Value of Pretreatment MRI Determined Parameters for Predicting Outcome After Radio-frequency Ablation of Hepatocellular Carcinoma

Recent pathological, immunohistochemical and molecular profiling have progressively revealed the complex classification of primary liver cancers and have resulted in a subclassification roughly separating mucinous cholangiocarcinomas from mixed tumors from "pure" hepatocellular carcinoma (HCC).

In addition, HCC is classified as keratin (K) 19 positive or K19 negative. K19 is a biliary/hepatic progenitor cell (HPC) marker only expressed in a subset of HCC with poor prognosis and high risk of early recurrence after treatment; particularly in radio-frequency ablation (RFA). These patients consequently show worse survival compared to patients with K19 negative HCC.

A recent publication has shown the value of pretreatment biopsy with K19 staining and suggests that the role of routine biopsies in potentially curable HCC should be reconsidered.

However, currently, pretreatment biopsies are rarely performed in the diagnosis of HCC due to the excellent performance of magnetic resonance imaging (MRI) in detection, diagnosis and staging of cirrhotic livers. This progression of MRI is based on technical refining of three-phase contrast-enhanced imaging, development of the cellular marker sequence diffusion-weighted MRI (DWI) and hepatospecific biliary contrast-agents. While pretreatment biopsies would have incremental value in treatment planning, its routine application in daily clinical practice is faced with multiple problems. First, K19 evaluation at histopathology can - in most cases - only reliably be performed on whole resection specimens and is more difficult on biopsies. Second, routine biopsies would lead to a dramatic increase in workload for radiologists and pathologists influencing daily practice and organization. Also, lesions are not always readily detectable at ultrasound and possibly not eligible for percutaneous biopsy requiring laparoscopy guided biopsy under anesthesia. Third, the risk of tract seeding remains controversial but has not been completely excluded.

Previous publications have indicated imaging patterns that may be associated with worse prognostic tumoral parameters. The corona sign seen at CT during arterial portography and hepatic arteriography (invasive imaging only performed in Japan) is associated with the presence of microvascular invasion. However, this sign has not yet been described at MRI. Also MRI derived parameters may predict the presence of HCC expressing progenitor cell markers, including K19. If MRI determined imaging parameters could indeed provide a surrogate marker for presence of K19 and/or microvascular invasion as potential important prognostic factors in RFA of HCC, these imaging parameters may thus hold prognostic information towards RFA treatment and possibly predict treatment outcome. The hypothesis is that lesions showing features correlating to K19 positivity such as persistent (rim) enhancement and rim like intensity at DWI may correlate to worse prognosis than HCC that don't express these imaging features.