Comparison of six fit algorithms for the intra-voxel incoherent motion model of diffusion-weighted magnetic resonance imaging data of pancreatic cancer patients

• DOI: 10.1371/journal.pone.0194590
• Publication Date: 2018-04-04T13:35:52Z
• Authors: Oliver J. Gurney-...; Remy Klaassen; Martijn Froeling; Sebastiano Barbieri; Jaap Stoker; Marc R. W. Engelb...; Johanna W. Wilmink; Marc G. Besselink; Arjan Bel; Hanneke W. M. van...; Aart J. Nederveen
• Keywords: Intravoxel incoherent motion

The intravoxel incoherent motion (IVIM) model for diffusion-weighted imaging (DWI) MRI data bears much promise as a tool visualizing tumours and monitoring treatment response. To improve the currently poor precision of IVIM, several fit algorithms have been suggested. In this work, we compared performance two Bayesian IVIM four other pancreatic cancer imaging. DWI were acquired in 14 patients during examinations. Three different measures fitting assessed: (i) uniqueness parameters (Spearman's rho); (ii) (within-subject coefficient variation, wCV); (iii) contrast between tumour normal-appearing tissue. For diffusivity D perfusion fraction f, (IVIM-Bayesian-lin) offered best trade-off precision. With exception IVIM-Bayesian-lin, all resulted very pseudo-diffusion D* with wCV more than 50%. approaches were, however, significantly correlated f. Therefore, added value was considered limited patients. easier implemented least squares fixed (IVIM-fixed) performed similar to IVIM-Bayesian-lin f D. conclusion, performing algorithm IVM-Bayesian-lin, but an implement performs similarly