Assessment of collimators in radium-223 imaging with channelized Hotelling observer: a simulation study

Research output: Contribution to journalArticle

Abstract

Objective: Radium-223 ( 223 Ra) is used in unsealed radionuclide therapy for metastatic bone tumors. The aim of this study is to apply a computational model observer to 223 Ra planar images, and to assess the performance of collimators in 223 Ra imaging. Methods: The 223 Ra planar images were created via an in-house Monte Carlo simulation code using HEXAGON and NAI modules. The phantom was a National Electrical Manufacturers Association body phantom with a hot sphere. The concentration of the background was 55 Bq/mL, and the sphere was approximately 1.5–20 times that of the background concentration. The acquisition time was 10 min. The photopeaks (and the energy window) were 84 (full width of energy window: 20%), 154 (15%), and 270 keV (10%). Each 40 images, with and without hot concentration, were applied to a three-channel difference-of-Gaussian channelized Hotelling observer (CHO), and the signal-to-noise ratio (SNR) of the hot region was calculated. The images were examined using five different collimators: two low-energy general-purpose (LEGP), two medium-energy general-purpose (MEGP), and one high-energy general-purpose (HEGP) collimators. Results: The SNR value was linearly proportional to the contrast of the hot region for all collimators and energy windows. The images of the 84-keV energy window with the MEGP collimator that have thicker septa and larger holes produced the highest SNR value. The SNR values of two LEGP collimators were approximately half of the MEGP collimators. The HEGP collimator was halfway between the MEGP and LEGP. Similar characteristics were observed for other energy windows (154, 270 keV). The SNR value of images captured via the 270-keV energy window was larger than 154-keV, although the sensitivity of the 270-keV energy window is lower than 154-keV. The results suggested a positive correlation between the SNR value and the fraction of unscattered photons. Conclusions: The SNR value of CHO reflected the performance of collimators and was available to assess and quantitatively evaluate the collimator performance in 223 Ra imaging. The SNR value depends on the magnitudes of unscattered photon count and the fraction of unscattered photon count. Consequently, in this study, MEGP collimators performed better than LEGP and HEGP collimators for 223 Ra imaging.

Original languageEnglish
Pages (from-to)649-657
Number of pages9
JournalAnnals of Nuclear Medicine
Volume32
Issue number10
DOIs
Publication statusPublished - Dec 1 2018

Fingerprint

Radium
Signal-To-Noise Ratio
Photons
Radioisotopes
Bone and Bones

All Science Journal Classification (ASJC) codes

  • Radiology Nuclear Medicine and imaging

Cite this

@article{920a1e9558164739893658a6cda6519f,
title = "Assessment of collimators in radium-223 imaging with channelized Hotelling observer: a simulation study",
abstract = "Objective: Radium-223 ( 223 Ra) is used in unsealed radionuclide therapy for metastatic bone tumors. The aim of this study is to apply a computational model observer to 223 Ra planar images, and to assess the performance of collimators in 223 Ra imaging. Methods: The 223 Ra planar images were created via an in-house Monte Carlo simulation code using HEXAGON and NAI modules. The phantom was a National Electrical Manufacturers Association body phantom with a hot sphere. The concentration of the background was 55 Bq/mL, and the sphere was approximately 1.5–20 times that of the background concentration. The acquisition time was 10 min. The photopeaks (and the energy window) were 84 (full width of energy window: 20{\%}), 154 (15{\%}), and 270 keV (10{\%}). Each 40 images, with and without hot concentration, were applied to a three-channel difference-of-Gaussian channelized Hotelling observer (CHO), and the signal-to-noise ratio (SNR) of the hot region was calculated. The images were examined using five different collimators: two low-energy general-purpose (LEGP), two medium-energy general-purpose (MEGP), and one high-energy general-purpose (HEGP) collimators. Results: The SNR value was linearly proportional to the contrast of the hot region for all collimators and energy windows. The images of the 84-keV energy window with the MEGP collimator that have thicker septa and larger holes produced the highest SNR value. The SNR values of two LEGP collimators were approximately half of the MEGP collimators. The HEGP collimator was halfway between the MEGP and LEGP. Similar characteristics were observed for other energy windows (154, 270 keV). The SNR value of images captured via the 270-keV energy window was larger than 154-keV, although the sensitivity of the 270-keV energy window is lower than 154-keV. The results suggested a positive correlation between the SNR value and the fraction of unscattered photons. Conclusions: The SNR value of CHO reflected the performance of collimators and was available to assess and quantitatively evaluate the collimator performance in 223 Ra imaging. The SNR value depends on the magnitudes of unscattered photon count and the fraction of unscattered photon count. Consequently, in this study, MEGP collimators performed better than LEGP and HEGP collimators for 223 Ra imaging.",
author = "Akihiko Takahashi and Shingo Baba and Masayuki Sasaki",
year = "2018",
month = "12",
day = "1",
doi = "10.1007/s12149-018-1286-4",
language = "English",
volume = "32",
pages = "649--657",
journal = "Annals of Nuclear Medicine",
issn = "0914-7187",
publisher = "Springer Japan",
number = "10",

}

TY - JOUR

T1 - Assessment of collimators in radium-223 imaging with channelized Hotelling observer

T2 - a simulation study

AU - Takahashi, Akihiko

AU - Baba, Shingo

AU - Sasaki, Masayuki

PY - 2018/12/1

Y1 - 2018/12/1

N2 - Objective: Radium-223 ( 223 Ra) is used in unsealed radionuclide therapy for metastatic bone tumors. The aim of this study is to apply a computational model observer to 223 Ra planar images, and to assess the performance of collimators in 223 Ra imaging. Methods: The 223 Ra planar images were created via an in-house Monte Carlo simulation code using HEXAGON and NAI modules. The phantom was a National Electrical Manufacturers Association body phantom with a hot sphere. The concentration of the background was 55 Bq/mL, and the sphere was approximately 1.5–20 times that of the background concentration. The acquisition time was 10 min. The photopeaks (and the energy window) were 84 (full width of energy window: 20%), 154 (15%), and 270 keV (10%). Each 40 images, with and without hot concentration, were applied to a three-channel difference-of-Gaussian channelized Hotelling observer (CHO), and the signal-to-noise ratio (SNR) of the hot region was calculated. The images were examined using five different collimators: two low-energy general-purpose (LEGP), two medium-energy general-purpose (MEGP), and one high-energy general-purpose (HEGP) collimators. Results: The SNR value was linearly proportional to the contrast of the hot region for all collimators and energy windows. The images of the 84-keV energy window with the MEGP collimator that have thicker septa and larger holes produced the highest SNR value. The SNR values of two LEGP collimators were approximately half of the MEGP collimators. The HEGP collimator was halfway between the MEGP and LEGP. Similar characteristics were observed for other energy windows (154, 270 keV). The SNR value of images captured via the 270-keV energy window was larger than 154-keV, although the sensitivity of the 270-keV energy window is lower than 154-keV. The results suggested a positive correlation between the SNR value and the fraction of unscattered photons. Conclusions: The SNR value of CHO reflected the performance of collimators and was available to assess and quantitatively evaluate the collimator performance in 223 Ra imaging. The SNR value depends on the magnitudes of unscattered photon count and the fraction of unscattered photon count. Consequently, in this study, MEGP collimators performed better than LEGP and HEGP collimators for 223 Ra imaging.

AB - Objective: Radium-223 ( 223 Ra) is used in unsealed radionuclide therapy for metastatic bone tumors. The aim of this study is to apply a computational model observer to 223 Ra planar images, and to assess the performance of collimators in 223 Ra imaging. Methods: The 223 Ra planar images were created via an in-house Monte Carlo simulation code using HEXAGON and NAI modules. The phantom was a National Electrical Manufacturers Association body phantom with a hot sphere. The concentration of the background was 55 Bq/mL, and the sphere was approximately 1.5–20 times that of the background concentration. The acquisition time was 10 min. The photopeaks (and the energy window) were 84 (full width of energy window: 20%), 154 (15%), and 270 keV (10%). Each 40 images, with and without hot concentration, were applied to a three-channel difference-of-Gaussian channelized Hotelling observer (CHO), and the signal-to-noise ratio (SNR) of the hot region was calculated. The images were examined using five different collimators: two low-energy general-purpose (LEGP), two medium-energy general-purpose (MEGP), and one high-energy general-purpose (HEGP) collimators. Results: The SNR value was linearly proportional to the contrast of the hot region for all collimators and energy windows. The images of the 84-keV energy window with the MEGP collimator that have thicker septa and larger holes produced the highest SNR value. The SNR values of two LEGP collimators were approximately half of the MEGP collimators. The HEGP collimator was halfway between the MEGP and LEGP. Similar characteristics were observed for other energy windows (154, 270 keV). The SNR value of images captured via the 270-keV energy window was larger than 154-keV, although the sensitivity of the 270-keV energy window is lower than 154-keV. The results suggested a positive correlation between the SNR value and the fraction of unscattered photons. Conclusions: The SNR value of CHO reflected the performance of collimators and was available to assess and quantitatively evaluate the collimator performance in 223 Ra imaging. The SNR value depends on the magnitudes of unscattered photon count and the fraction of unscattered photon count. Consequently, in this study, MEGP collimators performed better than LEGP and HEGP collimators for 223 Ra imaging.

UR - http://www.scopus.com/inward/record.url?scp=85051194837&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051194837&partnerID=8YFLogxK

U2 - 10.1007/s12149-018-1286-4

DO - 10.1007/s12149-018-1286-4

M3 - Article

C2 - 30073570

AN - SCOPUS:85051194837

VL - 32

SP - 649

EP - 657

JO - Annals of Nuclear Medicine

JF - Annals of Nuclear Medicine

SN - 0914-7187

IS - 10

ER -