Transcriptional regulatory remodeling shapes cancer cell states, yet its relationships with genomic and post-translational alterations remain incompletely understood. Here, we developed TregOmic to infer transcriptional regulator activity and integrate it with pan-cancer multi-omics profiles. Across cancer types, aggressive tumor states were associated with reduced lineage-associated activity and increased activity of epithelial-mesenchymal transition (EMT)-related regulators. Copy-number alterations were associated with lineage-associated regulator activity, while several established cancer drivers ranked among mutated genes associated with the largest numbers of regulator activities after covariate adjustment. Phosphosite measurements provided additional information about activity beyond protein abundance. Joint modeling identified distinct combinations of cis- and trans-acting mutations, protein abundance and site-specific phosphorylation associated with regulator activity across cancer contexts. RUNX2 activity was associated with tumor stage and EMT, while RUNX2-pS28 was recurrently associated with RUNX2 activity and showed elevated abundance in multiple tumor cohorts. Together, these findings connect cancer-associated regulatory states to specific molecular alterations and provide a framework for prioritizing candidate regulatory events for functional investigation.