A multi-modal prostate segmentation scheme by combining spectral clustering and active shape models

Robert Toth; Pallavi Tiwari; Mark Rosen; Arjun Kalyanpur; Sona Pungavkar; Anant Madabhushi

doi:10.1117/12.770772

11 March 2008 A multi-modal prostate segmentation scheme by combining spectral clustering and active shape models

Robert Toth, Pallavi Tiwari, Mark Rosen, Arjun Kalyanpur, Sona Pungavkar, Anant Madabhushi

Proceedings Volume 6914, Medical Imaging 2008: Image Processing; 69144S (2008) https://doi.org/10.1117/12.770772
Event: Medical Imaging, 2008, San Diego, California, United States

Abstract

Segmentation of the prostate boundary on clinical images is useful in a large number of applications including calculating prostate volume during biopsy, tumor estimation, and treatment planning. Manual segmentation of the prostate boundary is, however, time consuming and subject to inter- and intra-reader variability. Magnetic Resonance (MR) imaging (MRI) and MR Spectroscopy (MRS) have recently emerged as promising modalities for detection of prostate cancer in vivo. In this paper we present a novel scheme for accurate and automated prostate segmentation on in vivo 1.5 Tesla multi-modal MRI studies. The segmentation algorithm comprises two steps: (1) A hierarchical unsupervised spectral clustering scheme using MRS data to isolate the region of interest (ROI) corresponding to the prostate, and (2) an Active Shape Model (ASM) segmentation scheme where the ASM is initialized within the ROI obtained in the previous step. The hierarchical MRS clustering scheme in step 1 identifies spectra corresponding to locations within the prostate in an iterative fashion by discriminating between potential prostate and non-prostate spectra in a lower dimensional embedding space. The spatial locations of the prostate spectra so identified are used as the initial ROI for the ASM. The ASM is trained by identifying user-selected landmarks on the prostate boundary on T2 MRI images. Boundary points on the prostate are identified using mutual information (MI) as opposed to the traditional Mahalanobis distance, and the trained ASM is deformed to fit the boundary points so identified. Cross validation on 150 prostate MRI slices yields an average segmentation sensitivity, specificity, overlap, and positive predictive value of 89, 86, 83, and 93&percent; respectively. We demonstrate that the accurate initialization of the ASM via the spectral clustering scheme is necessary for automated boundary extraction. Our method is fully automated, robust to system parameters, and computationally efficient.

Citation Download Citation

Robert Toth, Pallavi Tiwari, Mark Rosen, Arjun Kalyanpur, Sona Pungavkar, and Anant Madabhushi "A multi-modal prostate segmentation scheme by combining spectral clustering and active shape models", Proc. SPIE 6914, Medical Imaging 2008: Image Processing, 69144S (11 March 2008); https://doi.org/10.1117/12.770772

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available