Background: Prostate Cancer (PC) is a common and highly variable disease. Using current diagnostic methods, the prostate specific antigen (PSA) blood test and histological grading of prostate tissue needle biopsies, it is often difficult to evaluate whether the patient has a PC that requires active treatment or not. The absolute majority of all 10,000 cases of PCs diagnosed annually in Sweden have tumours graded as Gleason score (GS) 6-7 and a PSA value in blood below 10. Many of these are harmless and can be left without active treatment and hence spared problematic post-therapy side-effects, others are highly malignant and require early diagnosis and treatment. Better prognostic markers are needed and the aim of this study was to evaluate prognostic markers and to test if these markers could identify patients with indolent tumours. Methods: We have studied tumour material from 419 men consecutively diagnosed with PC at transurethral resection (1975-1990). The majority of these patients (295) had no metastasis at diagnosis and was not given any curative treatment and only hormonal treatment upon symptoms from metastatic progression. Standard histological sections and tissue microarrays (TMA) from these tumours and surrounding normal prostate tissue were stained and evaluated for cell proliferation (Ki67), blood vessels (endoglin and von Willebrand factor, vWf) and the extracellular matrix component hyaluronan (HA). An orthotopic rat PC model was used to explore hyaluronan staining, hyaluronic acid synthase (HAS)-1 mRNA levels and the effect of local HA treatment on tumour growth. Results: Tumour cell proliferation (Ki67) and the density of intra-tumoural endoglin stained blood vessels were independent prognostic markers (i.e. they added prognostic information to the conventional prognostic markers; clinical stage and GS). None of the GS 6 patients with low staining for both Ki67 and endoglin died of PC within 15 years of follow-up. High HA staining in the tumour epithelium and stroma was a negative prognostic marker of cancer specific survival but they were not independent of GS. High HA staining and high vascular density in the stroma of the surrounding morphologically normal prostate were prognostic for short cancer specific survival. Implantation of tumour cells in the normal rat prostate resulted in an increase in HA and HAS-1 mRNA levels in the prostate tissue surrounding prostate tumours. Concurrently intra-prostatic injection of HA also stimulated tumour growth. Conclusions: By evaluating both tumour cell proliferation (Ki67) and vascular density, it is possible to identify patients with very low risk of cancer specific death in the absence of active treatment. Prostate tumours influence the surrounding non-malignant prostate tissue, for example they cause an increased angiogenesis and synthesis of hyaluronan. Such responses can possibly be used to diagnose PC and to evaluate PC aggressiveness.