Follicular lymphoma (FL) is a clinically heterogeneous disease requiring the need for easily quantifiable prognostic biomarkers. Micro-vessel density has been shown to have prognostic significance in some, but not all studies. Previous analyses has been based on simple numerical assessment of vessels within histological sections, providing a relatively thin (3–5μm) ‘snap-shot’ of what is a three-dimensional, branching network. This inherent limitation, coupled with methodological variation in its assessment, has lead to conflicting results and uncertainty of its prognostic value in many malignancies including follicular lymphoma. To determine if assessment of true tumor neovascularisation through angiogenic sprouting may be of more clinical relevance, we performed immunostaining with two routinely used endothelial cell markers (CD31 and CD34) in an established FL tissue microarray (TMA). After initial analysis, we focused attention on the vessels at the smallest end of the spectrum seen within routine thickness sections. These represent small, single staining structures no greater than 30 μm2 in area. We subsequently used extended focal imaging within thicker sections to trace these vascular structures and confirmed them to be blind-ending angiogenic sprouts. Diagnostic biopsies taken from patients at the extremes of survival of FL were analysed with respect to numbers of these sprouts, and revealed higher angiogenic activity in patients who died from lymphoma progression less than 5 years after diagnosis compared with those surviving greater than 15 years (p=0.025). This effect was only seen with CD31 and not CD34. Image overlay analysis of serial sections demonstrated that lymphatic vessels highlighted with LYVE-1, a specific lymphatic endothelial marker were positive with CD31 and negative for CD34. However, no differences between number or extent of sprouting of lymphatic vessels were seen in the two prognostic groups; therefore revealing true vascular angiogenic sprouting seen with CD31 analysis, and demonstrating that these vascular angiogenic sprouts express CD31, but less frequently express CD34. We further characterised these angiogenic sprouts using double-labelling immunofluorescence to assess pericyte coverage. Results indicated there was largely no pericyte coverage of these vascular structures, suggesting that these vessels may be targeted using anti-angiogenic therapy and not protected by pericytes. The increased angiogenic activity seen in the poorer prognostic subgroup was seen only in the inter-follicular regions and not in the neoplastic follicles. It is therefore unlikely that the increased vascularisation is a direct result of tumour cell- driven angiogenesis as a closer spatial relationship between tumour cells and vessels would be expected. Previous studies have highlighted that increased lymphoma-associated macrophages are associated with adverse outcome; their role in promoting angiogenesis has been well studied. We therefore used automated image analysis to assess numbers of CD163+, an M2 type macrophage marker identifying a subset of lymphoma-associated macrophages. Although there was no difference in absolute number of macrophages seen between the two groups, there was a positive correlation between number of these cells and extent of angiogenic sprouting. Last, we assessed the impact of angiogenic sprouting and time to transformation and identified a trend towards increased angiogenic activity in those patients who transformed within three years of diagnosis. In summary, we have used an improved gauge of angiogenic activity by quantifying angiogenic sprouts in TMA and in routine histological sections, and highlighted the impact of angiogenesis on survival and time to transformation in patients with FL. Further investigation into the mechanisms driving increased angiogenesis and its subsequent impact on survival is currently being undertaken in a validation series using a TMA of 450 patients with FL at our institution.
Disclosures: No relevant conflicts of interest to declare.