Predicting treatment success - depression

Professor Helen Mayberg introduces a series of studies to identify biomarkers for how a depressed patient will respond to treatment. The anterior cingulate is important in this regard.

[Predicting who responds to treatment is] one of the things we have been very interested in, because we’ve mapped out these areas over the last 20 years with brain imaging. It started with simple techniques using metabolism with PET scanning, where we inject small amounts of radioactive chemical to map the function in the brain to now using functional MRI scanning. We and others have identified that there are various combinations of these regions we’ve been speaking about. What we noticed back more than 10 years ago was that if we scanned patients when they presented with their illness and then they went on to be treated, that actually not everyone’s scan at baseline looked exactly the same. As we actually started to look at the data as a function of how the patients eventually did, we started to notice that certain brain regions were in different states in the scan and actually would predict how the patients eventually did. The regions that were most predictive of that were in a region called the anterior cingulate so that the anterior cingulate would be overactive in patients who would go on to do well but under active in patients who would go on to do poorly. We sort of filed that finding away, other people replicated it over time, and now we have a number of projects that are specifically looking at can we actually not just predict how you will do in a generic sense, but actually can we predict what treatment you will optimally respond to? So we have large studies now where we scan people who have never been treated, and we also look at genetics, immune markers. We take a number of different potential biomarkers and physiologic or clinical measurements. We then assign them randomly to either get drug or therapy. Then we follow them for 3 months and see how they did. Then we can look back at their initial measures and look to see is there a particular prototype with the biomarkers that will predict that you really would have done better with therapy than drug, or that you require drug, or that you actually don’t get better on either. This is very important because in our other experiments we are studying actually new treatments for patients who respond to nothing. If you have gone through many, many treatments, spent many, many years in a state of being ill, wouldn’t it be wonderful to actually, the first time you present, recognize that you may have a more serious or malignant form of the illness and that if we have really effective treatments for these most intractably ill patients, we can go straight to them instead of trying out things that take time, enhance suffering for our patient, where in fact we could have known it wasn’t going to work anyway.