Abstract Aim: To select tumor biomarkers, included in the AKT pathway, able to predict response to tamoxifen (TAM) in patients with estrogen receptor (ER) positive breast cancer. Patients & Methods: 157 endocrine treatment naïve patients with known ER and HER2 status were treated with TAM: tumour tissue was available either as tissue microarrays (TMA) (n=107) or core biopsies (n=50). After removal of duplicates, ER(-) BCs, & patients without full outcome data 103 patients were available. Ten biomarkers were evaluated by immunohistochemical staining (% positivity and/or H-Score): ER, HER2, Ki67, & phosphorylated forms of MAPK, AKT, ER (Ser118), PRAS40, and IGF1R, as well as pTEN & also androgen receptor (AR) expression. TMA cores were collected from three tumor sites, two central and one peripheral. Tumors were also analysed for AKT1 E17K somatic mutation using BEAMing technology. Patient outcome was assessed by clinical benefit (CB) rate & survival analyses (Time To Progression (TTP) and Time To Death (TTD)). Biomarkers were selected by regression trees to be included in multivariate models for clinical outcomes. False Discovery Rate(FDR) was applied to univariate results. Combinations of pAKT & pPRAS40, Ki67 & ER and ER/AR ratio-were also analysed. Results: Type of Biopsy (Core versus TMA): Generally % positivity was lower in TMAs than cores, for HER2, pER118, pIGF1R and cytoplasmic pAKT and pPRAS40. pAKT, pMAPK and Ki67 were largely unaffected by biopsy type. After FDR testing, significant differences remained for pER118 (p=0.002), pIGF1R (p<0.001), pPRAS40 (p<0.001) and HER2 (P=0.005) positivity levels. Biopsy Location (TMAs): There was no significant difference in % positivity between central & peripheral tumor sites for all biomarkers examined. After FDR correction modest differences were seen for HER2 & pER118 (both p=0.044) only between the two central samples. IHC endpoints versus clinical outcome: Univariate analyses identified only lower levels of pTEN and higher levels of Ki67 (% positivity) as predictive of poor outcome (TTP & TTD) following TAM treatment. In the luminal A (ER+/HER2-) subgroup only Ki67 was predictive of poor outcome (p=0.012). The predictive power of Ki67 for TTP was not improved by combination with pTEN or ER. An inverse ER H-Score association with TTP was confined to the 'luminal A' cohort. AR was not predictive. 71 patients had AR/ER ratio: 61 had ratio <2 (42 had CB & 19 no CB), and 10 had ratio ≥2 (3 CB & 7 no CB) (p<0.02). There was no significant difference in TTP or TTD by AR/ER ratio. Only 2 /146 (<2%) tumors had AKT1 E17K mutations. Conclusions: Biomarker staining cannot be assumed to be the same in TMAs & core biopsies: such data should be combined with caution. Importantly however, there was no evidence of difference in Ki67 between TMA & core samples. There is no evidence to suggest that biomarker staining (including Ki67) at different tumor sites shouldn't be combined, or that ranking of scores differs across such sites. The findings confirmed previous reports that Ki67 & reduced level of PI3K /AKT pathway regulator pTEN can predict resistance to TAM in ER+ BCs. The AKT mutation rate was low. Citation Format: Campbell C, Mathew J, Ellis I, Finlay P, Green A, Bradbury I, Borgquist S, Elebro K, Politz O, Haegebarth A, Gee JM, Robertson JF. Markers of PI3K/Akt, ER & AR pathways and Ki67 expression in relation to tamoxifen outcome in ER+ breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-03-02.