The PI3K-mTOR-AKT pathway regulates tumour proliferation, gene expression and metabolism, but pathway inhibition induces heterogeneous feedback reactivation, limiting anti-tumour responses. Measuring heterogeneity of pathway inhibition in tissues using protein biomarker phosphorylation or location is challenging. An integrated multi-modal imaging workflow was developed to assess the heterogeneity of AZD2014 (mTORC1/2 inhibitor) response in a PTEN-null renal cancer model. Spatial responses of metabolite biomarkers were analysed by mass spectrometry imaging (MSI). Control and treated tumours were classified according to metabolite-defined regions enriched in control versus AZD2014-treated tumours, respectively. Noticeably, AZD2014-treated tumours retained regions similar to regions dominant in untreated tumours. Imaging mass cytometry analysis of protein biomarkers in 'control-like' regions following AZD2014 treatment showed reduced phospho-S6, indicating suppression, but retained high expression of the glucose transporter GLUT1. Increasing PI3K-AKT inhibition by combining with AZD8186 (PI3Kβ inhibitor) further decreased the control-like metabolic signature, showing PI3K-dependent resistance. This demonstrates that MSI-based workflows yield novel insights into the pharmacodynamic effects of mTORC1/2 inhibition in tumours, which classical biomarkers do not resolve. Coupling these workflows with spatial-omics approaches can deliver greater insights into heterogeneity of treatment response.