3-D Multi-parametric Contrast-Enhanced Ultrasound for the Prediction of Prostate Cancer

Rogier R Wildeboer; Ruud J G van Sloun; Pintong Huang; Hessel Wijkstra; Massimo Mischi

doi:10.1016/j.ultrasmedbio.2019.05.017

3-D Multi-parametric Contrast-Enhanced Ultrasound for the Prediction of Prostate Cancer

Ultrasound Med Biol. 2019 Oct;45(10):2713-2724. doi: 10.1016/j.ultrasmedbio.2019.05.017. Epub 2019 Jul 10.

Authors

Rogier R Wildeboer¹, Ruud J G van Sloun², Pintong Huang³, Hessel Wijkstra⁴, Massimo Mischi²

Affiliations

¹ Lab of Biomedical Diagnostics, Department of Electrical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands. Electronic address: [email protected].
² Lab of Biomedical Diagnostics, Department of Electrical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands.
³ Department of Ultrasonography, Second Affiliated Hospital of Zhejiang University, Hangzhou, China.
⁴ Lab of Biomedical Diagnostics, Department of Electrical Engineering, Eindhoven University of Technology, Eindhoven, The Netherlands; Department of Urology, Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, The Netherlands.

PMID: 31300222
DOI: 10.1016/j.ultrasmedbio.2019.05.017

Abstract

Trans-rectal ultrasound-guided 12-core systematic biopsy (SBx) is the standard diagnostic pathway for prostate cancer (PCa) because of a lack of sufficiently accurate imaging. Quantification of 3-D dynamic contrast-enhanced ultrasound (US) might open the way for a targeted procedure in which biopsies are directed at lesions suspicious on imaging. This work describes the expansion of contrast US dispersion imaging algorithms to 3-D and compares its performance against malignant and benign disease. Furthermore, we examined the feasibility of a multi-parametric approach to predict SBx-core outcomes using machine learning. An area under the receiver operating characteristic (ROC) curve of 0.76 and 0.81 was obtained for all PCa and significant PCa, respectively, an improvement over previous US methods. We found that prostatitis, in particular, was a source of false-positive readings.

Keywords: 3-D; Contrast ultrasound dispersion imaging; Dynamic contrast-enhanced ultrasound; Machine learning; Prostate cancer; Systematic biopsy.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Contrast Media*
Humans
Image Enhancement / methods*
Imaging, Three-Dimensional / methods*
Male
Predictive Value of Tests
Prostate / diagnostic imaging
Prostatic Neoplasms / diagnostic imaging*
Reproducibility of Results
Ultrasonography / methods*

Substances

Contrast Media