Response to: "Comment on Hepatectomy Versus Sorafenib in Advanced Nonmetastatic Hepatocellular Carcinoma".

Abstract

We thank Quian et al for their interest about our article entitled “Hepatectomy Versus Sorafenib in Advanced Nonmetastatic Hepatocellular Carcinoma.”1 Their comments indeed give us the opportunity to point out some relevant concepts. The latest Barcelona Clinic Liver Cancer update enlarges the number of drugs available for the treatment of advanced stage HCC, with Sorafenib not being anymore, the first-line treatment.2 However, sorafenib and the subsequent systemic therapies for HCC have been introduced just for unresectable patients.3 Nevertheless, despite the resectability of a certain part of patients with advanced HCC was well established,4,5 the systemic therapy has been superimposed as the front-line and, in some way, the only therapeutic choice. Despite the robust evidence reporting a significant survival advantage of liver resection for advanced stages (whenever possible),6,7 the surgical option has remained ignored by the guidelines for these cases. Our article aimed to investigate what that has meant by comparing the previous first-line treatment (sorafenib) versus the surgical curative approach, in a very large national cohort, weighting all the anamnestic and oncologic differences between the 2 groups. About the potential confounding factors represented by adjuvant therapies after surgery as claimed by Quian et al, it is worth to mention that patients who underwent liver resection did not receive further treatment unless tumor recurrence occurred and, even in the case of recurrence, surgery remained the first option, if suitable. Conversely, patients undergoing systemic therapy prosecuted sorafenib or were shifted to the best supportive care in the event of progression. We agree with Quian et al that neo- and adjuvant approaches will progressively play an increasing role in the management of these patients that, in the case of neoadjuvant therapy, might subsequently be also operated.8,9 However, while waiting for the read out of ongoing trials testing combinations of different therapies, especially in the neoadjuvant setting, there is no reason to postpone surgery in resectable patients. The optimal survival observed after liver resection could also be explained as the consequence of a peculiar conceptual approach to those patients: in fact, when patients are managed with a curative intent, in the case of relapse other curative approaches are considered (and feasible) more often than in patients treated with palliative therapies. Otherwise, when patients are considered at the time of HCC diagnosis as not curable, uncommonly they are switched to other treatments in the case of, for example, a favorable response. This conceptual approach is supported by the treatment allocation proposed by the Barcelona Clinic Liver Cancer which is quite rigid, considering the different treatments available as alternatives rather than part of an sequential pathway to reach the cure or, at least, the stabilization of the tumor. From this point of view, a therapeutic hierarchy strategy,10 may ensure to tailor the clinical decision to the patient characteristics, without restricting the treatment to a particular stage, but offering the most suitable approach as the sum of the survival advantage and the feasibility. Quian et al emphasized the differences between the 2 groups (sorafenib vs resection) as a major drawback of our paper. Objectively, these differences exist. However, this is not surprising but somehow obvious, because the definition of advanced HCC covers a wide variety of disease presentations and liver resection could be a solution applicable just in a part of them. However, when that happens results are providing not negligible chances of cure,4–7 and of established safety since decades.11 Anyway, despite the retrospective nature of our analysis, the methodology we used reduced the risk of selection bias and the effect of confounding factors, by the matching or the weighting of all the differently distributed variables between groups. Indeed, large numbers and advanced statistical method, as per inverse probability weighting (IPW), are able to provide evidence comparable to that of randomized clinical trials.12 As pointed out in the discussion, we do not presume that our methodology is perfect, but the results we reported deserve consideration, strengthening more than an impression that surgery, when feasible, should be hierarchically considered a priority. Finally, Quian et al suggested evaluating the results according to different types of macrovascular invasion aimed at providing a therapeutic indication between surgery and systemic therapy. However, patients with invasion of the main portal trunk were excluded from our study. In the other cases, the invaded portal branch was resected together with the pertinent vascular area of the parenchyma. Because of this, we cannot provide a comparison among different management strategy.