Functional Magnetic Resonance Imaging for Distinguishing Type of Papillary Renal Cell Carcinoma: a Preliminary Study.

Qingqiang Zhu,Jing Ye,Wenrong Zhu,Jingtao Wu,Wenxin Chen,Jun Ling
DOI: https://doi.org/10.1259/bjr.20201315
2021-01-01
British Journal of Radiology
Abstract:OBJECTIVE:To investigate the feasibility of magnetic resonance diffusion kurtosis imaging (DKI) and intravoxel incoherent motion (IVIM) for distinguishing Type 1 and 2 of papillary renal cell carcinoma (PRCC). METHODS:A total of Type 1 (n = 20) and Type 2 (n = 16) of PRCC were examined by pathology. For DKI and IVIM, mean diffusivity (MD), fractional anisotropy (FA), mean kurtosis (MK), kurtosis anisotropy (KA), radial kurtosis (RK), diffusivity (D), pseudodiffusivity (D*) and perfusion fraction (f) were performed in assessment of type of PRCC. RESULTS:The mean SNRs of IVIM and DKI images at b = 1500 and 2000 s/mm2 were 8.6 ± 0.8 and 7.8 ± 0.6. Statistically significant differences were observed in MD and D values (1.11 ± 0.23 vs 0.73 ± 0.13, 0.91 ± 0.24 vs 0.49 ± 0.13, p < 0.05) between Type 1 and Type 2 of PRCC, while comparable FA, RK, D* and f values were found between Type 1 and Type 2 of PRCC (p > 0.05). Statistically significant differences were observed in MK and KA values (1.23 ± 0.16 vs 1.91 ± 0.26, 1.49 ± 0.19 vs 2.36 ± 0.39, p < 0.05) between Type 1 and Type 2 of PRCC. Areas of MD, MK, KA and D values under ROC curves for differentiating Type 1 and Type 2 of PRCC were 0.836, 0.818, 0.881 and 0.766, respectively. Using MD, MK, KA and D values of 0.93, 1.64, 1.94, 0.68 as the threshold value for differentiating Type 1 from Type 2 of PRCC, the best result obtained had a sensitivity of 85.0%, 80.0%, 90.0%, 85.0%, a specificity 75.0%, 68.7%, 87.5%, 81.2%, and an accuracy of 83.3%, 80.5%, 88.9%, 86.1%, respectively. CONCLUSION:DKI and IVIM are feasible techniques for distinguishing type of PRCC, given an adequate SNR of IVIM and DKI images. ADVANCES IN KNOWLEDGE:1. MD and D values are higher for Type 1 of PRCC and lower for Type 2 of PRCC.2. MK and KA values are higher for Type 2 of PRCC and lower for Type 1 of PRCC.3. DKI and IVIM can be used as clinical biomarker for PRCC type's differential diagnosis, given an adequate SNR.
What problem does this paper attempt to address?