*Purpose: The use of steatotic livers is controversial, but may represent an underutilized resource to increase the donor pool. Although <30% macrovesicular steatosis (MaS) is generally considered safe, there is little consensus on what constitutes a transplantable grossly steatotic liver. There is also significant variability in how MaS is measured or defined. We studied the variability in MaS measurement of biopsies between donor hospitals and transplant centers and compared these measurements with digital imaging software (DIS) for detailed quantitative and reproducible analysis of MaS.

*Methods: Using HALO Image Analysis Platform, a DIS, we analyzed 53 liver biopsy samples from steatotic donors between 2016-2018 from three transplant centers with MaS percentage (%MaS) read by donor hospitals, transplant centers, or both. We generated scatter plots and quantified Spearman’s correlation coefficients to determine the relationship between %MaS reads by transplant centers vs. donor hospitals as well as DIS vs. transplant centers and DIS vs. donor hospitals with various MaS thresholds. We excluded non-numerical readings on the original donor biopsies.

*Results: Using DIS, the median vacuole count was found to be 53531 (interquartile range [IQR]: 4048-16488), median average vacuole area per cell was 33.7µ² (IQR: 26.1-46.5), median average nuclear area was 20.9µ2 (IQR: 18.8-26.0), the median average number of vacuoles per cell was 1.2 (IQR: 0.6-1.6), and the median %MaS was 10.7% (IQR: 9.1-13.0). There was a moderate, positive significant correlation between %MaS reads between donor hospitals and transplant centers (Spearman’s rho [r_s]= 0.56, p=0.01). When the vacuole size threshold was set at or greater than the mean nuclear size, there was no evidence of correlation between DIS and transplant center %MaS reads (r_s= -0.27, p=0.31). The same was seen when vacuole size threshold was set at >2x the mean nuclear size, (r_s = -0.28, p=0.30), or >3x (r_s= -0.27, p=0.32). When comparing DIS and donor center %MaS reads, there was no evidence of statistical correlation when the threshold was set at or greater than the mean nuclear size (r_s = 0.27, p=0.60), but demonstrated a strong, negative statistical correlation when the vacuole size was set at >2x (r_s= -0.81, p=0.04) or >3x (r_s= -0.82, p=0.05) the mean nuclear size.

*Conclusions: DIS is a method that provides precise and reproducible quantification of steatosis. We showed poor correlation of DIS when compared to donor hospital and transplant center reads. The inverse relationship observed between donor center and DIS reads with large size droplets brings to light the significant variability in %MaS interpretation that can greatly impact organ utilization. DIS has the potential to standardize %MaS and serve as a tool to better study graft outcomes.

« Back to 2019 American Transplant Congress