Reticulin loss in benign fatty liver: an important diagnostic pitfall when considering a diagnosis of hepatocellular carcinoma.

Reticulin stains are commonly used in surgical pathology to assess mass lesions for the possibility of hepatocellular carcinoma. The loss of normal reticulin staining can help support a diagnosis of hepatocellular carcinoma, and this stain has proven to be particularly helpful on limited biopsies and fine-needle aspirates. However, an underappreciated diagnostic pitfall is that non-neoplastic liver tissue can also show reticulin loss when there is fatty change. To further characterize this important diagnostic pitfall, reticulin staining was studied in cases of nonalcoholic steatosis, nonalcoholic steatohepatitis, and hepatic adenomas with fatty change. A total of 112 cases with varying degrees of steatosis were collected from 4 academic centers, including 49 cases of steatosis, 49 cases of steatohepatitis, and 14 hepatic adenomas with fatty change. Steatosis was graded as mild (5% to 30% macrovesicular steatosis), moderate (31% to 60%), and marked (>60%). Reticulin stains were scored as the number of foci with diminished reticulin staining in 10 hpf. A focus of diminished reticulin was scored when the extent of reticulin loss was similar to that seen in hepatocellular carcinomas. In the total study set, 28 cases showed mild steatosis, 40 cases showed moderate steatosis, and 44 cases showed marked steatosis. Interestingly, increasing amounts of fat were associated with decreased reticulin staining. For mild steatosis, reticulin loss was rare, with the number of foci of reticulin loss per 10 hpf averaging 0.8 (range, 0 to 3); however, this increased for moderate steatosis, which showed a mean of 3.0 foci per 10 hpf (range, 0 to 5), and was most prominent with marked steatosis, which showed an average of 5.8 foci of reticulin loss per 10 hpf (range, 5 to 8). An almost identical pattern was seen in cases of nonalcoholic steatohepatitis. Overall, reticulin loss was not associated with the degree of inflammation or with the presence or absence of balloon cell change. Reticulin loss also did not correlate with fibrosis stage. In hepatic adenomas, reticulin loss was seen only in areas of fatty change, and decreased reticulin again paralleled the amount of steatosis, with more prominent reticulin loss in those cases with marked steatosis. In conclusion, reticulin loss that reaches levels seen in hepatocellular carcinoma can be seen focally in benign liver tissues with fatty change. Overall, loss of reticulin is more common and more extensive with marked fatty change and does not seem to be linked to inflammation or fibrosis stage. Loss of reticulin can also be seen in hepatic adenomas with fatty change. Increased awareness of this important diagnostic pitfall will help prevent overcalling of reticulin loss when evaluating biopsies and resections of hepatic neoplasms with fatty change.