MORE ON SPECIFIC CASE-USE IN KIDNEY DISEASE: FATTY KIDNEY AND KIDNEY TRANSPLANTATION

Fatty kidney

Renal adiposity represents a mechanistic extension of the early-injury paradigm, in which structural compression, microvascular dysfunction, and hypoxia precede measurable declines in eGFR. Imaging-based quantification of renal sinus and parenchymal fat therefore provides a direct link between metabolic injury, early risk stratification, and longitudinal prognostication (Fig. 3).

Figure 3:: Closing the CKD diagnostic blind spot: early imaging detection slows disease progression. Conceptualized by the authors, informed by prior work (7). Figure legend: Integration of imaging into CKD diagnosis and management. Imaging enables detection of structural and functional renal changes before biochemical alterations become evident (e.g. eGFR decline or albuminuria). Early imaging-based identification allows earlier initiation of therapeutic interventions, which can slow or even halt disease progression. Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; and UACR, urine albumin-to-creatinine ratio.

Metabolic and functional correlates

Community data from the Framingham Heart Study (n = 2923) showed that participants in the highest renal sinus fat decile had a 2.1–fold higher odds of hypertension and a 2.3–fold higher odds of CKD, after adjusting for visceral adipose tissue [50]. MRI studies further these findings: among 105 adults with T2DM, each 1% increase in cortical PDFF associated with a 12% higher odds of prevalent CKD after adjustment for body mass index and glycemic control [43]. Greater RSF has also been linked to albuminuria, hypertension and activation of the RAAS in primary aldosteronism, illustrating systemic-renal cross-talk [50–52].

Prognostic evidence

Prospective evidence, although still limited, indicates that renal adiposity is not only a passive marker of systemic obesity. In an 18–month randomized dietary trial (n = 278), higher baseline renal sinus fat predicted lower eGFR and greater micro–albuminuria, while overall renal sinus fat fell by 9% irrespective of diet [53]. Although long–term prospective data remains sparse, these observations suggest that renal adiposity is more than a passive marker of systemic obesity.

Therapeutic modulation

The impact of therapeutic modulation by lifestyle, surgery and medication on renal sinus fat has been assessed. Lifestyle. Weight-loss interventions consistently reduce renal sinus fat; the 18-month randomized dietary trial demonstrated a significant reduction in renal sinus fat (−9%; P < 0.05 vs. baseline) [53]. Bariatric surgery. Whole-body MRI in 74 severely obese patients revealed a ∼30 % decrease in renal sinus fat six months after sleeve gastrectomy, and the magnitude of renal sinus fat regression independently predicted hypertension remission [54]. Pharmacotherapy. In a 32-week randomized study of high-cardiovascular-risk T2DM, both semaglutide and empagliflozin reduced cortical apparent diffusion coefficient (ADC) and total kidney volume on diffusion-weighted MRI—changes interpreted as early micro-structural improvement before eGFR change [55]. These data support the inclusion of renal-fat metrics as adjunct end-points in intervention trials alongside albuminuria, C-reactive protein, and GFR slope.

AI and multimodal integration

Manual segmentation is a practical bottleneck. A 2024 European Radiology Experimental study reported that a 3-D nnU-Net achieved Dice scores ≥0.94 for renal segmentation on contrast- and non-contrast CT, enabling automated quantification of renal and perirenal fat within seconds [56]. Beyond segmentation, radiomic models that fuse non-contrast CT features with machine-learning algorithms now estimate split renal function with an AUC of 0.90, highlighting the potential of imaging-based phenotyping to complement eGFR [57].

Knowledge gaps and future directions in fatty kidney imaging

Despite growing evidence linking renal adiposity to kidney dysfunction, key uncertainties remain. These include whether renal-fat accumulation is causal or merely a risk marker, scanner-specific thresholds for defining “pathological” fat, and whether reductions in renal fat independently predict hard renal outcomes. Long-term, multicenter outcome studies are needed before renal-fat imaging can be fully integrated into clinical practice.

Imaging of allograft kidney in kidney transplant recipients

In kidney transplant recipients, Doppler US remains the cornerstone of routine graft surveillance due to its non-invasiveness, availability, and ability to assess vascular integrity and perfusion. Standard protocols typically involve Doppler US within the first week post-transplant to evaluate early complications such as renal artery thrombosis, urinary leaks, and perinephric fluid collections [58]. Subsequent imaging may be performed at 1, 3, 6, and 12 months, although this varies by center and is often symptom-driven beyond the first year, and if there is a need to evaluate for complications [58].

In KT, imaging similarly enables detection of subclinical injury trajectories—such as microvascular dysfunction and progressive fibrosis—long before functional decline is apparent, positioning graft imaging as a natural extension of early-injury-based risk prediction into the post-transplant setting (Fig. 3).

The most critical US parameters for assessing transplant kidney function are RI for vascular resistance, echogenicity and size for parenchymal integrity, perfusion for vascular patency, and evaluation for hydronephrosis or collections indicating mechanical or surgical complications. RI is commonly measured, with elevated values (>0.8) associated with rejection, obstruction, or acute tubular necrosis. In cases of graft dysfunction or delayed graft function, adjunctive imaging modalities such as nuclear renography (e.g. Mercaptoacetyltriglycine scans), CT, or MRI/MRA might be utilized selectively. While biopsy remains the gold standard for diagnosing rejection or chronic changes, non-invasive techniques such as SWE, MRE, and photoacoustic imaging (PAI) are under investigation for detecting fibrosis, offering potential future tools for longitudinal graft assessment. PAI combines US and laser technology to detect acoustic signals from tissues, allowing for the quantification of collagen content and assessment of renal fibrosis.

As a key driver of chronic allograft dysfunction, renal allograft fibrosis typically evolves subclinically and is shown to predict long-term graft loss [59]. Renal allograft fibrosis is the progressive accumulation of extracellular matrix in the transplanted kidney, reflecting chronic structural injury from immune-mediated damage (e.g. chronic antibody-mediated rejection), ischemia-reperfusion, calcineurin inhibitor toxicity, and recurrent or de novo disease [59]. It manifests histologically as interstitial fibrosis and tubular atrophy, often with glomerulosclerosis and vascular remodeling [60].

Early detection of renal allograft fibrosis is essential, as significant histologic injury can be present even when serum creatinine and eGFR appear normal, and timely identification enables earlier intervention to prevent irreversible damage and prolong graft survival [60].

Fibrosis cannot be reliably detected with conventional US, and early fibrotic changes might typically go unnoticed. Diagnosis traditionally depends on biopsy, which is limited by sampling bias and risk [59]; emerging advanced or quantitative noninvasive modalities such as renal H-scan aim to quantify whole-organ fibrotic burden more accurately and reproducibly.

Registry-based analyses reinforce the prognostic value of serial RI assessment. In a large retrospective cohort (n = 1685) from France, trends in RI between 1 and 3 months post-transplant best predicted long-term risk of death with a functioning graft (DWFG); values ≥0.70 at both time points conferred nearly fourfold higher hazard (HR ∼3.77) relative to low RI at both times [61]. Another multicenter study (n ≈ 425) demonstrated that a ≥10% increase in RI between 4- and 12-months post-transplant independently predicted death-censored graft loss (HR ∼6.2; AUC ∼0.75) [62]. These findings underline the relevance of trajectory over absolute measurements and support structured Doppler surveillance in the first post-transplant year, even in the absence of symptoms. These findings suggest that structured longitudinal Doppler surveillance may support future risk-stratified screening strategies in large transplant populations.