1.

Introduction

Dynamic fluorescence molecular tomography (FMT) is a powerful imaging tool that can resolve the three-dimensional (3D) distribution of fluorophores in small animals and has been extensively used for imaging studies of pharmacokinetics parameters.¹ Compartment modeling is a classic approach for pharmacokinetic analysis, and the exchange rates of materials between different compartments are defined as pharmacokinetic parameters.² The combination of dynamic FMT and compartmental modeling can provide 3D images of pharmacokinetic parameters in vivo for tumor detection, drug development, and metabolism studies.³ However, dynamic FMT remains challenging in imaging pharmacokinetic parameters of small animal organs due to the severe ill-posed nature of the inverse problem, making it challenging to determine the location of specific organs and perform the pharmacokinetics analysis.⁴ Furthermore, when different tissues or organs are assigned with the same average optical parameters, inaccurate FMT forward models can also be generated, further reducing the accuracy of the pharmacokinetic parameter analysis.

Previous studies have shown that the prior information of anatomical structures provided by structural imaging techniques such as computer tomography (CT) or magnetic resonance imaging (MRI) can be used to construct more accurate forward physical models by assigning heterogeneous optical parameters to different tissues or organs after segmenting the structural information; Moreover, the structural information can be used to construct a priori models, which can guide and constrain the solution process of the inverse problem and improve the reconstruction quality.^5–7 Therefore, a hybrid dual-modality FMT/CT system has been employed in this work to image the pharmacokinetic parameters.

The liver is an essential organ for studying metabolism parameters in dynamic FMT.⁸ Indocyanine green (ICG) is a clinically approved reagent for liver function assessment by the Food and Drug Administration. Meanwhile, ICG is a widely used optical fluorophore in the near-infrared range for FMT. Therefore, liver tumors can be distinguished in vivo according to the different metabolic properties of ICG in normal and liver cancer cells by pharmacokinetic parameters imaging with the help of dynamic FMT.

In this letter, dynamic FMT is conducted to detect liver tumors in vivo. First, the orthotopic liver tumor mouse model was constructed. Then, 3D imaging of ICG metabolic parameters of the mouse liver was achieved with the help of the FMT/CT dual-modality imaging system and a direct reconstruction algorithm, and liver tumors can be discriminated based on the reconstructed parametric images according to the different metabolic properties of ICG in liver cancer cells. Finally, the parametric imaging results were validated against in vitro anatomical results.

2.

Materials and Methods

2.1.

FMT/CT Dual-Modality Imaging System

The dual-modality system consists of two imaging modalities, FMT and CT, which are placed vertically at a 90-deg angle.⁹ The schematic diagram of the system is shown in Fig. 1. A laser (i) (Asahi Spectra, Torrance, California) is employed to excite the imaged small animal (viii). Full-angle measurements are implemented by a rotation stage working under a step-by-step mode. Data acquisition is performed by an electron-multiplying charge-coupled device camera (iii) (iXon DU-897, Andor Technologies, Belfast, Northern Ireland) coupled with a 35-mm $f/1.6$ lens (iv) (C3514-M, Pentax, Japan). Corresponding excitation filter (ii) and emission filter (v) pairs are employed for data acquisition. The x-ray source (vi) is an XTF5011 X-ray tube (Oxford Instrument, California), and the x-ray detector (vii) is a $1056\times 1056\mathrm{pixel}$ CMOS array detector (Hamamatsu, Shizuoka, Japan) for acquiring CT projection data.

Fig. 1

Schematic diagram of the FMT/CT dual-modality imaging system.

In the dual-modality system, the laser was used as the excitation light source, and the shape of the excitation light source was a linear excitation light with a length of 2 cm and a width of 2 mm. The field of view of the charged-coupled device (CCD) camera was $8.2\mathrm{cm}\times 8.2\mathrm{cm}$. The CCD pixels used for fluorescence projection data acquisition were $2\times 2$, the exposure time was set as 0.5 s, and the excitation light power was 1.5 mW.

3.

Results and Discussion

The anatomical imaging results of the mouse are shown in Fig. 2. Figure 2(a) shows a representative CT projection image of the tumor-bearing mouse. Figure 2(b) shows three crosssectional images of the CT reconstruction results of the tumor-bearing mouse with the heights corresponding to the positions of the red, green, and blue dashed lines in Fig. 2(a), respectively. The whole imaging area of the mouse was manually segmented into lung, liver, kidney, and other tissues [Fig. 2(c)] according to the gray values of the CT images, and these segmentation results will be used as structural a priori information for the metabolic parameters of the FMT reconstruction.

Fig. 2

CT imaging results of the tumor-bearing nude mouse. (a) A representative CT projection image of the tumor-bearing mouse. (b) Three cross-sectional images of the CT results. (c) CT segmentation results of lung, liver, kidney, and other tissues.

Figure 3 shows the dynamic FMT reconstruction results of liver metabolic parameters. Figure 3(a) shows a CT projection, and the green dotted lines indicate the corresponding heights of the reconstructed images in Figs. 3(b)–3(d). Figure 3(b) shows an image of a frozen mouse section obtained after imaging, the height of which was approximately matched to the height of the tomographic results by manual marking. The actual location of the tumor in the liver is shown by the green arrow in Fig. 3(b). The liver tumor region can be clearly identified from the imaging results of metabolic parameters shown in Fig. 3(d). Moreover, based on the reconstructed metabolic parameter images, four metabolic parameter images $A$, $B$, $\alpha$, and $\beta$ can be obtained. By substituting the parameter values into Eq. (2), we can quantitatively recover the ICG metabolic curve of the liver. As shown in Fig. 3(e), the blue curve represents the ICG metabolic curve at the location of the blue point in the normal liver region in Fig. 3(b), and the red curve represents the ICG metabolic curve at the location of the red point in the tumor region in Fig. 3(b). It can be found that compared with the normal liver tissue, the uptake rate of ICG is increased and the excretion rate of ICG is decreased in the tumor area, which is consistent with the theoretical expectation.

Fig. 3

Imaging results of metabolic parameters in liver tumors. (a) CT projection image. (b) Cryosection image of mouse with a green arrow indicating the tumor location. (c) Cross-sectional view of CT imaging results. (d) The metabolic parameter images A (a.u.), B (a.u.), $\alpha$ (${\min }^{-1}$), and $\beta$ (${\min }^{-1}$) of the mouse. (b)–(d) correspond to the heights represented by the green-dashed lines in (a). (e) ICG metabolic curves of the liver.

4.

Conclusion

Pharmacokinetic parameters can describe the dynamics of drug absorption, metabolism and excretion processes and enable quantitative monitoring of physiological properties such as permeability and perfusion rates of tissues and organs. Metabolic parameter imaging technology can visualize the pharmacokinetic parameters of each voxel in the imaging area, allowing a more intuitive description of the physiological processes of drug absorption and excretion in different tissues and organs, which is of great significance for physiological and pathological studies.

In this letter, the pharmacokinetic parameters of ICG have been monitored through a hybrid FMT/CT system and the direct reconstruction algorithm. Although liver tumors can also be directly observed from CT imaging results, as a structural imaging technique, the functional state of their tissues and organs cannot be reflected in the CT images. As a functional imaging technique, quantitative assessment of liver function can be obtained by pharmacokinetic parametric images derived from dynamic FMT. Furthermore, ICG metabolic curves at any location in the liver can be achieved according to the reconstructed pharmacokinetic parameters. Compared with CT structural imaging, the imaging of metabolic parameters with dynamic FMT can obtain quantitative metabolic parameter images, providing a more powerful research tool for organ function assessment, disease diagnosis, and new drug development. Future work will focus on imaging pharmacokinetic parameters of pathological livers and other organs.