Detection and measurement of urinary stones on virtual monoenergetic images derived from rapid tube voltage switching dual-energy CT

Abstract

Introduction

We aimed to assess urinary stone detection and measurement, which are important indicators for treatment, using virtual monoenergetic (VM) images derived from rapid tube voltage switching dual-energy CT (rsDECT).

Methods

Forty-eight urinary stones placed in a 32-cm diameter phantom filled with saline and 38 patients with 95 urinary stones underwent rsDECT scans with CTDIvol of 5 mGy for phantoms and 8.1 ± 2.5 mGy for patients. VM images at energies from 40 to 100 keV were generated. Stone detection rate, detection confidence level (1–4 points), and size measurement deviation (digital caliper as gold standards) on VM images were recorded and compared.

Results

All stones could be detected in phantoms on VM images of all energies with one urinary stone missed in patients on VM images above 70 keV. Stones with size equal to or greater than 2 mm were detectable with highest confidence (4 points) on all VM images, while the detection confidence for stones with size smaller than 2 mm was higher on the low-energy images (40–60 keV). In addition, stone length and width measurement values decreased with the increased energy level, and high-energy VM images provided better agreements with digital caliper.

Conclusion

VM images in low-dose rsDECT can be used to detect urinary stones with high efficacy. Low-energy VM images provide higher detection confidence for small stones, while higher-energy images are more accurate in size measurements.

Implications for practice

Low-dose DECT should be used for detecting and characterizing small urinary stones in clinical practice to ensure high efficacy, and the low-energy and high-energy VM images in DECT should be optimized for stone detection and size measurement, respectively.