Quantitative evaluation of diffusion-kurtosis imaging for grading endometrial carcinoma: a comparative study with diffusion-weighted imaging.

AIM

To evaluate the diagnostic performance of histogram analysis of diffusion kurtosis magnetic resonance imaging (DKI) and standard diffusion-weighted imaging (DWI) in discriminating tumour grades of endometrial carcinoma (EC).

MATERIALS AND METHODS

Seventy-three patients with EC were included in this study. The apparent diffusion coefficient (ADC) value from standard DWI, apparent diffusion for Gaussian distribution (D), and apparent kurtosis coefficient (K) from DKI were acquired using a 3 T magnetic resonance imaging (MRI) system. The measurement was based on an entire-tumour analysis. Histogram parameters (D, K, and ADC) were compared between high-grade (grade 3) and low-grade (grade 1 and 2) tumours. The diagnostic performance of imaging parameters for discriminating high- from low-grade tumours was analysed using a receiver operating characteristic curve (ROC).

RESULTS

The area under the ROC curve (AUC) of the 10th percentile of D, 90th percentile of K and 10th percentile of ADC were higher than other parameters in distinguishing high-grade tumours from low-grade tumours (AUC=0.821, 0.891 and 0.801, respectively). The combination of 10th percentile of D and 90th percentile of K improved the AUC to 0.901, which was significantly higher than that of the 10th percentile of ADC (0.810, p=0.0314) in differentiating high- from low-grade EC.

CONCLUSION

Entire-tumour volume histogram analysis of DKI and standard DWI were feasible for discriminating histological tumour grades of EC. DKI was relatively better than DWI in distinguishing high-grade from low-grade tumour in EC.