Impact of Anatomical Site on RNA-Based Molecular Subtypes in Paired High-Grade Serous Ovarian Carcinoma Samples

Background: High-grade serous ovarian carcinoma (HGSOC) can be subdivided into four prognostic molecular subtypes based on gene expression: C1/Mesenchymal (C1.MES), C2/Immunoreactive (C2.IMM), C4/Differentiated (C4.DIF) and C5/Proliferative (C5.PRO), each representing distinct biological characteristics with immune and stromal microenvironments. PrOTYPE enables prognosis and treatment guidance from biopsy material. Metastatic biopsies are often more accessible than primary adnexal sampling; their utility assumes stable tumor-intrinsic properties relative to the primary. Metastases may diverge due to microenvironmental pressure as well as the site-specific subtype dynamics. Methods: Treatment-naïve HGSOC specimens from 138 patients were profiled using the 55-gene nanostring PrOTYPE assay at adnexal, contralateral adnexal, and/or metastatic sites. Results: Adnexal PrOTYPE yielded expected distributions (21% C1.MES, 31% C2.IMM, 23% C4.DIF, 25% C5.PRO) with moderate reproducibility (κ = 0.49). Same-site replicate analysis showed substantial reproducibility (κ = 0.7). Non-adnexal sites were enriched for immune/mesenchymal subtypes (C1.MES/C2.IMM, 36/63 cases), most prominently at the omentum (24/32 C1.MES). C5.PRO was distinctly underrepresented at non-adnexal sites. Subtype shifts from adnexal to extra-adnexal sites were enriched for the second-place adnexal type prediction (p < 0.001). Detailed 55-gene analysis showed POSTN/CTSK were most commonly upregulated across metastatic sites. EMT pathway enrichment increased with metastatic distance (from adnexa to omentum, adj p < 0.05), paralleling—but independent of—C1.MES predominance. Conclusions: Adnexal PrOTYPE showed good stability. However, non-random subtype shifts and EMT enrichment at metastatic sites suggest dissemination selects pre-existing transcriptional plasticity rather than acquiring states de novo as HGSOC adapts to new microenvironments. Microenvironment changes may help predict metastatic potential and should be considered for precision medicine targeting.