Resection for hepatocellular carcinoma with hepatic vein tumour thrombus: Pushing the limits beyond the guidelines frontiers

Abstract

Surgical treatment of hepatocellular carcinoma associated with hepatic vein tumor thrombosisJournal of HepatologyVol. 61Issue 3PreviewPresence of hepatic vein tumor thrombosis (HVTT) in patients with hepatocellular carcinoma (HCC) is regarded as signaling an extremely poor prognosis. However, little is known about the prognostic impact of surgical treatment for HVTT. Full-Text PDF The paper by Kokudo et al. [[1]Kokudo T. Hasegawa K. Yamamoto S. Shindo J. Takemura N. Aoki T. et al.Surgical treatment of hepatocellular carcinoma associated with hepatic vein thrombosis.J Hepatol. 2014; 61: 583-588Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar] in this issue of the Journal of Hepatology timely shakes up the BCLC guidelines for the management of hepatocellular carcinoma (HCC), raised by certain as absolute dogma, despite the proponents of these guidelines themselves as shown by the successive modifications they regularly propose. The “Kokudos” (son and father) and co-workers analyse here a retrospective series of 187 consecutive liver resections for HCC extending to the hepatic veins. Precisely, the study includes 153 patients with microscopic invasion of the hepatic veins, by definition unknown before surgery, 21 with macroscopic major hepatic vein invasions, and 13 with tumour thrombus into the inferior vena cava. These 3 subgroups represent 10.0%, 1.3%, and 0.9%, respectively, of all consecutive HCC (1525 cases) operated during a 17 years period (1994–2011) at one of the most prestigious Japanese liver surgery tertiary centres. The main interesting subgroup concerns patients with macroscopic tumour thrombus in the main hepatic veins or the vena cava known before the decision to proceed for surgery (2.2% of operated HCC). Excellent results could be achieved for this highly selected subset of severe patients (38.2% with cirrhosis, 14.7% with portal hypertension) with advanced disease (88.2% with combined portal vein tumour thrombus and 26.5% with ⩾3 tumours): the 90 days mortality rate was 5% and the median survival time was 3.95 years! Multivariate analysis revealed tumour thrombus in the vena cava and R1/R2 resection as risk factors for the overall survival. These surgeons-authors must be commended for their pugnacity and their willingness to share this experience in a journal of hepatology, i.e., the eponym “Journal of Hepatology”. The latter point is important. Indeed in our numerous discussions at multidisciplinary team meetings with our colleagues and hepatologist friends, all will acknowledge that the latter write, read, and quote almost exclusively papers from the literature of their specialty and the very same “scotoma” is valuable for the liver surgeons. In other terms, daily, hepatologists and liver surgeons have a dialogue of deaf relying on an evidently-biased analysis of the so-called evidence-based medicine. Of course for a hepatologist the evidence level of the study by Kokudo et al. is low: What about the control group? Maybe the patients reported here are simply harvesting low progressing tumours with a spontaneous fair survival? … However, if the prospective randomised study comparing resection vs. sorafenib the reference treatment in this setting of BCLC C patients is ever done, its results will probably not be applicable to the real life for many well-known reasons. This means that we eagerly await new methodological tools maybe such as big health care data analysis or propensity-score methods to answer to the question with a higher level of evidence. An individualised prognostic score model might be one solution. While submitting this point of view, I will take the opportunity to share some additional comments. Since the Okuda classification reported in 1984, more than 10 classifications of HCC have been reported including the CLIP (1998), GRETCH and BCLC (1999), CUPI and sTNM (2002), JIS (2003), Tokyo (2005), AJCC/UICC 7th ed (2010), Taipe (2010). The absence of even one common criteria t