For a study, researchers sought to compare virtual unenhanced images (VUE) with true unenhanced images (TUE) for the identification and measurement of urinary stones during contrast-enhanced dual-energy CT urography (DECTU) at various phases and slice thickness.
They looked at 122 patients with urinary stones who needed triphasic DECTU. The following data were reconstructed: TUE images with a thickness of 1.25 mm (TUE portal venous phase VUE images with 1.25 and 5-mm thickness (VUE(VP)_1.25 mm and VUE(VP)_5 mm); and excretory phase VUE images with a thickness of 1.25 and 5 mm (VUE(EP)_1.25mm and VUE(EP)_5mm). There were two types of stones: large (≥5 mm) and small. Two radiologists evaluated and statistically compared the above groups’ stone detection rates, sizes, and CT values.
On TUE 1.25 mm pictures, 230 urinary stones (163 large and 67 small stones) were found. The detection rates for large stones were 100%, 96.9%, 85.9%, and 80.4% for VUE(VP)_1.25 mm, VUE(VP)_5 mm, VUE(EP)_1.25 mm, and VUE(EP)_5 mm, respectively. The rates for small stones were 77.6%, 37.3%, 46.3%, and 23.9%, respectively. At both slice thicknesses, VUE(VP) pictures considerably increased the stone detection rate compared to VUE(EP) images. Thinner slice thickness pictures improved the accuracy while still identifying stones with smaller sizes and lower Hounsfield units on VUE images. For both TUE and VUE photos, the stone detection inter-reader agreement indicated a k value range of 0.85 to 0.94.
Thin section VUE(VP) pictures with a 33% radiation dose decrease may accurately identify large stones (≥5 mm). TUE, however, continued to be superior for small stones. In VUE photos, stone size is overestimated.