On behalf of the Society of Interventional Radiology, our selected group representing multiple disciplines that treat hepatocellular carcinoma (HCC) has critically analyzed the Cochrane Library Report “Transarterial (Chemo) Embolisation for Unresectable Hepatocellular Carcinoma (Review).”1 In short, we believe that this review does not directly address the issue of survival benefit of chemoembolization as it is currently practiced for patients with unresectable HCC. Clearly, the three Cochrane authors are experts in evidence-based medicine, but to the best of our knowledge, none are experts on the topic of locoregional therapy of HCC. We believe that this lack of topic expertise led to inclusion of a number of prospective randomized controlled trials that do not reflect current practice and has led to bias that invalidates their conclusions. Specific comments follow.
Chemoembolization (TACE) as it is currently practiced involves administration of at least one cytotoxic chemotherapy agent plus solid embolic particles, plus-or-minus ethiodized oil. Bland embolization (particles without chemotherapy) operates by complete arterial occlusion resulting solely in tissue ischemia and infarction. Thus, the three studies that used only bland embolization should be excluded or analyzed completely separately.2-4 TACE that uses chemotherapy in suspension but without particulate embolization is more akin to hepatic artery infusional therapy than to true TACE, so those studies should also be excluded.5-7 Until recently, when sorafenib was documented to increase survival in patients with HCC and preserved liver function, TACE was used as a monotherapy. Use of TACE either as neoadjuvant therapy prior to potentially curative resection,8 or as adjuvant therapy following resection,5-7 is not the standard of care in North America or Europe, and thus is irrelevant to this review. One other study assessed focal ablative therapy (either percutaneous ethanol injection or radiofrequency ablation) ± bland embolization.4 Since all patients received potentially curative local ablative therapy, the effects of bland embolization are obscured; thus, this study should also be dismissed. Finally, the current standard practice for administration of chemoembolic agents is by way of the lobar or sublobar hepatic arteries in order to minimize delivery to nontumorous liver. Studies that used proper hepatic artery administration essentially treat the whole liver, and should be discarded as not representative of the state-of-the-art practice, and likely biases the results toward a poorer survival outcome.9-11
Of the original nine studies in the meta-analysis, three remain that legitimately reflect the standard practice of TACE: chemotherapy in suspension with particulate emboli, administered within a selective lobar or sublobar fashion, and used as either monotherapy or with sorafenib.3, 12, 13 When examining the hazard ratio for the group of studies examining TACE (rather than bland embolization) (Analysis 1.2), the hazard ratio is 0.79 (95% confidence interval: 0.63-1.00), which would suggest that TACE in fact is likely beneficial with regard to overall survival.
Several caveats need to be made with respect to the authors' conclusion that “there is not sufficient evidence to support or refute TACE or bland embolization for patients with unresectable HCC.” First, the technology continues to evolve. Drug eluting beads with ionically bound doxorubicin appears to significantly increase tumor response while reducing serious adverse events in subsets of patients with advanced HCC.14 Second, the preferred therapy in North America and Europe for patients with limited HCC (UNOS T2 or T3) and cirrhosis is liver transplantation. Five-year survival in patients who meet the Milan criteria and undergo cadaveric liver transplantation is ∼75%, compared to less than 10% in patients who receive supportive care, TACE, radioembolization, and/or sorafenib. TACE plays a key role in a “bridge-to-transplant” and “downstage-to-transplant” strategy. When considering recommendations regarding TACE (or other locoregional therapy), its role in augmenting liver transplantation should be addressed as well.
In summary, whereas the Cochrane review may have some methodological validity, the data presented do not properly analyze the dataset from the last two decades and the conclusions are not applicable or interpretable to HCC management in 2012. Based on the comments above, the Society of Interventional Radiology recommends that the Cochrane Review committee responsible for this report reconsider the prospective randomized controlled trials included in the review and revise the report. It is also our suggestion that the writing group for the review contain at least one individual experienced in the intervention being assessed (preferably both a Hepatologist who treats HCC and an Interventional Radiologist who perform TACE arterial chemoembolization in patients with HCCs) to assure that the methods used in the included studies reflect current standard practice.