response of CR if she had overall response of CR or PR on one visit and met the confirmation criteria per RECIST. ORR is defined as percentage of patients with objective response.

Each patient with measurable disease at baseline was assessed for OR from the sequence of Response Evaluation Criteria in Solid Tumors (RECIST) scan data up to data cut-off. RECIST scans were performed every 12 weeks (+/- 2weeks) from randomization

Time (in days) from randomization until objective disease progression or death (in the absence of objective progression). RECIST tumour assessments carried out every 12 weeks from randomization (+/- 2 weeks) until data cut-off on 19th March 2008.

Time from randomisation until objective progression or death (in the absence of objective progression), measured only in those patients who achieved a confirmed Complete Response or confirmed Partial Response.

A Clinical Benefit (CB) responder is defined as a patient having a best overall response of Complete response (CR), Partial Response (PR) or Stable disease (SD) provided SD (or better) was present = 154 days from randomization (ie SD = 24 weeks with the 2 week RECIST assessment time window allowed). The Clinical Benefit Rate is the percentage of patients with CB.

The measure of dispersion for mean population clearance is based on the estimated inter-individual variance

Patients who agreed to participate in the PK substudy. The results are based on 148, 122 and 140 plasma-concentration records from patients in the 250 mg, 250 mg + LD and 500 mg treatment arms respectively.

The measure of dispersion for volume of distribution is based on the inter-individual variance estimated for the apparent volume of plasma into which Fulvestrant distributes

Patients who agreed to participate in the PK substudy.