Prevalence of homologous recombination deficiency among 53,000 tumour samples

Bookmark and Share
Published: 5 Jun 2017
Views: 2438
Dr Arielle Heeke - Georgetown Lombardi Comprehensive Cancer Center, Washington, USA

Dr Heeke speaks with ecancer at ASCO 2017 about the prevalence of haematology recombination deficiency among 53,000 tumour types.


What we wanted to do was figure out exactly how many tumours out there have somatic mutations within the homologous recombination pathway. We’ve known for quite some time that BRCA mutations are important in terms of the risk of developing cancer during your lifetime but also how you respond to specific therapies, notably platinum, chemotherapy and PARP inhibitors. BRCA mutation carriers actually perform much better and there’s also evidence that that’s true in folks with somatic mutations as well.

There is growing evidence that there are many additional DNA repair proteins that might possibly have similar clinical significance but we haven’t really well defined those proteins yet and we certainly don’t know how prevalent they are or if patients with these mutations will respond similarly to, again, platinum chemotherapy and PARP inhibitors. So we wanted to just broadly look out there across all tumour types to see how often we could identify these newer mutations.

We ultimately had access to over 53,000 tumours. We partnered with Caris Life Sciences which is a commercial company that can perform extended molecular profiling for any patient across the United States that has obviously a tumour biopsy. In partnering with them we got to analyse quite a large number, certainly the largest that anyone has ever done before. Ultimately we were able to identify that 13% of the tumours that we tested had one of these mutations. We looked quite broadly; there’s ongoing debate about which mutations should be considered, which genes are of importance but we ultimately studied about twenty different genes.

Going for all tumour types, finding all the mutations, that’s surely just like a pin at an infinite number of dartboards?

Yes, but that’s also what’s exciting. A lot of the work right now has been focussed on breast cancer, ovarian cancer, now extending into prostate, pancreatic, a little bit into head and neck. We really wanted to see is this more of an issue all across the board and are there even more patients that could potentially benefit from specific therapies or potentially benefit from more intensive screening. That’s certainly a long way away but just to start that more broader search.

There are a lot of clinical trials ongoing right now that are starting to pull in these bigger groups of mutations, [?? 2:30] recombination deficient tumours. Most of these trials are looking at PARP inhibitors, some are now starting to draw in immunotherapy but we still have a ways to go but that’s the ultimate goal.

How does this influence the possible design of future trials having these more narrow targets but it seems to be just about everything everywhere? But having at least some sense of where to start looking and possibly even coming up with new drug designs, new targets that way.

The first step, and what we’re also trying to do at Georgetown, is designing trials where you’re shooting broadly and capturing all these different either germline mutation carriers or with evidence of the mutation within the tumour itself with the hopes that giving them these targeted therapies, like PARP inhibition, will work. But we don’t really know for sure yet that more broadly this is going to work beyond BRCA. It might take a few years to really identify additional genes that are as impactful a player as BRCA. Even PALB2 we don’t have all the data on that gene either.

You could try and sum this up with maybe a conclusion, a summary, a final message?

The biggest take home point from our study is that homologous recombination mutations are quite common. We identify mutations in 13% of the 53,000 tumours evaluated. Other groups have also published on homologous recombination deficiency and have had similar percentages or even higher, so it certainly is something important. What was unique to our study too is we saw these mutations very broadly so you could have a non-small cell lung cancer patient, even though now they have lots of treatment options, but potentially even though you wouldn’t think of a PARP inhibitor for a lung cancer patient maybe if they have evidence of homologous recombination deficiency they might benefit from that treatment as well.

Then just to give folks a quick look at what we found in terms of the specific genes. We evaluated for PTEN, there is a lot of controversy if that’s truly a gene that’s important in the homologous recombination pathway. But it was our most commonly mutated gene in 5.8% and then BRCA2 and BRCA1 followed at about 2.5-3% and then ATM and PALB2 were our next most commonly mutated genes. In terms of which cancer lineages were the most commonly affected, ovarian was the most commonly affected lineage which we all know from previous literature. For us we got mutation rates in 14.1% of all the tumours tested. Surprisingly bladder was next at 9.7% and then breast followed at 8%. So that was also fitting with current literature.

It just sounds like such an awful amount of space to take up, just the sheer physical storage of that much information.

But that is the future, trying to get these clouds with all of our information. If we could all collaborate a little bit better in these clouds because there’s a lot of information out there that we’re all doing separately and if we can get it together maybe we can start answering these questions a little bit better.