Dual-energy Computed Tomography (DECT) in Renal Masses: Nonlinear versus Linear Blending

Rational and Objectives: To investigate whether a nonlinear-blending algorithm improves tumor conspicuity and image quality in the evaluation of renal masses at dual-energy computed tomography (DECT) during nephrographic phase of enhancement. Materials and Methods: The Institutional Review Board approved this retrospective study from archival material from patients consenting to the use of medical records for research purposes. A retrospective review of contrast-enhanced abdominal DECT scans in 45 patients (mean age, 59.5 years; range, 24-84 years) was performed. DECT data were reconstructed using nonlinear and linear blending. A region of interest was located within tumors and adjacent normal parenchyma; attenuation differences and contrast-to-noise ratios (CNRs) were calculated for renal masses on nonlinear- and linear-blended images. The two datasets were subjectively compared in terms of tumor detection and image quality. An exact Wilcoxon's matched pairs signed rank and marginal homogeneity tests were used to test whether differences in attenuation, CNR, and subjective assessment were greater using nonlinear blending. Results: The mean difference in attenuation for renal masses and adjacent portion of renal parenchyma was 138.4 Hounsfield units +/- 28.9 SD using nonlinear blending, and 121.6 HU +/- 18.0 SD using linear blending (P < .001). Mean CNR was 12.6 +/- 2.5 SD using nonlinear blending, and 9.6 +/- 2.2 SD using 0.3 linear-blended (P < .001). No significant difference in tumor detection was observed between the two algorithms. Image quality was significantly better (P < .001) using nonlinear blending. Conclusion: Compared with standard linear blending, nonlinear-blending algorithm improves tumor conspicuity and image quality in renal masses at DECT evaluation during nephrographic phase of enhancement.