STEAP2 Is a Candidate Marker of a 7q/CNA-Associated Iron-Handling Malignant Epithelial State in Lung Adenocarcinoma

Background/Objectives: Single-gene markers may help translate complex malignant epithelial states into bulk and tissue studies, but they must be distinguished from copy-number surrogates and prognostic biomarkers. We evaluated STEAP2 as a candidate marker of a lung adenocarcinoma (LUAD) malignant epithelial state. Methods: We integrated TCGA-LUAD bulk expression, subtype, copy-number and survival data with GSE131907 author-annotated malignant cells and GSE148071 CNV-high epithelial-like cells. Analyses included sample-level paired comparisons, sequencing-depth-adjusted mixed-effects models, cutoff and continuous-expression sensitivity analyses, and a pre-specified 7q gene benchmark. Results: STEAP2-high TCGA tumors were enriched for the Squamoid expression subtype (59.6% versus 23.3%; odds ratio, 4.69; 95% confidence interval [CI], 2.06-10.69). Iron-handling was the most reproducible single-cell feature: paired STEAP2-positive minus STEAP2-negative differences were 0.779 (95% CI, 0.605-0.952) in GSE131907 and 0.340 (95% CI, 0.304-0.377) in GSE148071, with consistent direction after removing STEAP2 from the score. Metabolic, interferon/major histocompatibility complex and EMT/YAP-like programs were context-dependent. STEAP2 ranked in the top decile of 106 evaluable 7q-window genes (rank fraction, 0.099) and retained state-composite association after adjustment for a 7q copy-number proxy and canonical Squamoid markers. The survival association attenuated after stage adjustment (hazard ratio, 1.17; p = 0.359). Conclusions: STEAP2 is a candidate marker of a 7q/copy-number-associated iron-handling malignant epithelial state. The evidence supports state identification rather than prognosis, therapeutic targeting or causal driver claims.