Alexandra Reis Borges MD, Presenter: Nothing to Disclose

Ania Benitez, Abstract Co-Author: Nothing to Disclose

Pilar Lopez, Abstract Co-Author: Nothing to Disclose

Gemma Tarduchy, Abstract Co-Author: Nothing to Disclose

Jose Bravo Marques, Abstract Co-Author: Nothing to Disclose

Sebastian Cerdan, Abstract Co-Author: Nothing to Disclose

Laura Barrios, Abstract Co-Author: Nothing to Disclose

Gabriela Gasparinho, Abstract Co-Author: Nothing to Disclose

Manuela Mafra MD, Abstract Co-Author: Nothing to Disclose

To assess the value of MR based biomarkers in the prediction of high grade glioma (HGG) response to antiangiogenic agents and the impact of anti-angiogenic treatment in genetic, metabolic and pathologic profiles of HGG in a preclinical mouse model.

We have serially evaluated 44 mice, 17 during treatment with a mAb against VEGF and 27 controls on a dedicated 7T MR scanner using an orthotopic mouse model of HGG. After tumor implantation MR was performed with 4 days intervals and mice sacrificed at treatment completion or when showing signs of progression. Tumors and normal contralateral brain were studied for histopathology, RT-qPCR and 1H HRMAS. Treatment response was assessed using the RECIST criteria and based on Fischer's linear discriminant analysis predictive models were built of treatment response. Spearman's correlations were obtained between genetic and metabolic profiles of treated and untreated mice.

Among treated mice 10 responded and 7 did not respond to treatment. Response was associated with significant increase in survival and decrease in tumor growth. Decrease mADC, mCBV, mCBF and mMTT and increased T2* were identified as MR biomarkers of response. Fischer's discriminative analysis applied to T2 and DWI image datasets obtained before and on D2 after the 1st treatment separated responders from non-responders with an accuracy of 92%. Gene expression biomarkers of response included underexpression of survivin, caspase 3, HIF1α, hexokynase 2, EGF, integrin α5, VE-cadherin, galectin 3 and MMP13 and overexpression of CXCL12 and SOX1. VEGF-A expression in responders and non-responders did not show a statistically significant difference. Spectroscopic biomarkers of response included decreased levels of lactate, lipids, choline and its metabolites, myoinositol and inhibitory neurotransmitters and increased levels of NAA.

In a mouse model of HGG we identified MR imaging, genetic and spectroscopic biomarkers of response to antiangiogenic treatment. Using Fischer's discriminative analysis T2 and DWI image datasets discriminated responders from non-responders with 92% accuracy as soon as the second day after the 1st treatment.

Timely prediction of tretament response to biologically targeted drugs will allow appropriate selection of patients who will benefit from continued treatment and identify those who will need a different course of action.

Borges, A, Benitez, A, Lopez, P, Tarduchy, G, Marques, J, Cerdan, S, Barrios, L, Gasparinho, G, Mafra, M, Impact of Antiangiogenic Therapy on MR Biomarkers and How They Can Predict Treatment Response.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14009283.html