Opportunistic Detection of Chronic Kidney Disease Using CT-Based Measurements of Kidney Volume and Perirenal Fat

Abstract

Background/Objectives: Chronic kidney disease (CKD) is a prevalent condition with many cases remaining undiagnosed, although early detection is essential. Adipose tissue distribution-particularly perirenal fat thickness (PrFT)-has recently been linked to renal pathophysiology. This study assessed the association between CT-derived parameters of fat distribution and kidney morphology with CKD. Materials and Methods: This retrospective study included 237 patients (117 subjects, 120 controls) who underwent abdominal CT and had serum creatinine data. The dataset was randomly split (70% training, 30% test) to develop and evaluate a logistic regression model. CKD was defined as estimated Glomerular Filtration Rate (eGFR) < 60 mL/min/1.73 m². PrFT was measured as the distance from the posterior renal capsule to the posterior abdominal wall; renal hilum fat was segmented using a -195 to -45 HU range. Additional parameters (measured using automated segmentation tools) included kidney volume (KV), visceral/subcutaneous fat areas, skeletal muscle area and attenuation, and liver attenuation. Bilateral measurements were averaged. Results: KV (OR = 0.249, 95% CI: 0.146-0.422, p < 0.001) and PrFT (2nd tercile: OR = 7.720, 95% CI: 2.860-20.839; 3rd tercile: OR = 16.892, 95% CI: 5.727-49.822; both p < 0.001) were identified as independent predictors of CKD. These variables were used to construct a simplified model, which demonstrated moderate clinical applicability (AUC = 0.894) when evaluated on the test subset. Conclusions: KV and PrFT emerged as independent predictors of CKD, forming the basis of a simplified model with potential for opportunistic clinical application. This approach may facilitate earlier detection of CKD in patients undergoing CT imaging for unrelated clinical reasons. These imaging parameters are not intended to replace serum creatinine or eGFR but may serve as complementary predictors in specific clinical contexts.

Keywords: abdominal fat; chronic kidney disease; computed tomography; imaging biomarkers; kidney volume; opportunistic screening; perirenal fat; renal hilum attenuation; renal hilum fat; visceral fat.

Figure 1

Automated all organ segmentation using the TotalSegmentator extension in 3D Slicer software in a patient with stage V chronic kidney disease. The kidney parenchyma is segmented and shown in red bilaterally, with the renal hilum not included. Segmentation of the remaining organs, shown in various colors, is presented only to illustrate the software output.

Figure 2

Automated tissue-type segmentation in 3D Slicer using the TotalSegmentator extension at the level of the inferior endplate of the L3 vertebra. Visceral fat is shown in blue, subcutaneous fat in yellow, and skeletal muscle in red.

Figure 3

Manual segmentation of renal hilum fat is presented in coronal plane. The green area represents the right renal hilum, and the yellow area represents the left renal hilum. Fat was segmented using a Hounsfield unit threshold range of −195 to −45.

Figure 4

Illustration of posterior perirenal fat thickness measurement. The red line indicates the distance between the posterior renal capsule and the posterior abdominal wall at the level of the renal vein.

Figure 5

Kidneys segmentation performed in Exhibeon software using the 3D Smart Brush tool in healthy control. Only the kidney parenchyma was segmented. The renal hilum is not included in the segmentation. Blue contour represents right kidney, and green contour represents left kidney.

Figure 6

Receiver operating characteristic (ROC) curve of the simplified multivariable logistic regression model on the test dataset (AUC = 0.894).