The value of diffusion kurtosis imaging in evaluating pathological grade of cervical squamous cell carcinoma / 实用放射学杂志

ABSTRACT

Objective To investigate the value of diffusion kurtosis imaging (DKI)quantitative parameters in evaluating patholog-ical grade of cervical squamous cell carcinoma (CSCC).Methods The DKI images of 45 patients with CSCC were analyzed retrospec-tively.According to the results of pathology,22 cases were divided into poorly differentiated group and 23 cases well-moderately dif-ferentiated group.The DKI parameters of two groups were measured by two observers,which included mean kurtosis (MK),axial kurtosis (Ka),radial kurtosis (Kr),fractional anisotropy of kurtosis (FAk),mean diffusivity (MD),axial diffusivity (Da),radial diffusivity (Dr)and fractional anisotropy (FA).The intra-class correlation coefficients (ICC)was used to test the consistency of the parameters measured results on two observers.The two independent samples t test was used to compare the parameters of two groups,and the ROC curve was used to evaluate the effectiveness of each parameter in order to evaluate the poorly differentiated CSCC and find the boundary values.Results The data consistency of two observers were good (ICC>0.75).The MK,Ka and Kr values on poorly differentiated CSCC were greater than that on well-moderately differentiated (P0.05).Thearea un-der curve (AUC)of MK,Ka,Kr,MD,Da and Dr values to diagnose poorly differentiated CSCC were 0.914,0.831,0.865,0.850, 0.778 and 0.865,respectively.The boundary values of diagnosing poorly differentiated CSCC were MK≥0.973,Ka≥1.075,Kr≥0.823, MD≤0.974μm2/ms,Da≤1.185μm2/ms and Dr≤0.762μm2/ms,respectively.Conclusion DKI can effectively predict the patho-logical grading of CSCC,which has a good clinical application prospects.