A random walk-based segmentation framework for 3D ultrasound images of the prostate

Ling Ma, Rongrong Guo, Zhiqiang Tian, Baowei Fei

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Purpose: Accurate segmentation of the prostate on ultrasound images has many applications in prostate cancer diagnosis and therapy. Transrectal ultrasound (TRUS) has been routinely used to guide prostate biopsy. This manuscript proposes a semiautomatic segmentation method for the prostate on three-dimensional (3D) TRUS images. Methods: The proposed segmentation method uses a context-classification-based random walk algorithm. Because context information reflects patient-specific characteristics and prostate changes in the adjacent slices, and classification information reflects population-based prior knowledge, we combine the context and classification information at the same time in order to define the applicable population and patient-specific knowledge so as to more accurately determine the seed points for the random walk algorithm. The method is initialized with the user drawing the prostate and non-prostate circles on the mid-gland slice and then automatically segments the prostate on other slices. To achieve reliable classification, we use a new adaptive k-means algorithm to cluster the training data and train multiple decision-tree classifiers. According to the patient-specific characteristics, the most suitable classifier is selected and combined with the context information in order to locate the seed points. By providing accuracy locations of the seed points, the random walk algorithm improves segmentation performance. Results: We evaluate the proposed segmentation approach on a set of 3D TRUS volumes of prostate patients. The experimental results show that our method achieved a Dice similarity coefficient of 91.0% ± 1.6% as compared to manual segmentation by clinically experienced radiologist. Conclusions: The random walk-based segmentation framework, which combines patient-specific characteristics and population information, is effective for segmenting the prostate on ultrasound images. The segmentation method can have various applications in ultrasound-guided prostate procedures.

Original languageEnglish (US)
Pages (from-to)5128-5142
Number of pages15
JournalMedical physics
Volume44
Issue number10
DOIs
StatePublished - Oct 2017
Externally publishedYes

Fingerprint

Prostate
Seeds
Decision Trees
Population Characteristics
Population
Prostatic Neoplasms
Biopsy

Keywords

  • classification
  • context
  • image segmentation
  • prostate cancer
  • random walk
  • transrectal ultrasound (TRUS)

ASJC Scopus subject areas

  • Biophysics
  • Radiology Nuclear Medicine and imaging

Cite this

A random walk-based segmentation framework for 3D ultrasound images of the prostate. / Ma, Ling; Guo, Rongrong; Tian, Zhiqiang; Fei, Baowei.

In: Medical physics, Vol. 44, No. 10, 10.2017, p. 5128-5142.

Research output: Contribution to journalArticle

Ma, Ling ; Guo, Rongrong ; Tian, Zhiqiang ; Fei, Baowei. / A random walk-based segmentation framework for 3D ultrasound images of the prostate. In: Medical physics. 2017 ; Vol. 44, No. 10. pp. 5128-5142.
@article{9f26fe6213fb485288dea2cc4e6aa241,
title = "A random walk-based segmentation framework for 3D ultrasound images of the prostate",
abstract = "Purpose: Accurate segmentation of the prostate on ultrasound images has many applications in prostate cancer diagnosis and therapy. Transrectal ultrasound (TRUS) has been routinely used to guide prostate biopsy. This manuscript proposes a semiautomatic segmentation method for the prostate on three-dimensional (3D) TRUS images. Methods: The proposed segmentation method uses a context-classification-based random walk algorithm. Because context information reflects patient-specific characteristics and prostate changes in the adjacent slices, and classification information reflects population-based prior knowledge, we combine the context and classification information at the same time in order to define the applicable population and patient-specific knowledge so as to more accurately determine the seed points for the random walk algorithm. The method is initialized with the user drawing the prostate and non-prostate circles on the mid-gland slice and then automatically segments the prostate on other slices. To achieve reliable classification, we use a new adaptive k-means algorithm to cluster the training data and train multiple decision-tree classifiers. According to the patient-specific characteristics, the most suitable classifier is selected and combined with the context information in order to locate the seed points. By providing accuracy locations of the seed points, the random walk algorithm improves segmentation performance. Results: We evaluate the proposed segmentation approach on a set of 3D TRUS volumes of prostate patients. The experimental results show that our method achieved a Dice similarity coefficient of 91.0{\%} ± 1.6{\%} as compared to manual segmentation by clinically experienced radiologist. Conclusions: The random walk-based segmentation framework, which combines patient-specific characteristics and population information, is effective for segmenting the prostate on ultrasound images. The segmentation method can have various applications in ultrasound-guided prostate procedures.",
keywords = "classification, context, image segmentation, prostate cancer, random walk, transrectal ultrasound (TRUS)",
author = "Ling Ma and Rongrong Guo and Zhiqiang Tian and Baowei Fei",
year = "2017",
month = "10",
doi = "10.1002/mp.12396",
language = "English (US)",
volume = "44",
pages = "5128--5142",
journal = "Medical Physics",
issn = "0094-2405",
publisher = "AAPM - American Association of Physicists in Medicine",
number = "10",

}

TY - JOUR

T1 - A random walk-based segmentation framework for 3D ultrasound images of the prostate

AU - Ma, Ling

AU - Guo, Rongrong

AU - Tian, Zhiqiang

AU - Fei, Baowei

PY - 2017/10

Y1 - 2017/10

N2 - Purpose: Accurate segmentation of the prostate on ultrasound images has many applications in prostate cancer diagnosis and therapy. Transrectal ultrasound (TRUS) has been routinely used to guide prostate biopsy. This manuscript proposes a semiautomatic segmentation method for the prostate on three-dimensional (3D) TRUS images. Methods: The proposed segmentation method uses a context-classification-based random walk algorithm. Because context information reflects patient-specific characteristics and prostate changes in the adjacent slices, and classification information reflects population-based prior knowledge, we combine the context and classification information at the same time in order to define the applicable population and patient-specific knowledge so as to more accurately determine the seed points for the random walk algorithm. The method is initialized with the user drawing the prostate and non-prostate circles on the mid-gland slice and then automatically segments the prostate on other slices. To achieve reliable classification, we use a new adaptive k-means algorithm to cluster the training data and train multiple decision-tree classifiers. According to the patient-specific characteristics, the most suitable classifier is selected and combined with the context information in order to locate the seed points. By providing accuracy locations of the seed points, the random walk algorithm improves segmentation performance. Results: We evaluate the proposed segmentation approach on a set of 3D TRUS volumes of prostate patients. The experimental results show that our method achieved a Dice similarity coefficient of 91.0% ± 1.6% as compared to manual segmentation by clinically experienced radiologist. Conclusions: The random walk-based segmentation framework, which combines patient-specific characteristics and population information, is effective for segmenting the prostate on ultrasound images. The segmentation method can have various applications in ultrasound-guided prostate procedures.

AB - Purpose: Accurate segmentation of the prostate on ultrasound images has many applications in prostate cancer diagnosis and therapy. Transrectal ultrasound (TRUS) has been routinely used to guide prostate biopsy. This manuscript proposes a semiautomatic segmentation method for the prostate on three-dimensional (3D) TRUS images. Methods: The proposed segmentation method uses a context-classification-based random walk algorithm. Because context information reflects patient-specific characteristics and prostate changes in the adjacent slices, and classification information reflects population-based prior knowledge, we combine the context and classification information at the same time in order to define the applicable population and patient-specific knowledge so as to more accurately determine the seed points for the random walk algorithm. The method is initialized with the user drawing the prostate and non-prostate circles on the mid-gland slice and then automatically segments the prostate on other slices. To achieve reliable classification, we use a new adaptive k-means algorithm to cluster the training data and train multiple decision-tree classifiers. According to the patient-specific characteristics, the most suitable classifier is selected and combined with the context information in order to locate the seed points. By providing accuracy locations of the seed points, the random walk algorithm improves segmentation performance. Results: We evaluate the proposed segmentation approach on a set of 3D TRUS volumes of prostate patients. The experimental results show that our method achieved a Dice similarity coefficient of 91.0% ± 1.6% as compared to manual segmentation by clinically experienced radiologist. Conclusions: The random walk-based segmentation framework, which combines patient-specific characteristics and population information, is effective for segmenting the prostate on ultrasound images. The segmentation method can have various applications in ultrasound-guided prostate procedures.

KW - classification

KW - context

KW - image segmentation

KW - prostate cancer

KW - random walk

KW - transrectal ultrasound (TRUS)

UR - http://www.scopus.com/inward/record.url?scp=85024376407&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85024376407&partnerID=8YFLogxK

U2 - 10.1002/mp.12396

DO - 10.1002/mp.12396

M3 - Article

C2 - 28582803

AN - SCOPUS:85024376407

VL - 44

SP - 5128

EP - 5142

JO - Medical Physics

JF - Medical Physics

SN - 0094-2405

IS - 10

ER -