Genomic predictors of treatment response in prostate cancer

Bookmark and Share
Published: 4 Jul 2017
Views: 1974
Dr Robert Bristow - Princess Margaret Cancer Centre, Toronto, Canada

Dr Bristow speaks with ecancer at WIN 2017 about trials ongoing with patients in the Canadian Prostate Cancer Genome Network (CPC-GENE) to determine patient subgroups most at risk of disease progression and drug resistance.

He notes the use of imaging to guide treatment and gauge response, and considers the significance of the abiraterone data from earlier this year at ASCO 2017.

What we were doing today, really, was describing the results from a five year study, the Canadian Prostate Cancer Genome Network, which was used to really try to triage patients into aggressive and non-aggressive groups for potential intensification of therapy, some novel trials. What we did, actually, was look at the inherent genetics of these patients using whole genome sequencing and really found that there are specific groups of patients in which their clinical parameters are exactly the same and yet have a defined genetic profile that show they have aggressive disease and that they will fail surgery or radiotherapy within eighteen months of that therapy. Those patients probably have occult metastases and they need better therapy than we’re currently giving them, they need better adjuvant therapies. So designing a test based on this kind of discovery approach for whole genome sequencing allows us to pull patients apart and start this type of intensification trial.

So a patient group [?? 1:01] patients within groups. So right now a hundred patients will come through the door, they’ll have similar Gleason scores, they’ll have similar PSA tests and similar staging investigations and they’ll be non-metastatic. Yet we know anywhere from 30-40% of those patients will fail local therapy but we can’t tell them up front a priori who they are. That’s really the basis of the Canadian study was to try to do that, to pull the 30 or 40 patients out of the rest of the group and say that those are the patients now we really have to put a microscope on and actually get them into novel trials because a local therapy alone is not doing them justice. Even within the 30-40% that are failing there’s a further group that fail quite, quite quickly and we actually think those are the patients most at risk for metastatic disease and developing castrate resistance. So actually acting earlier with those patients can prevent the onset of metastatic disease, prevent the onset of castration resistance and therefore change the cure numerator to the denominator in this group of patients.

Was there any confirmatory imaging used alongside the sequencing?

That’s a really good question. Our imaging is getting better and better in terms of looking at patients that might have metastases. We used to use traditional bone scans, for example, to define whether or not there were widespread metastases but now we have PET scans, particularly PSMA, that might be much more sensitive for them. So, in concert really with the genetic tests, the new trials going forward have to have best imaging as part of that triaging equation. So we’re not there yet, that’s for the next prospective trials to prospectively validate the genetic tests but in the era of best imaging which now really would include a PSMA scan.

How does this compare to the I-SPY breast cancer trial?

These trials are just about to be started and there are a number of individual trials worldwide. What I was saying is that these trials now are going to be based on more multi-modal indices so patients that not only have genetic instability, for example, but might also have hypoxia. So more complex signatures up front and then to actually look at those agents, for example, that might hit both hypoxia and genetic instability such as a PARP inhibitor or an anti-hypoxia approach. So the similarity is the idea of a biomarker driven set of trials in an otherwise localised disease where you’re trying to really find best adjuvant therapy. A slight difference here would be that the types of signatures that we are developing are really ones of complex phenotypes and complex tumour types so the actual signatures might be more complex.

Could abiraterone become part of these combinations?

Yes, absolutely, we can’t not use abiraterone now based on the data, so primary chemotherapy and abiraterone are probably drugs that definitely will be used in this high risk setting. Then the question, though, is are there specific tumour types in which you wouldn’t use abiraterone because of course you can get resistance to that drug. There was also another trial that was at ASCO comparing the first use of enzalutamide and then a cross-over to abiraterone at failure or starting with abiraterone and crossover and actually there was no difference between the two arms suggesting that, again, the sequencing of the drugs and who is going to benefit starting one or the other we really don’t know. So I would say that the approach to very high risk disease where we believe that the occult metastatic burden is high would include chemotherapy and/or abiraterone. The question is can we target a little bit further based on the biology.