Early detection of subclinical pathology in patients with stable kidney graft function by arterial spin labeling
Wei Wang,Yuanmeng Yu,Xue Li,Jinsong Chen,Yong Zhang,Longjiang Zhang,Jiqiu Wen
DOI: https://doi.org/10.1007/s00330-020-07369-5
IF: 7.034
2020-01-01
European Radiology
Abstract:Objectives To evaluate the utility of arterial spin labeling (ASL) for the identification of kidney allografts with underlying pathologies, particularly those with stable graft function. Methods A total of 75 patients, including 18 stable grafts with normal histology (normal group), 21 stable grafts with biopsy-proven pathology (subclinical pathology group), and 36 with unstable graft function (unstable graft group), were prospectively examined by ASL magnetic resonance imaging. Receiver operating characteristic curves were generated to calculate the area under the curve (AUC), sensitivity, and specificity. Results Patient demographics among the 3 groups were comparable. Compared with the normal group, kidney allograft cortical ASL values decreased in the subclinical pathology group and the unstable graft group (204.7 ± 44.9 ml/min/100 g vs 152.5 ± 38.9 ml/min/100 g vs 92.3 ± 37.4 ml/min/100 g, p < 0.001). The AUC, sensitivity, and specificity for discriminating allografts with pathologic changes from normal allografts were 0.92 (95% CI, 0.83–0.97), 71.9%, and 100% respectively by cortical ASL and 0.82 (95% CI, 0.72–0.90), 54.4%, and 100% respectively by serum creatinine. The cortical ASL identified allografts with subclinical pathology among patients with stable graft function with an AUC of 0.80 (95% CI, 0.64–0.91), sensitivity of 57.1%, and specificity of 88.9%. Combined use of proteinuria and cortical ASL could improve the sensitivity and specificity to 76.2% and 100% respectively for distinguishing the subclinical pathology group from the normal group. Conclusions Cortical ASL is useful for the identification of allografts with underlying pathologies. More importantly, ASL showed promise as a non-invasive tool for the clinical translation of identifying kidney allografts with subclinical pathology. Key Points • Cortical ASL values were decreased in kidney allografts with subclinical pathologic changes as compared with normal allografts (152.5 ± 38.9 ml/min/100 g vs 204.7 ± 44.9 ml/min/100 g, p < 0.001). • Cortical ASL differentiated allografts with pathologic changes and subclinical pathology group from normal group with an AUC of 0.92 (95% CI, 0.83–0.97) and 0.80 (95% CI, 0.64–0.91) respectively. • Cortical ASL discriminated allografts with underlying pathologic changes from normal allografts with a specificity of 100%, and combined use of proteinuria and cortical ASL values could also achieve 100% specificity for discriminating allografts with subclinical pathology from normal allografts.