Detection of Non-Liver Incidentalomas on Sonogram in an Urban Veteran Hepatocellular Carcinoma Screening Program

Abstract

Purpose: Concern regarding the burden of unnecessary evaluations of non-liver findings on screening sonograms for hepatocellular carcinoma (HCC) is currently an area of debate. The purpose of this study was to evaluate the incidence of incidental findings on screening sonograms, as well as subsequent workup performed, in an electronic medical record (EMR)-based HCC screening program with automated clinical reminders in an urban VA Medical Center. Methods: A retrospective database of cirrhotic patients previously created by a search of the VA New York Harbor Health System electronic medical software (VistA) from 2008 until 2011 was utilized. Rates of radiological and AFP screening tests were calculated from February 8, 2012 through February 7, 2013. Rates of HCC detection and incidental findings leading to further work-up were calculated. Results: Three hundred eighty live, cirrhotic patients actively pursuing health care at the VA within the past 2 years were included in the study. Non-GI physicians were the primary directors of screening for 59.3% of the patients, and gastroenterologists were responsible for 40.7%. During the year of follow-up, 63.4% of patients had at least one imaging test or AFP level performed, 52.4% had at least one imaging study performed, 23.2% had two imaging studies performed, and 14.5% of patients had two imaging studies and two serum AFPs. There was a statistically significant difference in screening rates when comparing patients with ICD-9 codes for cirrhosis to those without. There were 210 screening sonograms performed. There were an additional 6 CTs and 1 MRI ordered for follow-up of non-liver pathology identified during these sonograms (2.9% and 0.5% respectively), including four complex renal cysts, two suspected portal venous thromboses, and one case of hydronephrosis. On these follow-up exams, three of these findings were confirmed; however, no non-liver malignancies were biopsy proven, due to concerns regarding the patient's suitability to tolerate invasive procedures given the severity their underlying liver disease. The follow-up MRI detected an HCC, which had not been previously visualized on the prior sonogram. Overall, 16 new cases of HCC were diagnosed, five of which were initially visualized on sonogram. Conclusion: An EMR based HCC surveillance program at a VA hospital provides an important avenue to detect HCC in high-risk patients that may otherwise be missed. Fear of surveillance sonograms leading to excessive expense and radiation exposure due to further work-up of incidental findings appears to be unfounded, as in our study rates; further studies were low, and many of the incidental findings were ultimately corroborated on follow-up imaging.