OC-067 An Expression Signature of the Angiogenic Response in Gastrointestinal Neuroendocrine Tumours: Correlation with Tumour Phenotype and Survival outcomes

Abstract

<h3>Introduction</h3> Gastroenteropancreatic neuroendocrine tumours (GEPNETs) are heterogeneous with respect to biologic behaviour and prognosis. Since angiogenesis is a renowned pathogenic hallmark as well as a therapeutic target, we aimed to investigate the prognostic and clinico-pathological role of tissue markers of hypoxia and angiogenesis in GEPNETs. <h3>Methods</h3> Tissue microarray (TMA) blocks were constructed with 86 tumours diagnosed from 1988 to 2010. TMA sections were immunostained for Hypoxia inducible factor 1α (Hif-1α), Vascular Endothelial Growth Factor A (VEGF-A), Carbonic Anhydrase IX (CaIX) and Somatostatin receptors (SSTR) 1 to 3 and Ki-67. Biomarker expression was correlated with clinico-pathological variables and tested for survival prediction using Kaplan-Meier and Cox regression methods. <h3>Results</h3> Eighty-six consecutive cases were included: 51% male, median age 51 (range 16–82), 68% presenting with a pancreatic primary, 95% well differentiated. Forty-four cases (51%) had distant metastases (liver 72%, lymph nodes 28%). Median tumour size was 3.0 cm (range 0.6–12.5). Vascular invasion and necrosis were present in 20 (23%) and 18 (21%) of the specimens. Median overall survival (OS) was 8.8 years (range 0.1–13.5). Overexpression of CaIX was observed in 10% of the specimens, VEGF-A in 78%, Hif-1α in 59%, SSTR1 in 17%, SSTR2 in 31% and SSTR3 in 1%. Ki-67 index was obtained in all cases and scored as G1 in 84%, G2 in 13% and G3 in 4%. SSTR2 overexpression was predominant in pancreatic NETs (p &lt; 0.01), whilst Hif-1α was predominant in non pancreatic NETs (p = 0.05). Higher Ki-67 labelling was associated with larger tumour size (p &lt; 0.001) and necrosis (p = 0.03). Overexpression of Hif-1α and VEGF-A correlated with the presence of liver metastases (p &lt; 0.001). Patients with Ki-67 count &gt; 1% (p = 0.02), high Hif-1α and low SSTR2 expression (p = 0.03) displayed significantly shorter OS times. <h3>Conclusion</h3> We have identified a coherent expression signature by immunohistochemistry that can be used for patient stratification and to optimise treatment decisions in GEPNETs. Tumours with low proliferation index, preserved SSTR2 and low Hif-1α expression have an indolent phenotype and may be offered less aggressive management and less stringent follow up. <h3>Disclosure of Interest</h3> None Declared