IMAGING IN EARLY RENAL INJURY

eGFR and albuminuria remain indispensable for the diagnosis, risk stratification, and monitoring of CKD, as clinical trials have used both eGFR and albuminuria as entry criteria and outcomes [14]. However, they oscillate during the day and from day to day, reveal little about the micro-environmental injury driving long-term decline and are late events: by the time CKD is diagnosed based on eGFR, around 50% of the functional kidney mass has been lost, and the albuminuria threshold is 10-fold higher than physiological levels. Histologically, low-grade fibrosis, microvascular rarefaction, and cortical hypoxia are established early in the disease course [3]. In fatty kidney, ectopic lipid accumulated in three main compartments of the kidney: the perirenal space, the renal sinus and, less frequently, the cortex and medulla, leading to tubular inflammation and fibrosis, while fat deposited in the renal sinus can compress low-pressure veins and lymphatics, increasing interstitial pressure and local hypoxia [15–17].

Advanced imaging techniques now allow these subclinical changes to be detected and quantified, offering the promise of far earlier diagnosis and more precise staging than conventional biomarkers (Fig. 2) [7]. Contemporary imaging pipelines interrogate oxygenation, perfusion, microstructural integrity, steatosis, and stiffness—often within a single session—enabling a truly multiparametric approach. Modern quantitative imaging bridges this gap by interrogating oxygenation, perfusion, micro-structural integrity, and stiffness in vivo, thereby providing complementary—and at times superior—prognostic information to laboratory tests alone. The location of the kidney graft confers certain advantages for imaging (Table 1).

Figure 2:: Translational and AI-integrated renal imaging: from quantitative biomarkers to clinical application. Advanced renal imaging extends beyond structure to function and prognosis. In hypertension and kidney transplantation, SWE, CEUS, and MRE assess fibrosis, perfusion, and rejection noninvasively, while AI and radiomics enhance feature extraction and graft outcome prediction. Integration with multi-omics approaches supports precision nephrology. Future priorities include longitudinal validation, imaging-informed clinical trials, and inclusion of imaging biomarkers in therapeutic decision-making. Abbreviations: ADC, apparent diffusion coefficient; AI, artificial intelligence; ARFI, acoustic radiation force impulse; ASL, arterial spin labeling; BOLD, blood oxygenation level-dependent; CEUS, contrast-enhanced ultrasound; CKD, chronic kidney disease; CVD, cardiovascular disease; DKD, diabetic kidney disease; DWFG, death with functioning graft; eGFR, estimated glomerular filtration rate; IVIM-DWI, intravoxel incoherent motion diffusion-weighted imaging; MRA, magnetic resonance angiography; MRE, magnetic resonance elastography; MRI, magnetic resonance imaging; MRS, magnetic resonance spectroscopy; PAI, photoacoustic imaging; PDFF, proton-density fat fraction; PRS, polygenic risk score; R2*, transverse relaxation rate; RI, resistive index; RRI, renal resistive index; RSF, renal sinus fat; SMI, superb microvascular imaging; SWE, shear-wave elastography; STE, strain-tensor elastography; T1/T2, relaxation times; and US, ultrasound.

Imaging method | Purpose/comparison focus | Key quantitative findings | Clinical applications | Level of evidence
Contrast-Enhanced Ultrasound (CEUS) | Detect microvascular perfusion deficits vs. Doppler US | Hypertensive patients show ∼28% lower cortical perfusion than controls (1476 vs. 2062 AU; P < 0.001) and blunted flow reserve (34 % vs. 56 %) | Early CKD microcirculatory assessment, renal mass characterization, transplant perfusion | Level II b (prospective controlled human studies)
Doppler Resistive Index (RRI) | Correlate vascular resistance with renal prognosis | RRI > 0.70 predicts faster eGFR decline; each 0.01 ↑ = 4 % ↑ mortality hazard & 6 % ↑ dialysis risk; ACE I ↓ RRI 0.61→0.56 (24 mo) | Routine CKD, hypertension, transplant evaluation | Level II a (meta-analysis + cohort)
Superb Microvascular Imaging (SMI) | Compare to Doppler US and CEUS for low-flow detection | Visualizes 300–500 μm vessels; qualitative detection of microvascular rarefaction mirroring ASL MRI findings | CKD cortical perfusion mapping, tumor vascularity, post-treatment monitoring | Level III b (pilot observational studies)
Ultrasound Elastography (SWE/ARFI) | Compare stiffness vs. biopsy-proven fibrosis | Meta-analysis (1394 pts): AUC 0.87 mild, 0.78 moderate, 0.86 severe; in DKD 19 kPa → 93 % sens/86 % spec; ARFI 2.37 vs. 2.96 m/s (HTN) | Noninvasive fibrosis quantification in CKD, DKD, HTN nephropathy | Level I a (systematic review + multiple RCT-validated cohorts)
BOLD MRI (Oxygenation) | Evaluate tissue oxygenation vs. eGFR/creatinine | Higher R₂* = 3 × risk ESKD; Dapagliflozin ↓ R₂* by 9 % (6 h) → ↑ oxygenation | Risk stratification, drug response (esp. SGLT2i effects) | Level II a (prospective cohort + intervention)
ASL MRI (Perfusion) | Quantify RBF vs. CKD stage | Threshold 143 ml/min/100 g distinguishes CKD (AUC 0.98); r ≈ 0.8 with eGFR | Quantitative non-contrast perfusion mapping for CKD & DKD | Level II a (prospective cohort + intervention)
IVIM-DWI (Diffusion) | Microstructure + microvascular flow vs. biopsy/fibrosis | Each 1 % ↓ in f = 21 % ↑ ESKD risk; AUC 0.99 for early DKD with D + FA combo | Structural injury marker for CKD progression prediction | Level II b (prospective controlled human studies)
T₁ Mapping (Fibrosis) | Quantify fibrosis vs. histology & outcomes | Highest T₁ = 3 × risk ESKD; radiomics AUC 0.93–0.94 for fibrosis >25 % | Non-contrast fibrosis biomarker for CKD/DKD monitoring | Level II a (prospective cohort + intervention)
Multiparametric MRI (mpMRI) | Integrate perfusion + oxygenation + diffusion + T₁ | 4-parameter model → only BOLD R₂* remains independent predictor of eGFR decline; repeatability CV ≈ 4%–11% | Comprehensive risk profiling in CKD and transplant | Level II b (prospective controlled human studies)
Computed tomography | Quantify renal sinus fat volume and validate reproducibility | Single APDRS correlates with RSF segmentation (r = 0.80); inter-reader ICC = 0.98; single-slice protocol ICC 0.93 (intra)/0.86 (inter) | Gold-standard for renal sinus fat quantification and validation of MR/US methods | Level II b (prospective controlled human studies)
Proton-Density Fat-Fraction (PDFF) Mapping | Quantify renal parenchymal fat fraction and assess its link to oxygenation | CV <5%; detects fat 0.4%–50%; PDFF correlates with BOLD R₂* (↑ fat = ↑ hypoxia); IDEAL-IQ: PDFF ↑ in T2DM vs. controls (P < 0.001), correlates negatively with eGFR (r = –0.437) and positively with creatinine (r = 0.421); AUC 0.857, repeatability r = 0.81 | Early metabolic phenotyping in T2DM, obesity, and HTN; longitudinal therapy monitoring; adjunct biomarker in multiparametric MRI | Level II a (prospective cohort + intervention)
Imaging + Omics Integration | Combine quantitative imaging with proteomics/genomics for prediction | CKD273 proteomic + T₁/ASL imaging models under validation—aim to enhance individualized risk stratification | Future precision medicine integration | Level V (conceptual/expert opinion)