Caractérisation des propriétés viscoélastiques du placenta par élastographie ultrasonore transitoire bidimensionnelle
THÈSE dirigée par :
REMENIERAS Jean-Pierre Ingénieur de recherche - HDR, Inserm, Université de Tours
PERROTIN Franck PU-PH, Inserm, Université de Tours
RAPPORTEURS :
ROZENBERG Patrick PU-PH, Université de Versailles
TSATSARIS Vassilis PU-PH, Université Paris Descartes
JURY :
CALLE Samuel MCU - HDR, CNRS, Université de Tours
GENNISSON Jean-Luc Chargé de recherche, CNRS, U-PSud, CEA, Orsay
PERROTIN Franck PU-PH, Inserm, Université de Tours
REMENIERAS Jean-Pierre Ingénieur de recherche - HDR, Inserm, Université de Tours
ROZENBERG Patrick PU-PH, Université de Versailles
TSATSARIS Vassilis PU-PH, Université Paris Descartes
WINER Norbert PU-PH, Université de Nantes
Abstract
Screening and diagnosis of placental insufficiency (PI), whether intrauterine growth restriction (IUGR) or preeclampsia (PE) are major public health issues. In clinical practice, the mechanical properties of the placenta are not explored; however changes in its tissue architecture could cause variations in elasticity. Among the ultrasound (US) elastography methods, transient elastography seemed suitable for such an application. This technique consists in calculating the shear wave speed (Cs) generated by an external vibration propagating in the medium under consideration. Elasticity values obtained from current US methods are calculated at a single frequency. As a structural modification of the tissue may correspond to a particular power law of frequency dispersion of Cs, we evaluated the interest of a multifrequency approach to distinguish the elasticity of healthy placentas and that of placentas with PI signs in the third trimester of pregnancy. We have developed a preliminary plane wave device (for ex vivo exploration) to validate the principle of the proposed method, and then a 2D transient elastography device (ex vivo and in vivo exploration). The data is adjusted using a fractional rheological model where the frequency behavior is modeled by a power law (exponent n of the model). We have shown that IUGR placentas have Cs and n values lower than those of healthy placentas or PE. This decrease in the n value could be explained by histopathological lesions of IUGR. As for the decrease of Cs in cases of IUGR, this result is consistent with the study of an IUGR rat model by uterine ligation. Finally, the frequency dispersion analysis is feasible in pregnant women. The added value of this method should now be tested in a large clinical study.