Pancreatic ductal adenocarcinoma (PDAC) exhibits profound molecular heterogeneity and poor prognosis, necessitating novel tailored therapies. The basal and classical subtypes - driven by glycolysis versus lipid metabolism - have distinct prognostic implications warranting further characterization of their underlying transcriptional mechanisms.

Using spatial RNA sequencing we mapped PDAC molecular subtype heterogeneity, capturing spatially-resolved gene expression signatures and generating a comprehensive high-resolution dataset of 42,035 spatial spots. Subtype assignments were validated via multiplex immunofluorescence and quantitative analyses in patient-derived organoids and xenografts.

Our analysis resolved cancer cell signatures, deconvoluted intra-tumoral heterogeneity, and delineated an evolutional classical-to-basal trajectory. We identified metabolically ‘hot’, high-grade tumor niches characterized by concurrent enrichment of glycolysis and lipogenesis across both subtypes, nominating them as subtype-agnostic therapeutic targets. Preclinical models demonstrated that despite the basal subtype’s glycolysis dependence, both classical and basal tumors are susceptible to glycolysis inhibition.

This work demonstrates that metabolic identity, spatial context, and tumor–stroma crosstalk are an inseparable triad that drives PDAC behavior. Our findings show that aggressive metabolic tumor niches can be targeted by glycolysis inhibition in a subtype-agnostic manner, challenging the dogma of subtype-specific therapeutic silos and highlighting highly adaptable energetic niches as reservoirs that drive tumor progression.