Abstract Breast cancer is a heterogeneous disease that has the highest incidence and mortality rate among cancers in women worldwide. Breast cancer patients are stratified into three clinical subtypes with different treatment strategies and prognostic values. The development of targeted therapies against the biomarkers that define these strata constitutes one of the precedents of precision oncology, which aims to provide tailored treatments to cancer patients by targeting the molecular alterations found in each tumour. Although this approach has increased patient outcomes, many treatment failure cases still exist. Drug ineffectiveness and relapse have been associated with the coexistence of several malignant subpopulations with different drug sensitivities within the same lesion, a phenomenon known as intratumor heterogeneity. This heterogeneity has been extensively studied from a tumour-centric view, but recently, it has become evident that the tumour microenvironment plays a crucial role in intratumor heterogeneity. However, few studies consider the tumour-microenvironment interplay and its influence on drug sensitivity. In this work, we predict the sensitivity of 10x Visium spatial transcriptomics data from 9 breast cancer patients to >1,200 drugs and verify different response patterns across the tumour, interphase and microenvironment regions. We uncover a sensitivity continuum from the tumour core to the periphery accompanied by a functional gradient. Moreover, we identify conserved therapeutic clusters with distinct response patterns within the tumour region. We link the specific drug sensitivities of each therapeutic cluster to different ligand-receptor interactions that underpin distinct biological functions. Finally, we demonstrate that genetically identical cancer spots may belong to different therapeutic clusters and that this therapeutic heterogeneity is related to their location at the edge or core of tumour ducts. These results highlight the importance of considering the distance to the tumour core and the microenvironment composition when identifying suitable treatments to target intratumor heterogeneity.