We would like to thank Gao et al. for their correspondence (Heterogeneity of intermediate-stage HCC necessitates personalized management including surgery. Nat. Rev. Clin. Oncol. doi:10.1038/nrclinonc.2014.122-c1)1 on our recent Review (Treatment of intermediate-stage hepatocellular carcinoma Nat. Rev. Clin. Oncol. 11, 525–535; 2014).2 In our Review, we outlined the current approaches to stratifying patients with intermediate-stage hepatocellular carcinoma (HCC), and discussed the potential heterogeneity that remains among this population. We stated that transarterial chemoembolization (TACE) is considered the standard treatment for intermediate-stage HCC, but emphasized that not all patients with intermediate-stage HCC are good candidates for TACE; those patients who are not candidates for TACE or patients in whom TACE has failed should be considered for other treatments, including sorafenib and radioembolization. In their correspondence, Gao et al.1 criticized our Review for stating that TACE and sorafenib should be the only standard treatments for intermediate-stage HCC (a paradigm that they disagree with), whereas this statement is in fact far removed from the message delivered in our Review. Furthermore, Gao et al.1 claim that the Barcelona Clinic Liver Cancer (BCLC) treatment recommendations3 are based on findings in populations of patients with intermediate-stage HCC as a whole and, therefore, do not provide guidance as to which modality will yield the best result in individual patients. As a result, they state that deviations from the guidelines are highly frequent in clinical practice.1 Guidelines are developed to help patients, health practitioners, health-care providers and governments decide which treatment option is most appropriate for specific conditions and/or under particular circumstances.4 Such recommendations should be based on the strongest scientific evidence, but of course their application to the treatment of individual patients should take into account several clinical factors that, for sure, are not detailed in guidelines.

Gao et al.1 go on to highlight the important role of hepatic resection in the recently published Hong Kong Liver Cancer (HKLC) classification.5 This proposed stratification system accepts that hepatic resection can be pursued in subgroups of patients with intermediate-stage or advanced-stage HCC.5 Regrettably, the HKLC classification has several limitations, as summarized in an editorial that accompanied its publication.6 The most-relevant concern is that the study on which the HKLC classification is based was retrospective and, therefore, was probably subject to an unintentional selection bias; patients who were selected for hepatic resection rather than TACE must have had features that gave the surgeon confidence that a good outcome would be achieved, whereas those selected for TACE probably lacked such characteristics, immediately introducing a bias against TACE.7

Gao et al.1 also comment on a recent randomized controlled trial conducted by Yin et al.8 in Shanghai, China, which compared hepatic resection and TACE. Although the results were clearly in favour of surgical treatment rather than TACE (median overall survival of 41 months versus 14 months; P <0.001), and the multivariate analysis demonstrated that type of treatment was an independent risk factor associated with overall survival,8 this trial has some issues that deserve consideration. The major concern is the very poor outcome observed for the patients treated with TACE (that is, a median overall survival duration of only 14 months). The poor survival outcome did not seem to be related to treatment-related toxicity, as no grade 5 adverse events were reported.8 However, this outcome could reflect the inexplicable mediocre response rate (objective response rate of 23%, comprising 10 complete responses and only 11 partial responses or stable disease), despite the fact that the modified Response Evaluation Criteria In Solid Tumors (mRECIST) response criteria were used.8 Contemporary conventional TACE procedures, using a super-selective approach for patient stratification, are associated with objective response rates >50%.9 Moreover, in Japan10 and Europe,11,12 TACE is associated with median survival durations of 40–48 months in patients with intermediate-stage HCC with preserved liver function—who according to Yin et al.8 would be potential candidates for hepatic resection. Another issue with the trial by Yin and colleagues8 is the ill-defined inclusion criteria, expressed simply as 'good surgical risk patient' and 'resectable HCC', which prevents any attempt to reproduce and validate the results. In addition, although not reaching statistical significance, patients in the TACE group had slightly larger tumours and slightly poorer liver function than those in the surgical cohort, and the authors did not report on performance status.8 More interestingly, only 7% of the patients screened (180 out of 2,502 patients) met the inclusion criteria for being randomized,8 confirming that just a minor proportion of patients with intermediate-stage HCC might be considered for hepatic resection.

Finally, regarding the suggested treatment allocation for patients with BCLC B (intermediate-stage) HCC summarized in Figure 1 of the correspondence by Gao and co-workers,1 the BCLC proposals have always stated that patients with a solitary HCC larger than >5 cm that has expansive growth, and who remain free of symptoms, vascular invasion and tumour dissemination after proper imaging evaluation might benefit from surgical resection; disease in such patients should be classified as BCLC stage A.3,13,14,15

Additional trials comparing hepatic resection with state-of-the-art TACE are needed in order to evaluate the potential benefit of hepatic resection in patients with intermediate-stage HCC. The current evidence, however, does not support the recommendation of this treatment approach in patients with BCLC B disease.