We use analysis of cell free DNA routinely now in oncology care. In our lung cancer patients when we’re looking for genotype we send off a blood test, we look for EGFR or resistance mutation and we can target that. What we’re looking at here is can we use that same method of looking at free DNA instead of for genotyping but for cancer detection. And that requires a different focus – you have to look broadly for any cancer signal but maybe in doing so you can fill an untapped need. We know that lung cancer screening can make a difference, we know that lung cancer detection is important but CT-based screening is not widely adopted. Is this an untapped opportunity that we can leverage? So that is the motivation behind what was launched, the CCGA study.
This is a huge effort, 15,000 individuals will be enrolled, currently over 12,000, 70% with cancer, 30% not cancer and the simple question is can you find the cancers without finding false positives in your non-cancer patients. In this report we focus in on 181 lung cancer subjects and about 560 non-cancer participants and we’re using extensive sequencing to look for cancer signal. This is three sequencing methods – a deep targeted approach, 507 genes, 60,000x depth of coverage, and then two genome-wide approaches - genome wide sequencing of cell-free DNA and genome wide bisulfite or methylation analysis of cell free DNA. These approaches have not been published before. They’re extremely comprehensive and in doing so you find every genomic feature that could be associated with cancer. But at the same time you sequence the white blood cells and the white blood cells, actually, are rich with mutations, clonal hematopoiesis. These mutations shed into the blood and they pollute the DNA with all this artefact, all these mutations that you think might be cancer. But by winnowing out that signal from the white cells you can actually push the false positive rate down. Here we focussed on a 2% false positive rate, 98% specificity, and looked at the detection rate of these assays with that high specificity.
We all want perfect specificity but the truth is that you’re not going to get 100% specificity in relatively comparable non-cancer patients because these non-cancer patients are going to get cancer. We already know that some of our false positives, with follow-up, have actually gone on to develop cancer. So we’re still following those patients. But right now for this early effort we’re tweaking up the specificity to 98% and at that point we can detect about 50% of early stage curable lung cancers and about 90% of advanced lung cancers. You can detect cancers across histologies, in smokers and in non-smokers, symptomatic and screening cancers, it really has a broad potential for lung cancer detection.
That was one of the main take-aways from the question session at the end was that this has potential to be a blood test but it’s not a blood test yet.
So, let’s think about the diagnostic development here. This is a discovery effort; we’ve figured out that within the genome, if it’s done right, there’s rich signal which can point to cancer. Now we focus these efforts into a diagnostic. It’s not like these three sequencing approaches are going to be used, these will be refined and focussed using the thousands of patients left in this CCGA study. We have all those specimens and so now we further optimise, focus the assay and turn it into a diagnostic. And when that is validated it then needs to be studied as a utility tool – can it make a difference in at risk individuals. So right now there’s a study that’s currently enrolling, the STRIVE study which is looking at a breast cancer screening population in this population that is at risk of breast cancer and other cancers. Can we find those cancers with a blood test earlier than without? So that’s one kind of utility study but those eventual utility analyses are needed for us to really understand how this method can make an impact and those are planned or ongoing.
Any idea when there might be these further refinements that you’ve mentioned? Any planned follow-up that we would be expecting in the next few years?
People want to know what’s the timeframe here and I guess the way I think about this is two years ago I did not believe this at all was feasible, I was a total sceptic. Plasma genotyping I believed, cancer detection I don’t think so. And yet here we are and this is real progress, this is real proof of principle. So we’re making a lot of headway, there are a lot of specimens left. Patients are enrolling into these trials enthusiastically, doctors are enrolling enthusiastically. There’s so much enthusiasm about this kind of approach that I think these trials, even though they’re going to be large, are very much feasible based upon this proof of principle. So I don’t know when we’re going to get there but we are making major headway and this feels like it’s within reach.