The influence of Holmium-166 loaded microspheres on diffusion weighted imaging : an ex-vivo study

Introduction Intra-arterial radioembolization (RE), using microspheres that contain a β-emitting radioisotope, is an emerging technique for the treatment of either primary or metastatic liver cancer [1]. One type of microspheres that has been developed for RE and that is currently under investigation in a phase I clinical trial [2] is composed of neutron activated holmium-166 embedded in poly L-lactic acid microspheres (Ho-PLLA-MS). Because of the highly paramagnetic character of holmium, these microspheres are suggested to be an ideal device for MR guided therapy [3]. One part of such an MR guided approach would consist of MRI based treatment follow up to measure tumor response. A promising MR based approach to assess tumor response is functional imaging by means of Diffusion Weighted Imaging (DWI). It has been shown that DWI enables to distinguish between benign and malignant hepatic lesions, between hepatic abscess and tumor and between necrotic and viable tumor tissue and that DWI can assess tumor response [4]. To measure diffusion, characterized by the apparent diffusion coefficient (ADC), usually a pulsed-gradient spin-echo sequence [5], is used. In this sequence, a pair of (identical) linear gradients is applied symmetric around the π-pulse of a conventional spin-echo sequence. Static spins that are dephased by the first gradient will be rephased by the second gradient but diffusing spins will not entirely be rephased by the second gradient as a consequence of their motion through the applied gradients, resulting in signal loss. The amount of diffusion weighting is controlled by the b-factor: b = (γGδ) · (Δ δ/3), where γ is the gyromagnetic ratio, G is the gradient strength, δ is the duration of the gradients and Δ the interval between the pulse centers. Using a range of b-values and fitting the mono-exponential function S = S0 · exp[-b · ADC] to the acquired signal intensities, the ADC can be determined. However, assessment of the ADC following this procedure is only valid if there are no other gradients present in the tissue. Susceptibility variations in tissue induced by the paramagnetic Ho-PLLA-MS will cause field variations leading to non-linear local gradients. Due to these background gradients which add to the diffusion-sensitizing gradients, the MR signal will reduce because of spins diffusing in the local gradients and this can influence the ADC measurements as has been demonstrated using a phantom setup [6]. The purpose of this work was to explore if this influence is large enough to hamper treatment response measurements after radioembolization. This was done using an ex-vivo pig liver into which a non-radioactive therapeutic amount of Ho-PLLA-MS was administered after which ADC changes were measured.