At least one extra-hepatic finding was documented in 56.5% of exams, the vast majority of which were determined to be minor significance incidental findings. The most common extra-hepatic findings occur in the gallbladder, with 39.9% of exams document an incidental gallbladder finding, most commonly gallstones (24.4%). Gallstones are very common, with vast majority asymptomatic, although approximately 10–15% will become symptomatic over a period of 10 to 15 years . Although the presence of gallstones may be clinically relevant in the future, no patients with acute gallbladder disease were identified in this study. Further, the majority of patients had prior right upper quadrant imaging (92.2%), including prior ultrasound, so the presence of cholelithiasis was likely already known in the vast majority of subjects. The gallbladder was thickened in 10.3%, all of which were attributed to hepatic dysfunction, although multiple radiology reports suggested clinical correlation for right upper quadrant pain if there was suspicion of acute cholecystitis.
Following the gallbladder, the second most frequent incidental findings were identified in the right kidney, most commonly renal cysts, all of which were anechoic and demonstrated no complexity, therefore requiring no follow up .
Potentially clinically significant incidental findings during HCC screenings in our patient population are rare, with only four moderate clinically significant incidental findings identified (0.9%). No major clinical significance findings were identified. Only one incidental finding identified required specific additional evaluation (0.2%).
There is limited available data describing the prevalence of incidental findings on ultrasound, however our data are similar to rates of incidental findings described in the literature. Orme et al reviewed incidental findings discovered during imaging research, with 9.2% ultrasounds identifying incidental findings, compared to 39.8% on all modalities . In their study, none of the incidental findings were determined to be clinically significant, compared to 2.5% of all modalities. Our reported rate of clinically significant incidental findings is similar to the study by Orme et al, which identified no clinically significant findings. However, that study reviewed examinations performed for research purposes and likely was conducted in a much healthier population compared to this study population.
Incidence of incidental findings has been studied in a different patient population undergoing screening imaging studies, namely CT colonography for screening of colon cancer . A retrospective review of extra-colonic findings on CT colonography by Pooler et al identified that 88.3% of screening exams are normal or identify benign or unimportant incidental findings, such as simple renal cysts or cholelithiasis. In that study, 2.2% of all patients had a clinically significant incidental finding which required treatment or surveillance, suggesting that CT imaging may identify more clinically significant incidental findings than right upper quadrant US.
Given the low rates of significant incidental findings identified on HCC screening ultrasound, these data support a more limited HCC screening ultrasound protocol. Therefore, we recommend the ultrasound protocol be limited to the liver, gallbladder, and biliary tree, with partial inclusion of the pancreatic head, right kidney, aorta, and inferior vena cava on liver images, but with no dedicated evaluation necessary.
Further, a targeted exam may increase sensitivity for nodule detection, as shown in a study by da Silva et al . This prospective study evaluated liver nodule detection in patients evaluated by a targeted liver ultrasound compared to a complete upper abdominal ultrasound protocol. A significantly higher frequency of nodules was detected when a dedicated protocol was used, possibly due the more dedicated time focusing on relevant liver findings .
Many institutions currently use protocols which do not specifically include the spleen, lower quadrants, main portal vein spectral waveform, or color Doppler evaluation of the portal and hepatic veins, which are optional per the US LI-RADS guidelines . Routine imaging of the hepatic vasculature, spleen, and evaluation for ascites may provide added value for screening studies, including suggesting the presence of portal hypertension .
In our study, LI-RADS category was US-1 in 91.4%, US-2 in 5.4%, and US-3 in 3.3%. Our rates of LI-RADS categories are similar to the rates reported by Millet et al, which reported US-1 in 90.4%, US-2 in 4.6%, and US-3 4.9% . Evaluation of the outcomes of the US-3 exams in our study is limited by small sample size. Compared to a study by Sevco et al, follow up of US-3 observations identified no abnormality or benign observation (LR-1 or LR-2) in 57.1% of our cases vs 73.9% in the Sevco study, 0% LR-3 vs. 5.5%, 35.7% LR- 4 or LR-5 vs. 18.8%, and 7.1% LR-M vs. 1.8% .
There are multiple limitations to this study. Sample size included a 2.5-year period of studies with a relatively small sample size of 432 exams in 294 patients. Therefore, it is probable that very rare significant incidental findings may have not been included in the evaluated data set. Further, patients were identified by chart review which required a documented diagnosis of cirrhosis or chronic hepatitis B in the medical chart, which may not be present in all patients undergoing HCC screening.
Another limitation was the retrospective nature of the study and reliance on the original clinical report for identification of incidental findings. There were 10 different reporting radiologists who had 0.5 to 35 years post-training clinical experience. This may have resulted in differences in rates of reporting incidental findings when present. Finally, a high percentage of reports did not include visualization score or document LI-RADS categories (60.8%), likely a reflection of known slow adoption of standardized reporting in radiology .