03.04.2020 | Urogenital | Ausgabe 8/2020
Protocol analysis of dual-energy CT for optimization of kidney stone detection in virtual non-contrast reconstructions
- Matthias Lazar, Helmut Ringl, Pascal Baltzer, Daniel Toth, Christian Seitz, Bernhard Krauss, Ewald Unger, Stephan Polanec, Dietmar Tamandl, Christian J. Herold, Michael Toepker
Previous studies have shown that split-bolus protocols in virtual non-contrast (VNC) reconstructions of dual-energy computed tomography (DE-CT) significantly decrease radiation dose in patients with urinary stone disease. To evaluate the impact on kidney stone detection rate of stone composition, size, tube voltage, and iodine concentration for VNC reconstructions of DE-CT.
In this prospective study, 16 kidney stones of different sizes (1.2–4.5 mm) and compositions (struvite, cystine, whewellite, brushite) were placed within a kidney phantom. Seventy-two scans with nine different iodine contrast agents/saline solutions with increasing attenuation (0–1400 HU) and different kilovoltage settings (70 kV/150 kV; 80 kV/150 kV; 90 kV/150 kV; 100 kV/150 kV) were performed. Two experienced radiologists independently rated the images for the presence and absence of stones. Multivariate classification tree analysis and descriptive statistics were used to evaluate the diagnostic performance.
Classification tree analysis revealed a higher detection rate of renal calculi > 2 mm in size compared with that of renal calculi < 2 mm (84.7%; 12.7%; p < 0.001). For stones with a diameter > 2 mm, the best results were found at 70 kV/Sn 150 kV and 80 kV/Sn 150 kV in scans with contrast media attenuation of 600 HU or less, with sensitivity of 99.6% and 96.0%, respectively. A higher luminal attenuation (> 600 HU) resulted in a significantly decreased detection rate (91.8%, 0–600 HU; 70.7%, 900–1400 HU; p < 0.001). In our study setup, the detection rates were best for cystine stones.
Scan protocols in DE-CT with lower tube current and lower contrast medium attenuation show excellent results in VNC for stones larger than 2 mm but have limitations for small stones.
• The detection rate of virtual non-contrast reconstructions is highly dependent on the surrounding contrast medium attenuation at the renal pelvis and should be kept as low as possible, as at an attenuation higher than 600 HU the VNC reconstructions are susceptible to masking ureteral stones.
• Protocols with lower tube voltages (70 kV/Sn 150 kV and 80 kV/Sn 150 kV) improve the detection rate of kidney stones in VNC reconstructions.
• The visibility of renal stones in virtual non-contrast of dual-energy CT is highly associated with the size, and results in a significantly lower detection rate in stones below 2 mm.