ProCAID: a phase I clinical trial to combine the AKT inhibitor AZD5363 with docetaxel and prednisolone chemotherapy for metastatic castration resistant prostate cancer

Eligible patients were 18 years or older with histologically or cytologically proven mCRPC and an Eastern Cooperative Oncology Group performance status of 0 or 1. All patients had disease progression based on PSA and/or radiographic criteria defined by the Response Evaluation Criteria in Solid Tumors (RECIST, version 1.1) and the Prostate Cancer Working Group 2 [20, 21]. Radiologically measurable and/or evaluable disease was acceptable. Patients were required to have a serum testosterone <1.7 nmol/L and ongoing LHRH analogue or antagonist therapy was permitted to maintain a castrate state. Other therapies for prostate cancer, other than ongoing bisphosphonates or denosumab, were discontinued ≥4 weeks prior to commencing study treatment and anti-androgen withdrawal response was excluded where relevant. Haematological parameter requirements were: haemoglobin ≥9 g/dL, platelets ≥100 × 10⁹/L, neutrophils ≥1.5 × 10⁹/L, bilirubin ≤ the institutional upper limit of normal (ULN), alanine (ALT) and aspartate (AST) aminotransferase ≤1.5 x ULN and sodium and potassium within the normal range for the treating institution. Patients were excluded if they had received previous treatment with cytotoxic chemotherapy but were permitted the prior use of second generation hormonal therapies e.g., abiraterone or enzalutamide. Other exclusion criteria included prior malignancy with an estimated ≥30% chance of relapse within 2 years, previously identified brain metastases, or spinal cord compression unless treated with full functional recovery, prior radiotherapy to >30% of bone marrow, another investigational agent within 30 days of study medication, type I or II diabetes mellitus requiring either insulin or oral hypoglycaemics for routine management, gastrointestinal conditions that might affect drug absorption, significant cardiac disease within the last 6 months, a left ventricular ejection fraction < the institutional lower limit of normal, a QTc interval of >480 msec, or recent exposure to potent inhibitors or inducers of CYP3A4 or substrates of CYP3A4 and CYP2D6. The complete eligibility criteria are listed in the Supplementary Appendix.

The treatment cycle is summarised in Fig. 1 and comprised DP for up to 10 cycles of 21 days. All patients received docetaxel 75 mg/m² by one hour intravenous infusion on day 1 and prednisolone 5 mg BID, orally, on days 1 to 21 of each cycle. In addition, patients received AZD5363 which was continued until either disease progression, the commencement of new anti-prostate cancer systemic therapy or unacceptable toxicity. Planned AZD5363 dose levels were 320 mg (dose level 1), 400 mg (dose level 2) and 480 mg (dose level 3), BID, orally, given according to a 4 days on and 3 days off schedule which commenced from cycle 1, day 2. Dexamethasone pre-medication was recommended at 8 mg, orally, at 12, 3 and 1 h prior to each docetaxel infusion and anti-emetics were given according to local institutional protocols. Body surface area calculations and re-calculations due to changes in weight and docetaxel dose banding (up to +/− 5%) were permitted in accordance with local institutional practices.

We utilised a conventional 3 + 3 dose escalation phase I clinical trial design. An evaluable patient was defined as one that completed cycle 1 and received at least 80% of the total cumulative doses of AZD5363 and prednisolone, and the full dose of docetaxel, with no more than a 14-day delay in starting cycle 2, or who had experienced a dose limiting toxicity (DLT). AE data was recorded for all patients who commenced study treatment.

The National Cancer Institute Common Toxicity Criteria for Adverse Events version 4.03 (CTCAEv4.03, http://​evs.​nci.​nih.​gov/​ftp1/​CTCAE/​CTCAE_​4.​03_​2010-06-14_​QuickReference_​5x7.​pdf) was used to characterise AEs. DLTs were predefined and based on recorded AEs as: a greater than 14 day delay in administration of docetaxel for cycle 2 due to drug toxicity; grade 4 neutropenia ≥7 days duration; grade 3 or 4 neutropenia associated with a temperature ≥ 38.5 °C; grade 3 or 4 neutropenia associated with bacteriologically proven sepsis; any grade 4 thrombocytopenia; grade 3 thrombocytopenia associated with non-traumatic bleeding (except where this could be explained by therapeutic anticoagulation); ≥ grade 3 hyperglycaemia for more than 1 week despite optimal intervention; grade 4 hyperglycaemia; AST or ALT >10 x ULN where AZD5363 was considered the most likely cause; AST or ALT >8 x ULN when combined with a doubling of bilirubin from baseline and where AZD5363 was considered the most likely cause; QTc (Fridericia’s or Bazett’s correction) interval > 500 msec or QTc increase >60 msec from baseline on two ECGs at least 30 min apart that could not be attributed to another cause; symptomatic congestive cardiac failure (New York Heart Association class III/IV) and a drop in LVEF that could not be attributed to another cause; a decrease in LVEF of ≥20% to a level below the institutional lower limit of normal range; clinically significant rash that despite optimal treatment remained grade ≥ 3 for 5 days or longer and that could not be attributed to another cause; grade ≥ 3 nausea, vomiting or diarrhoea, despite optimal anti-emetic or anti-diarrhoeal therapy and which could not be attributed to another cause; any other grade ≥ 3 toxicity which in the opinion of the investigator was clinically significant and related to AZD5363. Designation of a DLT excluded isolated laboratory changes of any grade (except as specified above) without clinical sequelae or clinical significance.

Intra-patient dose escalation was not permitted. In addition to protocol defined DLTs, the protocol allowed for the Safety Review Committee to consider other toxicities including those emerging during subsequent treatment cycles in prior patients in making recommendations. Dose escalation was undertaken if none of 3, or 1 of 6, evaluable patients experienced a DLT at the current dose level. If 2 or more patients experienced a DLT then the preceding dose level was established as the maximum tolerated dose (MTD). The recommended phase II dose (RP2D) was the MTD or dose level three if this was tolerated.

The primary endpoint was the determination of an RP2D for AZD5363, using a four days on/three days off administration schedule, in combination with full dose DP. Secondary endpoints included safety and tolerability profiles using CTCAEv4.03 and AZD5363 pharmacokinetics when combined with DP. Results were summarised descriptively.

Trial conduct was consistent with Good Clinical Practice guidelines and the Declaration of Helsinki and following national ethics and regulatory approvals. Informed consent was obtained from all individual participants included in the study. The trial was coordinated by the Cancer Research UK Southampton Clinical Trials Unit and sponsored by University Hospital Southampton NHS Foundation Trust.

Ten patients with a median age of 67.5 (range 56–72) were recruited from two UK centres with 4 patients in dose level 1, and 6 in dose level 2. Nine patients (90%) had bony metastases and 5 (50%) had visceral metastases. One patient in dose level 1 was considered to be non-evaluable for DLT assessment due to a compliance error, resulting from misunderstanding rather than treatment related AEs, in cycle 1 resulting in only 54.2% of the cumulative AZD5363 dose being taken. This patient subsequently tolerated treatment at full dose from cycle 2 onwards. Patient characteristics are shown in Table 1.

A median of 8.5 cycles (range 3–10) of DP and 4 cycles (range 1–13) of AZD5363 was administered. The actual number of cycles administered for each patient is shown in Table 2. In dose level 1, all 3 evaluable patients received an equal or greater number of cycles of AZD5363 compared to DP. In dose level 2 however, this was the case for only 2 of 6 patients. The 4 other patients in dose level 2 received only 1 or 2 cycles of AZD5363 whilst receiving between 3 and 10 cycles of docetaxel. No DLTs were seen in dose level 1. Two patients in dose level 2 experienced DLTs. These were due to grade 3 rash for ≥5 days and grade 3 diarrhoea despite optimal anti-diarrhoeal therapy. The severity and number of AEs overall and by dose level is shown in Table 3. Five patients in cycle 1, and 8 (80%) across all cycles had at least one grade 3 or 4 AE.

Rates of specific AEs, occurring in at least one patient, are shown for cycle 1 (Table 4) and across all cycles (Table 5). Across all cycles, the most common grade 3 or 4 AEs considered by the investigator to be related to AZD5363 were maculopapular rash and diarrhoea (3 patients and 2 patients respectively). In most cases of rash and diarrhoea, utilisation of systemic anti-histamines and loperamide respectively allowed for successful management and to allow ongoing administration of study medication. Febrile neutropenia occurred in 2 (20%) of patients across all treatment cycles. There were no treatment related deaths.

Transient hyperglycaemia is an expected AE for both AZD5363 and corticosteroids. Dexamethasone was used as chemotherapy pre-medication and prednisolone is routinely administered with docetaxel for mCRPC [2]. We saw hyperglycaemia in every patient. The mean random glucose on cycle 1 day 2 (first dose of AZD5363) was 6.0 mmol/L pre-dose and then 8.7 mmol/L at 2 h, 9.5 mmol/L at 4 h, and 6.5 mmol/L at 8 h post AZD5363 dose. Insulin and C-peptide levels also rose and fell in parallel to these glucose changes in all patients. Individual patient data are shown in Fig. 2. No patient experienced symptomatic complications as a result of hyperglycaemia or required intervention to correct a high glucose level (an algorithm, including criteria for metformin administration, was provided within the trial protocol).

Pharmacokinetic analyses indicated that AZD5363 exposure for patients receiving concomitant DP chemotherapy was in keeping with that found in patients receiving mono-therapy in the AZD5363 development programme.

PSA reduction from baseline level to <50% (PSA₅₀) at 12 weeks of treatment was seen in 7 (70%) of patients (Fig. 3).