Richard S. Finn, MD: The challenge with things like Y-90 is they have a different level of evidence for use than systemic treatment, for example. In the United States, they’re considered a device, and so we don’t have strong randomized data against chemoembolization. Would you agree with that?
Richard H. Marshall, MD: I agree.
Richard S. Finn, MD: But to their credit, you mentioned that there has been a lot of activity in this space in the past few weeks. The SARAH trial, you started to talk about it. Why don’t you give us some background about what we do know from randomized data?
Richard H. Marshall, MD: We do know, when we talk about the comparison of TACE to Y-90, the overall survivals are not different. What we do know about the differences between these is that Y-90 is usually better tolerated by patients based on quality-of-life scores. There are subgroups who do better with TACE, and there are subgroups who do better with Y-90. But from the more recent data, specifically the SARAH trial, we didn’t see a difference in Y-90 and sorafenib, a systemic therapy, in an overall survival endpoint. What we did see, again, were better quality-of-life scores in the patients who were treated with Y-90. We also saw fewer adverse events in the patients treated with Y-90. So, we’re learning a lot about this therapy and how it fits in locally advanced hepatocellular carcinoma.
Richard S. Finn, MD: You say that there wasn’t any difference between the 2. I think some of us, looking at these clinical trials, get concerned that they were geared to show that Y-90 was better than sorafenib and they didn’t meet that endpoint. They would probably be negative trials. Jordi, can you comment, briefly?
Jordi Bruix, MD: Yes. This is methodology and statistical understanding. You can design a trial for superiority and, if it is positive, it’s positive. If it is negative, it’s negative. Being negative, you cannot say that because it was not superior, it’s similar. So, the question to show that something is similar has to be a completely different result.
Richard S. Finn, MD: Right.
Jordi Bruix, MD: This is something that is difficult to understand when you are not an expert in statistics and the message is something that, maybe, not better is not meaning to be similar, and the concepts are philosophical. But it is relevant to understand that even with the suggestions that there is activity, because activity is there without embolization, you kill the tumor. You cannot say that now we can replace sorafenib by radioembolization because the data are not there.
The quality of life that is always emphasized, and what you choose to define as quality of life, is quite controversial, to be politically correct. You can have reductions in the scores, but then you have also tools to determine if they are clinically meaningful. And this is all the confusion around quality of life. When you look at the graph, it looks like it goes down—more or less intense. But this doesn’t mean that it’s relevant. And then, with all the noise we have in the field, of radioembolization being known to be active against cancer, we still do not know if this has a niche that benefits a subgroup of patients. I think that what comes from the SARAH trial, and the SIRveNIB trial that was just presented at the 2017 ASCO Annual Meeting, was also negative. When you put the trials together, this will just be hypothesis-generating. The data will not really bring robust evidence.
Richard H. Marshall, MD: As you mentioned before, we really need to provide individualized care, and the Barcelona Clinic Liver Cancer system does help categorize care. But on a per-patient level, there are patients who may do better with one therapy versus the other.
Jordi Bruix, MD: You can recommend whatever, but this cannot be a level of recommendation.
Richard H. Marshall, MD: Correct.
Jordi Bruix, MD: For example, you can decide that, in a patient, the best thing you can do is Reiki. This does not mean that Reiki is effective.
Richard S. Finn, MD: OK.
Jordi Bruix, MD: So, this does not mean that this is effective. And in the real patient, you can decide, but it cannot become a level of recommendation.
Transcript Edited for Clarity