Monte Carlo Bias Field Correction in Endorectal Diffusion Imaging

• Published in: IEEE transactions on biomedical engineering Vol. 61; no. 2; pp. 368 - 380
• Main Authors: Lui, Dorothy, Modhafar, Amen, Glaister, Jeffrey, Wong, Alexander, Haider, Masoom A.
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.02.2014; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Bias; Bias correction; Coils; Diffusion Magnetic Resonance Imaging - methods; endorectal diffusion imaging; Humans; Image Processing, Computer-Assisted - methods; Magnetic resonance imaging; Male; Medical research; Monte Carlo Method; NMR; Nonhomogeneous media; Nuclear magnetic resonance; Phantoms; Phantoms, Imaging; Principal component analysis; Prostate - pathology; Prostate cancer; prostate cancer (PCa); Prostatic Neoplasms - pathology; Rectum; Signal to noise ratio; Statistics, Nonparametric
• ISSN: 0018-9294; 1558-2531; 1558-2531
• DOI: 10.1109/TBME.2013.2279635

Prostate cancer is one of the leading causes of cancer death in the male population. The detection of prostate cancer using imaging has been challenging until recently. Multiparametric magnetic resonance imaging (MRI) has been shown to allow accurate localization of the cancers and can help direct biopsies to cancer foci, which is required to plan the treatment. The interpretation of MRI, however, requires a high level of expertise and review of large multiparametric datasets. An endorectal receiver coil is often used to improve signal-to-noise ratio and aid in detection of smaller cancer foci. Moreover, computed high b-value diffusion-weighted imaging show improved delineation of tumors but is subject to strong bias fields near the coil. Here, a nonparametric approach to bias field correction for endorectal diffusion imaging via Monte Carlo sampling is introduced. It will be shown that the delineation between the prostate gland and the background and intensity inhomogeneity may be improved using the proposed approach. High b-value generated results also show improved visualization of tumor regions. The results suggest that Monte Carlo bias correction may have potential as a preprocessing tool for endorectal diffusion images for the prostate cancer detection and localization or segmentation.