Multiparametric MR Imaging for Evaluating Renal Function and Microstructure Authors: Akira Yamamoto, Yu Ueda, Hideki Kagioka, Tsutomu Tamada Published: 2026 ABSTRACT Chronic kidney disease (CKD) is a major global health burden, encompassing a heterogeneous group of disorders with diverse etiologies and pathophysiological mechanisms. Conventional biomarkers such as serum creatinine, estimated glomerular filtration rate (eGFR), and proteinuria are indispensable for detection and staging, yet they may be insensitive to early CKD including diabetic kidney disease (DKD) and can be discordant with underlying structural damage. Over the past 15 years, functional renal MRI has gained attention as a promising noninvasive approach to quantify complementary domains of renal pathology without ionizing radiation or gadolinium-based contrast agents, enabling repeated assessments even in patients with impaired renal function. This review provides a comprehensive overview of multiparametric MRI techniques for evaluating renal microstructure and physiology across CKD, including T1 and T2 mapping for interstitial expansion and inflammatory activity, blood oxygenation level dependent MRI for oxygenation and corticomedullary gradients, diffusion-based methods (apparent diffusion coefficient [ADC], diffusion tensor imaging, intravoxel incoherent motion) for microstructural injury and microvascular components, and arterial spin labeling for perfusion. Corticomedullary differentiation (CMD) is highlighted as an integrative feature across modalities, including signal-intensity-based CMD using noncontrast steady-state free precession (SSFP) with spatially selective inversion recovery and individualized optimal inversion time (multi-TI). Across CKD populations, many MRI-derived metrics correlate with contemporaneous kidney function, and diffusion-derived indices—particularly cortical ADC—have been associated with subsequent eGFR decline, suggesting potential value for progression-risk assessment. DKD is discussed as an important illustrative subtype in which complex pathophysiology may highlight the added value of multiparametric approaches. Clinical translation will depend on standardized acquisition, quality control, and multicenter validation to establish reproducibility and outcome-based thresholds. Introduction Chronic kidney disease (CKD) is a major global health issue,1 affecting a large and growing population worldwide and encompassing a wide spectrum of etiologies. Diabetes has become its leading cause,2,3 but CKD also includes hypertensive, glomerular, and hereditary diseases. Although conventional biomarkers such as serum creatinine, estimated glomerular filtration rate (eGFR), and proteinuria are useful for detecting and staging kidney damage, they primarily reflect global renal function and may not fully capture early tubulointerstitial injury or the underlying structural and physiological heterogeneity of CKD. As a result, there is an increasing need for more comprehensive and multiparametric approaches to assess kidney disease.4–7 Renal biopsy provides the most accurate assessment of renal pathology; however, its invasive nature limits its use in routine clinical practice, particularly for longitudinal monitoring. In this context, MRI has emerged as a promising noninvasive alternative. Advanced functional MRI techniques enable quantitative assessment of renal microstructure, perfusion, oxygenation, and interstitial changes, thereby providing domain-specific information that is not captured by conventional biomarkers such as eGFR or proteinuria alone. By integrating multiple parameters, MRI has been proposed as a “virtual biopsy,” with the potential to detect subclinical abnormalities, characterize key pathophysiological processes such as fibrosis/edema, microvascular dysfunction, hypoxia, and tubular disorganization, and provide a quantitative baseline for longitudinal monitoring. Importantly, many of these techniques can be performed without ionizing radiation or gadolinium-based contrast agents, making them suitable for repeated evaluation, particularly in patients with impaired renal function. A growing body of evidence suggests that multiparametric MRI can detect subtle renal alterations, correlate with contemporaneous renal function, and in some cases show associations with subsequent functional decline, although the strength of evidence varies across parameters.4,5,8,9 Within the spectrum of CKD, diabetic kidney disease (DKD) represents a particularly important and well-studied subtype, characterized by complex and heterogeneous pathophysiology, including early hemodynamic changes and later structural injury. DKD is therefore frequently used as an illustrative model for evaluating advanced imaging biomarkers, while also highlighting the broader challenges of assessing heterogeneous kidney disease. In this review, we provide an overview of the current state of functional multiparametric MRI for the assessment of renal structure and function across CKD, with emphasis on key imaging parameters, their biological correlates, and their potential clinical applications. MRI Parameters as Indicators of Histological and Physiological Changes T1 mapping Quantitative T1 mapping measures the longitudinal relaxation time of tissue water on a pixel-by-pixel basis. Renal T1 values are highly sensitive to tissue composition. In healthy kidneys, cortex and medulla have distinct T1 values, cortical T1 is typically much shorter than medullary T1 due to lower water content and higher protein content in the cortex.10 Pathological changes that increase tissue water content or alter the macromolecular environment tend to prolong T1. In particular, expansion of the interstitial space due to fibrosis or edema leads to a marked increase in T1 relaxation time. Fibrosis, characterized by the accumulation of collagen and extracellular matrix, increases both free and bound water content and alters water–molecule interactions, thereby prolonging T1. Similarly, inflammatory edema, through an increase in free water, exerts a T1-prolonging effect. Thus, T1 mapping provides a marker of renal structural integrity, often dubbed a “fibrosis index” in kidney MRI (Table 1). Empirical studies support T1’s association with histopathology. In animal models, T1 mapping can clearly distinguish normal vs. fibrotic kidneys. For example, in a rat model of tubulointerstitial nephritis, cortical T1 correlated with the degree of interstitial fibrosis and inflammation on histology.11 Studies in patients have similarly shown that cortical T1 elevation corresponds to fibrosis severity.12–14 Collectively, T1 mapping serves as a robust, noninvasive imaging biomarker of renal tissue health. Low native T1 values in normal kidneys correspond to tightly packed, healthy parenchyma, while elevated T1 values are suggestive of interstitial expansion resulting from fibrosis and/or edema, reflecting pathological remodeling. T2 mapping T2 mapping measures the transverse relaxation time of tissue, which depends on water–molecule interactions. Renal T2 is influenced by tissue water content and microstructure, somewhat overlapping with T1’s sensitivities but in a different manner. In general, T2 increases with free water content in conditions such as edema or inflammation (Table 1), while fibrosis-related T2 changes are more variable and less consistent than T1 changes.15 Renal T2 changes are less consistently reported than T1 changes. Cortical T2 elevation has been described in some kidney transplant studies, but its functional significance remains unclear.10 Early disease is generally associated with T2 prolongation, whereas chronic end-stage disease tends to show T2 shortening.15 In summary, T2 mapping is particularly sensitive to acute or active pathological changes characterized by edema and inflammation, while T2 may normalize or decrease in the setting of chronic fibrosis.10 It could thus complement T1 mapping: for instance, a kidney with high T1 but low T2 likely has irreversible fibrosis with little active inflammation, whereas a kidney with simultaneously high T1 and high T2 might have ongoing inflammatory injury. Blood oxygenation level dependent (BOLD) imaging BOLD MRI, which measures T2* relaxation (or its inverse R2*), is a technique originally developed in brain imaging that has been applied to kidneys to estimate tissue oxygenation. BOLD MRI exploits the fact that deoxygenated hemoglobin is paramagnetic and causes local magnetic field distortions that shorten T2*. In essence, lower tissue oxygenation leads to more deoxyhemoglobin in blood, producing a stronger T2* shortening effect (i.e., higher R2*). Thus, a high renal R2* (or low T2*) is interpreted as relative hypoxia, whereas a low R2* (longer T2*) suggests better oxygenation6,16 (Table 1). Physiologically, the renal medulla operates at borderline hypoxia even in health due to countercurrent exchange and high oxygen consumption for salt reabsorption. This is reflected by a higher R2* in medulla than cortex under normal conditions.16 In disease states, the “chronic hypoxia hypothesis” posits that sustained hypoperfusion and increased oxygen demand in nephrons lead to medullary hypoxia, which in turn triggers fibrosis and CKD progression.17 BOLD MRI has been used to investigate this, but results have been somewhat variable. Several factors may explain these inconsistencies. Acute edema can lengthen T2* and mask underlying hypoxia effects, and patient hydration, anemia, or differences in scanning protocols can all influence BOLD readings.16 Despite these nuances, BOLD MRI remains valuable for insight into renal physiology. It has confirmed that losing the normal corticomedullary oxygen gradient is a feature of CKD—in many CKD patients the medulla is not much more deoxygenated than the cortex, implying altered oxygen utilization.18,19 BOLD MRI findings have also been correlated with histological changes: Inoue et al. reported that lower cortical oxygenation (higher cortical R2*) was associated with greater interstitial fibrosis and higher HIF-1α (hypoxia-inducible factor-1α) expression in kidneys.17 Overall, BOLD MRI indicates that tissue hypoxia and altered oxygen handling are part of the pathophysiology of renal dysfunction, even if exact R2* values must be interpreted with caution. With ongoing improvements in technical standards and growing consensus on standardizing renal BOLD methods across centers,18,20 BOLD imaging may provide a noninvasive means to monitor renal oxygenation and guide hypoxia-targeted therapies. Diffusion tensor imaging (DTI) DTI extends conventional diffusion MRI by measuring not only the magnitude of water diffusion, as reflected by the apparent diffusion coefficient (ADC), but also its directionality. In the kidney, water diffusion is anisotropic—meaning it varies by direction—due to the organized structure of renal tubules, vessels, and collecting ducts that run radially in the medulla (Table 1). DTI yields metrics like fractional anisotropy (FA), which ranges from 0 (completely isotropic diffusion) to 1 (diffusion restricted to one direction). Healthy kidneys exhibit a high FA in the medulla (water preferentially diffuses along the aligned tubules) and a lower FA in the cortex (more isotropic tissue architecture).21 The corticomedullary difference in FA is a hallmark of normal renal microstructure, reflecting the highly organized tubular architecture of the medulla. Pathological processes that disrupt tubular organization or integrity, such as tubulointerstitial fibrosis, tubular atrophy, or edema, reduce anisotropy by making water diffusion more random, leading to a decline in medullary FA. Evidence from both human and animal studies supports the histologic relevance of DTI, demonstrating that FA decreases in proportion to tubular loss and fibrosis severity in interstitial fibrotic disease.22–24 Thus, DTI provides a unique window into renal tissue architecture and is particularly sensitive to microstructural integrity, with decreasing medullary FA indicating tubular destruction or disarray. Although DTI requires high image quality and is susceptible to motion artifacts, ongoing improvements in acquisition protocols have enhanced reproducibility and robustness. Intravoxel incoherent motion (IVIM) diffusion IVIM MRI is an advanced diffusion technique that separates true tissue diffusion from “pseudodiffusion” caused by microcirculation of blood. The apparent diffusion coefficient represents a composite measure influenced by both true molecular diffusion and perfusion-related effects. By acquiring diffusion-weighted images at multiple b-values and fitting a bi-exponential model, IVIM imaging separates these contributions and yields distinct parameters: D, which reflects the pure diffusion of tissue water; f, the perfusion fraction representing the signal contribution from blood flow; and D*, the pseudodiffusion coefficient associated with capillary blood flow velocity25–28 (Table 1). Physiologically, the renal cortex has higher ADC, D, D*, and f values than the medulla, reflecting greater free water motion within the interstitium and a higher perfusion fraction associated with its rich capillary network.29,30 In pathological conditions, diffusion metrics may tend to decrease in the presence of fibrosis, where collagen deposition and extracellular matrix expansion can restrict water mobility, leading to lower D and ADC. In addition, microvascular rarefaction may be associated with reductions in the perfusion fraction f and possibly D*. Therefore, CKD, particularly at advanced stages, tends to show reduced cortical ADC and D reflecting fibrosis, along with reduced f and D* consistent with capillary loss.11,14,29 This makes IVIM a comprehensive but complex biomarker. The main limitations are technical (it requires longer scan times and careful modeling), but continued improvements in MR hardware and image analysis are addressing these issues.31 Arterial spin labeling (ASL) perfusion ASL MRI is a non-contrast perfusion imaging technique that uses magnetically “labeled” blood water as an endogenous tracer of renal blood flow. In ASL, arterial blood is inverted (magnetically) just before entering the kidneys, and after a delay the MR image is acquired; subtracting this from a control image (no inversion) yields a perfusion-weighted image. This can be quantified in physiological units (e.g. mL/100 g/min) for renal cortical perfusion32,33 (Table 1). ASL predominantly reflects cortical perfusion in practice, as approximately 90% of renal blood flow is directed to the cortex. Biologically, ASL reflects how much blood the kidneys are receiving—a reduced ASL signal means reduced renal blood flow, which might result from narrowed renal arteries, afferent arteriole constriction, capillary loss, or other hemodynamic changes. Because it uses no exogenous contrast, ASL is safe in CKD patients even at low GFR. Research has established that declining renal function is accompanied by declining perfusion. Many ASL studies have shown that CKD patients have significantly lower cortical perfusion than healthy subjects.34–36 The physiological interpretation is that as kidneys incur damage (from diabetes, hypertension, etc.), there is loss of peritubular capillaries and dysregulation of blood flow, leading to renal hypoperfusion. Histologically, severe DKD kidneys often show arteriosclerosis and capillary rarefaction, which align with reduced ASL perfusion signals. Another advantage is that ASL can be repeated longitudinally without contrast or radiation, making it ideal for monitoring chronic disease. Technical challenges include lower signal-to-noise ratio and motion sensitivity, but continued improvements have made renal ASL feasible in 5 minute scan times.32 In summary, ASL MRI provides a window into renal perfusion—a key physiological parameter that declines with disease. It complements other markers like diffusion and oxygenation, since perfusion deficits often precede or accompany fibrosis and hypoxia in CKD. Corticomedullary differentiation (CMD) of optimal inversion time (TI) on steady-state free precession (SSFP) with inversion-recovery (IR) pulse using multi-T CMD of optimal TI using noncontrast SSFP with a spatially selective IR pulse (Time-SLIP concept) is a technique designed to enhance corticomedullary differentiation by exploiting intrinsic T1 differences between cortex and medulla. Multiple SSFP images are acquired at different inversion times (multi-TI), and for each individual the inversion time that yields the highest corticomedullary contrast ratio is selected as the optimal TI (Fig. 1). This individualized approach improves visualization of the corticomedullary junction compared with conventional unenhanced T1-weighted imaging and enables more reliable assessment of CMD, particularly in patients with impaired renal function in whom corticomedullary contrast is often attenuated. MRI Parameters for Predicting Kidney Function One practical goal of renal MRI research is to estimate kidney function (especially GFR) or to predict its decline. While direct GFR measurement by imaging is challenging, many MRI biomarkers show correlations with current GFR or with future loss of function. This section reviews how each MRI parameter relates to kidney function in cross-sectional studies, and whether it can help predict functional outcomes. T1 mapping Renal T1 mapping shows a moderate to strong inverse relationship with GFR, with higher cortical T1 values generally observed in patients with poorer kidney function (Fig. 2a, Fig. 3a). Across studies encompassing diverse CKD etiologies, cortical T1 has been reported to correlate with eGFR in the range of r = −0.4 to −0.8, indicating that prolonged T1 is associated with reduced renal function10,12,37–39 (Table 2). This association has been demonstrated in both native and transplanted kidneys,38 as well as in specific diseases such as IgA nephropathy12 and chronic kidney disease,39 supporting the link between elevated T1 values and functional impairment. However, the relationship between T1 and GFR may be less straightforward in DKD. In early diabetic nephropathy, glomerular hyperfiltration can preserve or even increase GFR despite ongoing structural injury, potentially weakening the association between T1 values and eGFR. Consistent with this, studies in diabetic nephropathy have reported minimal differences in absolute cortical or medullary T1 values compared with healthy controls, while alterations in corticomedullary T1 differentiation appear to better capture disease-related changes.4,9 Overall, although the strength and pattern of the association vary by disease type and stage, renal T1 mapping is generally related to kidney function across a wide spectrum of renal diseases. While T1 mapping does not directly measure filtration, it serves as an indirect indicator of renal functional status, as higher T1 values are commonly linked to structural damage and reduced renal function. Although factors such as anemia or hydration can influence T1 measurements (Table 3), the association between T1 values and GFR has been sufficiently consistent to support the use of T1 mapping as a complementary tool alongside blood-based assessments of renal function. T2 mapping The relationship between renal T2 values and GFR is less well defined than that observed for T1 values (Table 2). Several studies have shown no consistent linear correlation between T2 values and eGFR,10 likely because T2 values may either increase or decrease depending on the balance between acute inflammatory changes and chronic fibrosis (Table 3). From a pathological and physiological perspective, advanced CKD with severe fibrosis may be associated with reduced cortical T2, whereas moderate CKD with active inflammation may show elevated cortical or medullary T2 due to edema, resulting in a non-monotonic relationship with GFR. Consequently, T2 mapping is currently not a reliable standalone predictor of GFR and is better regarded as an adjunct parameter (Fig. 2b, Fig. 3b). BOLD imaging BOLD MRI is an indirect surrogate of renal tissue oxygenation and does not assess glomerular filtration function directly. Several studies have reported an inverse correlation between cortical R2* and GFR (Table 2), likely because reduced cortical perfusion and hypoxia accompany advanced renal disease16,17 (Fig. 2c, Fig. 3c). The relationship between medullary R2* and renal function remains controversial: while some studies have suggested that medullary R2* may provide information related to renal function or disease severity,40 others have reported weak or no significant correlation between medullary R2* and eGFR in CKD patients.16 This inconsistency may in part reflect the fact that the renal medulla is physiologically hypoxic even in healthy individuals with preserved GFR, as well as the sensitivity of R2* to acute physiological and technical factors such as supplemental oxygen administration or anemia at the time of imaging, which can influence R2* independently of underlying GFR (Table 3). Experimental studies nonetheless indicate that disease-related alterations in renal oxygenation can be detected by BOLD MRI. In diabetic rodent models, medullary R2* has been shown to increase early in the course of nephropathy, suggesting reduced oxygenation before overt structural damage becomes apparent.41 Similarly, dos Santos et al. reported early increases in medullary R2* on BOLD MRI in diabetic rats, consistent with hypoxia-related changes preceding established nephropathy.42 In human diabetes, however, findings have been heterogeneous: some studies reported higher medullary R2* in diabetic kidney disease, whereas others found no significant differences or even paradoxical trends compared with healthy controls.16,43–45 For example, a study of high-risk patients with type 2 diabetes observed no clear difference in renal R2* despite near-normal GFR, although subtle changes in the distribution of R2* values were noted.43 Overall, while BOLD MRI cannot replace direct GFR measurements, markedly abnormal findings—such as attenuation or loss of the normal corticomedullary oxygenation gradient—may be associated with more advanced disease and has been proposed as a marker of impaired oxygen handling.17,19 DTI DTI has been increasingly applied to evaluate renal microstructural integrity, and several studies have demonstrated an association between DTI-derived parameters and renal function.23,46,47 Among these parameters, FA, particularly in the renal medulla, has shown a moderate positive correlation with eGFR in several kidney disease settings, with correlation coefficients on the order of approximately r = 0.55 to 0.70 reported in some studies (Table 2). Reduced medullary FA has often been associated with impaired renal function, which is biologically plausible given that FA reflects the directional organization of renal tubules. This organization can be disrupted by tubulointerstitial fibrosis, tubular atrophy, and interstitial expansion, leading to more isotropic water diffusion. However, renal DTI measurements are technically demanding and may be influenced by motion, ROI selection, and tissue edema, which can complicate the interpretation of FA changes (Table 3). Importantly, DTI may also be sensitive to early microstructural alterations that precede overt functional decline. Several investigators have suggested that a reduction in medullary FA may serve as a potential early marker of tubulointerstitial injury, even before a measurable decrease in GFR occurs. Supporting this concept, studies in patients with early diabetes and preserved GFR have reported mildly reduced medullary FA compared with healthy controls, suggesting the presence of incipient microstructural changes that may be related to hyperglycemia and early tubular injury.46,48–50 Taken together, these findings indicate that renal DTI, and medullary FA in particular, may provide complementary information on structural damage underlying functional impairment and may be especially valuable for detecting early renal involvement. IVIM diffusion Diffusion-weighted MRI using the ADC has consistently shown an association with renal function. Higher cortical ADC values generally correlate with better eGFR, reflecting less restricted water diffusion in relatively preserved renal tissue (Table 2). Across diabetic kidney disease and other chronic kidney disease etiologies, numerous studies have reported a significant association between renal ADC values and GFR, with lower cortical ADC generally corresponding to reduced renal function37,51,52 (Fig. 2e, Fig. 3e). A recent systematic review further identified ADC as one of the most reproducible MRI-derived markers of CKD severity and progression risk.8 IVIM imaging extends diffusion MRI by separating true tissue diffusion (D) from perfusion-related effects (D* and f). The diffusion coefficient D behaves similarly to ADC and tends to decrease with worsening renal function, consistent with restricted water mobility caused by fibrosis and extracellular matrix expansion (Fig. 2f, Fig. 3f). However, perfusion-related parameters show a more complex relationship with GFR. Notably, early in diabetic nephropathy there appears to be a phase of increased perfusion related to glomerular hyperfiltration. In a study of normoalbuminuric diabetic patients, diffusion D was already lower than normal—suggesting incipient microstructural alterations—while the perfusion fraction f was higher than in healthy controls.48 This pattern was interpreted as reflecting renal hyperemia or hyperfiltration in early diabetes, which subsequently transitions to reduced perfusion as disease progresses. Consistent with this concept, advanced CKD is typically characterized by microvascular rarefaction and reduced cortical blood flow, leading to lower f and D*(Fig. 2g, h, Fig. 3g, h). As a result, IVIM perfusion parameters may exhibit weak or non-monotonic correlations with GFR across heterogeneous patient populations (Table 3). Overall, while ADC remains the most robust diffusion-based marker of renal function, IVIM provides complementary information by disentangling diffusion and perfusion components, offering additional insight into the pathophysiological mechanisms underlying renal dysfunction. ASL perfusion ASL MRI provides a quantitative, non-contrast estimate of renal perfusion and has generally shown an association with renal function across CKD populations. Because renal blood flow and glomerular filtration are physiologically related at the whole-organ level, ASL-derived cortical perfusion is often reduced in CKD and tends to show a positive association with renal function (Table 2). Across multiple studies, cortical perfusion is lower in CKD than in healthy kidneys, and the magnitude of reduction generally increases with more advanced disease stages (Fig. 2d, Fig. 3d). Reported correlations between ASL-measured cortical perfusion and eGFR are typically moderate, commonly ranging from approximately r ≈ 0.4 to 0.8, indicating that reduced renal perfusion often parallels declining kidney function.32,34–36,53,54 ASL may be particularly attractive for longitudinal monitoring, as it can be repeated without gadolinium-based contrast agents and is therefore applicable even in patients with reduced kidney function.32,54 Nevertheless, the strength of the ASL–eGFR relationship can vary because ASL is sensitive to physiological and technical factors, including hydration status, arterial transit time, labeling efficiency, respiratory motion, and the choice of acquisition/readout scheme (Table 3).32,34 Accordingly, statement papers emphasize standardized protocols and careful quality control to improve reproducibility and facilitate clinical translation.32 Overall, ASL-derived cortical perfusion represents a physiologically meaningful marker that complements diffusion- and oxygenation-based MRI metrics and provides noninvasive information linked to current renal functional status.32,34,37 CMD for Predicting Kidney Function CMD refers to the physiological and structural contrast between the renal cortex and medulla observed on MRI, and its usefulness has been reported using parameters such as ΔT1, ΔADC, and ΔR2*. CMD is attractive as a functional biomarker because it reflects the integrity of corticomedullary compartmentalization, potentially influenced by differences in tubular density, perfusion, oxygenation, and water content, rather than absolute parameter values alone. Importantly, CMD can deteriorate before overt end-stage morphological changes become apparent, and loss of CMD has been consistently associated with reduced kidney function across CKD cohorts.10,37,55 T1-based CMD A substantial body of work supports CMD on T1-weighted imaging and/or T1 mapping as a clinically meaningful correlate of renal dysfunction. Early studies suggested that diminished corticomedullary contrast is a common imaging feature in renal insufficiency and may reflect underlying alterations in renal tissue composition and water handling37,55 (Table 4). Quantitative measures such as corticomedullary ΔT1 (T1-CMD) may capture disease-related changes even when absolute cortical or medullary T1 values show limited separation; this may be relevant in DKD, where absolute cortical or medullary T1 values can show limited separation and ratio-based indicators like CMD measures may be more sensitive.10,56 Multimodal CKD studies have further shown that reduced T1-CMD accompanies declining eGFR and complements other functional MRI markers in characterizing CKD severity.37 Because T1-CMD depends on acquisition and analysis methods, standardized methodology and attention to reproducibility are important for its reliable application.10,56 Oxygenation-gradient CMD on BOLD MRI CMD can also be conceptualized as the corticomedullary oxygenation gradient on BOLD MRI, reflected by differences in R2* between cortex and medulla. Physiologically, the renal medulla is more hypoxic than the cortex, producing a characteristic gradient. Attenuation or flattening of this gradient suggests altered oxygen handling and may be associated with CKD severity18,19,40 (Table 4). Although BOLD-derived CMD is sensitive to physiological and technical confounders such as hydration status, anemia, and oxygen administration, careful interpretation and standardized acquisition have been emphasized in consensus and statement papers when assessing corticomedullary oxygenation patterns.16,18 Diffusion-based CMD Diffusion MRI provides another CMD dimension: the cortex–medulla difference in diffusivity (commonly ΔADC) is reduced or may be lost in CKD compared with healthy kidneys37 (Table 4). In settings with advanced cortical injury—particularly in the presence of high cortical fibrosis—diffusion-based CMD can even invert, with cortical ADC becoming lower than medullary ADC (i.e., negative ΔADC).57 This pattern is consistent with marked cortical microstructural restriction associated with fibrosis and tubular damage, and diffusion-based CMD can complement absolute ADC/D values by capturing altered corticomedullary separation not fully reflected by global eGFR alone.11,37,57,58 CMD of optimal TI on SSFP with IR pulse using multi-TI Using noncontrast SSFP MRI with a spatially selective inversion recovery pulse, this method can improve visualization of the corticomedullary junction, which may facilitate ROI placement when corticomedullary borders are less conspicuous.59,60 In sequence comparisons, FASE (fast advanced spin echo) may provide a higher corticomedullary contrast ratio, whereas SSFP can offer favorable overall visualization depending on imaging conditions and disease severity.61,62 With this approach, the corticomedullary contrast ratio on SSFP images acquired at optimal TI has been reported to correlate with renal function as reflected by eGFR (r = 0.314, P = 0.011)59 (Figs. 4, 5) (Table 4). Although these findings suggest that signal-intensity–based CMD using optimal-TI SSFP may provide a practical, non-contrast imaging marker complementary to parametric CMD approaches, the current evidence remains limited in scope and external validation. Compared with more established renal MRI biomarkers such as T1 mapping, diffusion-based metrics, ASL perfusion, and BOLD imaging, this technique is less widely standardized and has been evaluated in relatively fewer cohorts. Therefore, it should be considered a promising but still evolving method that may benefit from further validation in larger and multicenter studies before broader clinical application can be established. MRI Parameters for Predicting Prognosis A central unmet need in CKD and DKD management is improved prediction of future renal function decline beyond conventional clinical markers. Although eGFR and albuminuria remain foundational, progression is heterogeneous—particularly in DKD, where hyperfiltration, variable inflammatory activity, and mixed vascular/tubulointerstitial injury can lead to discordance between functional indices and structural damage.4–6,9 In this context, multiparametric renal MRI is attractive because it can quantify complementary domains—fibrosis/edema (T1/T2), microstructure (diffusion/DTI/IVIM), perfusion (ASL), and oxygenation (BOLD)—that may precede or accompany irreversible nephron loss.5,8,9 T1 mapping Native T1 mapping has strong histopathologic relevance as a marker of interstitial expansion due to fibrosis and/or edema,11–14 and it correlates with current kidney function across many CKD etiologies.10,12,37–39 This makes T1 a biologically plausible prognostic candidate, as fibrosis is a recognized substrate for future decline. At present, however, fewer longitudinal studies have directly established T1 mapping as an independent predictor of eGFR slope or hard renal outcomes compared with diffusion MRI. However, emerging evidence suggests that native T1 may also predict kidney outcome events in CKD cohorts,63 and that DKD may show weaker cross-sectional T1–eGFR associations in early stages due to hyperfiltration and heterogeneous injury.4,9 Accordingly, in prognosis-oriented frameworks, T1 mapping may be most useful when integrated with other domains (e.g., perfusion and diffusion), rather than as a standalone predictor (Table 5). BOLD imaging Renal BOLD MRI is conceptually linked to the “chronic hypoxia hypothesis,” in which impaired oxygen handling may contribute to tubulointerstitial fibrosis and CKD progression.17 Histology-linked clinical work has shown associations between BOLD-derived indices and fibrotic/hypoxia-related tissue changes,17 and longitudinal studies have suggested that baseline renal BOLD metrics may help predict subsequent eGFR decline independent of conventional markers64 (Table 5). Nevertheless, BOLD prognostication requires cautious interpretation because R2* is an indirect surrogate influenced by confounders such as hydration status, anemia, and oxygen administration, and protocols vary across centers.16,18,20 As a result, BOLD MRI is best positioned as a complementary prognostic domain, particularly when interpreted alongside perfusion and diffusion markers. Diffusion-based imaging Among renal MRI biomarkers, diffusion-derived parameters currently have some of the strongest longitudinal evidence for outcome prediction. In CKD cohorts, lower baseline cortical ADC has been consistently associated with faster subsequent eGFR decline, and diffusion MRI metrics have been linked to future functional loss in both native CKD and renal allograft populations.52 These findings suggest that diffusion restriction reflects microstructural injury—such as fibrosis and tubular damage—that is directly relevant to progression risk rather than merely contemporaneous renal function. Beyond conventional ADC, DTI- and IVIM-derived parameters may further refine prognostic assessment (Table 5). Although these metrics are strongly associated with renal function and disease severity cross-sectionally,22,23,29,30 prospective data directly linking baseline FA or IVIM parameters to hard renal outcomes remain comparatively limited. Taken together, diffusion MRI—particularly cortical ADC—appears to be one of the most promising MRI-based prognostic markers to date. A pragmatic approach is to incorporate diffusion-, DTI-, and IVIM-derived parameters within multiparametric MRI models, where they may improve progression-risk stratification beyond reliance on single imaging or clinical markers alone.5,8 CMD metrics as prognostic features Although the literature strongly supports CMD as a correlation of current renal function than as an independent longitudinal predictor, CMD may plausibly contribute to prognosis when integrated with quantitative metrics such as ADC and perfusion that have stronger outcome-based evidence (Table 5). This supports the broader concept that multiparametric MRI can provide risk-relevant information not captured by eGFR alone and may eventually enable MRI-based stratification aligned with established clinical frameworks. Clinical Implications of Functional Renal MRI Functional renal MRI has the potential to complement conventional clinical assessment by providing noninvasive, domain-specific information on renal microstructure, perfusion, and oxygenation that is not captured by serum creatinine, eGFR, or proteinuria alone. In routine practice, eGFR reflects global filtration but remains insensitive to early tubulointerstitial injury and may be discordant with structural disease burden, particularly in DKD where hyperfiltration and heterogeneous injury patterns can obscure ongoing damage.4,9 Multiparametric MRI therefore offers a “virtual biopsy” concept that may help identify subclinical abnormalities, characterize dominant pathophysiologic pathways such as fibrosis/edema, microvascular dysfunction, hypoxia, and tubular disorganization, and provide a quantitative baseline for longitudinal monitoring.5,8,9 A potential clinical implication is improved risk stratification. Several MRI markers correlate with contemporaneous kidney function, while diffusion-derived metrics—particularly cortical ADC—have been associated with subsequent eGFR decline in longitudinal CKD cohorts, suggesting possible value for progression risk assessment beyond a single time-point eGFR.52,57 CMD-based measures, whether derived from T1 mapping, diffusion, BOLD, or signal-intensity techniques, further reflect disruption of corticomedullary compartmentalization and may serve as indicators of disease severity; however, their role in predicting long-term outcomes remains less well established compared with diffusion-derived parameters.10,18,37,57 In this context, combining complementary MRI domains—for example, diffusion for microstructural injury, ASL for perfusion, and T1/T2 for interstitial expansion and inflammatory activity—may provide a more comprehensive assessment than reliance on any single parameter, although the incremental prognostic value of such multiparametric approaches requires further validation.5,8 Another potential implication is the ability to track disease evolution and treatment response without gadolinium-based contrast agents. Non-contrast approaches such as ASL, diffusion/DTI/IVIM, BOLD, and T1/T2 mapping can be repeated safely, which is particularly relevant for CKD populations in whom contrast administration is often avoided.32 This repeatability may support longitudinal follow-up and could facilitate clinical studies by providing intermediate imaging biomarkers that reflect tissue-level changes. In DKD, where early hemodynamic changes and later microvascular rarefaction coexist, perfusion- and diffusion-based markers may offer complementary insights into evolving injury, although their precise role in predicting irreversible functional decline remains to be clarified.4,9,53 From a practical standpoint, clinical translation requires careful consideration of standardization, quality control, and interpretability. Many functional MRI measures are influenced by physiological and technical confounders—such as hydration status, anemia, motion, and protocol variability—necessitating harmonized acquisition and analysis strategies across centers to improve reproducibility and generalizability.10,18,32 At present, functional renal MRI is best considered an adjunct to existing clinical frameworks, with potential to support pathophysiology-informed phenotyping and contribute to risk assessment. Further multicenter studies and outcome-based validation will be required to determine how these quantitative MRI biomarkers may be integrated into routine CKD and DKD management.5,8 Conclusion Multiparametric renal MRI represents a promising, noninvasive approach for assessing kidney structure and function beyond conventional biomarkers. By integrating information on fibrosis/edema, microstructure, perfusion, and oxygenation, it may provide a physiology-informed “virtual biopsy” without the need for ionizing radiation or gadolinium. Many MRI-derived parameters have been shown to correlate with current renal function across CKD etiologies, and CMD is consistently associated with reduced eGFR. With regard to prognosis, diffusion-derived metrics currently have the most consistent longitudinal support, whereas native T1 and BOLD indices may also provide additional information in selected settings, although the available evidence remains limited. ASL offers a repeatable, non-contrast method for assessing renal perfusion and may be useful for longitudinal monitoring, particularly in DKD where interpretation of eGFR can be complex. Despite these advances, several challenges remain for clinical translation, including the influence of physiological confounders, variability in acquisition protocols, and the relative lack of large prospective outcome studies. At present, the added value of multiparametric integration over individual biomarkers has not been fully established, although combining complementary metrics may provide a more comprehensive assessment of renal pathophysiology. Further efforts toward standardization, reproducibility, and multicenter validation will be important to better define the role of renal MRI in clinical practice.