Prostate Cancer Detection Using Image-Based Features in Dynamic Contrast Enhanced MRI

Liping Wang*, Yuanjie Zheng, Andrik Rampun, Reyer Zwiggelaar

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference Proceeding (Non-Journal item)

90 Downloads (Pure)

Abstract

Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE MRI) provides valuable information in prostate cancer detection. Existing computer-aided detection methods focus on estimating the DCE curves as pharmacokinetic models and directly calculating the perfusion-related measurements from the DCE signals. Substantial image content contained in DCE MRI series, which captures the spatio-temporal pattern receives less attention. This work aims to investigate the performance of the image-based features extracted from DCE MRI on prostate cancer detection. Various image-based features are extracted from DCE MRI series. Their performance on prostate cancer detection is compared with features extracted from the pharmacokinetic models and the perfusion-related measurements. Features are concatenated and feature selection is applied to reduce the feature dimensionality and improve cancer detection performance. Evaluation is based on a publicly available dataset. Using image-based features outperforms using either the features extracted from the pharmacokinetic models or the perfusion-related measurements. By applying feature selection to the aggregation of all features, the performance of prostate cancer detection achieves 0.821, for the area under the receiver operating characteristics curve. This study demonstrates that compared with the commonly used pharmacokinetic models and the perfusion-related features, image-based features provide an additional contribution to prostate cancer detection and can potentially be used as an alternative approach to model DCE MRI.

Original languageEnglish
Title of host publicationMedical Image Understanding and Analysis - 25th Annual Conference, MIUA 2021, Proceedings
EditorsBartłomiej W. Papież, Mohammad Yaqub, Jianbo Jiao, Ana I. Namburete, J. Alison Noble
PublisherSpringer Nature
Pages43-55
Number of pages13
Volume12722
ISBN (Print)9783030804312
DOIs
Publication statusPublished - 12 Jul 2021
Event25th Annual Conference on Medical Image Understanding and Analysis, MIUA 2021 - Virtual, Online
Duration: 12 Jul 202114 Jul 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12722 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference25th Annual Conference on Medical Image Understanding and Analysis, MIUA 2021
CityVirtual, Online
Period12 Jul 202114 Jul 2021

Keywords

  • DCE MRI
  • Image-based features
  • Prostate cancer detection

Fingerprint

Dive into the research topics of 'Prostate Cancer Detection Using Image-Based Features in Dynamic Contrast Enhanced MRI'. Together they form a unique fingerprint.

Cite this