Fulvestrant +/- Akt Inhibition in Advanced Aromatase Inhibitor Resistant Breast Cancer

Phase 1 (n=9-12) As fulvestrant and AZD5363 have not previously been administered to this population, we have incorporated an initial phase I dose escalation: * 3 patients will receive fulvestrant 500mg on day 1 and 400mg AZD5363 oral capsules or tablets bd 4 days on and 3 days off. They will be assessed for dose limiting toxicity (DLT) on day 1, 15 and 28 of cycle 1. The safety review committee (SRC) will meet to review DLT reports when the third patient finishes cycle 1. * If 400mg is tolerable (0/3-6 or 1-6 has a DLT) then the dose of AZD5363 will be escalated to 480mg. * If 480mg is tolerable in a cohort of 6 patients then 480mg will be the Maximum tolerated dose (MTD). * If 480mg is not tolerable (2/6 patients have a DLT), then the 400mg dose will be the MTD. * If 400mg is not tolerable (2 or more patients have a DLT) then the dose will be reduced to 320mg for the next cohort of 6 patients. * if 320mg is tolerable in a cohort of 6 patients, then 320mg will be the MTD. * if 320mg is not tolerable, then the trial will not proceed. If patients are withdrawn for reasons other than toxicity during Stage 1 before DLT assessments then they will be replaced. Phase 2 (n=136) Recruitment in to Phase 2 will commence after all 6 patients in Stage 1 have completed at least 1 cycle of therapy and the SRC have given the go-ahead. Patients will be randomised to one of two arms: Arm A (control arm): Fulvestrant + placebo The control arm will consist continuous 28 day cycles of fulvestrant 500mg on day 1, plus placebo capsule or tablet bd 4 days on and 3 days off. Arm B (experimental arm): Fulvestrant + AZD5363 The experimental arm will consist continuous 28 day cycles of fulvestrant 500mg on day 1, plus AZD5363 capsule or tablet bd 4 days on and 3 days off. A further safety assessment will be made after 20 patients have been randomised to receive AZD5363 or placebo and have at least 1 cycle of protocol treatment in stage 2. Phase 2 will have 3 stages, this is because we want to investigate whether tumours with certain characteristics are more likely to respond to the AZD5363. As mentioned above, we will test the tumour tissue for the presence of a PIK3CA mutation (a gene encoding catalytic subunit of class 1 PI3K) and for low levels of phosphatase and tensin homolog (PTEN). Laboratory research has shown that tumours with these characteristics may respond better to the AZD5363. Tumours which do not have these characteristics are called wild-type, and we want to make sure we don't recruit too many of these patients if they might not benefit from the treatment. We will start off by recruiting all eligible patients. However, when we know that we have 40 patients with wild-type tumours, and they have been on trial treatment for eight weeks, we will stop recruiting patients with wild-type tumours. This means that we will do a blood test to confirm the tumour has the PIK3CA mutation before a patient can be entered onto the study. We will perform a review of the tumour measurements in this wild-type group to determine whether the tumours of the patients receiving AZD5363 have shrunk more than patients receiving placebo. If there is no evidence that the tumour has shrunk in the AZD5363 group compared to placebo, we will not recruit any more patients with wild-type tumours. If there is evidence that that the tumours have shrunk in the AZD5363 group, we will start recruiting patients with wild-type tumours again.