Evaluation of Variation in Corticomedullary Attenuation in CT Urography

Author

Lu Mao

Published

February 13, 2026

Statistical analysis

Categorical variables were summarized by frequency and percentage, and continuous variables were summarized by median and interquartile range (IQR). The Wilcoxon rank-sum test was used to compare continuous variables between groups. Linear regression, with Pearson’s correlation coefficient, was used to assess the correlation between log corticomedullary ratio (CMR) and continuous variables such as creatinine, eGFR, and age. A p-value < 0.05 was considered statistically significant. All statistical analyses were performed using R version 4.1.1 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Patient characteristics

Paitent characteristics are summarized in Table 1.

Table 1: Patient characteristics
Characteristic Female, N = 521 Male, N = 981 Overall, N = 1501
Age 72 (64, 77) 70 (64, 75) 70 (64, 76)
Creatinine (mg/dL) 0.74 (0.68, 0.89) 0.96 (0.83, 1.14) 0.89 (0.75, 1.07)
e-GFR (mL/min/1.73sqm) 81 (66, 90) 80 (64, 90) 80 (64, 90)
Right Corticomedullary Ratio 2.28 (1.94, 2.56) 2.51 (2.18, 2.79) 2.43 (2.07, 2.75)
Left Corticomedullary Ratio 2.29 (1.94, 2.64) 2.55 (2.17, 2.91) 2.49 (2.08, 2.81)
Indication


    Gross hematuria 18 (35%) 51 (52%) 69 (46%)
    Hematuria unspecified 17 (33%) 30 (31%) 47 (31%)
    Microscopic hematuria 17 (33%) 17 (17%) 34 (23%)
1 Median (IQR); n (%)

Comparison between right and left corticomedullary ratio

Figure 1: Right vs left CMR (log scale).
Table 2: Coefficient of variation of CMR (between left and right) within patients.
CoV
3.27%

CMR vs Creatinine

As shown in Figure 2, there is a little correlation relationship between CMR and creatinine.

Figure 2: Relationship between CMR (log scale) vs creatinine.

CMR vs eGFR

As shown in Figure 3, there is a little correlation relationship between CMR and eGFR.

Figure 3: Relationship between CMR (log scale) vs eGFR.
Table 3: Summary statistics of CMR by eGFR quartiles.
Characteristic 1, N = 381 2, N = 381 3, N = 371 4, N = 371 p-value2
Right Corticomedullary Ratio 2.29 (1.95, 2.66) 2.67 (2.26, 2.90) 2.48 (2.15, 2.78) 2.40 (2.06, 2.55) 0.062
Left Corticomedullary Ratio 2.29 (1.95, 2.72) 2.70 (2.25, 2.83) 2.52 (2.12, 2.79) 2.35 (2.08, 2.77) 0.14
Right Corticomedullary Ratio 2.29 (1.94, 2.68) 2.69 (2.25, 2.85) 2.48 (2.09, 2.74) 2.40 (2.08, 2.71) 0.082
1 Median (IQR)
2 Kruskal-Wallis rank sum test
Figure 4: Boxplot of CMR by eGFR quartiles.

CMR by gender

As shown in Table 4 and Figure 5 below, males have significantly higher CMR than females.

Table 4: Summary statistics of CMR by gender
Characteristic Female, N = 521 Male, N = 981 p-value2
Right Corticomedullary Ratio 2.28 (1.94, 2.56) 2.51 (2.18, 2.79) 0.009
Left Corticomedullary Ratio 2.29 (1.94, 2.64) 2.55 (2.17, 2.91) 0.010
Mean Corticomedullary Ratio 2.29 (1.95, 2.57) 2.58 (2.16, 2.83) 0.010
1 Median (IQR)
2 Wilcoxon rank sum test
Figure 5: Boxplot of CMR by gender.
Table 5: Percent increase in CMR in males vs females, with 95% CI.
Kidney Percent_Difference CI
Right 8.9% (1.6%, 16.7%)
Left 8.4% (0.7%, 16.6%)
Mean 8.6% (1.2%, 16.6%)
Table 6: Multivariate regression of log-CMR on gender, age, and eGFR.
Characteristic Beta 95% CI1 p-value
Gender


    Female — —
    Male 0.08 0.01, 0.16 0.023
Age 0.00 0.00, 0.00 0.8
e-GFR (mL/min/1.73sqm) 0.00 0.00, 0.00 >0.9
1 CI = Confidence Interval

Adjusting for age and eGFR, males have on average 8.7% higher CMR than females (p = 0.023).

CMR by age

As shown in Figure 6 below, there is little correlation between age and CMR.

Figure 6: Relationship between CMR (log scale) vs age.