PD-L1 is overexpressed by dendritic cells of mCRPC patients (pts) progressing on androgen receptor antagonist therapy. Anti-PD-1 agents lead to infrequent but durable responses. We tested the hypothesis that dual checkpoint blockade of CTLA-4 with tremelimumab (T) and PD-L1 with durvalumab (D) enhances immune mediated activity in mCRPC.
In this multicenter randomized phase II study, mCRPC pts (measurable disease, prior abiraterone and/or enzalutamide, no more than one taxane for mCRPC) were randomized to D 1500mg IV Q4 weeks +/- 4 doses of T 75mg IV. The primary endpoint was ORR (RECIST1.1 and iRECIST) using a Simon 2-stage design. Key secondary endpoints were PSA response rate (RR) and time to progression (TTP). A treatment arm would be considered of interest if ≥ 4 ORs (null 5% or less, alt 20% or more). Correlative testing was done for PD-L1/CD8 IHC on mandatory tumour biopsies and 74-gene panel (∼1Mb) sequencing of plasma cell-free DNA (cfDNA) both collected at baseline.
52 pts were enrolled: median age 70 yrs (50-83), ECOG 0/1 (13/39 pts), taxane for CRPC (25 pts/48%). Table below shows details of responses. In stage 1, 13 pts were randomized to D with 0% ORR. D+T advanced to stage 2 with a total of 39 pts enrolled who received a median of 3 cycles (1-27). D+T related AEs were mainly grade 2 or less: fatigue (46%), anorexia (28%), rash (24%), diarrhea (23%), nausea/vomiting (21/18%) and thyroid dysfunction (15%). Most common grade 3/4 AEs: LFTs (8%) and diarrhea (8%). Six pts discontinued treatment due to AEs. There were no grade 5 AEs. There were six ORs (16% (95% CI: 6-32); OR median duration not reached, longest ongoing 25mos+ with PSA < 0.2 ng/ml); all six remain on D, and pain improved in three of four pts. Correlation with full cfDNA panel to be presented.Table:
|Durvalumab||Durvalumab + Tremelimumab|
|ORR||0% (95% CI, 0-25%) 0/13||16% (95% CI, 6-32%) 6/37|
|TMB , 11 mts/Mb ctDNA||≥||0/1||1/2 (50%)|
|CD8 density median||≥||0/5||5/16 (31%)|
|PSA RR||0% (0-25%) 0/13||16% (6-32%) 6/37|
|TTP, median (95% CI), mos||2.1 (1.4, 2.6)||2.6 (1.8, 2.8)|
Based on prespecified criteria, D did not show sufficient clinical activity, but further studies incorporating patient selection by biomarkers are warranted for D+T.
Clinical trial identification
Legal entity responsible for the study
Canadian Cancer Trials Group (CCTG).
S.J. Hotte: Research grant / Funding (institution): AstraZeneca. E. Winquist: Honoraria (self): AstraZeneca. K.N. Chi: Advisory / Consultancy: AstraZeneca. S. Sridhar: Advisory / Consultancy: AstraZeneca. U. Emmenegger: Research grant / Funding (institution): AstraZeneca. C. Canil: Advisory / Consultancy: AstraZeneca. A.R. Hansen: Research grant / Funding (institution): AstraZeneca. L.K. Seymour: Research grant / Funding (institution): AstraZeneca. All other authors have declared no conflicts of interest.