Value of intravoxel incoherent motion and diffusion kurtosis imaging in predicting peritumoural infiltration of soft-tissue sarcoma: a prospective study based on MRI-histopathology comparisons

AIM: To investigate the performance of intravoxel incoherent motion (IVIM) and diffusion kurtosis imaging (DKI) in the identification of peritumoural infiltration of soft-tissue sarcoma (STS). MATERIALS AND METHODS: From July 2018 to January 2020, 34 STS patients who underwent 3-T magnetic resonance imaging (MRI), including IVIM and DKI, were reviewed. The standard apparent diffusion coefficient (ADC), true diffusion (D), pseudo-diffusion coefficient (D*), perfusion fraction (f), mean kurtosis (MK), and mean diffusion (MD) of each lesion were analysed independently by two observers. An MRI-histopathology control method was used to ensure the correspondence of MRI sections with histopathological sections. Differences in STS with and without infiltration were evaluated. The area under the curve (AUC) was used to determine the best cut-off point for different parameters. Interobserver agreement was assessed using the intraclass correlation coefficient. RESULTS: Standard ADC, D, MK, and MD values reliably distinguished STS that had positive and negative infiltration. The MD value had the best diagnostic performance. Use of an MD cutoff value of 2.35 x 10(-3) mm(2)/s to distinguish positive and negative infiltration had an AUC of 0.85, accuracy of 88.2%, sensitivity of 94.4%, and specificity of 81.3%. The two independent observers had nearly perfect agreement for all parameters. CONCLUSION: The standard ADC and D value of IVIM, and the MK and MD values of DKI reliably identify the presence of peritumoural infiltration of STS. (C) 2021 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

Automated summary

Standard ADC and D values of IVIM, as well as MK and MD values of DKI, reliably identify peritumoural infiltration of STS. An MD cutoff value of 2.35 x 10(-3) mm(2)/s showed the best diagnostic performance with an AUC of 0.85, accuracy of 88.2%, sensitivity of 94.4%, and specificity of 81.3%. Interobserver agreement for all parameters was nearly perfect.