The Abstract Database hosted by SPL
|
All Abstracts
Upload
Advanced Search
Gallery View
Download Statistics
Help
Import
Log in
Machine Learning for Target Selection in MR-Guided Prostate Biopsy: A Preliminary Study
|
|
Institution: |
1Department of Radiology, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA. 2GE Global Research Center, Niskayuna, New York, USA. |
Publication Date: |
May-2012 |
Citation: |
The International Society for Magnetic Resonance in Medicine 20th Scientific Meeting & Exhibition, 2012, Melbourne, Australia. |
Appears in Collections: |
ISMRM |
Sponsors: |
R01 CA111288 (BRP) P41 RR019703 (NCIGT) P01 CA067165 U01 CA151261 US Army Medical Research and Materiel Command under W81XWH-10-1-0201 |
Generated Citation: |
Moradi M., Fedorov A., Wells III W.M., Tuncali K., Gupta S., Fennessy F., Tempany C.M. Machine Learning for Target Selection in MR-Guided Prostate Biopsy: A Preliminary Study. The International Society for Magnetic Resonance in Medicine 20th Scientific Meeting & Exhibition, 2012, Melbourne, Australia. |
| Downloaded: | 28 times. [view map] |
| Paper: | Download, View online |
| Export citation: |
With an estimated 240,890 cases to be diagnosed in 2011, prostate cancer (PCa) remains among the most common male malignancies with an annual death toll of 33,270 men. Prostate tumors are not visible in the transrectal ultrasound (TRUS) B-mode images used for the guidance of routine biopsies. Therefore, TRUS biopsy is not targeted. There is strong evidence that multiparametric MRI (mpMRI) including Diffusion Weighted (DW) and Dynamic Contrast-Enhanced imaging (DCE) in addition to the T2-weighted MRI can improve the accuracy of PCa localization and staging [1]. We have reported transperineal MR-guided prostate biopsy with highly accurate targeting as a solution to improve the accuracy of the biopsy process in detection of prostate cancer [2, 3]. The targets are selected by a team of three radiologists, based on pre-operative mpMRI (Fig. 1). The objective of the work presented here is to assess the feasibility of quantitative analysis of mpMRI aided by a machine learning approach to assist the target selection process. The idea is to train a classifier on the mpMRI data from radical prostatectomy patients with histologically confirmed PCa. The resulting classifier is used retrospectively on the data from patients that have undergone MR-guided prostate biopsy, in the areas selected as targets by the radiologists, to compute a measure of the probability of cancer and compare with the histological analysis of the core biopsy samples.
Additional Material
1 File (189.495kB)
ISMRM2012_ Moradi_Figure1.jpg (189.495kB)