PURPOSE: To analyze the correlation between the ultrasonographic (US) grading system of fatty liver (FL) and histologic grading system in living liver donor candidates and to investigate the clinical significance of this qualitative US grading system in the selection of living donor candidates. MATERIALS AND METHODS: For a recent 21-month period, ninety three living donor candidates who underwent both preoperative US and parenchymal biopsy of the liver were consecutively selected. FL was ultrasonographically graded using the well-known three-point grading system (ie, mild, moderate and severe degrees) whereas histologic grade of FL was divided into minimal (<10%), mild (11-30%), moderate (31-60%) and severe (>60%) degrees depending upon the percentages of each of macrovesicular, microvesicular and total fat-containing hepatocytes. US grade and histologic grade of FL in each patient were retrospectively correlated according to the US and pathologic records in their databases. Statistical analysis was conducted with the chi-square test and linear by linear association. RESULTS: US findings included the normal liver, mild FL, and moderate FL in 63, 23 and 7 patients, respectively. Analyzed with the total fat content, 38 of 63 patients (60%) whose US finding was normal proved to have FL of various histologic grades. Meanwhile, US grade of FL correlated well with the histologic grade in 16 (53%) of 30 patients who showed mild or moderate FL on US, and in the remaining patients, US grade was more commonly underestimated compared to the histologic grade. All patients with moderate FL on US proved to have either moderate or severe FL at histology. US grade statistically correlated well with the histologic grade classified by either the total or macrovesicular fat contents (p<.001) while a poor correlation was seen when histologic grade using the microvesicular fat content was used. CONCLUSION: The well-known qualitative US grading system of fatty liver seems to show a relatively good correlation with the histologic grade, but it has a tendency to underestimate compared to the histologic grade, particularly when histologic grade using the total fat content was used. Therefore, it seems unnecessary to undergo preoperative liver biopsy when the US grade is greater than moderate FL in living donor candidate. However, the normal US finding might require liver biopsy.