### Abstract

Purpose To assess the adequacy of a statistical model based on the gamma distribution for diffusion signal decays of prostate cancer (PCa) using b-values ranging up to 2000 sec/mm^{2}, and to evaluate the differences in gamma model parameters for PCa, benign prostatic hyperplasia (BPH), and peripheral zone (PZ). Materials and Methods Twenty-six patients with histologically proven PCa underwent diffusion-weighted magnetic resonance imaging using five b-values (0, 500, 1000, 1500, 2000 sec/mm^{2}). The acquired signal decay curves were fit with both gamma and truncated Gaussian models and a statistical comparison between the two fits was performed. The acquired parameters using the gamma model (mean, standard deviation, the area fraction for D<1.0 mm^{2}/s [Frac<1.0], the area fraction of D>3.0 mm^{2}/s [Frac>3.0]) were compared between PCa, BPH, and PZ. Results The gamma model provided a statistically improved fit over the truncated Gaussian model in PCa. The mean and the standard deviation were significantly lower in PCa than in BPH and PZ (P<0.01). Frac<1.0 was significantly higher in PCa than in BPH and PZ, and Frac>3.0 was significantly lower in PCa than in BPH and PZ (P<0.01). Conclusion A statistical model based on the gamma distribution proved suitable for describing diffusion signal decay curves of PCa. This approach may provide better correlation between diffusion signal decay and histological information in the prostate gland.

Original language | English |
---|---|

Pages (from-to) | 56-62 |

Number of pages | 7 |

Journal | Journal of Magnetic Resonance Imaging |

Volume | 42 |

Issue number | 1 |

DOIs | |

Publication status | Published - 2015 Jul 1 |

### Keywords

- diffusion-weighted imaging
- magnetic resonance imaging
- non-Gaussian diffusion
- prostate neoplasms
- statistical model

### ASJC Scopus subject areas

- Radiology Nuclear Medicine and imaging

## Fingerprint Dive into the research topics of 'Diffusion-weighted imaging of prostate cancer using a statistical model based on the gamma distribution'. Together they form a unique fingerprint.

## Cite this

*Journal of Magnetic Resonance Imaging*,

*42*(1), 56-62. https://doi.org/10.1002/jmri.24761