1691

Differentiating Between Malignant and Benign Bone Tumors: Combined Diagnostic Value of DWI, IVIM and DKI

Ying Li ¹, Cuiping Ren¹, Wenhua Zhang¹, Jingliang Cheng¹, Dandan Zheng², and Liangjie Lin²
¹The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China, ²Clinical & Technical Support, Philips Healthcare, Beijing, 100102, China, Beijing, China

Synopsis

Keywords: Other Musculoskeletal, Tumor

Motivation: The integration of various diffusion models can yield more precise insights into tumor diffusion and perfusion.

Goal(s): To investigate the utility of combining DWI, IVIM and DKI in the differential diagnosis of benign and malignant bone tumors.

Approach: Six parameters derived from three diffusion models of 107 patients were statistically compared through either the independent samples t-test or Mann-Whitney U test. Diagnostic performance was assessed using ROC curves for both individual examinations and their combined analysis in distinguishing between benign and malignant tumors.

Results: The combination of ADC, D, D*, MK, and MD values had better diagnostic efficacy than these parameters alone.

Impact: Statistically significant differences were observed in the values of ADC, D, D*, MD, and MK between benign and malignant bone tumors. Furthermore, the combination of ADC, D, D*, MK, and MD values had better diagnostic efficacy than these parameters alone.

Introduction

Despite being among the rarest cancer types, malignant bone tumors present a significant threat to human health due to their elevated mortality, disability, and metastasis rates. DWI captures biological information and quantify the dispersion of water molecules within tissue. Nonetheless, due to the intricate cellular microstructural barriers within tumors, water diffusion within tumors can be complex. Intravoxel Incoherent Motion (IVIM) is a bi-exponential model that concurrently reflects genuine water molecule diffusion within tissue and microvascular perfusion. This reveals the intricate microstructural properties of tumor tissue. Diffusion kurtosis imaging (DKI) can quantitively estimate the skewed distribution of water diffusion through a non-Gaussian distribution model which were more sensitive to heterogeneous and irregular tumor intracellular environments^[1]. The objective of this study is to investigate the utility of combining DWI, IVIM and DKI in the differential diagnosis of benign and malignant bone tumors.

Methods

107 bone tumor patients (70 males and 37 females, age ranged from 7 to 73 years) including 41 benign and 66 malignant cases diagnosed based on pathological biopsy were included in this study. MR imaging was performed on a 3.0-T clinical scanner (Ingenia CX, Philips Healthcare, Best, the Netherlands) with an eighteen-channel phased-array coil. All patients underwent STIR, DWI, IVIM (b=0、10、30、50、75、100、150、300、500、800、1000、1500s/mm², NEX=4) and DKI (b=0、800、1500s/mm², diffusion gradient directions=15) MRI. Considering the heterogeneity of the tumor, the radiologist manually defined the region of interest (ROI) of the tumor contour based on DWI images of the largest plane of the tumor. The ROI was then automatically replicated to other parameter maps, including apparent diffusion coefficient (ADC) for DWI, pseudo-diffusion (D*), perfusion fraction (f) and diffusion coefficient (D) for IVIM, and mean diffusion coefficient (MD) and mean kurtosis (MK) for DKI. These parameters were compared using either the independent samples t-test or Mann-Whitney U test according to whether the data follows a normal distribution. Pathological diagnosis served as the standard for our analysis. Additionally, Receiver Operating Characteristic (ROC) curves were generated to assess the combined discrimination of benign and malignant bone tumors. A P value < 0.05 was considered statistically significant.

Results

The ADC, D, D* and MD of malignant bone tumors exhibited lower values compared to benign tumors, whereas the f and MK values were marginally elevated. Statistically significant differences were observed in the values of ADC, D, D*, MD, and MK between benign and malignant bone tumors (Table 1). The representative images were shown in Figure 1 and 2. ROC analysis demonstrated that the area under the curve (AUC) for ADC, D, D*, MK, and MD in distinguishing between benign and malignant bone tumors was 0.817, 0.765, 0.624, 0.745, and 0.718, respectively. The integrative analysis combining D, D*, MK, MD, and ADC values enhanced the diagnostic accuracy, yielding an AUC of 0.854, sensitivity 92.68% and specificity 62.12%(Table 2 and Figure 3).

Discussion

The utility of D* and f for diagnosis has been contentious in prior research^[2]. The variance observed in the f within this study may stem from the substantial number of malignant bone tumors, the diversity of their pathological types, the uneven distribution, and the marked disparities in the vascular architecture among tumors of differing malignancies^[3]. Combining multi-diffusion models yields a higher diagnostic efficacy than any single one. Such an integrative approach more accurately captures the nuances of true diffusion dynamics and the intricate water molecule movement within biological tissues, offering a detailed representation of the tumor microstructure's complexity and its subtle alterations^[4].

Conclusions

In this study, statistically significant differences were observed in the values of ADC, D, D*, MD, and MK between benign and malignant bone tumors. Furthermore, the combination of ADC, D, D*, MK, and MD values had better diagnostic efficacy than these parameters alone.

Acknowledgements

No acknowledgement found.

References

[1] Le Bihan D, Iima M. Diffusion Magnetic Resonance Imaging: What Water Tells Us about Biological Tissues. PLoS Biol ,2015, 13(7): e1002203.

[2] Polverari G, Ceci F, Bertaglia V, et al. 18 F-FDG pet parameters and radiomics features analysis in advanced NSCLC treated with immunotherapy as predictors of therapy response and survival[J]. Cancers, 2020, 12(5): 1163.

[3] Suzuki A, Maeda T, Baba Y, et al. Acidic extracellular pH promotes epithelial mesenchymal transition in Lewis lung carcinoma model[J].Cancer Cell Int, 2014, 14(1): 129.

[4] Sevtap Arslan, Fatma Bilge Ergen, Güzide Burça Aydın, et al. Different Attenuation Models of Diffusion-Weighted MR Imaging for the Differentiation of Benign and Malignant Musculoskeletal Tumors[J].J Magn Reson Imaging,2022,55(2):594-607.

Figures

Table 1 The differences of ADC, IVIM and DKI quantitative parameters between malignant and benign lesions.

Figure 1 A 12-year-old female with a benign bone tumor in the left tibia exhibits a compromised bony fibrous structure. A, an axial STIR image displays a distinct, round-like hyperintensity within the anterior cortical region; B, the lesion exhibits pronounced and uniform hyperintensity on DWI with a b-value of 1000; C, the lesion shows uniform signal intensity on ADC map; D~F, the lesion illustrate a slightly elevated signal on D and D* images and reduced signal on the f image; G, a slight increase in signal intensity on the MD image; H, a decrease signal on the MK image within the lesion.

Figure 2 A 55-year-old female was diagnosed with a malignant bone tumor, specifically a plasmacytoma in the right upper tibial bone. A, an axial STIR image reveals a hyperintense lesion located in the upper right tibia.B, the lesion exhibits conspicuous uniform hyperintensity in DWI; C, the lesion exhibits low signal in ADC image; D ~F, the lesion displays a diminished signal on the D and f images, whereas it exhibits an elevated signal on the D* image; G, the lesion demonstrates reduced signal intensity on the MD image; H, the lesion slightly elevates signal intensity on the MK image.

Table 2 Diagnostic efficiency of quantitative parameters.

Figure 3 Receiver operating characteristic curves (ROCs) for ADC, D, D*, MD, MK and their combined in distinguishing malignant bone tumors from benign lesions.

Proc. Intl. Soc. Mag. Reson. Med. 32 (2024)

1691

DOI: https://doi.org/10.58530/2024/1691