Supplementary Material for: Evaluation of Renal Fibrosis by Mapping Histology and Magnetic Resonance Imaging

Background: Renal fibrosis is a key driver of progression in chronic kidney disease (CKD). Recent advances in diagnostic imaging techniques have shown promising results for the noninvasive assessment of renal fibrosis. However, the specificity and accuracy of these techniques are controversial because they indirectly assess renal fibrosis. This limits fibrosis assessment by imaging in CKD for clinical practice. To validate magnetic resonance imaging (MRI) assessment for fibrosis, we derived representative models by mapping histology-proven renal fibrosis and imaging in CKD. Methods: Ninety-seven adult Chinese CKD participants with histology were studied. The kidney cortex interstitial extracellular matrix volume was calculated by the Aperio ScanScope system using Masson’s trichrome slices. The kidney cortex microcirculation was quantitatively assessed by peritubular capillary density using CD34 staining. The imaging techniques included intravoxel incoherent motion diffusion-weighted imaging and magnetic resonance elastography (MRE) imaging. Relevant analyses were performed to evaluate the correlations between MRI parameters and histology variables. Multiple linear regression models were used to describe the relationships between a response variable and other variables. The best-fit lines, which minimize the sum of squared residuals of the multiple linear regression models, were generated. Results: MRE values were negatively associated with the interstitial extracellular matrix volume (Rho = −0.397, p < 0.001). The best mapping model of extracellular matrix volume with the MRE value and estimated glomerular filtration rate (eGFR) we obtained was as follows: Interstitial extracellular matrix volume = 218.504 – 14.651 × In(MRE) – 18.499 × In(eGFR). DWI-fraction values were positively associated with peritubular capillary density (Rho = 0.472, p < 0.001). The best mapping model of peritubular capillary density with DWI-fraction value and eGFR was as follows: Peritubular capillaries density = 17.914 + 9.403 × (DWI – fraction) + 0.112 × (eGFR). Conclusions: The study provides histological evidence to support that MRI can effectively evaluate fibrosis in the kidney. These findings picture the graphs of the mapping model from imaging and eGFR into fibrosis, which has significant value for clinical implementation.

背景：肾纤维化是慢性肾脏病（chronic kidney disease, CKD）病情进展的关键驱动因素。近年来诊断成像技术的进展为肾纤维化的无创评估带来了颇具前景的研究成果。然而，由于这些成像技术仅能间接评估肾纤维化，其特异性与准确性尚存争议，这限制了影像学评估肾纤维化在慢性肾脏病临床实践中的应用。为验证磁共振成像（magnetic resonance imaging, MRI）对肾纤维化的评估效能，本研究通过匹配慢性肾脏病患者经组织学证实的肾纤维化与影像学数据，构建了代表性模型。
方法：本研究纳入97例经组织学确诊的中国成人慢性肾脏病受试者。采用Aperio ScanScope系统对马松三色染色切片进行分析，计算肾皮质间质细胞外基质容积；通过CD34染色检测肾小管周围毛细血管密度，对肾皮质微循环进行定量评估。本次研究采用的影像学技术包括体素内不相干运动扩散加权成像与磁共振弹性成像（magnetic resonance elastography, MRE）。开展相关分析以评估磁共振成像参数与组织学变量之间的相关性，采用多重线性回归模型分析因变量与其他变量之间的关联关系，并构建了使多重线性回归模型残差平方和最小的最优拟合线。
结果：磁共振弹性成像数值与间质细胞外基质容积呈负相关（Rho=-0.397，p<0.001）。本研究构建的基于磁共振弹性成像数值与估算肾小球滤过率（estimated glomerular filtration rate, eGFR）预测细胞外基质容积的最优匹配模型如下：间质细胞外基质容积=218.504 – 14.651×ln(MRE) – 18.499×ln(eGFR)。扩散加权成像分数值（DWI-fraction）与肾小管周围毛细血管密度呈正相关（Rho=0.472，p<0.001）。本研究构建的基于扩散加权成像分数值与估算肾小球滤过率预测肾小管周围毛细血管密度的最优匹配模型如下：肾小管周围毛细血管密度=17.914 + 9.403×(DWI-fraction) + 0.112×eGFR。
结论：本研究提供了组织学证据，证实磁共振成像可有效评估肾脏纤维化。本研究构建的基于影像学参数与估算肾小球滤过率的肾纤维化匹配模型，可为其临床转化应用提供重要参考价值。