We’ve been following sorafenib (nexavar), a multi-kinase inhibitor with anti-angiogenic activity (see prior post). This oral agent, which is already approved as an effective treatment for cancers of the liver and kidney. It’s been studied as a single agent and appears to perhaps be associated with prolonged stable disease even if not with clear tumor shrinkage, but a randomized phase III trial of chemo with or without sorafenib showed no benefit and a particular lack of benefit (actually, an increase in mortality, as in risk of death) in patients with squamous NSCLC (see prior post).

Another study was presented at ASCO this year by Dr. Joan Schiller, now at Univ. of Texas – Southwestern in Dallas. She heads the lung committee for Eastern Cooperative Oncology Group, or ECOG, one of the main North American collaborative cancer research groups, which ran this multi-center trial, known as ECOG 2501 (or E2501)(abstract here).

This trial had a very novel design, which was used specifically to study an agent that may be beneficial primarily for lung cancer patients with slowly growing disease, if sorafenib is indeed an agent that slows progression more than shrinks tumors dramatically. The study used what is called a randomized discontinuation design, in which all patients started by receiving sorafenib for two months, and those with a response (very few) continued on sorafenib, while those who progressed in that time came off of the study. But the patients who showed stable disease after two months on the trial were then randomized to continue on sorafenib or go onto a placebo. The trial design looks like this:

This may be the most complex thing we’ve ever discussed here, so if you understand this, you can understand anything here. If you don’t understand it, don’t stress – it’s not critical to get the gist of things. But it makes it a little easier that the trial focuses just on “step 2”, the patients who had stable disease after two initial months on sorafenib, who were randomized to continue it or switch to a placebo:

The trial was done this way really just to look for a “signal” of whether this agent is helping some proportion of people with relatively indolent NSCLC, so the primary goal was to compare the proportion of patients with “disease control” (responding or stable) on ongoing sorafenib vs. placebo.

This study was for patients with advanced NSCLC who had received at least two prior lines of chemotherapy, and EGFR inhibitor or anti-angiogenic drug therapy didn’t count as a line of therapy.

The trial started with 342 patients enrolled for the run-in phase, of whom 107 had stable disease and were able to be randomized after two months of sorafenib, so more than 2/3 of patients dropped off (6 patients, or 2%, had a response). From this small group, another 24 weren’t analyzed because they were found to be ineligible or just never managed to get treated (that’s a high percentage), leaving just 84 to be randomized to continue active drug or switch to a placebo.

As it turned out, there was a relatively minor error in how the trial was conducted, due to a disconnect between lists of patients vs. pills given, and several patients got pills that were wrong for what they were assigned (truly a blinded study). They were able to keep track of what patients actually received and analyzed results on that basis, but because of that error they ended up with 41 patients receiving sorafenib, 12 receiving a combination of sorafenib and placebo at different points, and 30 receiving placebo only. This certainly wasn’t great, but since patients were able to cross over to active drug if they received placebo, it probably wasn’t too compromising, except to our ability to interpret the results of the trial.

When comparing the results of the recipients of sorafenib vs. placebo after the run in phase, it appeared that the patients on active drug did get some benefit. Stable disease was recorded in 47% of the sorafenib recipients vs. only 16% of those who received placebo (p = 0.01), and also a better progression-free and overall survival:

Sorafenib wasn’t associated with any really surprising side effects. Looking at everyone enrolled (including the two-month run-in phase), there was a less than 1% risk of serious bleeding in the lungs (1 case fatal out of 331 evaluated), 11% with moderate fatigue, 9% with moderate hand-foot syndrome (blistering and pain on the palms and soles), 3% with moderately severe high blood pressure, and otherwise just a few rare events. Four patients had strokes, a couple developed blood clots, and one person died of kidney failure in addition to the one who died from a pulmonary bleed.

What can we say about this trial overall? It’s fair to say that there’s some evidence of activity, that some patients benefit, but this was by looking at the quarter of patients who hadn’t already progressed after two months. I know people who participated on this trial who are convinced that the drug was meaningfully helpful for some patients, who saw patients progress on placebo after remaining stable on the active drug. But the real question is whether this agent is really beneficial enough and truly well tolerated compared to other options, either already available or still out there as investigational agents. That’s still a very open question.