Conventional transarterial chemoembolization versus drug-eluting bead transarterial chemoembolization for the treatment of hepatocellular carcinoma

Since 2006, when the first drug-eluting beads became commercially available, DEB-TACE has become the de facto standard in many centers, gradually replacing lipiodol-based cTACE as the standard treatment for patients with inoperable HCC. Nonetheless, the scientific basis for that paradigm shift is poor. The PRECISION V trial was by far the biggest study to compare TACE with DEB-TACE, including 212 patients [8]; its primary endpoint was tumor response measured by MRI after 6 months. However, the lack of hard endpoints, such as OS or progression-free survival (PFS), was criticized afterwards [28]. At least some benefit was shown for DEB-TACE over cTACE regarding objective response in the subgroup of more advanced patients (Child-Pugh B, ECOG 1, bilobar disease, and recurrent disease). Reyes et al. published another prospective randomized trial in 2009 [9]. Unfortunately, this phase II trial included only 20 patients, too few to allow a definitive assessment of the secondary endpoints, PFS and OS. Sacco et al. published a series of 67 patients in which they primarily investigated safety, toxicity, and tumor response after 1 month; survival only served as a secondary endpoint [10]. The study by Van Malenstein et al. was also restricted regarding the evaluation of toxicity and safety; tumor response was measured only once 6 weeks after the first treatment [11]. All other available studies were retrospective [12‐17]. Three of them relied primarily on surrogate endpoints based on cross-sectional imaging [13, 14, 16], and two of these reported OS as a secondary endpoint [13, 14]. One retrospective trial only reported complications [29]. Only the studies by Dhanasekaran et al., Scartozzi et al., and Wiggermann et al. [12, 15, 17] primarily analyzed hard endpoints such as PFS or OS. Nonetheless, the results of these three studies differed considerably. Dhanasekaran et al. and Wiggermann et al. each reported significant survival benefits of DEB-TACE, while the results of Scartozzi et al. indicated the opposite. Notably, none of these studies considered tumor grading, although it has been reported to have a significant impact on the outcome of patients with HCC after liver transplantation [23‐25], and TACE [11].

Llovet et al. and Lo et al. were able to demonstrate a survival benefit of TACE versus symptomatic treatment in 2002 [30, 31]. Nonetheless, in 2011 this statement in general was heavily attacked by the Cochrane Review, which concluded that a clear survival benefit of any type of TACE has not yet been demonstrated [32]. Hence, we decided to employ OS as a primary endpoint and believe that future prospective studies should also be based on hard endpoints such as OS or PFS [5, 33].

Altogether, the present study showed no significant difference in OS between cTACE and DEB-TACE. Additional analyses were conducted to take into account Child-Pugh stages and the status of portal invasion; both methods led to comparable survival times. During the preparation of our manuscript, a prospective analysis was published. This trial was stopped early for futility and found no significant difference between cTACE and DEB TACE in the 2-year survival – further supporting our OS data presented herein [34]. The main difference in our study was the significantly lower number of treatments needed in the DEB-TACE group compared with cTACE, which is likely to enhance patient comfort. We consider that these findings are highly relevant for clinical decision-making. Patients can be exposed to a lower number of treatment sessions, thus lowering their risk of procedure-associated complications. Additionally, the need for less treatment sessions makes DEB-TACE more cost efficient than cTACE despite the considerable price difference between lipiodol and drug-eluting beads.

The main weakness of this study is its retrospective, single-centered and non-randomized design. We tried to eliminate any bias by defining clear drop-out criteria before analysis, which led to the exclusion of 63 % of patients (Fig. 1). The remaining cTACE and DEB-TACE cohorts exhibited no significant difference in any of the factors known to significantly affect outcome. A relatively high percentage of patients were classified as BCLC C. According to the BCLC scheme, stage C patients should receive sorafenib therapy. Nonetheless, this is not reflected in real clinical practice in most major liver disease centers worldwide. Especially if a patient is classified as BCLC C because of having ECOG stage 1, TACE is often preferred over sorafenib. As this was the case in most of our BCLC C patients, we believe that our collective represents the typical TACE patient quite well. Another limitation is the use of two different chemotherapeutic agents. Nonetheless, we believe that this difference is negligible. To this day, no drug or combination of drugs has been proven better than any other for the treatment of HCC in a randomized trial [35]. Only one non-randomized retrospective trial found that cisplatin was superior to doxorubicin, but at the cost of higher side effects [36]. We did not investigate complications as this was not the focus of our study, and this is generally difficult to assess in a retrospective analysis. Furthermore, we used DC Beads® ranging in diameter from 100–300 μm to 500–700 μm. The larger drug-eluting beads were primarily used at the beginning of the treatment period; from 2010 on, we completely switched to smaller beads of 100–300 μm, which we believe preserve the patency of the arterial feeding vessels, thereby allowing more repetitive treatments. From 2012 on, we switched to the even smaller DC Beads M1® (70–150 μm). Nonetheless, it is questionable whether the assumed positive effect of smaller beads might have changed the present results.