Deep learning image reconstruction method for limited-angle ultrasound tomography in prostate cancer

Alexis Cheng, Younsu Kim, Emran M.A. Anas, Arman Rahmim, Emad Boctor, Reza Seifabadi, Bradford J. Wood

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Problem: The gold standard for prostate cancer diagnosis is B-mode transrectal ultrasound-guided systematic core needle biopsy. However, cancer is indistinguishable under ultrasound and thus additional costly imaging methods are necessary to perform targeted biopsies. Speed of sound is a potential biomarker for prostate cancer and has the potential to be measured using ultrasound tomography. Given the physical constraints of the prostate's anatomy, this work explores a simulation study using deep learning for limited-angle ultrasound tomography to reconstruct speed of sound. Methods: A deep learning-based image reconstruction framework is used to address the limited-angle ultrasound tomography problem. The training data is generated using the k-wave acoustic simulation package. The general network structure is composed of a series of dense fully-connected layers followed by an encoder and a decoder network. The basic idea behind this neural network is to encode a time of flight map into a lower dimension representation that can then be decoded into a speed of sound image. Results and Conclusions: We show that limited-angle UST is feasible in simulation using an auto-encoder-like DL framework. There was a mean absolute error of 7.5 ± 8.1 m/s with a maximum absolute error of 139.3 m/s. Future validation on experimental data will further assess their ability in improving limited-angle ultrasound tomography.

Original languageEnglish (US)
Title of host publicationMedical Imaging 2019
Subtitle of host publicationUltrasonic Imaging and Tomography
EditorsBrett C. Byram, Nicole V. Ruiter
PublisherSPIE
ISBN (Electronic)9781510625570
DOIs
StatePublished - Jan 1 2019
EventMedical Imaging 2019: Ultrasonic Imaging and Tomography - San Diego, United States
Duration: Feb 17 2019Feb 18 2019

Publication series

NameProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume10955
ISSN (Print)1605-7422

Conference

ConferenceMedical Imaging 2019: Ultrasonic Imaging and Tomography
CountryUnited States
CitySan Diego
Period2/17/192/18/19

Fingerprint

Computer-Assisted Image Processing
image reconstruction
Image reconstruction
learning
Tomography
Prostatic Neoplasms
tomography
Ultrasonics
cancer
Learning
Acoustic wave velocity
coders
acoustics
Biopsy
acoustic simulation
Large-Core Needle Biopsy
Aptitude
biomarkers
decoders
anatomy

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Atomic and Molecular Physics, and Optics
  • Biomaterials
  • Radiology Nuclear Medicine and imaging

Cite this

Cheng, A., Kim, Y., Anas, E. M. A., Rahmim, A., Boctor, E., Seifabadi, R., & Wood, B. J. (2019). Deep learning image reconstruction method for limited-angle ultrasound tomography in prostate cancer. In B. C. Byram, & N. V. Ruiter (Eds.), Medical Imaging 2019: Ultrasonic Imaging and Tomography [1095516] (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 10955). SPIE. https://doi.org/10.1117/12.2512533

Deep learning image reconstruction method for limited-angle ultrasound tomography in prostate cancer. / Cheng, Alexis; Kim, Younsu; Anas, Emran M.A.; Rahmim, Arman; Boctor, Emad; Seifabadi, Reza; Wood, Bradford J.

Medical Imaging 2019: Ultrasonic Imaging and Tomography. ed. / Brett C. Byram; Nicole V. Ruiter. SPIE, 2019. 1095516 (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 10955).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Cheng, A, Kim, Y, Anas, EMA, Rahmim, A, Boctor, E, Seifabadi, R & Wood, BJ 2019, Deep learning image reconstruction method for limited-angle ultrasound tomography in prostate cancer. in BC Byram & NV Ruiter (eds), Medical Imaging 2019: Ultrasonic Imaging and Tomography., 1095516, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 10955, SPIE, Medical Imaging 2019: Ultrasonic Imaging and Tomography, San Diego, United States, 2/17/19. https://doi.org/10.1117/12.2512533
Cheng A, Kim Y, Anas EMA, Rahmim A, Boctor E, Seifabadi R et al. Deep learning image reconstruction method for limited-angle ultrasound tomography in prostate cancer. In Byram BC, Ruiter NV, editors, Medical Imaging 2019: Ultrasonic Imaging and Tomography. SPIE. 2019. 1095516. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE). https://doi.org/10.1117/12.2512533
Cheng, Alexis ; Kim, Younsu ; Anas, Emran M.A. ; Rahmim, Arman ; Boctor, Emad ; Seifabadi, Reza ; Wood, Bradford J. / Deep learning image reconstruction method for limited-angle ultrasound tomography in prostate cancer. Medical Imaging 2019: Ultrasonic Imaging and Tomography. editor / Brett C. Byram ; Nicole V. Ruiter. SPIE, 2019. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE).
@inproceedings{3dde76bd384b4e8db28168bb58bcacbe,
title = "Deep learning image reconstruction method for limited-angle ultrasound tomography in prostate cancer",
abstract = "Problem: The gold standard for prostate cancer diagnosis is B-mode transrectal ultrasound-guided systematic core needle biopsy. However, cancer is indistinguishable under ultrasound and thus additional costly imaging methods are necessary to perform targeted biopsies. Speed of sound is a potential biomarker for prostate cancer and has the potential to be measured using ultrasound tomography. Given the physical constraints of the prostate's anatomy, this work explores a simulation study using deep learning for limited-angle ultrasound tomography to reconstruct speed of sound. Methods: A deep learning-based image reconstruction framework is used to address the limited-angle ultrasound tomography problem. The training data is generated using the k-wave acoustic simulation package. The general network structure is composed of a series of dense fully-connected layers followed by an encoder and a decoder network. The basic idea behind this neural network is to encode a time of flight map into a lower dimension representation that can then be decoded into a speed of sound image. Results and Conclusions: We show that limited-angle UST is feasible in simulation using an auto-encoder-like DL framework. There was a mean absolute error of 7.5 ± 8.1 m/s with a maximum absolute error of 139.3 m/s. Future validation on experimental data will further assess their ability in improving limited-angle ultrasound tomography.",
author = "Alexis Cheng and Younsu Kim and Anas, {Emran M.A.} and Arman Rahmim and Emad Boctor and Reza Seifabadi and Wood, {Bradford J.}",
year = "2019",
month = "1",
day = "1",
doi = "10.1117/12.2512533",
language = "English (US)",
series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",
publisher = "SPIE",
editor = "Byram, {Brett C.} and Ruiter, {Nicole V.}",
booktitle = "Medical Imaging 2019",

}

TY - GEN

T1 - Deep learning image reconstruction method for limited-angle ultrasound tomography in prostate cancer

AU - Cheng, Alexis

AU - Kim, Younsu

AU - Anas, Emran M.A.

AU - Rahmim, Arman

AU - Boctor, Emad

AU - Seifabadi, Reza

AU - Wood, Bradford J.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Problem: The gold standard for prostate cancer diagnosis is B-mode transrectal ultrasound-guided systematic core needle biopsy. However, cancer is indistinguishable under ultrasound and thus additional costly imaging methods are necessary to perform targeted biopsies. Speed of sound is a potential biomarker for prostate cancer and has the potential to be measured using ultrasound tomography. Given the physical constraints of the prostate's anatomy, this work explores a simulation study using deep learning for limited-angle ultrasound tomography to reconstruct speed of sound. Methods: A deep learning-based image reconstruction framework is used to address the limited-angle ultrasound tomography problem. The training data is generated using the k-wave acoustic simulation package. The general network structure is composed of a series of dense fully-connected layers followed by an encoder and a decoder network. The basic idea behind this neural network is to encode a time of flight map into a lower dimension representation that can then be decoded into a speed of sound image. Results and Conclusions: We show that limited-angle UST is feasible in simulation using an auto-encoder-like DL framework. There was a mean absolute error of 7.5 ± 8.1 m/s with a maximum absolute error of 139.3 m/s. Future validation on experimental data will further assess their ability in improving limited-angle ultrasound tomography.

AB - Problem: The gold standard for prostate cancer diagnosis is B-mode transrectal ultrasound-guided systematic core needle biopsy. However, cancer is indistinguishable under ultrasound and thus additional costly imaging methods are necessary to perform targeted biopsies. Speed of sound is a potential biomarker for prostate cancer and has the potential to be measured using ultrasound tomography. Given the physical constraints of the prostate's anatomy, this work explores a simulation study using deep learning for limited-angle ultrasound tomography to reconstruct speed of sound. Methods: A deep learning-based image reconstruction framework is used to address the limited-angle ultrasound tomography problem. The training data is generated using the k-wave acoustic simulation package. The general network structure is composed of a series of dense fully-connected layers followed by an encoder and a decoder network. The basic idea behind this neural network is to encode a time of flight map into a lower dimension representation that can then be decoded into a speed of sound image. Results and Conclusions: We show that limited-angle UST is feasible in simulation using an auto-encoder-like DL framework. There was a mean absolute error of 7.5 ± 8.1 m/s with a maximum absolute error of 139.3 m/s. Future validation on experimental data will further assess their ability in improving limited-angle ultrasound tomography.

UR - http://www.scopus.com/inward/record.url?scp=85066621352&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066621352&partnerID=8YFLogxK

U2 - 10.1117/12.2512533

DO - 10.1117/12.2512533

M3 - Conference contribution

AN - SCOPUS:85066621352

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2019

A2 - Byram, Brett C.

A2 - Ruiter, Nicole V.

PB - SPIE

ER -