Liver Imaging Reporting and Data System in Patients at High Risk for Hepatocellular Carcinoma in the Memphis Veterans Affairs Population

Hepatocellular carcinoma (HCC) is the third most common cause of death from cancer worldwide.¹ Liver cancer is the fifth most common cancer in men and the seventh in women.² The highest incidence rates are in sub-Saharan Africa and Southeast Asia where hepatitis B virus is endemic. The incidence of HCC in western countries is increasing, particularly due to the rise of hepatitis C virus (HCV) as well as alcoholic liver disease and nonalcoholic fatty liver disease. The incidence of HCC has tripled in the US in the past 2 decades.^1-3

HCC can be diagnosed by radiographic images without the need for biopsy if the typical imaging features are present.³ The European Association for the Study of Liver Disease (EASL) and the American Association for the Study of Liver Diseases (AASLD) recommend screening abdominal ultrasonography at 6-month intervals for high-risk patients.^3,4 High-risk patients include patients with cirrhosis, especially those with hepatitis B or C.³

If screening ultrasonography detects a nodule, size determines whether a follow-up ultrasound is needed vs obtaining a contrast-enhanced dynamic computed tomography (CT) scan or a magnetic resonance image (MRI).³ If ultrasonography detects a nodule > 1 cm in diameter, then a dynamic CT or MRI is performed. Characteristic hyperenhancement during later arterial phase and washout during the venous or delayed phase is associated with a nearly 100% specificity for HCC diagnosis.⁵ Arterial-enhancing contrast is required when using CT and MRI because HCC is a hypervascular lesion.⁶ The portal venous blood dilutes the majority of the liver’s arterial blood; therefore, the liver does not enhance during the arterial phase, while HCC will show maximum enhancement.⁷ Furthermore, HCC should demonstrate a “washout” of contrast during the venous phase on CT and MRI.⁴ Standard imaging protocol dictates that 4 phases are needed to properly diagnose HCC including unenhanced, arterial, venous, and delayed.⁴

Regular surveillance increases the likelihood of detecting HCC before the presentation of clinical symptoms and facilitates receipt of curative therapy.^8-10 Patients with viral hepatitis and cirrhosis with HCC found on screening are more likely to have earlier-stage disease and survive longer from the time of diagnosis.¹¹ Furthermore, it has been observed that HCC detected by surveillance is significantly more likely to undergo curative therapy compared with incidental or symptomatic detection of HCC.⁹

Technical improvements in imaging techniques include advancement in contrast agents, multidetector row helical CT, and the flexibility/range of pulse sequences available in MRI.⁷ Even with technical improvements in all modalities used in HCC imaging, detecting HCC remains difficult, especially when detecting the small (< 2 cm) lesions in a cirrhotic liver.⁷ Interpretation of imaging also remains a challenge as HCC does not always fit strict criteria: lack of “washout” in a hypervascular lesion, determining small HCC lesions from benign nodules, and hypovascular/isovascular HCC.⁵ Radiologic differentials in the diagnosis of HCC include transient hepatic intensity difference (THID)/transient hepatic attenuation difference (THAD), arterio-portal shunt, and regenerative nodules.¹² In the common clinical setting, patients undergo multiple imaging studies that are interpreted by multiple radiologists, which can add to the difficulty in the diagnosis of HCC.¹³

The radiology community recognized the inconsistencies and complexities of HCC imaging. Therefore, the American College of Radiology endorsed the Liver Imaging Reporting and Data System (LI-RADS), which had the goal of reducing variability in lesion interpretation through standardization and improving communication with clinicians.¹⁴ LI-RADS uses a diagnostic algorithm for CT and MRI that categorizes observed liver findings in high-risk individuals based on the probability or relative risk of HCC without assigning a formal diagnosis.¹⁴ LI-RADS takes into account arterial phase enhancement, tumor size, washout appearance, the presence and nature of a capsule, and threshold growth.¹⁵ LI-RADS categorizes an observed liver finding on a scale of 1 to 5, with 1 corresponding to a definitely benign finding and 5 with definitive HCC.¹⁴ Furthermore, LI-RADS sought to limit the technical variabilities among institutions.

LI-RADS was launched in 2011 and has been utilized by many clinical practices while continuing to be expanded and updated.¹⁶ Recent studies examined the specificity of LI-RADS as well as interreader variability.^17,18 For nodules viewed on MRI, both LI-RADS categories 4 and 5 had high specificity for HCC.¹⁷ When looking at interreader repeatability, LI-RADS showed moderate agreement among experts using the diagnostic algorithm.¹⁹ Further studies have compared LI-RADS with the AASLD guidelines and the Organ Procurement and Transplantation Network (OPTN) guidelines.¹⁶ When compared with other guidelines, LI-RADS expands the definition of indeterminate findings into probably benign, intermediate probability of HCC, and probably HCC, which corresponds to LI-RADS categories 2, 3, and 4.¹⁶

We looked retrospectively at a group of patients previously diagnosed with HCC to see whether utilizing the LI-RADS scoring system within our screening system might have allowed an earlier prediction of HCC and a timelier intervention. Prior to this investigation the LI-RADS system was not used for HCC screening at our US Department of Veterans Affairs (VA) facility. We examined screened patients at the Memphis VA Medical Center (MVAMC) in Tennessee who were subsequently diagnosed with HCC to see which LI-RADS category the last surveillance CT prior to diagnosis would fall into, 6 months to a year prior to the diagnosis of HCC. Our control population was a group of patients screened with CT for their liver nodules who were found not to have HCC.