Ultra-low-dose unenhanced chest CT: Prospective comparison of high kV/low mA versus low kV/high mA protocols

Purpose

To qualitatively and quantitatively compare unenhanced ultra-low-dose chest computed tomography (ULD-CT) acquired at 80 kVp and 135 kVp.

Materials and methods

Fifty-one patients referred for unenhanced chest CT were prospectively included. There were 29 men and 22 women, with a mean age of 64.7 ± 11.6 (SD) years (range: 35–91 years) and a mean body mass index of 26.2 ± 6.3 (SD) (range: 17–54.9). All patients underwent two different ULD-CT protocols (80 kVp-40 mA and 135 kVp-10 mA). Image quality of both ULD-CT examinations using a 5-level scale as well as assessability of 6 predetermined lung parenchyma lesions were blindly evaluated by three radiologists and compared using a logistic regression model. Image noise of the two protocols was compared with Wilcoxon signed-rank test.

Results

The mean dose-length product at 80 kVp and at 135 kVp were 14.7 ± 1.8 (SD) mGy.cm and 15.6 ± 1.9 (SD) mGy.cm, respectively (P < 0.001). Image noise was significantly lower at 135 kVp (58.9 ± 12.4) than at 80 kVp (74.7 ± 14.5) (P < 0.001). For all readers and for all examinations, the 135 kVp protocol yielded better image quality than 80 kVp protocol, with a mean qualitative score of 4.5 ± 0.7 versus 3.9 ± 0.8 (P < 0.001). The 135 kVp protocol was significantly more often of diagnostic quality than the 80 kvp protocol (92.3% versus 77.8%, respectively) (P < 0.001) and was less prone to image quality deterioration in obese patients. Parenchymal lesions were never better depicted on the 80 kVp protocol than with the 135 kVp protocol.

Conclusion

Unenhanced chest ULD-CT should be acquired at a high kilovoltage and low current, such as 135 kVp-10 mA, over a low kilovoltage and high current protocol.