Computer-aided diagnosis (CAD) systems are being increasingly used in clinical setting to report multi-parametric magnetic resonance imaging (mp-MRI) of the prostate. Usually, CAD systems automatically highlight cancer suspicious regions to the radiologist, reducing reader variability and interpretation errors. Nevertheless, implementing this software requires selection of which mp-MRI parameters could better discriminate between malignant and non-malignant regions. To exploit functional information, some parameters are derived from dynamic contrast-enhanced (DCE) acquisitions. In particular, many CAD software employed pharmacokinetic features, as Ktrans and kep, derived from Tofts model, to estimate a likelihood map of malignancy. However, non-pharmacokinetic models could be also used to describe DCE-MRI curves, without involving any prior knowledge, or measuring the arterial input function, which could lead to potentially large errors in parameters estimation. In this work, we implemented an empirical function derived from the phenomenological universalities (PUN) class to fit DCE-MRI. Parameters of the PUN model are used in combination with T2-weighted and diffusion-weighted acquisitions to fed a support vector machine classifier to produce a voxelwise malignancy likelihood map of the prostate. All results were compared to a CAD system based on Tofts pharmacokinetic features to describe DCE-MRI curves, using different quality aspects of image segmentation, also evaluating number and size of false positive (FP) candidate regions. This study included 61 patients with 70 biopsy-proven prostate cancers (PCa). The metrics used to evaluate segmentation quality between the two CAD systems were not statistically different, although the PUN-based CAD reported a lower number of FP, with reduced size compared to the Tofts-based CAD. In conclusion, the CAD software based on PUN parameters is feasible to detect PCa, without affecting segmentation quality, hence, it could be successfully applied in clinical settings, improving the automated diagnosis process, and reducing computational complexity.
Physics in medicine and biology. 2018 Mar 23 [Epub ahead of print]
Simone Mazzetti, Valentina Giannini, Filippo Russo, Daniele Regge
Department of Surgical Sciences , Università degli Studi di Torino, Turin, ITALY., Department of Surgical Sciences, Università degli Studi di Torino, Turin, ITALY., Radiology, Candiolo Cancer Institute - FPO, IRCCS, Candiolo, Turin, ITALY.