Inter- and intra-patient variations in tumor microenvironment of serous ovarian cancer are largely unexplored. We aimed to explore potential co-regulation of tumor stroma characteristics, analyze their concordance in primary and metastatic lesions, and study their impact on survival. A tissue microarray (TMA) with 186 tumors and 91 matched metastases was subjected to immunohistochemistry double staining with endothelial cell marker CD34 and fibroblast and pericyte markers α-SMA, PDGFβR and desmin. Images were digitally analyzed to yield "metrics" related to vasculature and stroma features. Intra-case analyses showed that PDGFβR in perivascular cells and fibroblasts were strongly correlated. Similar findings were observed concerning α-SMA. Most stroma characteristics showed large variations in intra-case comparisons of primary tumors and metastasis. Large PDGFβR-positive stroma fraction and high PDGFβFR positive perivascular intensity were both significantly associated with shorter survival in uni- and multi-variate analyses (HR 1.7, 95% CI 1.1-2.5; HR 1.7, 95% CI 1.1-2.8). In conclusion, we found PDGFβR- and α-SMA-expression to be largely independent of each other but concordantly activated in perivascular cells and in fibroblasts within the primary tumor. Stromal characteristics differed between primary tumors and metastases. PDGFβR in perivascular cells and in fibroblasts may be novel prognostic markers in serous ovarian cancer.
Keywords: cancer associated fibroblasts; pericytes; prognosis; serous ovarian cancer; tumor microenvironment.