• 3D ultrasound
• 3D volume
• colordoppler
• diabetic kidney disease
• renal volume
• ultrasound

Abstract
Objective
To correlate in a cohort of type 2 diabetic subjects, different diabetic kidney disease (DKD) phenotypes classified into 4 groups according to the level of estimated glomerular filtration rate (e-GFR) and presence of albuminuria (ACR, albumin to creatinine ratio) to biometric, morphological and vascular parameters obtained with 2D/3D kidney ultrasound.
To assess the agreement between 3D ultrasound and standard 2D (ellipsoid formula) methods in measuring renal volumes.
Methods
We planned a cross-sectional study in a target population with type 2 diabetes mellitus who underwent 2D/3D echo-color-Doppler US within a period of 6 months. Morphological and vascular US indexes were compared in subjects with different DKD phenotypes. A single skilled and trained operator recorded bilaterally both kidney diameters by means of 2D standard US examination and 3D volume with X-matrix array technology. Than the renal volumes were calculated applying the ellipsoid formula for 2D US method and with manual segmentation for 3D US method. Also renal blood flow (RBF) and renal resistive index (RI) measurements were recorded bilaterally at the same time. To estimate the differences in renal dimension, renal volume, renal vascular indexes between the 4 patient groups the ANOVA and Kruskal-Wallis tests with multiple comparisons were considered appropriate for normally and not-normally distributed variables respectively. Two-tailed tests and P-values <=.05% were considered significant. Measuring agreement between the acquisition of renal volume methods (2D versus 3D) was evaluated by using Bland Altman test.
Results
We enrolled 70 diabetic individuals, 18 (26%, group 1, i.e. eGFR > 60 ml/min/1.73 m2 and ACR < 30 mg/g), 16 (23%, group 2, i.e. eGFR > 60 ml/min/1.73 m2 and ACR ≥ 30 mg/g), 19 (27%, group 3, i.e. eGFR ≤ 60 ml/min/1.73 m2 and ACR < 30 mg/g) and 17 (24%, group 4, i.e. eGFR ≤ 60 ml/min/1.73 m2 and ACR ≥ 30 mg/g).
In the overall population 3D total renal volume was 381±92 ml, ellipsoid total renal volume was 369±104 ml, 3D parenchymal renal volume was 312±76 ml, ellipsoid parenchymal renal volume was 313±87 ml. The total RBF was 691±478 cc/min and 751±449 cc/min in the left and right side respectively. Mean±SD of resistivity index was 0.77±0.10 at both sites.
By bivariate analysis a linear growing trend of RI according to DKD phenotypes from 1 to 4 was demonstrated (p=.06), suggesting its high correlation with eGFR. By multivariate analysis a significant correlation between eGFR and RI (P<.0001) was found.
At bivariate analysis the 3D total renal volume was 416±78 ml, 419±115 ml, 327±74 ml and 366±69 ml for groups 1, 2, 3 and 4 respectively, p=.004. The ellipsoid total renal volume was 421±60 ml, 402±148 ml, 303±81 ml and 356±73 ml for groups 1, 2, 3 and 4 respectively, p=.001. The 3D parenchymal renal volume was 344±64 ml, 343±96 ml, 263±54 ml and 304±59 ml for groups 1, 2, 3 and 4 respectively, p=.001. The ellipsoid parenchymal renal volume was 358±58 ml, 341±1118 ml, 259±69 ml, 301±63 ml for groups 1, 2, 3 and 4 respectively, p=.001.
In measuring agreement, Bland Altman test revealed a bias of 12 ml [-6.23;30.1 95%CI], lower 95%CI levels of -137.5 [-168.7;-106.3 95%CI], and upper 95%CI levels of 161.4 [130.17;192.6 95%CI], between 3D and ellipsoid total volume, and a bias of -1.1 ml [-17.4;15.2 95%CI], lower 95%CI levels of -134.8 [-162.8;-106.9 95%CI], and upper 95%CI levels of 132.6 [104.7;160.6 95%CI], between 3D and ellipsoid parenchymal volume.
Conclusions
Our study demonstrates a significant correlation between RI and both DKD phenotypes and eGFR. Among different morphological and vascular parameters the 3D total renal volume, ellipsoid total renal volume, 3D parenchymal renal volume and ellipsoid parenchymal renal volume are influenced by DKD phenotypes. 2D US measurement of total and parenchymal renal volumes yields a significant underestimation when compared to 3D US method.