Abstract 4891: Gene expression analysis of paired baseline (BL) and on-treatment tumor samples from FAIRLANE, a double-blind placebo (PBO)-controlled randomized phase II trial of neoadjuvant ipatasertib (IPAT) plus paclitaxel (PAC) in early triple-negative breast cancer (eTNBC)

Purpose: In the FAIRLANE trial (NCT02301988) in eTNBC, adding the oral AKT inhibitor IPAT to neoadjuvant PAC led to numerical increases in rates of pathologic complete response (pCR; primary endpoint) and complete response (CR) by MRI in unselected patients (pts), with a numerically greater treatment effect in pts with PIK3CA/AKT1/PTEN-altered tumors [Oliveira et al., AACR 2018]. There was no association between PIK3CA/AKT1/PTEN alterations and response in the PBO arm. Gene expression analysis was performed to evaluate on-treatment changes in the tumor microenvironment and association with response in both treatment arms. Methods: Gene expression at BL and cycle 1 day 8 (C1D8) was evaluated by RNA sequencing using TruSeq RNA Access (Illumina, Inc., San Diego, CA, USA) at Expression Analysis (Morrisville, NC, USA). Paired samples (pts with RNA-seq at both BL and C1D8) were compared to assess on-treatment changes in gene expression. Using RNA-seq data, in silico cell-type enrichment analysis was performed using xCell to calculate the percentage of infiltrating immune cells [Aran et al., Genome Biol 2017]. Results: Paired samples were available from 87 (58%) of the 151 treated pts; response results in this subset were representative of the overall population. Calculated infiltrating immune cell scores (xCell) increased from BL to C1D8 in both treatment arms, exemplified by increases in multiple immune cell biomarkers (eg CD8A, GZMA, CXCL9, CD274/PD-L1). In the PBO arm, higher C1D8 immune score showed a significant association with better response (pCR or response by MRI); the association was stronger for C1D8 than for BL or change from BL immune score. However, there was no association between immune score (BL, C1D8, or change from BL) and response in the IPAT arm. At C1D8, IPAT-treated pts with a CR by MRI had lower average immune scores than PBO-treated pts with a CR. Among the 14 IPAT-treated pts who achieved a CR by MRI, all 7 pts with immune scores below the median had PIK3CA/AKT1/PTEN-altered tumors compared with only 3 of 7 pts with immune scores above the median. Conclusions: This exploratory gene expression analysis revealed differential associations of immune infiltration and tumor response depending on treatment. PAC treatment appears to increase immune cell infiltration in the eTNBC setting, and the presence of a T cell-rich environment (at C1D8) is strongly associated with improved outcomes in pts treated with single-agent PAC. However, the addit