An innovative methodology for the automated morphometric and quantitative estimation of liver steatosis

Abstract

The management of liver steatosis, due to its potential evolution towards severe diseases, requires accurate diagnosis. Fatty infiltration in liver diseases is commonly assessed histologically by semi-quantitative methods, which can be subjective. Automated computerized procedures using commercial software for image analysis have also been recently employed. The aim of the study was to develop an innovative automated computerized procedure to accurately evaluate both the morphometry and degree of lipid accumulation in liver. Fatty infiltration was assessed in paraffin- and resinembedded samples of steatotic livers from rats undergoing 0, 3, 7, 14, and 30-day choline-deficient diet, and from liver biopsy of a morbidly obese patient undergoing bariatric surgery. Specific software was developed, which works with a morphological operator, in addition to a chromatic one to select lipid droplets. The choline-deficient diet induced steatosis with a gradual shift from micro- to macro-vesicular. In paraffin sections, the macrovesicles-to-microvesicles ratio and the degree of steatosis, when using only the chromatic operator, produced overestimates. Results were consistent in both rat and human samples. An improvement of topographic, morphometric and quantitative estimation of fatty liver infiltration is obtained with our software, working with a morphological operator and using semi-thin sections from resin-embedded samples. This innovative procedure may be applied to human liver samples, offering promising diagnostic and prognostic perspectives.