Elsevier

Magnetic Resonance Imaging

Volume 47, April 2018, Pages 111-117
Magnetic Resonance Imaging

Original contribution
Diffusion kurtosis imaging of a human nasopharyngeal carcinoma xenograft model: Initial experience with pathological correlation

https://doi.org/10.1016/j.mri.2017.12.012Get rights and content

Abstract

Purpose

The aim of this study was to investigate the relationship between diffusion kurtosis imaging (DKI)-related parameters and pathological measures using human nasopharyngeal carcinoma (NPC) xenografts in a nude mouse model.

Materials and methods

Twenty-six BALB/c-nu nude mice were divided into two groups that were injected with two different nasopharyngeal squamous cell carcinoma cell lines (CNE1 and CNE2). DK magnetic resonance (MR) imaging was performed on a 3.0 Tesla MR scanner. DWI and DKI-related parameters, including apparent diffusion coefficient (ADC), mean diffusivity (MD) and mean kurtosis (MK) were measured. Mice were euthanatized when the maximum diameter of the primary tumor reached 1.5 cm after MR scanning. Tumor tissues were then processed for hematoxylin and eosin staining. The pathological images were analyzed using a computer-aided pixel-wise clustering method to evaluate tumor cellular density, nuclei portion, cytoplasm portion, extracellular space portion, the ratio of nuclei to cytoplasm and the ratio of nuclei to extracellular space. The relationships between DWI and DKI-related parameters and pathological features were analyzed statistically.

Results

The ADC and MD values of the CNE1 group (1.16 ± 0.24 × 10 3 mm2/s, 2.28 ± 0.29 × 10 3 mm2/s) was higher than that of the CNE2 group (0.82 ± 0.14 × 10 3 mm2/s, 1.53 ± 0.24 × 10 3 mm2/s, P < 0.001), but the MK values between the two groups were not significantly different (CNE1: 0.55 ± 0.14; CNE2: 0.47 ± 0.23; P > 0.05). A Pearson test showed that the ADC and MD values were significantly correlated with cellular density, nuclei portion, extracellular space portion and the ratio of nuclei to extracellular space (r =  0.861; − 0.909, P < 0.001; r =  0.487; 0.591, P < 0.05; r = 0.567; 0.625, P < 0.05; r =  0.645; − 0.745, P < 0.001, respectively). The MK values were significantly correlated with nuclei portion, cytoplasm portion and the ratio of nuclei to cytoplasm (r =  0.475, P < 0.05; r = 0.665, P < 0.001; r =  0.494, P < 0.05, respectively).

Conclusion

The preliminary animal results suggest that DKI findings can provide valuable bio-information for NPC tissue characterization. DKI imaging might be utilized as a surrogate biomarker for the non-invasive assessment of tumor microstructures.

Introduction

In recent decades, diffusion-weighted imaging (DWI) has become a well-established and widely used magnetic resonance (MR) functional imaging technique that enables the quantitative assessment of the water diffusion behavior in tissue. DWI is routinely used in daily MR exams as a conventional protocol, especially in oncology [1], [2]. The apparent diffusion coefficient (ADC), which is a quantitative biomarker of water diffusion in a microscopic environment, can be used to detect malignant lesions, assess tumor invasiveness, and even predict early treatment responses [1], [2], [3].

The theoretical premise of the DWI model was to assume that the water molecules in the body showed a normal distribution of diffusion, with the water motion in tissues following a Gaussian diffusion behavior. In fact, the diffusion of water molecules in biological tissues is more complex than the diffusion of free water due to the influence of diffusion barriers such as cell membranes and intracellular spaces. The displacement of water molecules in tissues deviates from the Gaussian distribution. Therefore, an advanced diffusion method based on a non-Gaussian diffusion model, which is known as diffusion kurtosis imaging (DKI), has been developed to account for the non-Gaussian diffusion behavior. This emerging technique was first described by Jensen et al. in 2005 [4]. Subsequent studies have suggested that DKI may better assess the complexity of microstructural environments and may provide more detailed information on tissue heterogeneity, vascularity and cellularity than conventional DWI [4], [5], [6], [7]. Although this Non-Gaussian diffusion model was primarily and largely applied for neural applications in the past [6], [7], increased studies have recently explored DKI to oncology [6], [7], [8], [9], [10], [11], [12], including Nasopharyngeal Carcinoma (NPC), which is one of the most common malignancies in Southeast Asia. Yuan et al. [13] attempted DKI using a maximal b value of 1500 s/mm2 in NPC, showing better fitting of DWI signal than the mono-exponential model. Lu et al. [14] and Wang et al. [15] have evaluated the use of DKI in early detection of radiation-induced temporal lobe necrosis in patients of NPC. Both studies supported that DKI might be a more sensitive imaging marker for detecting the microstructural abnormalities in temporal lobe before they were visible on conventional MRI. Another study showed a significant correlation between histogram parameters derived from DKI and the clinical stage of NPC [16]. In our previous study, we had successfully applied DKI in the pre-treatment evaluation of the NPC, and found that DKI might be superior to mono-exponential DWI in the early prediction of neoadjuvant chemotherapy (NAC) response in patients with locally advanced NPC [17].

Although these DKI-related researches on oncology all suggest that DKI can quantify changes in non-Gaussian water distribution to evaluate pathophysiological changes, few of them have investigated the relationships between DKI parameters and pathophysiological features. The aim of this study was to investigate the relationship between DKI-related parameters and pathological measures using human NPC xenografts in a nude mouse model.

Section snippets

Ethics statement

All mouse procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Fujian Cancer Hospital and performed according to institutional policies.

Human nasopharyngeal carcinoma xenograft models

Male BALB/c-nu nude mice that were four weeks old (Slack Laboratory Animal Co., Shanghai, China) were used for all the experiments. Twenty-six mice were divided into two groups that were injected with two different nasopharyngeal squamous cell carcinoma cell lines (CNE1 and CNE2). The human nasopharyngeal carcinoma cell lines

Results

All 26 mice had successful MR scans. For DK images, the tumor lesions were clearly shown as an increased signal intensity corresponding to the T2-weighted images(Fig. 1). It can be clearly seen that the non-Gaussian kurtosis analysis fits the data points better than does the monoexponential model. For pathological image analyses, all the HE-stained tissues sections were eligible and qualified for further analysis. Fig. 2 shows a typical HE tissue sample with the automated micro-anatomic

Discussion

Connecting DKI findings with tissue microstructures and pathophysiology can help us better understand the biological relevance of DKI parameters. Our research explored the potential utility of DKI for the characterization of tissue pathological features. The results of our study showed that the ADC and MD values were strongly and negatively correlated with CD and the N/ECS ratio, moderately and negatively correlated with the nuclei area portion, and positively correlated with ECS. The MK value

Limitations

Our study has several limitations. The ROIs we placed on DW/DK images were drawn manually on the maximum cross-sectional area of the tumor, but the HE images were selected randomly by the pathologists. Therefore, we could not associate the HE sections directly with the DW/DKI sections, which may affect the representativeness of tumor heterogeneity. Another limitation is the uneven color distribution of the HE images, and bias may occur on the boundaries between the nuclei, cytoplasm and ECS

Conclusions

In conclusion, the preliminary animal results demonstrated significant correlations between DKI parameters and tumor pathological features. These results provide insights into how the tissue microenvironment affects the MD and MK values, which also suggests that DKI findings can provide valuable bio-information for NPC tissue characterization. DKI can be utilized as a surrogate biomarker for the non-invasive assessment of tumor microstructures.

Acknowledgements

This work is a joint research project co-operated by Fujian Cancer Hospital and Fujian Normal University. The authors would like to thank every laboratory member involved in this cooperation. The project was supported by the Fujian Provincial Department of Science and Technology (NO. 2016Y0018 to Jing Zhong, NO. 2014Y0013 to Yunbin Chen), the National Science Foundation of China under Grant (NO. 61501121 to Peng Shi), and the National Clinical Key Specialty Construction Program of China.

The

References (32)

  • J. Yu et al.

    Correlation of standard diffusion-weighted imaging and diffusion kurtosis imaging with distant metastases of rectal carcinoma

    J Magn Reson Imaging

    (2015)
  • D. Wu et al.

    Characterization of breast tumors using diffusion kurtosis imaging (DKI)

    Plos One

    (2013)
  • J. Yuan et al.

    Non-Gaussian analysis of diffusion weighted imaging in head and neck at 3T: a pilot study in patients with nasopharyngeal carcinoma

    Plos One

    (2014)
  • L. Liyan et al.

    Diffusion kurtosis as an in vivo imaging marker of early radiation-induced changes in radiation-induced temporal lobe necrosis in nasopharyngeal carcinoma patients

    Clin Neuroradiol

    (2017)
  • W. Dan et al.

    Diffusion kurtosis imaging study on temporal lobe after nasopharyngeal carcinoma radiotherapy

    Brain Res

    (2016)
  • X.Q. Xu et al.

    Histogram analysis of diffusion kurtosis imaging of nasopharyngeal carcinoma: correlation between quantitative parameters and clinical stage

    Oncotarget

    (2017)
  • 1

    The first two authors contributed equally to this work.

    View full text