Anthony B. El-Khoueiry, MD: We’re going to shift to discussing second-line therapy and beyond. There’s a new reality in the treatment of advanced HCC [hepatocellular carcinoma]. We have multiple first-line options. We’ve had a slew of agents that had been previously approved for second line and beyond that we’re going to discuss. But it’s becoming more of a reality that we’re able to sequence therapies almost as we do for other solid tumors. Let’s talk about this and the best way to do it. As a reminder, I’ll start by reviewing some of the data for tyrosine kinase inhibitors [TKIs] or VEGF [vascular endothelial growth factor] agents in second line and beyond, and then Dr Gholam will summarize the immunotherapy options that have been studied in second line and beyond. Then we’ll discuss how to deploy them in clinical practice.

I’ll start with the RESORCE trial, which was the first second-line approval. That’s the study that took patients who are post-sorafenib [Nexavar] and compared regorafenib [Stivarga] to placebo in the second line. The study had the important caveat that patients must have tolerated a dose of 400 mg of sorafenib or higher for 20 of the 28 days. It required minimum tolerability of sorafenib in a nutshell. The primary end point was median overall survival [OS]. It met its primary end point of improvement of OS with regorafenib vs placebo, with the median OS being around 10 months with regorafenib in the second line.

Then came the CELESTIAL study, which looked at cabozantinib [Cabometyx]. This is a multi-targeted tyrosine kinase inhibitor. Among its targets are VEGF receptor 2 as well as MET, which is the receptor to hepatocyte growth factor. There are other targets, like AXL and MER, which may have some immunomodulatory effects. The CELESTIAL trial allowed up to 2 prior lines of therapy. About a quarter to a third of the patients were third-line patients. The study required sorafenib exposure, so all of the patients were post-sorafenib. It showed an improvement in median overall survival with cabozantinib vs placebo. It was around 10 months with cabozantinib. The hazard ratio was even better when it looked at only second-line patients vs second- and third-line patients. This received approval and became an option that we used in the second or third line.

Let’s talk about the REACH-2 study, which evaluated ramucirumab [Cyramza], the VEGF receptor 2 antibody. This study was done as a follow-up to the original REACH trial, which evaluated ramucirumab vs placebo and was a negative trial. However, in subgroup analysis that was done in retrospect, patients who had a high AFP [alpha-fetoprotein] of 400 ng/mL or higher appeared to benefit. A REACH-2 study was done as a follow-up. One of the inclusion criteria was having an AFP of 400 ng/mL or higher. These patients were all post-sorafenib. They were randomized to ramucirumab or placebo. There was a survival benefit with ramucirumab compared with placebo in that trial as well. One of the differentiating factors, besides the fact that the AFP had to be 400 ng/mL or higher, is that ramucirumab didn’t have the adverse effect profile of TKIs, specifically skin toxicity and diarrhea, but had some other things to be concerned about, like proteinuria and some bleeding, like a VEGF-directed antibody. All these trials were limited to Child-Pugh A patients.

Of note, in the CELESTIAL trial, we went back and did a subgroup analysis on patients who by week 8 went from Child-Pugh A to Child-Pugh B. Child-Pugh B was an exclusion, but a subset of patients went from Child-Pugh A to Child-Pugh B at week 8. When we analyzed their outcome, they also seemed to have a benefit from cabozantinib vs placebo. The tolerability wasn’t very different. These aren’t all necessarily related to the study therapy, but when you look at the adverse event profile, in the Child-Pugh B group, there were a little more liver adverse events and ascites, which isn’t unexpected for patients who become Child-Pugh B. But they still seemed to derive a benefit. This requires prospective validation. These aren’t patients who started with Child-Pugh B, but it’s worth mentioning this data set.

Transcript Edited for Clarity