ASL, DTI, and T1 mapping in identifying immune and non-immune pathogenic factors of chronic allograft injury

Abstract

Background

Chronic allograft injury (CAI) has a complex etiology and various clinical and pathological manifestations, posing challenges for clinical diagnosis and treatment. This study aimed to explore the value of arterial spin labeling (ASL), diffusion tensor imaging (DTI), and T1 mapping in identifying the causes of CAI.

Methods

This study included 69 kidney transplant recipients. Based on the estimated glomerular filtration rate (eGFR) and biopsy results, the recipients were classified into three groups: Group A, recipients with good renal allograft function (eGFR ≥60 mL/min/1.73 m²); Group B, recipients with immune-mediated factors; and Group C, recipients with non-immune-mediated factors. ASL, DTI, and T1 mapping were performed in all patients, and magnetic resonance imaging (MRI) parameters were calculated and analyzed.

Results

Renal blood flow (RBF) was higher in the recipients with normal renal allograft function than in those with CAI. Cortical and medullary apparent diffusion coefficient (ADC) values were lower in the immune factors group (2.05±0.22/2.06±0.22 ×10⁻³ mm²/s) than in the non-immune factors group (2.24±0.23/2.25±0.24 ×10⁻³ mm²/s). Conversely, cortical T1 values were higher in in the immune-mediated group than in the non-immune-mediated group. The area under the curve (AUC) value for cortical ADC, medullary ADC, and cortical T1 values in identifying CAI pathogenic factors was 0.716, 0.687, and 0.806, respectively. The combination of DTI and T1 mapping achieved better diagnostic efficiency for distinguishing between patients in Group B and Group C than DTI or T1 mapping alone [AUC =0.825, 95% confidence interval (CI): 0.699–0.952, P<0.001].

Conclusions

ASL, DTI, and T1 mapping can distinguish between recipients with normal renal allograft function and those with CAI. Further, DTI and T1 mapping can differentiate between immune and non-immune pathogenic factors of CAI.

Introduction

Kidney transplantation is the optimal treatment for end-stage renal disease (1). However, chronic allograft injury (CAI) remains a critical long-term complication that requires particular attention (2). CAI is characterized by the progressive deterioration of renal allograft function, directly affecting graft survival and recipient prognosis. The causes of CAI are complex and can be divided into two categories: surgical factors and medical factors. Surgical factors mainly include graft renal artery stenosis. In clinical practice, medical factors are the most common causes of CAI and exert more intricate effects. Medical factors can be further classified as immune-mediated and non-immune-mediated injuries. The former includes cellular and antibody-mediated rejection, while the latter includes ischemia–reperfusion injury, calcineurin inhibitor nephrotoxicity, recurrence of primary glomerular diseases, viral infections, and post-transplant hypertension (3).

Currently, the differential diagnosis of CAI primarily relies on laboratory tests and renal allograft biopsy (4). Clinical indicators can provide preliminary diagnostic evidence; however, their sensitivity in distinguishing among different etiologies is limited. Renal biopsy can provide a pathological classification diagnosis; however, it is invasive and lacks repeatability (2). The etiology of CAI is complex, with diverse clinical and pathological manifestations of varying severity, posing challenges for its diagnosis and treatment.

Functional magnetic resonance imaging (fMRI) is a non-invasive assessment method capable of simultaneously providing morphological and functional information on transplanted kidneys. It provides important imaging evidence for the non-invasive assessment of renal dysfunction, tissue oedema, fibrosis, and microcirculatory changes (5-8). Arterial spin labeling (ASL) uses arterial blood water protons as endogenous tracers for imaging. It quantifies renal blood flow (RBF) perfusion values based on the signal differences between labeled and unlabeled images (9). Diffusion tensor imaging (DTI) constructs models of tissue microstructure based on the anisotropic characteristics of water molecule diffusion by applying multidirectional diffusion-sensitizing gradient fields. Its quantitative parameters—apparent diffusion coefficient (ADC) and fractional anisotropy (FA)—can be used to assess renal fibrosis and renal insufficiency (10). T1 mapping enables sensitive detection of renal allograft edema and fibrosis progression by directly measuring tissue T1 relaxation time (11).

This study preliminarily applied fMRI to the etiological differentiation of CAI to assess its potential diagnostic value in distinguishing between immune and non-immune factors. The study findings may provide novel imaging solutions to long-standing challenges in pathologic staging. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0380/rc).

Methods

Study participants

This prospective study of kidney allograft recipients was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Ethics Committee of The First Affiliated Hospital of Soochow University (No. 2022-412), and written informed consent was obtained from all the participants.

A total of 69 recipients who underwent allogeneic kidney transplantation at The First Affiliated Hospital of Soochow University between May 2022 and May 2025 were consecutively enrolled in the study, among whom 45 had biopsy-confirmed CAI. The donor sources for kidney transplantation in this study included two categories: living related donors and deceased donors (organ donation after death). All recipients underwent ASL, DTI, and T1 mapping. The inclusion criterion for the control group (Group A) was an estimated glomerular filtration rate (eGFR) >60 mL/min/1.73 m². Based on the biopsy results and the medical pathogenic factors of CAI, the remaining recipients were categorized into an immune factors group (Group B) and a non-immune factors group (Group C). The exclusion criteria were as follows: (I) patient refusal; and/or (II) contraindications to magnetic resonance imaging (MRI).

Patient demographic characteristics, clinical information, and transplantation-related data were collected. Laboratory parameters, including hemoglobin, serum creatinine, and cystatin C, were obtained within 24 hours before and after MRI examination. The eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation (12).

MRI examination

Before MRI examination, recipients were required to fast and refrain from fluid intake for 4 hours and to complete breath-holding training (>15 s). All patients underwent MRI on a 1.5-T scanner (Ingenia Ambition, Philips Healthcare, Best, the Netherlands) equipped with a 28-channel phased-array coil. The imaging coverage extended from the level below the costal arch to below the pubic symphysis. The MRI protocol included routine sequences, as well as ASL, DTI, and T1 mapping. Conventional MRI comprised coronal T2-weighted imaging (T2WI) and axial T1-weighted imaging (T1WI) for morphological assessment.

The T1WI parameters were as follows: repetition time (TR), 417 ms; echo time (TE), 10 ms; voxel size, 0.99×1.22×5.00 mm³; field of view (FOV), 350×350 mm²; resolution, 352×286; and 20 slices. The total acquisition time for T1WI was approximately 2 minutes.

The T2WI parameters were as follows: TR, 1,086 ms; TE, 80 ms; voxel size, 1.45×1.45×5.00 mm³; FOV, 400×400 mm²; resolution, 276×276; and 30 slices. The total acquisition time for T2WI was approximately 3 minutes.

The ASL parameters were as follows: TR, 3,963 ms; TE, 15 ms; voxel size, 3.75×3.75×8 mm³; FOV, 240×240×88 mm³; turbo spin echo factor, 20; post-label delay, 1,800 ms; and 11 slices. The total acquisition time for ASL was approximately 5 minutes.

The T1 mapping parameters were as follows: modified Look-Locker inversion-recovery sequence; TR, 2.8 ms; TE, 1.3 ms; flip angle, 35°; voxel size, 2.2×2.2 mm²; FOV, 300×300 mm²; slice thickness, 5 mm; and 5 slices. The total acquisition time for T1 mapping was approximately 1 minute.

The DTI parameters were as follows: diffusion directions, 15; b values, 0 and 600 s/mm²; TR, 2,300 ms; TE, 67 ms; voxel size, 2.5×2.5×5 mm³; FOV, 320 × 400 mm²; slice thickness/gap: 5.0/1.0 mm; and 20 slices. The total acquisition time for DTI was approximately 3 minutes.

Image analysis

All images were post-processed on a Philips workstation. For ASL analysis, the axial T1WI image closest to the plane of the kidney’s maximal transverse diameter was selected. The entire renal cortical contour was manually traced and copied onto the RBF map. For DTI and T1 mapping, the coronal T2WI image closest to the plane of the kidney’s maximal diameter was selected. Six circular regions of interest (ROIs) were placed in the cortical and medullary regions at the upper, middle, and lower poles on coronal T2WI images. Each ROI comprised approximately 12 voxels. Areas containing blood vessels, hematomas, or cysts were avoided, and the ROIs were then propagated to the DTI and T1 mapping images. The corresponding T1, ADC, and FA values were recorded for each ROI, and mean values were subsequently calculated. Image analysis was independently performed in a blinded manner by two radiologists with 8 and 3 years of experience in abdominal imaging, respectively.

Renal allograft biopsy

In total, 45 patients underwent ultrasound-guided percutaneous renal transplant biopsy. The interval between the biopsy and MRI examination was less than 48 hours. Kidney transplant biopsies were evaluated according to the Banff classification (13). The Banff pathological scoring system primarily includes tubular atrophy (ct score), interstitial fibrosis (ci score), peritubular capillaritis (ptc score), glomerulitis (g score), interstitial inflammation (i score), and tubulitis (t score). The score ranges from 0 to 3, with 0 indicating no lesion and 3 representing the most severe condition (14).

Statistical analysis

Statistical analyses were performed using SPSS 27 statistical software. Measures were tested for normality, and those that conformed to a normal distribution were expressed as mean ± standard deviation. One-way analysis of variance was used to compare the renal cortical and medullary indicators among the groups, followed by pairwise comparisons using the Bonferroni correction. Non-normally distributed data are presented as median [interquartile range (IQR)] and were compared using the Kruskal-Wallis H test. Categorical variables were compared using the chi-squared test or Fisher’s exact test, as appropriate. An independent-samples t-test was used to compare the differences in each index between the cortex and medulla in each group. Interobserver reproducibility of MRI measurements was evaluated using the intraclass correlation coefficient (ICC). Pearson correlation analysis was used to examine the relationship between MRI parameters and eGFR, while Spearman rank correlation analysis was used to assess the association between MRI parameters and Banff pathological scores. The diagnostic efficiency of MRI parameters was evaluated using receiver operating characteristic (ROC) curve analysis. The DeLong test was used to compare areas under the curve (AUCs). A two-sided P value <0.05 was considered statistically significant.

Results

Patient characteristics

A total of 69 kidney transplant recipients were included in this study (44 males and 25 females). The median age of the recipients was 41.0 years (IQR, 32.5–48.5 years). The recipients were divided into three groups based on their eGFR and biopsy findings: Group A, recipients with good renal allograft function (eGFR ≥60 mL/min/1.73 m²); Group B, recipients with immune factors; and Group C, recipients with non-immune factors. The characteristics of the study population are summarized in Table 1.

Table 1. Summary of study characteristics.

Characteristics	Group A (n=24)	Group B (n=21)	Group C (n=24)	P
Gender (male/female)	24 (10/14)	21 (14/7)	24 (20/4)*	0.012
Age (years)	35.7±11.6	46.6±12.4*	40.6±9.0	0.006
BMI (kg/m2)	20.17±3.52	22.17±4.00	23.43±4.13*	0.018
SCr (μmol/L)	78.62±21.41	186.34±93.40*	148.10±94.23*	<0.001
eGFR (mL/min/1.73 m2)	97.65±20.49	43.12±21.43*	63.83±30.62*#	<0.001
CysC (mg/L)	1.41±0.28	2.63±1.57*	1.71±0.99	0.003
Albumin (g/L)	43.96±11.93	37.87±3.69*	37.63±3.14*	0.007
Hemoglobin (g/L)	107.37±24.13	108.43±23.77	135.46±20.31*#	<0.001
Cortical RBF (mL/100 g/min)	272.74±44.58	170.61±63.00*	192.33±78.08*	<0.001
Cortical ADC (×10−3 mm2/s)	2.30±0.20	2.05±0.22*	2.24±0.23#	<0.001
Medullary ADC (×10−3 mm2/s)	2.40±0.29	2.06±0.22*	2.25±0.24#	<0.001
Cortical FA	0.17±0.01	0.15±0.03*	0.13±0.02*	<0.001
Medullary FA	0.34±0.04	0.24±0.06*	0.28±0.06*	<0.001
Cortical T1 (ms)	1,283.63±133.53	1,439.89±119.34*	1,319.85±106.89#	<0.001
Medullary T1 (ms)	1,746.78±68.45	1,798.07±127.75	1,773.70±95.03	0.226

Data are presented as mean ± standard deviation or number. Group A, recipients with good renal allograft function (eGFR ≥60 mL/min/1.73 m²); Group B, recipients with CAI caused by immune factors; Group C, recipients with CAI caused by non-immune factors. *, compared with Group A (P<0.05); ^#, compared with Group B (P<0.05). ADC, apparent diffusion coefficient; BMI, body mass index; CAI, chronic allograft injury; CysC, cystatin C; eGFR, estimated glomerular filtration rate; FA, fractional anisotropy; eGFR, estimated glomerular filtration rate; RBF, renal blood flow; SCr, serum creatinine; T1, longitudinal relaxation time.

Interobserver reproducibility for MRI measurements

All MRI parameters demonstrated good consistency. The ICC values for cortical RBF, cortical ADC, medullary ADC, cortical FA, medullary FA, cortical T1, and medullary T1 were 0.919, 0.903, 0.910, 0.899, 0.901, 0.923, and 0.911, respectively.

Differences in MRI parameters of the renal cortex and medulla between groups

The cortical ADC values of Groups A, B, and C were 2.30±0.20, 2.05±0.22, and 2.24±0.23 ×10⁻³ mm²/s, respectively, while the corresponding medullary ADC values were 2.40±0.29, 2.06±0.22, and 2.25±0.24 ×10⁻³ mm²/s, respectively. No statistically significant differences were observed between the cortex and medulla in any of the groups (P=0.158, 0.923, and 0.821, respectively).

The FA values for the cortex of Groups A, B, and C 0.17±0.01, 0.15±0.03, and 0.13±0.02, respectively) were lower than those for the medulla 0.34±0.04, 0.24±0.06, and 0.28±0.06, respectively). Similarly, the T1 values for the cortex of Groups A, B, and C (1,283.63±133.53, 1,439.89±119.34, and 1,319.85±106.89 ms, respectively) were lower than those for the medulla (1,746.78±68.45, 1,798.07±127.75, and 1,773.70±95.03 ms, respectively). There were statistically significant differences between the cortex and medulla in each group (all P<0.001).

Differences in MRI parameters among the groups

The cortical RBF of Group A (272.74±44.58 mL/100 g/min) was significantly higher than that of Group B (170.61±63.00 mL/100 g/min, P<0.001) and Group C (192.33±78.08 mL/100 g/min, P<0.001). The cortical and medullary ADC values of Group A (2.30±0.20/2.40±0.29 ×10⁻³ mm²/s) were higher than those of Group B (2.05±0.22/2.06±0.22 ×10⁻³ mm²/s, all P<0.001) and Group C (2.24±0.23/2.25±0.24 ×10⁻³ mm²/s, all P<0.001). Additionally, Group C exhibited significantly higher cortical and medullary ADC values than Group B (P=0.018/0.043).

Cortical and medullary FA demonstrated similar patterns, but no statistically significant differences were observed between Group B and Group C (P=0.458/0.126). The cortical T1 value of Group A (1,283.63±133.53 ms) was significantly lower than that of Group B (1,439.89±119.34 ms, P<0.001) and Group C (1,319.85±106.89 ms, P=0.904). Conversely, Group B exhibited a significantly higher cortical T1 value than Group C (P=0.003). There were no statistically significant differences in the medullary T1 values among the three groups (Table 1 and Figures 1,2).

Figure 1.

ASL, DTI, and T1 mapping images of different groups. mpMRI in a 26-year-old woman (Patient A from Group A) with normal eGFR (eGFR, 110.5 mL/min/1.73 m²), a 54-year-old man (Patient B from Group B) with CAI caused by immune factors, and a 53-year-old woman (Patient C from Group C) with CAI caused by non-immune factors. Representative T2-weighted images (A,G,M), ADC maps (B,H,N), FA maps (C,I,O), T1 maps (D,J,P), T1-weighted images (E,K,Q), and RBF maps (F,L,R). Group A, recipients with good renal allograft function (eGFR ≥60 mL/min/1.73 m²); Group B, recipients with CAI caused by immune factors; Group C, recipients with CAI caused by non-immune factors. ADC, apparent diffusion coefficient; ASL, arterial spin labeling; eGFR, estimated glomerular filtration rate; FA, fractional anisotropy; mpMRI, multiparametric magnetic resonance imaging; RBF, renal blood flow; T1, longitudinal relaxation time.

Figure 2.

Comparison of MRI parameters among different groups. Comparison of cortical RBF (A), cortical ADC (B), medullary ADC (C), cortical FA (D), medullary FA (E), cortical T1 (F), and medullary T1 (G) among Group A, Group B, and Group C. *, P<0.05; **, P<0.01; ***, P<001; ns, not significant. Group A, recipients with good renal allograft function (eGFR ≥60 mL/min/1.73 m²); Group B, recipients with CAI caused by immune factors; Group C, recipients with CAI caused by non-immune factors. ADC, apparent diffusion coefficient; CAI, chronic allograft injury; eGFR, estimated glomerular filtration rate; FA, fractional anisotropy; MRI, magnetic resonance imaging; RBF, renal blood flow; T1, longitudinal relaxation time.

Relationships between MRI parameters, eGFR, and Banff histopathological scores

Renal cortical RBF, ADC, and FA showed positive correlations with eGFR (r=0.753, 0.561, and 0.377, respectively; P<0.001). Cortical T1 was negatively correlated with eGFR (r=−0.492, P<0.001). Medullary ADC and FA showed positive correlations with eGFR (r=0.559 and 0.744, respectively; P<0.001), while medullary T1 showed no significant correlation with eGFR (P=0.958) (Figure 3).

Figure 3.

Correlation between ASL, DTI, T1 mapping, and eGFR. ADC, apparent diffusion coefficient; ASL, arterial spin labeling; DTI, diffusion tensor imaging; eGFR, estimated glomerular filtration rate; FA, fractional anisotropy; RBF, renal blood flow; T1, longitudinal relaxation time.

Cortical RBF was negatively correlated with the cv score (r=−0.445, P=0.002). The cortical and medullary ADC values showed negative correlations with the cg and C4d scores (r=−0.457/−0.529, and −0.402/−0.494, respectively; P<0.05). Medullary FA showed negative correlations with the cv, ah, and C4d scores (r=−0.437, −0.603, and −0.401, respectively; P<0.05). Cortical T1 exhibited positive correlations with the g, ptc, cg, and C4d scores (r=0.470, 0.513, 0.510, and 0.481, respectively; P<0.05). Medullary T1 showed negative correlations with the ci, ct, ti, and i-IFTA scores (r=−0.475, −0.515, −0.446, and −0.432, respectively; P<0.05) (Figure 4).

Figure 4.

Correlation between ASL, DTI, T1 mapping, and Banff histopathological scores. In the correlation heatmap, darker colors represent higher degrees of correlation. ADC, apparent diffusion coefficient; ah, arteriolar hyalinosis; ASL, arterial spin labeling; cg, glomerular basement membrane double contours; ci, interstitial fibrosis; ct, tubular atrophy; cv, vascular fibrous intimal thickening; DTI, diffusion tensor imaging; FA, fractional anisotropy; g, glomerulitis; i, interstitial inflammation; i-IFTA, inflammation in areas of fibrosis; ptc, peritubular capillaritis; RBF, renal blood flow; t, tubulitis; t-IFTA, tubulitis in areas of interstitial fibrosis; T1, longitudinal relaxation time; ti, total inflammation; v, intimal arteritis.

Diagnostic performances of ASL, DTI, and T1 mapping

The cortical and medullary ADC values and the cortical T1 values could differentiate between the immune and non-immune groups (AUC =0.716, 0.687, and 0.806, respectively). The combined diagnostic performance of DTI and T1 mapping (AUC =0.825, P<0.001) was higher than that of either method alone, indicating improved diagnostic efficacy (Table 2 and Figure 5).

Table 2. Diagnostic efficiency of MRI parameters.

Parameters	Group A vs. Group B + C			Group A vs. Group B			Group A vs. Group C			Group B vs. Group C
	AUC (95% CI)	P		AUC (95% CI)	P		AUC (95% CI)	P		AUC (95% CI)	P
Cortical RBF	0.853 (0.765–0.941)	<0.001		0.905 (0.806–1.000)	<0.001		0.807 (0.678–0.937)	<0.001		–
Cortical ADC	0.678 (0.552–0.804)	<0.05		0.813 (0.687–0.940)	<0.001		–			0.716 (0.563–0.869)	<0.001
Medullary ADC	0.743 (0.617–0.868)	<0.001		0.834 (0.715–0.954)	<0.001		–			0.687 (0.532–0.841)	<0.001
Cortical FA	0.861 (0.772–0.950)	<0.001		0.793 (0.642–0.944)	<0.001		0.921 (0.830–1.000)	<0.001		–
Medullary FA	0.840 (0.749–0.932)	<0.001		0.899 (0.792–1.000)	<0.001		0.786 (0.657–0.915)	<0.001		–
Cortical T1	0.718 (0.582–0.853)	<0.05		0.827 (0.699–0.956)	<0.001		–			0.806 (0.674–0.937)	<0.001
mpMRI	0.906 (0.836–0.977)	<0.001		0.978 (0.945–1.000)	<0.001		0.924 (0.848–0.999)	<0.001		0.825 (0.699–0.952)	<0.001

Group A, recipients with good renal allograft function (eGFR ≥60 mL/min/1.73 m²); Group B, recipients with CAI caused by immune factors; Group C, recipients with CAI caused by non-immune factors. ADC, apparent diffusion coefficient; AUC, area under the curve; CAI, chronic allograft injury; CI, confidence interval; FA, fractional anisotropy; mpMRI, multiparametric magnetic resonance imaging; MRI, magnetic resonance imaging; RBF, renal blood flow; T1, longitudinal relaxation time.

Figure 5.

Diagnostic value of MRI parameters in different groups. (A) Group A vs. Group B + Group C; (B) Group A vs. Group B; (C) Group A vs. Group C; (D) Group B vs. Group C. Group A, recipients with good renal allograft function (eGFR ≥60 mL/min/1.73 m²); Group B, recipients with CAI caused by immune factors; Group C, recipients with CAI caused by non-immune factors. ADC, apparent diffusion coefficient; ASL, arterial spin labeling; CAI, chronic allograft injury; DTI, diffusion tensor imaging; eGFR, estimated glomerular filtration rate; FA, fractional anisotropy; RBF, renal blood flow; T1, longitudinal relaxation time.

Discussion

CAI typically occurs during the mid- to long-term phase following kidney transplantation, leading to a gradual decline in graft function. CAI has diverse etiologies and complex clinical manifestations. These characteristics pose challenges in the diagnosis and treatment of CAI (15). To date, fMRI studies of transplanted kidneys have largely focused on assessing renal function, distinguishing between normal and abnormal function, and analyzing correlations between MRI parameters and eGFR and Banff pathological scores (16-18). However, research on imaging-based differentiation of the various etiological factors of CAI, particularly immune and non-immune factors, is limited. The unique value of this study lies in its use of multiparametric fMRI for the non-invasive differential diagnosis of the etiological factors of CAI. The results revealed that DTI and T1 mapping provide valuable information for the clinical differentiation of immune and non-immune factors in CAI. This offers a new imaging approach for determining the etiology of CAI and guiding its clinical management.

Renal allograft injury is commonly accompanied by changes in renal blood perfusion, which can be non-invasively detected by ASL (19). Consistent with previous studies (20), we found that ASL can differentiate between patients with normal renal allograft function and those with CAI (P<0.001), and that RBF is positively correlated with eGFR (r=0.753, P<0.001). As Chhabra et al. (21). reported, when renal allograft dysfunction occurs, capillary injury leads to reduced renal cortical blood flow. However, this study found that ASL cannot further differentiate between immune-related and non-immune-related causes of CAI (P=0.668). ASL primarily measures tissue perfusion by tagging water protons in arterial blood as an endogenous tracer (22). It provides information on the overall RBF perfusion, reflecting the hemodynamic status of the kidney at a macroscopic level. However, there are differences in the microscopic pathological changes between immune-related and non-immune-related CAI, and these microscopic alterations may not be reflected in overall perfusion parameters.

The FA value in the renal medulla of the transplanted kidney was significantly greater than that in the cortex. However, consistent with previous findings (23), the ADC value could not accurately distinguish between the renal cortex and medulla. Similar to previous studies (24), the present study found that the ADC and FA values were higher in patients with normal graft function than in those with CAI (regardless of immune or non-immune factors), and these values decreased as eGFR decreased. Further, this study revealed that ADC values could differentiate between immune and non-immune factors (P=0.018 and 0.043, respectively), while FA values failed to demonstrate such discriminative capability.

FA measures the degree of anisotropy in water molecule diffusion, reflecting the orderliness and directionality of tissue microstructure, while ADC reflects the degree of water molecule diffusion restriction (25). The pathological alterations induced by immune and non-immune factors in CAI are complex and diverse. These alterations can disrupt the typical diffusion environment of water molecules. However, it may be that not all of them affect the directionality of water molecule diffusion. Therefore, it is difficult to accurately differentiate between immune-mediated and non-immune-mediated CAI based on FA values.

We found that cortical ADC values can distinguish the etiology of CAI. The primary pathological alterations in immune-mediated CAI are interstitial fibrosis and tubular atrophy (26). Immune cell infiltration, inflammatory response, and fibrous tissue proliferation lead to increased cortical tissue density. The extracellular space is reduced, significantly restricting water molecule diffusion (27). Non-immune-mediated CAI, such as the recurrence of glomerular diseases, is primarily characterized by glomerular lesions. These lesions are primarily confined to the cortical region where the glomeruli are located, leading to localized alterations in the cortical tissue architecture. Different types of glomerular diseases exhibit distinct pathological manifestations, such as mesangial proliferation and basement membrane thickening (28); however, they generally exert a more direct impact on the cortex. In the early stages of the disease, glomerular disease has a relatively minor effect on the medulla. The structure and function of the medulla thus remain relatively intact, with water molecule diffusion within the medulla appearing relatively normal. However, as the disease progresses, the medulla may also be affected to some extent.

The value of T1 mapping in assessing cardiac and hepatic fibrosis has been well established (29-31). In this study, no significant difference in medullary T1 values was observed among the three patient groups. It may be that renal injury decreases water content in the renal tubules, while inflammation and edema increase water content, resulting in minimal changes in the T1 value of the renal medulla (32). Jiang et al. reported a correlation between cortical T1 values and renal allograft function, with T1 values increasing as eGFR decreased (33). Further, the present study revealed that cortical T1 values can also differentiate between immune-mediated and non-immune-mediated CAI. The pathological alterations induced by these two factors differ significantly in cortical microstructure and function, which may influence the longitudinal relaxation time of water protons in the cortex. In immune-mediated CAI, interstitial fibrosis, edema, and inflammation lead to hydronephrosis in the renal allograft interstitium, resulting in prolonged cortical T1 relaxation time (34). Damage caused by non-immune factors has a relatively minor impact on the overall cortical structure and exerts less significant effects on T1 values than immune-mediated injury. Based on the measured cortical T1 values, a significant increase indicates immune-mediated CAI, while values that show no significant increase or remain close to the normal range are more consistent with non-immune-mediated CAI.

CAI is a significant factor affecting the long-term survival of renal transplant recipients (35). Our study found that MRI parameters and Banff histopathological scores showed moderate negative and positive correlations, respectively, in the diagnosis of chronic damage to transplanted kidneys. This suggests that MRI can, to some extent, reflect chronic structural damage and rejection-related pathological changes in transplanted kidneys. Several approaches are available for the differential diagnosis of CAI. It should be noted that laboratory tests for CAI caused by different etiologies may yield similar results. Therefore, biopsy and pathological examination remain indispensable for diagnosis and differential diagnosis (4). The characterization of pathological changes in transplanted kidneys can aid in distinguishing between different etiologies (4). Therefore, multiparametric MRI (mpMRI), in conjunction with Banff histopathological scores, may provide some valuable information for the clinical management of renal allografts. Research on the correlation between MRI findings and pathological rejection in renal transplantation is limited (36). Jiang et al. reported that cortical T1 values from T1 mapping were positively correlated with i and t scores, while RBF from ASL was negatively correlated with ptc scores (33).

Similar to previous studies (37-39), this study found that ASL, DTI, and T1 mapping demonstrated high diagnostic value in assessing renal allograft function. Further, DTI and T1 mapping exhibited high diagnostic efficacy in identifying causative factors of CAI. Their combined application facilitates non-invasive differentiation between immune and non-immune factors of CAI from the perspective of tissue microstructure. This approach enables long-term longitudinal follow-up, offering essential information for early intervention in the clinical management of kidney transplants.

Notably, the clinical application of mpMRI still requires careful consideration of scan duration and patient tolerance. Although combined ASL, DTI, and T1 mapping scans can provide more comprehensive functional information, they also prolong the examination time. In routine clinical workflows, excessively long scan times may affect scheduling efficiency and clinical accessibility. For patients with poor renal function following transplantation, those in poor general health, or those who find it difficult to cooperate during prolonged examinations, extended scan times may also increase discomfort and the risk of motion artefacts. Consequently, future research should seek to balance diagnostic performance with scanning efficiency. This includes further optimizing scanning protocols, selecting core sequences with the highest diagnostic value, and developing rapid acquisition techniques, respiratory motion control strategies, and automated post-processing methods. These measures will enhance the feasibility of mpMRI in clinical practice.

Further, the field of renal MRI continues to evolve. ASL, DTI, and T1 mapping can be combined with other complementary MRI techniques to provide a more comprehensive picture of the pathophysiological state of the transplanted kidney (40). In addition to the perfusion, diffusion, and T1 relaxation parameters employed in this study, blood-oxygenation-level-dependent (BOLD) MRI can be used to assess renal tissue oxygenation and distinguish between acute renal injury and acute tubular necrosis (41). Magnetic resonance elastography can detect changes in tissue shear stiffness and assess fibrosis in the transplanted kidney (42). Magnetization transfer imaging can assess the macromolecular content in tissue and detect the degree of renal fibrosis (43). Phase-contrast MRI can be used to measure renal arterial blood flow (44). In the future, the appropriate combination of these techniques with ASL, DTI, and T1 mapping may help to establish a more comprehensive evaluation system for transplanted kidney function, covering multiple dimensions such as perfusion, oxygenation, diffusion, fibrosis, and hemodynamics.

However, mpMRI requires further standardization before it can be widely adopted in clinical practice. Existing international standardization efforts have put forward technical recommendations regarding data acquisition, post-processing, and reporting for renal ASL, DTI, T1/T2 mapping, and BOLD techniques (45-48). Future research should further standardize patient preparation, scanning parameters, respiratory motion control methods, image registration, ROI or whole-kidney segmentation methods, parameter calculation workflows, and reporting guidelines. Multicentre, large-scale studies are required to validate the stability and reproducibility of measurement results across different devices and centers. Only once standardization and reproducibility have been fully validated can renal mpMRI be more reliably applied to long-term follow-up of transplanted kidneys, efficacy assessment, and clinical decision support.

This study had some limitations. First, the sample size was relatively small. Future studies should increase the sample size to further validate the reliability of the findings. Second, the non-immune factors group in this study primarily comprised patients with glomerular diseases, while the number of patients with other non-immune factors was limited. Future studies should further refine the grouping for assessment. Third, T1WI, T2WI, and functional quantitative sequences were not entirely consistent in terms of spatial resolution and voxel size, which may lead to minor spatial registration errors between sequences. This may affect the accuracy of ROI localization and the measurement of quantitative parameters. Fourth, no image registration was performed for ASL, DTI, and T1 mapping. Due to differences in acquisition parameters, spatial resolution, and the extent of respiratory motion effects among the various fMRI sequences, minor spatial misalignment may exist between sequences, which may further affect the consistency of ROI positioning and the accuracy of parameter measurements. ROI analysis was primarily based on the maximum cross-section of the kidney, with small circular ROIs used for parameter measurement. A three-dimensional whole-kidney volumetric analysis method covering the entire transplanted kidney was not employed. Consequently, the overall functional status of the transplanted kidney may not be fully reflected. Future studies should incorporate automated or semi-automated segmentation, image registration, and whole-kidney volume ROI analysis to enhance the accuracy and reproducibility of results. Finally, building upon optimized scan duration and patient tolerance, future research should explore a more streamlined and standardized combination of mpMRI sequences to further advance their clinical application in the assessment of transplanted kidney function.

Conclusions

ASL, DTI, and T1 mapping can differentiate between patients with normal renal allograft function and those with CAI. DTI and T1 mapping can further differentiate between immune and non-immune pathogenic factors in CAI. These techniques provide clinicians with more reliable and comprehensive information for the management of transplanted kidneys, facilitating improved monitoring and maintenance of graft function.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This prospective study of kidney allograft recipients was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Ethics Committee of The First Affiliated Hospital of Soochow University (No. 2022-412), and written informed consent was obtained from all participants.

Data Sharing Statement

Available at https://qims.amegroups.com/article/view/10.21037/qims-2026-1-0380/dss

qims-16-08-639-dss.pdf ^{(69.3KB, pdf)}

DOI: 10.21037/qims-2026-1-0380