From its approval in 2008, sorafenib is the recommended treatment option for advanced-stage patients and its safety and efficacy has been confirmed by several studies. However, its mechanism of action is not completely understood and many efforts have been dedicated to investigating possible treatment response predictors. Dermatological adverse events occurring within the first 2 months of treatment are predictors of longer survival, while the same role for hypertension and diarrhea still needs a prospective confirmation. This association is opposite to the strategy of starting at a low dose as it may imply suboptimal drug exposure. In case of radiological progression, the appearance of new extrahepatic metastasis or vascular invasion significantly worsens life expectancy if compared to other patterns of progression. To date no genetic or biologic marker is available to predict response, even if some encouraging results have been reported by the study of polymorphism of VEGF and its receptor. Currently, data are conflicting about the possible predictive role of α-fetoprotein. Due to failure or the progression of therapies for earlier evolutionary stages (BCLC B) some patients in such a clinical profile may be treated with sorafenib. Indeed, almost 50% of the sorafenib-treated patients belong to this class. Patients with severely decompensated liver disease (jaundice, ascites in need of intense diuretic therapy/paracentesis) may not benefit from treatment. The use of sorafenib in the waiting list for liver transplantation is controversial, while its use at an advanced age requires careful evaluation of existing comorbidities that may increase the risk of adverse events. Many strides have been made in the field of hepatocellular carcinoma systemic therapy, and many remain to be realized. Considering the disappointing results of the trials conducted on new agents, a more dynamic interpretation of events together with the development of new strategies is key to enriching new and hopefully more successful trials.