A clinical study of using ROC to compare the efficiency of ASL and BOLD in diagnosis of renal allograft function

Jin Peng; Yajun Hong; Feipeng Zhu; Yanjun Li; Song Luo; Guangming Lu

doi:10.21037/tau-23-136

A clinical study of using ROC to compare the efficiency of ASL and BOLD in diagnosis of renal allograft function

Transl Androl Urol. 2023 Apr 28;12(4):612-621. doi: 10.21037/tau-23-136. Epub 2023 Apr 17.

Authors

Jin Peng^#^{1

2}, Yajun Hong^#³, Feipeng Zhu⁴, Yanjun Li⁴, Song Luo⁵, Guangming Lu^{1

5}

Affiliations

¹ The First School of Clinical Medicine, Southern Medical University, Guangzhou, China.
² Department of Interventional Therapy, Chenggong Hospital, Xiamen University, Xiamen, China.
³ Office of Medical Case Statistics, Department of Medical Affairs, Zhongshan Hospital (Xiamen), Fudan University, Xiamen, China.
⁴ Department of Radiology, The First Affiliated Hospital with Nanjing Medical University, Nanjing, China.
⁵ Department of Radiology, Jinling Hospital, Medical School of Nanjing University, Nanjing, China.

^# Contributed equally.

Abstract

Background: This retrospective study aims to evaluate the effectiveness of renal transplantation function by comparing arterial spin labeling (ASL) and blood oxygen level dependent (BOLD) imaging with the receiver operating characteristic (ROC) curve.

Methods: According to the estimated glomerular filtration rate (eGFR) values, 42 patients with normal kidney grafts (the normal kidney graft group, eGFR <60 mL/min/1.73 m²) and 93 patients with injured grafts (the kidney graft injury group, eGFR <60 mL/min/1.73 m²) were included in the present study. Renal blood flow (RBF) and the effective transverse relaxation rate (R2*) were calculated by comparing ASL and BOLD imaging. The ROC curve and the youden index were used to evaluate the diagnostic performance of ASL, BOLD, and the combination of them.

Results: The results showed that all the clinical features of the patients, except for gender, differed significantly between the 2 groups (P<0.05). The mean RBF value of the renal transplant injury group (104.33±54.76 mL/100 g/min) was significantly lower than that of the normal group (191.84±63.96 mL/100 g/min, P<0.01). The mean medullary R2* value of the renal transplant injury group (27.91±3.35 1/s) was significantly higher than that of the normal group (25.22±2.94 1/s, P<0.01). Negative correlations were found between R2* and eGFR (r=-0.44), and RBF and R2* (r=-0.54; both P<0.01). The ROC analysis showed that both RBF and R2* reflected injured renal function [area under the curves (AUC) =0.86 and 0.72, respectively]. In addition, the AUC of RBF and R2* combined was 0.86, which was comparable to that of RBF alone (P=0.95), while combining R2* and RBF improved the diagnostic performance of R2* alone (AUC =0.86 versus 0.72, respectively; P<0.01). The youden index analysis showed that the diagnostic accuracy of ASL was 80.00%, better than 71.85% of BOLD; the sensitivity and specificity of ASL in diagnosing renal allograft dysfunction were 79.57% and 80.95%, superior to 77.42% and 59.52% of BOLD.

Conclusions: Our results showed that the non-invasive assessment of ASL in clinical kidney transplant function is a more promising imaging technique than BOLD.

Keywords: Arterial spin labeling; blood oxygen level-dependent imaging; glomerular filtration rate; magnetic resonance imaging; renal allograft.