Compton Background Elimination for in Vivo X-Ray Fluorescence Imaging of Gold Nanoparticles Using Convolutional Neural Network

Jung, Sung-Eun; Lee, Jimin; Cho, Hyungjoo; Kim, Taeyun; Ye, Sung-Joon

doi:10.1109/tns.2020.3023420

Cited by 7 publications

(9 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…19,20 In the same study, about 0.5 wt% of GNP accumulated in the kidneys in mice were detected by the XRF imaging system. 19,20 To summarize the experimental results of K-XFCT system or K-XRF imaging system, 0.10-0.25 wt% GNPs were detectable in 2.5-6.4-cm-diam. phantom with 120-140 kVp X-rays.…”

Section: Introductionmentioning

confidence: 81%

“…vial in a 2.5‐cm‐diam. phantom using 140 kVp fan‐beam X‐rays and two‐dimensional (2D) pixelated detector 19,20 . In the same study, about 0.5 wt% of GNP accumulated in the kidneys in mice were detected by the XRF imaging system 19,20 .…”

Section: Introductionmentioning

confidence: 90%

“…phantom using 140 kVp fan‐beam X‐rays and two‐dimensional (2D) pixelated detector 19,20 . In the same study, about 0.5 wt% of GNP accumulated in the kidneys in mice were detected by the XRF imaging system 19,20 . To summarize the experimental results of K‐XFCT system or K‐XRF imaging system, 0.10–0.25 wt% GNPs were detectable in 2.5–6.4‐cm‐diam.…”

Section: Introductionmentioning

confidence: 93%

See 2 more Smart Citations

Monte Carlo modeling of gold nanoparticles detection limits of benchtop three‐dimensional L‐ and K‐shell X‐ray fluorescence mapping systems

Jung

2022

X-Ray Spectrometry

Self Cite

View full text Add to dashboard Cite

This study aims to investigate the detection limits of gold nanoparticle (GNP) concentrations by Monte Carlo (MC) modeling of benchtop polychromatic Kand L-shell X-ray fluorescence mapping system. In Monte Carlo N-Particle (MCNP version 6.1) simulations, a 0.25-cm-diameter cylinder containing GNPs of various concentrations (i.e., 0.005%-1.0% gold by weight, wt%) was assumed to be located at the center of a cylindrical water phantom of various diameters (1.0-10 cm). Two different sets of incident pencil beam X-rays and detectors were modeled to stimulate X-ray fluorescence (XRF) of GNPs: (1) 62 kVp and silicon drift detector for L-XRF, (2) 105 kVp and cadmium telluride detector for K-XRF. The detection limits were calculated for given radiation doses to the center of phantom (375-1500 mGy). When the diameter of the phantom was 1 cm, the detection limits for L-XRF and K-XRF were an order of 0.001 wt % and of 0.01 wt%, respectively. The detectability of K-XRF turned out to be superior to that of L-XRF for the phantoms greater than or equal to 3 cm in diameter. The MC results will provide a guide for developing an optimal benchtop XRF imaging system for in vivo preclinical imaging, depending on the sizes of GNP-loaded objects, GNP concentrations, and radiation doses.

show abstract

Section: Introductionmentioning

confidence: 81%

Section: Introductionmentioning

confidence: 90%

Section: Introductionmentioning

confidence: 93%

See 1 more Smart Citation

Monte Carlo modeling of gold nanoparticles detection limits of benchtop three‐dimensional L‐ and K‐shell X‐ray fluorescence mapping systems

Jung

2022

X-Ray Spectrometry

Self Cite

View full text Add to dashboard Cite

show abstract

“…The current pinhole XRF imaging system can acquire XRF images within 2 min, but the detection limit needs to be improved to at least 0.01 wt%. In order to overcome the limitations of the direct subtraction method, the 2D convolutional neural network (CNN) model for Compton background elimination can be used to acquire XRF images without preinjection scanning 37 . Therefore, the image acquisition time and radiation dose can be reduced by half while maintaining the quality of XRF images.…”

Section: Discussionmentioning

confidence: 99%

Dual imaging modality of fluorescence and transmission X‐rays for gold nanoparticle‐injected living mice

et al. 2022

Self Cite

View full text Add to dashboard Cite

Background: X-ray fluorescence (XRF) imaging for metal nanoparticles (MNPs) is a promising molecular imaging modality that can determine dynamic biodistributions of MNPs. However, it has the limitation that it only provides functional information. Purpose: In this study, we aim to show the feasibility of acquiring functional and anatomic information on the same platform by demonstrating a dual imaging modality of pinhole XRF and computed tomography (CT) for gold nanoparticle (GNP)-injected living mice. Methods: By installing a transmission CT detector in an existing pinhole XRF imaging system using a two-dimensional (2D) cadmium zinc telluride (CZT) gamma camera, XRF and CT images were acquired on the same platform. Due to the optimal X-ray spectra for XRF and CT image acquisition being different, XRF and CT imaging were performed by 140 and 50 kV X-rays, respectively. An amount of 40 mg GNPs (1.9 nm in diameter) suspended in 0.20 ml of phosphate-buffered saline were injected into the three BALB/c mice via a tail vein. Then, the kidney and tumor slices of mice were scanned at specific time points within 60 min to acquire time-lapse in vivo biodistributions of GNPs. XRF images were directly acquired without image reconstruction using a pinhole collimator and a 2D CZT gamma camera. Subsequently, CT images were acquired by performing CT scans. In order to confirm the validity of the functional information provided by the XRF image, the CT image was fused with the XRF image. After the XRF and CT scan, the mice were euthanized, and major organs (kidneys, tumor, liver, and spleen) were extracted. The ex vivo GNP concentrations of the extracted organs were measured by inductively coupled plasma mass spectrometry (ICP-MS) and L-shell XRF detection system using a silicon drift detector, then compared with the in vivo GNP concentrations measured by the pinhole XRF imaging system. Results: Time-lapse XRF images were directly acquired without rotation and translation of imaging objects within an acquisition time of 2 min per slice. Due to the short image acquisition time,the time-lapse in vivo biodistribution of GNPs was acquired in the organs of the mice. CT images were fused with the XRF images and successfully confirmed the validity of the XRF images. The difference in ex vivo GNP concentrations measured by the L-shell XRF detection system and ICP-MS was 0.0005-0.02% by the weight of gold (wt%). Notably, the in vivo and ex vivo GNP concentrations in the kidneys of three mice were comparable with a difference of 0.01-0.08 wt%. Conclusions: A dual imaging modality of pinhole XRF and CT imaging system and L-shell XRF detection system were successfully developed. The developed systems are a promising modality for in vivo imaging and ex vivo

show abstract

“…The application of a convolutional neural network to in vivo XRF images of Au NPs obtained using a benchtop instrument to eliminate Compton-scattered photons was reported by Jung et al 302 The XRF imaging system comprised a 2D CdZnTe gamma camera, a pinhole collimator and fan-beam polychromatic X-rays. An architecture of the 2D convolutional neural network model for Compton background elimination was optimally designed and trained with data sets obtained by the measurements of water only and Au NP-embedded imaging phantoms.…”

Section: Nanostructuresmentioning

confidence: 99%

Atomic spectrometry update: review of advances in the analysis of metals, chemicals and materials

Carter

Clough

Fisher

et al. 2021

J. Anal. At. Spectrom.

View full text Add to dashboard Cite

show abstract

Compton Background Elimination for in Vivo X-Ray Fluorescence Imaging of Gold Nanoparticles Using Convolutional Neural Network

Cited by 7 publications

References 25 publications

Monte Carlo modeling of gold nanoparticles detection limits of benchtop three‐dimensional L‐ and K‐shell X‐ray fluorescence mapping systems

Monte Carlo modeling of gold nanoparticles detection limits of benchtop three‐dimensional L‐ and K‐shell X‐ray fluorescence mapping systems

Dual imaging modality of fluorescence and transmission X‐rays for gold nanoparticle‐injected living mice

Atomic spectrometry update: review of advances in the analysis of metals, chemicals and materials

Contact Info

Product

Resources

About