Abstract CT323: Accelerated phase I trial of two schedules of the combination of the PARP inhibitor olaparib and AKT inhibitor AZD5363 using a novel intrapatient dose escalation design in advanced cancer patients

Background: There is an urgent need for better trial designs to assess targeted drug combinations. We proposed a novel intrapatient (intrapt) dose escalation trial design to optimize drug exposures, minimize pharmacokinetic (PK) variability and reduce patient (pt) numbers needed (Yap et al, JCO 2013). In vivo synergy between PARP and PI3K pathway inhibition was seen in BRCA1-related and sporadic cancers (Juvekar et al; Ibrahim et al, Cancer Discov 2012), providing rationale for this study. Methods: Two-stage investigator initiated phase I trial: a) Intrapt dose escalation; b) Recommended phase II combination dose (RP2CD) expansion. Advanced cancer pts received escalating doses of AZD5363 (AZD) BID in 2 parallel arms (4 days on 3 days off [4/7 arm] at 320, 400, 480mg; 2 days on 5 days off [2/7 arm] at 480, 560, 640mg) with Olaparib (Ola) at 300mg BID in 3 weekly cycles. AZD was escalated after each cycle in each pt if drug related toxicities were ≤CTCAE G2. Dose limiting toxicities (DLT) were assessed during the 1st cycle of each dose level (DL). ≥6 evaluable pts were required at each DL. RECIST assessment was done every 3 cycles. Prior PARP or PI3K/AKT inhibitor use was allowed. PK and pharmacodynamics (PD) were assessed in tumor and normal tissue. Targeted +/- whole exome next generation sequencing was assessed in tumor and serial plasma DNA samples in all pts for predictive biomarkers of response. Results: Dose escalation was completed in 7.5 months (m) in 20 pts in 1 center; ≥6 evaluable pts were treated at each of the 3 DLs in both arms. Common (>15%) G1-2 toxicities were nausea, vomiting, fatigue, diarrhea and anemia. A DLT of G3 rash was seen at 480mg BID 4/7 AZD + 300mg BID Ola. Non DLT G3 anemia (n = 2), diarrhea (n = 2), fatigue (n = 1) and vomiting (n = 1) were seen in 4/7 arm; G3 hyperglycemia (n = 1), transaminitis (n = 1) and fatigue (n = 2) in 2/7 arm. No significant PK interactions were seen between Ola and AZD. Intrapt dose escalation of AZD showed dose dependent increases in PK exposures. Platelet-rich plasma PD showed significant decreases in pSer9 GSK3β post-therapy at all DLs (mean ≥55% [p