Well, we’ve noticed in the conference this year and in recent years that personalised therapy is a big deal. There are a lot of breakthrough therapies but they only work in a minority of the patients. So if you are able to identify those minority of the patients who still win and then you can use combinations for the other people for other therapies but we need biomarkers. There aren’t that many FDA approved biomarkers to use for cancer so we’ve been on the hunt for other ones. Our client, Bioarray Genetics which is based out of Farmington Connecticut, has developed a list of 325 genes that came from normal breast cancer development. They have found that those genes are predictive of taxane based therapy response in breast cancer. In partnership with Rancho, we are trying to find if we can use them in other cancers. So what I did is I looked in breast cancer, colon cancer, lung, prostate, pancreatic, leukaemia, gastric cancer, I think that’s all of them, to see whether these genes have any power and, to my surprise, they’re expressed in all these different tissues even though they came out of breast but they’re expressed in lung at detectable levels, sometimes some at high levels. Some of them are tissue specific, not even for breast but tissue specific for the other types of cancer. Some of them are housekeeping genes, which is a big surprise, but housekeeping genes can be apparently predictive of response, that’s what we’re seeing.

Then the last one, we tried to compare tumour versus normal in gastric and ovarian and lung cancer and things like that. These genes can cleanly separate samples into those and we also did it for subtypes of lung cancer, breast cancer and colon cancer. Surprisingly in several of these different cases they are involved enough in the disease progression or mechanism that we can divide subtypes, even though they didn’t come from those tissues originally.

Why is it increasingly important to understand these biomarkers? 

Every now and then you see in several of the talks today there are great immunotherapy markers and you get an anecdotal response where, ‘I had complete response and it’s been years. I was supposed to be dead but I’m fine,’ and you think, ‘Wow, this is a miracle drug,’ but it didn’t work in 70% of the people. So that’s why we really need some way to predict and advance the 30% or 20% who the responders are going to be. That might be gene expression biomarkers, it could be mutations, it could be epigenetics, we don’t know yet, but the easiest ones to get to now are gene expression biomarkers.

So what are the next steps for these biomarkers?

The next steps would be can we predict outcomes in the way that they’ve been used for breast. We know about tumour normal and subtypes but really we want to know drug response. We do know drug response in breast but now we want to move on to find some good big public or private datasets that are lung cancer only or colon cancer only where complete response is known or other markers of clinical outcome - overall survival - and then see if the genes can be predictive in that capacity as well.

Is this an exploratory process?

In our case, in Rancho’s and Bioarray’s, we know that it works in breast and so the next steps there are to get it closer to external validation and then commercialisation. But in the other cancers it’s exploratory right now.