Monday, October 6, 2014

Memantine and WBRT

After the whole brain radiation review that my collaborators and I just published came out, my friend and colleague, Jack West, put a nice post up on his website Cancer Grace about it and asked a follow up question on twitter:

He was referring to the recent Radiation Therapy Oncology Group's trial of memantine (a drug for dementia) given during whole brain radiation: RTOG0614.  So, to answer, I polled a couple of my friends who do nothing but neuro-oncology and reread the results of the trial (not out in published form yet, just as an abstract (paper #1 on this page) and the full talk from ASTRO).

As a quick summary: Whole brain radiation therapy (WBRT) is a treatment given to patients who have metastatic cancer in their brains.  There are a number of situations in which WBRT is given and it improves survival and significantly improves the life of patients.  On the downside, it has been shown to cause cognitive decline with as many as 60% of patients exhibiting measurable decline at 4 months after WBRT.  So, to combat this, a trial of memantine during and immediately after WBRT, an NMDA-receptor blocker used to treat Alzheimer's dementia was proposed and carried out.

In this trial, about 500 patients were enrolled, stratified by their RTOG brain metastasis RPA class (they only enrolled class 1 and 2 patients) and given either 20mg of memantine or placebo daily for 24 weeks.  A lot of their patients weren't able to be properly analyzed because of issues with survival (sadly, to be expected in this population), but those who were evaluable (~150) took a battery of 6 different cognitive tests.  The primary endpoint was performance on a specific test at the 24 week point, the e Hopkins Verbal Learning Test-Revised Delayed Recall (HVLT-R DR) and there was NOT a significant difference in the results of this (but it 'teetered on the edge of significance', with a p-value of 0.059).  Now, don't get me started on p-values.  Oops, too late.

The all holy p<0.05 is an arbitrary cutoff level by which we determine 'significance'.  I say again, arbitrary.  What it has become, I fear, is a gold-standard for a 'positive study'.  So in the case of this one, which technically did not meet its primary end point, many people are not swayed, because p wasn't less than or equal to 0.05.  the secondary endpoint, cognitive decline (a measure using several other of the tests) was met with p=0.01 and, crucially, there were no difference in side effects or survival. My *suspicion* is that the primary endpoint will be met if they can accrue (and analyze) another 50 patients, and this debate will end. However, until then, the jury is out.

Personally though, even though it failed to meet its primary end point, I will be recommending memantine to patients in a favorable RPA class getting WBRT, or at the very least having this discussion with my patient.

As an aside, the two Neuro specialists I polled had opposite answers, so it is fair to say that this remains in the undetermined category. But, the drug has been shown to be safe, and now *likely* efficacious. In a situation where we don't have other options, I'm sold.