Dynamic changes in circulating tumour DNA (ctDNA) levels may predict long-term outcome. We utilised samples from a phase I/II randomised trial (BEECH) to assess ctDNA dynamics as a surrogate for progression free survival (PFS) and early predictor of drug efficacy.
Patients with oestrogen receptor positive advanced metastatic breast cancer (ER+ mBC) in the BEECH study, paclitaxel plus placebo versus paclitaxel plus AKT inhibitor capivasertib, had plasma samples collected for ctDNA analysis at baseline and at multiple timepoints in the development cohort (safety run-in, part A) and validation cohort (randomised, part B). Baseline sample ctDNA sequencing identified mutations for longitudinal analysis, and mutation specific digital droplet PCR (ddPCR) assays were utilised to assess change in ctDNA abundance (allele fraction) between baseline and 872 on-treatment samples. Primary objective was to assess whether early suppression of ctDNA, based on pre-defined criteria from the development cohort, independently predicted outcome in the validation cohort.
In the development cohort, suppression of ctDNA was apparent after 8 days of treatment (p=0.014), with cycle 2 day 1 (4 weeks) identified as the optimal timepoint to predict PFS from early ctDNA dynamics. In the validation cohort, median PFS was 11.1 months in patients with suppressed ctDNA at 4 weeks and 6.4 months in patients with high ctDNA (HR = 0.20, 95% CI 0.083 – 0.50, p<0.0001). There was no difference in the level of ctDNA suppression between patients randomised to capivasertib or placebo overall (p=0.904) nor in the PIK3CA mutant subpopulation (p=0.071). Clonal haematopoiesis of indeterminate potential (CHIP) was evident in 30% (18/59) baseline samples, although CHIP had no effect on tolerance of chemotherapy nor on PFS.
Early on-treatment ctDNA dynamics are a surrogate for PFS. Dynamic ctDNA assessment has the potential to substantially enhance early drug development.