Stephanie Pitre-Champagnat, Abstract Co-Author: Nothing to Disclose

Ingrid Leguerney, Abstract Co-Author: Nothing to Disclose

Jacques Bosq, Abstract Co-Author: Nothing to Disclose

Fabian Kiessling MD, Abstract Co-Author: Nothing to Disclose

Benedicte Coiffier, Presenter: Nothing to Disclose

Nathalie Brigitte Lassau MD, PhD, Abstract Co-Author: Speaker, Toshiba Corporation Speaker, Bracco Group Speaker, Novartis AG Speaker, Pfizer Inc Speaker, F. Hoffmann-La Roche Ltd

Parametric maps from raw linear data can be performed in short process time with moving average model, and reflect reliably the heterogeneous histological measures within tumor by considering the contribution of the vessel size in the variations of intratumoral blood volume.

Parametric maps from Dynamic Contrast-Enhanced Ultrasonography (DCE-US) appear as a useful tool to describe the intratumoral vasculature and its heterogeneity. This study was designed to identify the best processing of parametric maps from raw data and to compare the results to histologic vascularity measurement.

DCE-US was performed on 17 melanoma-bearing nude mice after a 0.1mL bolus injection of SonoVue® (Bracco, Italy). Parametric maps treated time intensity curves (TIC) from raw linear data to extract pixelwise two parameters related to blood volume that were area under the curve (AUC) and peak intensity (PI). Three mathematical models were compared to fit the TIC in each pixel: a polynomial model used in clinical routine, a moving average model and a combination of two linear regressions. Parametric maps performed from the best fit approach were compared with histology for both region of interest (ROI) of whole tumor and several subROIs of 15mm2 within each tumor to reflect intratumoral vascular heterogeneity. As ground truth correlate, microvessel densities (MVD) were determined, and vessels size only for subROIs.

The moving average approach was the best compromise between values determination and processing pixelwise time (<1min) than the clinical model (about 2 hours). Among the 17 studied mice, a total of 64 subROIs were analyzed. For the whole tumor ROI set, a significant correlation was found between AUC and PI values and MVD with r=0.51 (p=0.044) and r=0.63 (p =0.008) respectively. In the case of subROIs, a strong correlation was observed between the DCE-US parameters and the MVD from 0-10µm (rAUC=0.74 (p <0.0001) ; rPI=0.83 (p <0.0001)) ; 10-40µm (rAUC=0.94 (p=0.004) ; (rPI=0.96 (p=0.002)) and >40µm (rAUC=0.90 (p=0.012) ; rPI=0.83 (p=0.041)).

Pitre-Champagnat, S, Leguerney, I, Bosq, J, Kiessling, F, Coiffier, B, Lassau, N, Dynamic Contrast-enhanced Ultrasound Parametric Maps for the Evaluation of Intratumoral Vasculature: Preclinical Study.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13019832.html