Abstract Background : Studies in experimental models have implied perivascular cells (PC) as regulators of multiple aspects of tumor biology, including tumor growth, metastasis, immune cell activity and response to treatment. However, the extent of inter- and intra-case variability of perivascular cells in clinical samples, and potential of perivascular cells as biomarkers, remain poorly characterized. Methods: We have developed an integrated analytical pipeline to investigate vascular and PC status in clinical samples. The analytical procedure uses double-staining with one endothelial cell marker (CD34) and one of four PC markers; platelet-derived growth factor receptor-alpha and -beta (PDGFR-α, PDGFR-β;), smooth muscle α actin (ASMA) and desmin. Digital image-analyses are subsequently used to quantitate PC-related metrics including intensity of marker expression in perivascular areas, and fraction of marker-covered vessels. The approach has been applied to multiple cohorts of clinically well-annotated tumor collections from three cancer types; colorectal cancer (CRC), ovarian cancer (OC) and renal cell cancer (RCC). Results: Analyses of PC status uncovered previously un- recognized independent and heterogeneous expression of the analyzed perivascular markers on different levels (perivascular cells, vessels and tumors). Notably, expression of perivascular markers was largely independent of vessel size and density in all three cancer types. Comparative analyses across tumor types identified differences including e.g. higher abundance of PDGFR-β; positive perivascular cells in CRC. A comparison of perivascular status in matched primary tumors and distant metastases uncovered a large degree of dis-concordance of most vascular characteristics except perivascular PDGFR-β; (CRC and OC cohorts). Importantly, this finding implies that vascular status in metastases cannot, in general, be predicted by analyses of primary tumors. Efforts to uncover the basis for inter-case variations in PC status of CRC revealed strong associations between tumor PDGFR-β; PC status and both PC status in adjacent normal tissue and BRAF status, but not RAS mutations. These findings suggest both host features and cancer mutations as determinants of PC characteristics. Concerning relationships between PC features and survival, we observed significant associations between high perivascular PDGFR-β; status and shorter survival in OC and RCC. These associations remain significant in multivariate analyses including standard clinicopathological characteristics. Impact on response to chemotherapy (CT) treatment was analyzed in both early stage and metastatic CRC. Ongoing analyses indicate that low perivascular PDGFR-βdefines a sub-group with shorter survival in the CT-treated group. Explorative analyses of a small set of bevacizumab-treated metastatic CRC also demonstrated significantly shorter progression free survival in the subset with low perivascular PDGFR-β expression. Conclusion: Digital-image-analyses-supported quantifications of PC features in three different tumor types have revealed heterogeneous expression of PC markers in a manner independent of vessel density and vessel size. Host characteristics and oncogenic status have been identified as candidate determinants of perivascular PDGFR-β; status in CRC. Notably, perivascular PDGFR-β; was identified as a novel independent predictor of survival in OC, RCC and CRC. Treatment-stratified analyses in CRC also implied perivascular PDGFR-β; as a determinant of benefit of chemotherapy and VEGF-directed anti-angiogenic drugs. Citation Format: Sara corvigno, artur mezheyeuski, Magnus Frodin, Maja Bradic Lindh, Tormod Kyrre Guren, Per Pfeiffer, H. Elin Kure, Tone Ikdahl, Eva Skovlund, Kjell Magne Tveit, Kristian Pietras, Anna Portyanko, Ina Hrynchyk, David Edler, Anca Dragomir, friedrik Ponten, Jan Mulder, Camilla Qvortrup, Maria Hallström, Katarina Öhrling, G. Bea A. Wisman, Elisabeth Åvall-Lundqvist, Martin Johansson, Ulrika Harmenberg, Hans Nijman, Per Sandström, Hanna Dahlstrand, Halfdan Sorbye, Bengt Gimelius, Ate Van Der Zee, Lars Egevad, Arne Östman. Marker-defined perivascular cells predict prognosis and response to treatment. [abstract]. In: Proceedings of the AACR Special Conference: Function of Tumor Microenvironment in Cancer Progression; 2016 Jan 7–10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2016;76(15 Suppl):Abstract nr C38.