Sensitivity Analysis of Reference-Free Quantitative Ultrasound Tissue Classification

Trong Nguyen, Minh N Do, Michael L Oelze

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

Abstract

To implement spectral-based quantitative ultrasound (QUS), currently it is necessary to take a reference scan from a well-characterized tissue-mimicking material for each scanner setting that is used. The purpose of this study was to evaluate if a reference-free approach could be adopted, which would eliminate the need for multiple reference scans while still maintaining the ability to objectively classify different tissue states. Specifically, we utilized a convolutional neural network (CNN) to classify tissues and tissue-mimicking phantoms without taking a reference for each setting used and compared the performance to conventional QUS approaches using a reference phantom. Rabbits that were maintained on a high fat diet for 0, 1, 2, 3 or 6 weeks with five rabbits per diet group (total N = 30) were scanned ultrasonically and classified into two groups based on their liver lipid levels: low fat (8%) and high fat (> 8%). An array transducer L9-4 with center frequency of 4 MHz was used to gather RF backscattered data in vivo from the rabbits. In the conventional QUS approach the RF signals were calibrated from a reference phantom and used to estimate an average BSC for each rabbit. In the reference-free approach, a CNN was trained on the time domain RF signals to classify the rabbit livers. To assess the reliability of the CNN to classify when the settings of the scanner were adjusted, five tissue-mimicking phantoms with different but known properties were scanned under different system settings: power, time gain compensation and number of transmit foci. The CNN was first trained on one system setting and then tested on data acquired from each phantom with the other settings. This was repeated for each individual setting. The testing accuracy of in vivo rabbit liver classification using the CNN without a reference was 73% compared to 60% when using the conventional QUS. The results demonstrated that the CNN can provide accurate and robust classification without having to use a reference for each setting. This work was supported by a grant from the NIH (R21 EB020766).

Original languageEnglish (US)
Title of host publication2018 IEEE International Ultrasonics Symposium, IUS 2018
PublisherIEEE Computer Society
ISBN (Electronic)9781538634257
DOIs
StatePublished - Dec 17 2018
Event2018 IEEE International Ultrasonics Symposium, IUS 2018 - Kobe, Japan
Duration: Oct 22 2018Oct 25 2018

Publication series

NameIEEE International Ultrasonics Symposium, IUS
Volume2018-October
ISSN (Print)1948-5719
ISSN (Electronic)1948-5727

Other

Other2018 IEEE International Ultrasonics Symposium, IUS 2018
CountryJapan
CityKobe
Period10/22/1810/25/18

Fingerprint

sensitivity analysis
rabbits
fats
liver
diets
scanners
lipids
transducers
estimates

Keywords

  • backscatter analysis
  • convolutional neural networks
  • quantitative ultrasound

ASJC Scopus subject areas

  • Acoustics and Ultrasonics

Cite this

Nguyen, T., Do, M. N., & Oelze, M. L. (2018). Sensitivity Analysis of Reference-Free Quantitative Ultrasound Tissue Classification. In 2018 IEEE International Ultrasonics Symposium, IUS 2018 [8579885] (IEEE International Ultrasonics Symposium, IUS; Vol. 2018-October). IEEE Computer Society. https://doi.org/10.1109/ULTSYM.2018.8579885

Sensitivity Analysis of Reference-Free Quantitative Ultrasound Tissue Classification. / Nguyen, Trong; Do, Minh N; Oelze, Michael L.

2018 IEEE International Ultrasonics Symposium, IUS 2018. IEEE Computer Society, 2018. 8579885 (IEEE International Ultrasonics Symposium, IUS; Vol. 2018-October).

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

Nguyen, T, Do, MN & Oelze, ML 2018, Sensitivity Analysis of Reference-Free Quantitative Ultrasound Tissue Classification. in 2018 IEEE International Ultrasonics Symposium, IUS 2018., 8579885, IEEE International Ultrasonics Symposium, IUS, vol. 2018-October, IEEE Computer Society, 2018 IEEE International Ultrasonics Symposium, IUS 2018, Kobe, Japan, 10/22/18. https://doi.org/10.1109/ULTSYM.2018.8579885
Nguyen T, Do MN, Oelze ML. Sensitivity Analysis of Reference-Free Quantitative Ultrasound Tissue Classification. In 2018 IEEE International Ultrasonics Symposium, IUS 2018. IEEE Computer Society. 2018. 8579885. (IEEE International Ultrasonics Symposium, IUS). https://doi.org/10.1109/ULTSYM.2018.8579885
Nguyen, Trong ; Do, Minh N ; Oelze, Michael L. / Sensitivity Analysis of Reference-Free Quantitative Ultrasound Tissue Classification. 2018 IEEE International Ultrasonics Symposium, IUS 2018. IEEE Computer Society, 2018. (IEEE International Ultrasonics Symposium, IUS).
@inproceedings{e0257b90b1ec49eab46b06fcfd9556ae,
title = "Sensitivity Analysis of Reference-Free Quantitative Ultrasound Tissue Classification",
abstract = "To implement spectral-based quantitative ultrasound (QUS), currently it is necessary to take a reference scan from a well-characterized tissue-mimicking material for each scanner setting that is used. The purpose of this study was to evaluate if a reference-free approach could be adopted, which would eliminate the need for multiple reference scans while still maintaining the ability to objectively classify different tissue states. Specifically, we utilized a convolutional neural network (CNN) to classify tissues and tissue-mimicking phantoms without taking a reference for each setting used and compared the performance to conventional QUS approaches using a reference phantom. Rabbits that were maintained on a high fat diet for 0, 1, 2, 3 or 6 weeks with five rabbits per diet group (total N = 30) were scanned ultrasonically and classified into two groups based on their liver lipid levels: low fat (8{\%}) and high fat (> 8{\%}). An array transducer L9-4 with center frequency of 4 MHz was used to gather RF backscattered data in vivo from the rabbits. In the conventional QUS approach the RF signals were calibrated from a reference phantom and used to estimate an average BSC for each rabbit. In the reference-free approach, a CNN was trained on the time domain RF signals to classify the rabbit livers. To assess the reliability of the CNN to classify when the settings of the scanner were adjusted, five tissue-mimicking phantoms with different but known properties were scanned under different system settings: power, time gain compensation and number of transmit foci. The CNN was first trained on one system setting and then tested on data acquired from each phantom with the other settings. This was repeated for each individual setting. The testing accuracy of in vivo rabbit liver classification using the CNN without a reference was 73{\%} compared to 60{\%} when using the conventional QUS. The results demonstrated that the CNN can provide accurate and robust classification without having to use a reference for each setting. This work was supported by a grant from the NIH (R21 EB020766).",
keywords = "backscatter analysis, convolutional neural networks, quantitative ultrasound",
author = "Trong Nguyen and Do, {Minh N} and Oelze, {Michael L}",
year = "2018",
month = "12",
day = "17",
doi = "10.1109/ULTSYM.2018.8579885",
language = "English (US)",
series = "IEEE International Ultrasonics Symposium, IUS",
publisher = "IEEE Computer Society",
booktitle = "2018 IEEE International Ultrasonics Symposium, IUS 2018",

}

TY - GEN

T1 - Sensitivity Analysis of Reference-Free Quantitative Ultrasound Tissue Classification

AU - Nguyen, Trong

AU - Do, Minh N

AU - Oelze, Michael L

PY - 2018/12/17

Y1 - 2018/12/17

N2 - To implement spectral-based quantitative ultrasound (QUS), currently it is necessary to take a reference scan from a well-characterized tissue-mimicking material for each scanner setting that is used. The purpose of this study was to evaluate if a reference-free approach could be adopted, which would eliminate the need for multiple reference scans while still maintaining the ability to objectively classify different tissue states. Specifically, we utilized a convolutional neural network (CNN) to classify tissues and tissue-mimicking phantoms without taking a reference for each setting used and compared the performance to conventional QUS approaches using a reference phantom. Rabbits that were maintained on a high fat diet for 0, 1, 2, 3 or 6 weeks with five rabbits per diet group (total N = 30) were scanned ultrasonically and classified into two groups based on their liver lipid levels: low fat (8%) and high fat (> 8%). An array transducer L9-4 with center frequency of 4 MHz was used to gather RF backscattered data in vivo from the rabbits. In the conventional QUS approach the RF signals were calibrated from a reference phantom and used to estimate an average BSC for each rabbit. In the reference-free approach, a CNN was trained on the time domain RF signals to classify the rabbit livers. To assess the reliability of the CNN to classify when the settings of the scanner were adjusted, five tissue-mimicking phantoms with different but known properties were scanned under different system settings: power, time gain compensation and number of transmit foci. The CNN was first trained on one system setting and then tested on data acquired from each phantom with the other settings. This was repeated for each individual setting. The testing accuracy of in vivo rabbit liver classification using the CNN without a reference was 73% compared to 60% when using the conventional QUS. The results demonstrated that the CNN can provide accurate and robust classification without having to use a reference for each setting. This work was supported by a grant from the NIH (R21 EB020766).

AB - To implement spectral-based quantitative ultrasound (QUS), currently it is necessary to take a reference scan from a well-characterized tissue-mimicking material for each scanner setting that is used. The purpose of this study was to evaluate if a reference-free approach could be adopted, which would eliminate the need for multiple reference scans while still maintaining the ability to objectively classify different tissue states. Specifically, we utilized a convolutional neural network (CNN) to classify tissues and tissue-mimicking phantoms without taking a reference for each setting used and compared the performance to conventional QUS approaches using a reference phantom. Rabbits that were maintained on a high fat diet for 0, 1, 2, 3 or 6 weeks with five rabbits per diet group (total N = 30) were scanned ultrasonically and classified into two groups based on their liver lipid levels: low fat (8%) and high fat (> 8%). An array transducer L9-4 with center frequency of 4 MHz was used to gather RF backscattered data in vivo from the rabbits. In the conventional QUS approach the RF signals were calibrated from a reference phantom and used to estimate an average BSC for each rabbit. In the reference-free approach, a CNN was trained on the time domain RF signals to classify the rabbit livers. To assess the reliability of the CNN to classify when the settings of the scanner were adjusted, five tissue-mimicking phantoms with different but known properties were scanned under different system settings: power, time gain compensation and number of transmit foci. The CNN was first trained on one system setting and then tested on data acquired from each phantom with the other settings. This was repeated for each individual setting. The testing accuracy of in vivo rabbit liver classification using the CNN without a reference was 73% compared to 60% when using the conventional QUS. The results demonstrated that the CNN can provide accurate and robust classification without having to use a reference for each setting. This work was supported by a grant from the NIH (R21 EB020766).

KW - backscatter analysis

KW - convolutional neural networks

KW - quantitative ultrasound

UR - http://www.scopus.com/inward/record.url?scp=85060611310&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85060611310&partnerID=8YFLogxK

U2 - 10.1109/ULTSYM.2018.8579885

DO - 10.1109/ULTSYM.2018.8579885

M3 - Conference contribution

AN - SCOPUS:85060611310

T3 - IEEE International Ultrasonics Symposium, IUS

BT - 2018 IEEE International Ultrasonics Symposium, IUS 2018

PB - IEEE Computer Society

ER -