Purpose: This study aimed to evaluate Japanese patient preferences regarding features of intermediate or advanced (Progressed) hepatocellular carcinoma (HCC) treatments: transarterial chemoembolization (TACE), hepatic arterial infusion chemotherapy (HAIC), and oral anti-cancer therapy. Methods: Patients with HCC, recruited from clinical sites and a patient panel in Japan, completed a cross-sectional web-based survey. Preferences were quantified using best-worst scaling, where patients identified the best and worst among 13 treatment features. Direct elicitation was used to identify preference for TACE, HAIC, or oral therapy, including the likelihood of trying each. Additional items asked for the willingness to try an oral medication that delays progression by six months but has an 8% or 21% risk of severe hand-foot skin reaction (HFSR). Results: The sample (N=119; 29 early stage; 90 Progressed) most preferred "oral medication", "artery branches plugged", and "prevents formation of new blood vessels", and least preferred "risk of liver damage" and "risk of catheter-related complications". Overall, 51%, 40%, and 8% preferred oral therapy, TACE, and HAIC, respectively (p<0.05), and the mean likelihood of trying each were 59%, 52%, and 35%, respectively (p<0.001). Patients with sorafenib or TACE experience most preferred what they had received; however, both groups were equally willing to try the other treatment. Patients preferring oral therapy favored "oral medication" over "artery branches plugged", "surgery is repeated as required when the cancer grows again", and "risk of liver damage", compared to those preferring TACE (p<0.05). Sixty-eight percent would probably try therapy with an 8% risk of severe HFSR, compared to 50% with a 21% risk. Conclusion: Treatment type, mode of action, and risks may drive HCC patient preferences. Such features likely should be incorporated into physician-patient interactions regarding treatment decision-making.
Keywords: best-worst scaling; hepatocellular carcinoma; patient preference.