Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Nov 18;12(12):7703-7716.
doi: 10.1364/BOE.443517. eCollection 2021 Dec 1.

Attention mechanism-based locally connected network for accurate and stable reconstruction in Cerenkov luminescence tomography

Xiaoning Zhang  1   2   3 , Meishan Cai  2   4   3 , Lishuang Guo  1   2 , Zeyu Zhang  1   2 , Biluo Shen  2   4 , Xiaojun Zhang  5 , Zhenhua Hu  2   4   6 , Jie Tian  1   2   4   7
Affiliations

Attention mechanism-based locally connected network for accurate and stable reconstruction in Cerenkov luminescence tomography

Xiaoning Zhang et al. Biomed Opt Express. .

Abstract

Cerenkov luminescence tomography (CLT) is a novel and highly sensitive imaging technique, which could obtain the three-dimensional distribution of radioactive probes to achieve accurate tumor detection. However, the simplified radiative transfer equation and ill-conditioned inverse problem cause a reconstruction error. In this study, a novel attention mechanism based locally connected (AMLC) network was proposed to reduce barycenter error and improve morphological restorability. The proposed AMLC network consisted of two main parts: a fully connected sub-network for providing a coarse reconstruction result, and a locally connected sub-network based on an attention matrix for refinement. Both numerical simulations and in vivo experiments were conducted to show the superiority of the AMLC network in accuracy and stability over existing methods (MFCNN, KNN-LC network). This method improved CLT reconstruction performance and promoted the application of machine learning in optical imaging research.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflicts of interest.

Figures

Fig. 1.
Fig. 1.
The overall pipeline of AMLC network. The blue blocks represent the fully connected (FC) sub-network. The yellow blocks represent the locally connected (LC) sub-network. Attention matrix is constructed based on coarse results.
Fig. 2.
Fig. 2.
CLT reconstruction results of single-source. (a-c) show simulation results in different methods, respectively. 3D views and 2D cross sections visually display the reconstruction results. (d) represents BCE of the reconstructed sources, while (e) represents Dice of the reconstructed sources. The mean BCE and Dice of three models are shown in (d) and (e), respectively.
Fig. 3.
Fig. 3.
CLT reconstruction results of distance experiment. (a-c) show simulation results in different gaps varied from 3.5 to 2.5 mm. Both 3D views and 2D cross sections visually display the reconstruction results. (d-e) represent S1 BCE and S2 BCE, respectively.
Fig. 4.
Fig. 4.
CLT reconstruction results of the big-source. (a) shows the 3D views and 2D cross sections in different methods. Quantitative analysis is shown in (b-c).
Fig. 5.
Fig. 5.
CLT reconstruction results of the anti-noise experiment. (a) shows the 3D views and 2D cross sections in different gradients. G0 represents the reconstruction results without noise. G1, G2 and G3 represent independent simulations with 5%,10% and 15% Gaussian noise. Quantitative analysis is shown in (b-c).
Fig. 6.
Fig. 6.
In vivo experimental results. (a-c) show the white light (WL) image, CLI and BLI results, respectively. (d-f) show the in vivo PET results. The yellow arrows show the tumor area. (g-h) represent GFP fluorescence image and the H&E stain result. (i-l) show the reconstruction results merged with MRI to evaluate reconstruction performance.
See this image and copyright information in PMC

Similar articles

See all similar articles

References

    1. Xu Y., Liu H., Cheng Z., “Harnessing the power of radionuclides for optical imaging: Cerenkov luminescence imaging,” J Nucl Med 52(12), 2009–2018 (2011).10.2967/jnumed.111.092965 - DOI - PubMed
    1. Cai M., Zhang Z., Shi X., Yang J., Hu Z., Tian J., “Non-negative iterative convex refinement approach for accurate and robust reconstruction in cerenkov luminescence tomography,” IEEE Trans Med Imaging 39(10), 3207–3217 (2020).10.1109/TMI.2020.2987640 - DOI - PubMed
    1. Ruggiero A., Holland J. P., Lewis J. S., Grimm J., “Cerenkov luminescence imaging of medical isotopes,” J Nucl Med 51(7), 1123–1130 (2010).10.2967/jnumed.110.076521 - DOI - PMC - PubMed
    1. Grootendorst M. R., Cariati M., Kothari A., Tuch D. S., Purushotham A., “Cerenkov luminescence imaging (CLI) for image-guided cancer surgery,” Clin Transl Imaging 4(5), 353–366 (2016).10.1007/s40336-016-0183-x - DOI - PMC - PubMed
    1. Lu M., Zheng S., Zhang X., Guo H., Shi X., Kang X., Qu Y., Hu Z., Tian J., “Cerenkov luminescence imaging on evaluation of early response to chemotherapy of drug-resistant gastric cancer,” Nanomed. Nanotechnol. 14, 205–213 (2016).10.1016/j.nano.2017.10.001 - DOI - PubMed

LinkOut - more resources