Yuan Meng Yu1,2, Wei Wang3,4, Jiqiu Wen3, Yong Zhang5, Guang Ming Lu1,6, Long Jiang Zhang7,8. 1. Department of Medical Imaging, Jinling Hospital, Clinical School of Southern Medical University, Nanjing, 210002, Jiangsu, China. 2. Department of MRI, The First People's Hospital of Yunnan Province, The Affiliated Hospital of Kunming University of Science and Technology, No. 157 Jinbi Road, Kunming, 650032, Yunnan, China. 3. National Clinical Research Center of Kidney Diseases, Jinling Hospital, Nanjing Medical University, 305 East Zhong Shan Road, Nanjing, 210002, China. 4. Department of Nephrology, Shanghai General Hospital, Shanghai JiaoTong University School of Medicine, Shanghai, China. 5. MR Research, GE Healthcare, Shanghai, 201203, China. 6. Department of Medical Imaging, Jinling Hospital, Medical School of Nanjing University, Nanjing, 210002, Jiangsu, China. 7. Department of Medical Imaging, Jinling Hospital, Clinical School of Southern Medical University, Nanjing, 210002, Jiangsu, China. kevinzhlj@163.com. 8. Department of Medical Imaging, Jinling Hospital, Medical School of Nanjing University, Nanjing, 210002, Jiangsu, China. kevinzhlj@163.com.
Abstract
OBJECTIVE: To compare the value of reduced field-of-view (FOV) intravoxel incoherent motion (IVIM) diffusion-weighted imaging (DWI) and arterial spin labeling (ASL) for assessing renal allograft fibrosis and predicting long-term dysfunction. METHODS: This prospective study included 175 renal transplant recipients undergoing reduced FOV IVIM DWI, ASL, and biopsies. Renal allograft fibrosis was categorized into ci0, ci1, ci2, and ci3 fibrosis according to biopsy results. A total of 83 participants followed for a median of 39 (IQR, 21-42) months were dichotomized into stable and impaired allograft function groups based on follow-up estimated glomerular filtration rate. Total apparent diffusion coefficient (ADCT), pure diffusion ADC, pseudo-perfusion ADC, perfusion fraction f from IVIM DWI, and renal blood flow (RBF) from ASL were calculated and compared. The area under the receiver operating characteristic curve (AUC) was calculated to assess the diagnostic and predictive performances. RESULTS: RBF was different in ci0 vs ci1 (147.9 ± 46.3 vs 126.0 ± 49.4 ml/min/100 g, p = .02) and ci2 vs ci3 (92.9 ± 46.9 vs 70.8 ± 37.8 ml/min/100 g, p = .03). RBF in the stable group was higher than that in the impaired group (144.73 ± 49.33 vs 102.19 ± 47.58 ml/min/100 g, p < .001). AUCs in distinguishing renal allograft fibrosis and predicting long-term allograft dysfunction for RBF were higher than cortical ADCT (ci0 vs ci1-3, 0.76 vs 0.59, p < .001; ci0-1 vs ci2-3, 0.79 vs 0.68, p = .01; ci0-2 vs ci3, 0.79 vs 0.68, p = .01; 0.76 vs 0.60, p = .04, respectively). CONCLUSION: Compared to reduced FOV IVIM DWI, ASL was a more promising technique for noninvasively distinguishing renal allograft fibrosis degree and predicting long-term allograft dysfunction. KEY POINTS: • Compared to total ADC from rFOV IVIM DWI, RBF from ASL can distinguish no fibrosis (ci0) vs mild fibrosis (ci1) (p = .02) and moderate fibrosis (ci2) vs severe fibrosis (ci3) (p = .04). • RBF had superior performance than diffusion parameters in discriminating fibrosis (no fibrosis [ci0] vs fibrosis [ci1-3], mild fibrosis [ci0-1] vs moderate to severe fibrosis [ci2-3], non-severe [ci0-2] vs severe [ci3] fibrosis; AUC = 0.76 vs 0.59, p < .001; 0.79 vs 0.68, p = .01; 0.79 vs 0.68, p = .01). • Compared to reduced FOV IVIM DWI, ASL was a more promising technique for noninvasively predicting long-term allograft dysfunction (AUC = 0.76 vs 0.60, p = .04).
OBJECTIVE: To compare the value of reduced field-of-view (FOV) intravoxel incoherent motion (IVIM) diffusion-weighted imaging (DWI) and arterial spin labeling (ASL) for assessing renal allograft fibrosis and predicting long-term dysfunction. METHODS: This prospective study included 175 renal transplant recipients undergoing reduced FOV IVIM DWI, ASL, and biopsies. Renal allograft fibrosis was categorized into ci0, ci1, ci2, and ci3 fibrosis according to biopsy results. A total of 83 participants followed for a median of 39 (IQR, 21-42) months were dichotomized into stable and impaired allograft function groups based on follow-up estimated glomerular filtration rate. Total apparent diffusion coefficient (ADCT), pure diffusion ADC, pseudo-perfusion ADC, perfusion fraction f from IVIM DWI, and renal blood flow (RBF) from ASL were calculated and compared. The area under the receiver operating characteristic curve (AUC) was calculated to assess the diagnostic and predictive performances. RESULTS: RBF was different in ci0 vs ci1 (147.9 ± 46.3 vs 126.0 ± 49.4 ml/min/100 g, p = .02) and ci2 vs ci3 (92.9 ± 46.9 vs 70.8 ± 37.8 ml/min/100 g, p = .03). RBF in the stable group was higher than that in the impaired group (144.73 ± 49.33 vs 102.19 ± 47.58 ml/min/100 g, p < .001). AUCs in distinguishing renal allograft fibrosis and predicting long-term allograft dysfunction for RBF were higher than cortical ADCT (ci0 vs ci1-3, 0.76 vs 0.59, p < .001; ci0-1 vs ci2-3, 0.79 vs 0.68, p = .01; ci0-2 vs ci3, 0.79 vs 0.68, p = .01; 0.76 vs 0.60, p = .04, respectively). CONCLUSION: Compared to reduced FOV IVIM DWI, ASL was a more promising technique for noninvasively distinguishing renal allograft fibrosis degree and predicting long-term allograft dysfunction. KEY POINTS: • Compared to total ADC from rFOV IVIM DWI, RBF from ASL can distinguish no fibrosis (ci0) vs mild fibrosis (ci1) (p = .02) and moderate fibrosis (ci2) vs severe fibrosis (ci3) (p = .04). • RBF had superior performance than diffusion parameters in discriminating fibrosis (no fibrosis [ci0] vs fibrosis [ci1-3], mild fibrosis [ci0-1] vs moderate to severe fibrosis [ci2-3], non-severe [ci0-2] vs severe [ci3] fibrosis; AUC = 0.76 vs 0.59, p < .001; 0.79 vs 0.68, p = .01; 0.79 vs 0.68, p = .01). • Compared to reduced FOV IVIM DWI, ASL was a more promising technique for noninvasively predicting long-term allograft dysfunction (AUC = 0.76 vs 0.60, p = .04).