Circulating tumor DNA (ctDNA) monitoring in patients with breast cancer receiving neoadjuvant palbociclib and endocrine therapy: A secondary analysis of the NeoRHEA phase 2 study

601 Background: Personalized circulating tumor DNA (ctDNA) assays are being evaluated in early breast cancer (BC). We investigated the value of ctDNA monitoring in the NeoRHEA study. Methods: NeoRhea (NCT03065621) is a single arm phase 2 study in which patients with estrogen receptor (ER)+/HER2- early breast cancer were treated with neoadjuvant palbociclib + endocrine therapy (ET) for 4 months. Plasma samples were collected at four timepoints: pre-treatment (BL), after the first treatment cycle (C1D28), before surgery (Surgery) and one month post Surgery (End of Study). ctDNA detection was evaluated using the personalized RaDaR assay. Whole-exome sequencing (WES) was performed on BL tumor biopsies followed by a personalized assay development using tumor, plasma and normal DNA in order to track up to 48-patient specific somatic variants in plasma cell-free DNA (cfDNA) using next generation sequencing. Associations between ctDNA detection and BL clinical-pathological characteristics, cell cycle arrest (CCCA) defined as Ki67 ≤2.7% at surgery, residual cancer burden (RCB) rate and breast cancer free survival (BCFS) were investigated. Results: Of 100 patients enrolled, 78 patients and 302 plasma samples were successfully profiled by the RaDaR assay. The number of variants targeted by the assay ranged between 7-48 (median = 25). The median estimated variant allele fraction for ctDNA positive samples was 0.02% (range 0.0009%-0.91%). A total of 42/76 patients were ctDNA positive at BL, 4/76 at C1D28, 4/75 at Surgery and 0/75 at End of Study. Out of 78 patients, 68% were postmenopausal, 78 % had cT2 and 69% cN0 tumors, 20% had multifocal/multicentric tumors, 18% had histological grade 3 tumors, 76% had CCCA (Ki67 not available in 20 patients) and 34% RCB 3 at surgery. ctDNA detection at BL was higher in histological grade 3 tumors (p = 0.03), lower in multifocal/multicentric tumors (p = 0.01), higher in RCB III tumors, (p = 0.01) but was not associated with CCCA. With a median follow-up of 3.8 years (range 1-5 years), 4 patients developed distant and one patient locoregional recurrences. ctDNA detection after one month of treatment (logrank p = 0.02), but not at baseline (p = 0.59) nor at surgery (p = 0.67) was associated with worse BCFS. Conclusions: Detection of ctDNA in early-stage breast cancer is challenging. Our data suggests association of ctDNA detection with some pathological and clinical variables. Independent validation is needed. Clinical trial information: NCT03065621 .