Patient Population
Ultimately,
46 men and 19 women (65 total patients) were included in this study.
The mean age of patients was 49.6 years and the range of the age was 19-83 years.
The mean body mass index (BMI) was 24.3 kg/m2 and the range of BMI was 16.2 to 36.4.
Radiation Dose
The effective radiation doses of each protocol were 6.29 mSv ± 1.76 (RD),
1.48 mSv ± 0.44 (LD),
and the assumed effective dose of the full scan range (same range from the other two protocols) in the ULD protocol was 0.64 mSv ± 0.12 (ULD) (Figure 3).
There were statistically significant differences (p < 0.001) between the effective doses of each protocol.
The estimated dose reductions compared to RD were 76.4 % in LD and 89.8% in the assumed ULD.
Image Quality Assessment
Objective image noise assessment
The objective image noise was significantly lower in the ULD-IMR set (both measured on muscle and fat) than in any other image sets,
including RD.
The other results were significantly different from each other even between the RD and LD image sets (Table 1 and Figure 4).
Subjective image assessment
The subjective image assessments are summarized in Table 2 and Figure 5.
There was no statistically significant difference in the subjective noise and diagnostic confidence between the RD and LD image sets,
but the quality,
noise,
and diagnostic confidence of ULD-IMR were lower than RD and LD (p < 0.05).
In case of noise,
the scores of ULD-IMR,
LD-iDOSE and RD-FBP were 4.2,
4.8 and 5 respectively (Good – Excellent).
Among the three image sets reconstructed from the ULD protocol,
IMR (ULD-IMR) showed better subjective quality,
noise,
and diagnostic confidence than iDose (ULD-iDose) and FBP (ULD-FBP).
Diagnostic Performance and Interobserver Agreement
A total of 178 stones,
including16 right ureter stones (9.0%),
29 left ureter stones (16.3%),
58 right kidney stones (32.6%),
and 75 left kidney stones (42.1%) were diagnosed in the RD image sets as a standard reference.
Among the stones,
98 (55.1%) were smaller than 3mm.
The concordant rates of the two reviewers examining the LD images compared to the standard reference are summarized in Table 3.
The overall concordant rates of two reviewers were 88.2% and 84.3%,
respectively.
In cases of stones ≥ 3mm,
the staff radiologist detected 79 of 80 (98.8%) and the resident radiologist detected 78 of 80 (97.5%).
However,
in cases of stones < 3mm,
the staff radiologist detected 78 of 98 (79.6%) and the resident radiologist detected only 72 of 98 (73.5%) (Figure 6).
For each location,
the staff radiologist and radiology resident achieved concordant rates in the ureter of 97.8% and 91.1%,
respectively (Figure 7) and in kidney of 85.0% and 81.2%,
respectively.
For the ureter,
both reviewers had 100% concordant rates in case of stones ≥ 3mm.
There were four false positive cases of the staff radiologist’s results that were all very tiny or subtle.
The inter-observer agreement was substantial (kappa value = 0.61).