Background: The treatment of HER2+ breast cancer (BC) has improved substantially with the use of targeted therapies, and patients achieving pathological complete response (pCR) following neoadjuvant therapy have significantly improved disease free survival. However, there is currently no way to determine which patients are most likely to respond to neoadjuvant HER2-targeting treatment. Growing evidence indicates tumour infiltrating lymphocytes influence HER2+ BC outcome. We hypothesize that a deeper investigation of the immune landscape of HER2+ patient tumours may improve patient stratification and identify an immune profile associated with pCR. Methods: 3plex immunohistochemistry panels identifying immune cell subpopulations and tertiary lymphoid structures were applied to HER2+ BC tumors (pre-treatment N=19, post-treatment N=9) of patients of the neoadjuvant TCHL clinical trial (NCT01485926), who received HER2-targeting therapy. Quantitative analysis of CD3+, CD20+, DC-Lamp+, CD4+, and CD8+ immune cells on multi-labelled whole-slide images of TCHL tumor sections was performed using artificial intelligence image analysis models. Patient-specific densities of each cell type in tumor regions were determined to facilitate assessment of pre-treatment differences in target immune subpopulations in complete, partial, and non-responders. Results: Statistical analysis demonstrated a trend for higher densities of each immune subpopulation in pre-treatment biopsies of pCR vs. non-pCR patients. Patient categorisation based on median threshold demonstrated patients achieving pCR are more frequently high for each immune cell type assessed, while non-pCR patients are more frequently in the low category. Investigation of association of pre-treatment immune infiltrate with clinicopathological features demonstrated trends for higher levels of T-cell infiltration in later stage and hormone receptor negative patients. Conclusions: Multiplex IHC is an optimal technique for the concurrent assessment of multiple target markers in patient tissue, which can be quantitatively analyzed in a high throughput manner using AI models. The findings of this preliminary study support recent evidence regarding the role of tumor infiltrating immune cells in HER2+ BC outcome and warrant further investigation in a larger cohort.