Diagnostic value of diffusion kurtosis imaging in differentiating indeterminate solitary pulmonary nodules: comparison with diffusion weighted imaging

Abstract

Objective To investigate the role of diffusion kurtosis imaging (DKI) in indeterminate solitary pulmonary nodules (SPN) diagnosis and to compare with conventional diffusion weighted imaging (DWI). Methods From March 2016 to Dec 2017, forty-three consecutive patients (30 male, 13 female, age: 56±11 years) with indeterminate SPNs were included. All patients underwent axial multi-b factor DWI (with b values=0, 50, 200, 400, 800, 1 400, 2 000 s/mm2) examination and were divided into benign group (19 cases) and malignant group (24 cases) according to pathological results of SPN. ADC Kurtosis (K) and Diffusivity (Dk) values were compared between malignant and benign group and among different subtypes of lung cancer using independent t test (normal distribution and homogeneity of variance) and Mann-Whitney U test (skewed distribution or variance). Receiver operating characteristic (ROC) curves were employed to evaluate the diagnostic performance. Results K values were significantly higher for malignant SPNs than for benign SPNs (0.839±0.197 vs. 0.718±0.120; t=2.359, P=0.023). ADC values were found to be significantly higher in benignity than malignant SPNs [(1.605±0.422) ×10-3mm2/s vs. (1.278±0.210)×10-3mm2/s; t=-3.089, P=0.005). No difference was observed in Dk between the two groups (P=0.922). All parameters cannot differentiate subtypes of lung cancer. The ADC value had higher AUC (area under ROC curve) than that of K value. The sensitivity (70.8%) and accuracy (72.1%) of ADC value was higher than K value, the specificity of both methods was equal. Conclusion DKI is a feasible non-invasive tool which has comparable capability of conventional DWI in SPNs differentiation, although with lower sensitivity and accuracy. DKI can provide additional information for SPNs characterization and has a potential to be a robust way in SPNs interpretation. Key words: Magnetic resonance imaging; Solitary pulmonary nodule