Using digital whole-slide images to evaluate renal amyloid deposition and its association with clinical features and outcomes of AL amyloidosis.

Abstract

Few data are available quantifying the proportion of amyloid deposition in renal biopsy specimens. The aim of the study is to investigate the correlation between the proportion of amyloid deposition in renal biopsy and clinical characteristics of Chinese patients with immunoglobulin light-chain amyloidosis (AL amyloidosis). 259 patients diagnosed with renalAL amyloidosis between 2003 and 2015 were studied retrospectively. We developed a digital, automated quantification method to evaluate amyloid deposits in glomeruli, vessels and interstitium on digital whole-slide images (WSIs). The associations between the proportion of amyloid-positive area in the renal biopsy and clinical manifestations were analyzed. The proportion (%) of amyloid-positive area in glomeruli, vessels, interstitium and the whole renal tissue were 11.81 ± 11.38, 14.14 ± 14.05, 3.34 ± 5.36 and 4.25 ± 5.77, respectively. The proportion of amyloid deposition in glomeruli, vessels and interstitium was positively correlated with serum creatinine (Scr), estimated glomerular filtration rate (eGFR) and urinary retinol binding protein (RBP). The proportion of glomerular amyloid deposition, age, urinary N-acetyl-b-D-glucosaminidase (NAG) and urinary RBP could independently predict the risk for overall death. The proportion (%) of amyloid-positive area in blood vessels, interstitium and the whole renal tissue, Scr, and urinary RBP were independent risk factors associated with renal survival. A novel digital analysis algorithm was firstly developed to quantify the proportion of amyloid deposits in renal tissues based on digitalWSIs. The degree and localization of amyloid deposits in the kidney evaluated by digitalWSIs may have predictive value in assessing risk of outcome of AL amyloidosis.