New investigational targeted therapeutic, entrectinib, shows early promise
Dr Alexander Drilon - Memorial Sloan Kettering Cancer Center, New York, USA
On behalf of my co-investigators I’m very proud to present the combined updated results of two phase I studies of entrectinib, an oral pan-Trk, Ros1 and Alk inhibitor in TKI treatment naïve patients with advanced solid tumours harbouring gene rearrangements. These are my disclosures.
Earlier on Dr had mentioned tasty targets and that reminded me of the fantastic gumbo that I had yesterday. But I think beyond that it really resonates with me when we talk about recurrent gene rearrangements which are appealing targets for the small molecule world. So, as all of you know, recurrent gene fusions are oncogenic drivers across a variety of cancers. You’ll see relevant to today’s discussion on the upper left here you have a schematic of the TrkA, B or C wildtype protein. If I can call your attention to the lower left here you’ll see that in the wildtype state downstream signalling is dependent on the binding of ligand that causes hetero- or homodimerization, thereby activating the MAP kinase and PI3 kinase and phospholipase C-gamma pathways. In the rearranged state here you have the three prime end of the gene with the intact tyrosine kinase domain that hooks up or fuses with an upstream partner. What that really does is that it provides dimerization domains that then result in ligand independent dimerization and activation of these downstream growth pathways. It’s important to point out that these fusions are detectable in the clinic right now by a variety of different platforms including FISH, DNA based NGS, which many centres are very quickly moving towards, and also RNA-seq, specifically anchored multiplex PCR. Select fusions, including the five drivers that we’re going to talk about today, are clinically actionable and responses to targeted therapy can be very dramatic and also can be durable.
These five genes are NTRK1, 2 and 3, Ros1 and Alk fusions and they’re identified across more than thirty different histologies altogether. These, in terms of prevalence, can be grouped into two general baskets so first of all they can be lower prevalence events in more common cancers, here you’ll see spitzoid melanomas, cholangiocarcinomas, papillary thyroid carcinomas, and found at a lower incidence in other tumours such as non-small cell lung cancer, colon cancer, sarcoma, melanoma and primary brain tumours. In contrast, they can also be very high prevalence events in rare adult and paediatric cancers and here you’ll see that these tumours include mammary analogue secretory carcinoma or MASC, breast secretory carcinoma, congenital mesoblastic nephroma and infantile or congenital fibrosarcoma.
Here off to the right you’ll see a schematic of a patient from our institution with a mammary analogue secretory carcinoma, actually she was misdiagnosed initially because her tumour looked like a histotype called acinic cell carcinoma and she received several therapies for her cancer and then finally we got her tumour to next generation sequencing and then discovered an ETV6-NTRK3 rearrangement which is pathognomonic for most of these tumours that was complemented by a positive break apart FISH test seen here and over-expression via IHC.
When talking about how we might go for these targets, these five different drivers, it’s important that we have a discussion about CNS disease because of the fact that up to 20-40% of all cancer patients can develop brain metastases and specifically lung cancer, to pick out an example. Lung cancers which harbour Alk, Ros1 and NTRK fusions, up to half of cases along the natural history of treatment for their disease can develop brain metastases. But beyond mets to the brain we also have to consider patients with primary brain tumours. So you’ll see here that glioblastomas, astrocytomas, harbour NTRK1 or 2 fusions and paediatric cancers, cancers that we find in infants and children, these paediatric gliomas can harbour NTRK3 fusions. Obviously an optimal therapy would address both the extent of extracranial and intracranial disease.
Today we’re going to talk about a drug that does just that. Entrectinib, or RXDX-101, is a highly potent orally available ATP competitive tyrosine kinase inhibitor. It has low to sub-nanomolar efficacy against the five drivers we discussed; you’ll see the IC50s off to the right. This drug is active in vitro and in vivo in a wide variety of models including NTRK1, 2 and 3, Ros1 and Alk containing models. Specifically it was designed to cross the blood-brain barrier. You’ll see preclinical data here with good blood-brain barrier penetration in three mammalian species but also we’ll show you clinical responses later on.
To really understand what data we’re presenting today I think it’s important to very briefly summarise the phase I development of entrectinib. The first trial that was undertaken was the ALKA-372-001 trial initiated in October of 2012 by Nerviano, and later Ignyta assumed responsibility in November of the following year, in the dose escalation phase this had explored both intermittent and continuous dosing and took patients as long as they had a genomic alteration involving NTRK, Ros1 or Alk, so both mutations and amplification or rearrangement events.
The second trial is the STARTRK-1 trial, a very catchy name, that explored continuous dosing which was subsequently determined to be the most optimal way of giving this drug, a similar population initiated in July of 2014, and based on data from both trials the recommended phase II dose of 600mg daily orally was recommended. Today we’re going to talk about that experience in its entirety, so 119 patients. Later today I’m going to talk about the updated safety data in all of these patients and for the purposes of this talk I’m going to focus on the efficacy data with a data cut-off of March of this year.
This is the waterfall plot of best response via RECIST 1.1 in all patients with an NTRK, Ros1 or Alk rearranged tumour. These are patients specifically that had not received tyrosine kinase inhibition before. The reason we’re focussing on these patients is that essentially these were patients that had the most meaningful responses on trial. In fact, outside of this population the only other patient that had a response was a neuroblastoma patient with an Alk mutation and no responses were seen in the rest of the population. You see here that approximately 80% of all patients had a response to therapy. If you pull out specifically NTRK all of the cases had disease shrinkage and a confirmed partial response with treatment. For Ros1 it was 86% of cases and for Alk it was 57% of cases. There was one additional NTRK rearranged astrocytoma patient that had stable disease via RECIST which, I should point out, is not a validated criteria for the assessment of brain tumours and so the site actually went on to do exploratory 3D volumetric assessment and did find 45% shrinkage of that patient’s tumour.
Moving on, we here look at the swimmer plot of all 25 patients, including the astrocytoma patient, and maybe to call your attention first to the black diamonds which essentially mark time to response. The first take-home point here is that responses to entrectinib were brisk and the majority of them were really seen at the first time point when patients were imaged within the first 4-8 weeks. The second point here is that responses were durable so you’ll note that the majority of patients have responses that are ongoing, half of patients had exceeded six months with ongoing responses and if you look at that top bar that’s a Ros1 rearranged lung cancer patient whose response is approaching two years and three months.
We’re going to show you a selection of these responses today. This was a 46 year old male with metastatic SQSTM1-NTRK1 rearranged non-small cell lung carcinoma diagnosed in November of 2013 with widely metastatic disease, very heavily pre-treated. Until finally, again, just like that MASC case I presented earlier, an NTRK1 rearrangement was found by anchored multiplex PCR. The patient at that point was actually on hospice and was doing extremely poorly on supplemental oxygen. Because this fusion was found and we thought that this was a very clinically actionable driver the patient was enrolled onto STARTRK-1. So these are the baseline scans of that patient, you’ll see very extensive disease in the left lung. Within a few weeks the patient had a dramatic clinical response to therapy and you’ll see here the accompanying radiologic response. At day 26 there was almost a 50% reduction in tumour burden and now approaching a year, so day 317, further response to targeted therapy, approximately 80% shrinkage of total tumour burden. This extracranial response was accompanied by a substantial intracranial response. At baseline this patient had 15-20 brain metastases, you see two of them here; on day 26 the patient had a complete response to therapy and, again, at day 317 a continued complete response to therapy with entrectinib. And the patient remains on therapy and is progression free at more than twelve months, so over a year.
This last case serves to complement the cases that we’re presenting at today’s talk. This was, I should highlight, a patient that was treated outside of the STARTRK-1 and the ALKA trials. This was a baby boy, 20 months old, who was diagnosed with metastatic infantile fibrosarcoma harbouring an ETV6-NTRK3 rearrangement presented at birth with a large left leg mass. Unfortunately it required through the knee amputation and thereafter the cancer came back requiring chemotherapy and resection in the brain and despite those measures recurrent CNS disease was noted in the right frontal and temporal lobes as well as leptomeningeal involvement. This baby boy started treatment approximately two months ago. You’ll see here the baseline scans which show very massive peritumoral oedema, midline shift, transtentorial herniation. This baby was progressively lethargic and actually the treating doctor had commented that death was likely imminent. He was very quickly put on this targeted therapy and with just 35 days of treatment there’s a substantial decrease in tumour and oedema with normalisation of the midline. We’re very happy to report that this baby is now back to eating and crawling and is at baseline in terms of his alertness.
So just to bring things to a close, this is our summary of the efficacy of entrectinib in NTRK rearranged cancers. It’s a very potent TrkA, B and C inhibitor, huge safety experience, 119 patients. Responses can be very rapid and durable. Responses were actually seen in all of the patients I had mentioned earlier, which include colorectal, primary brain tumour, astrocytoma, the fibrosarcoma case I showed you, the lung cancer case which you saw, and mammary analogue secretory carcinoma. And dramatic intracranial activity, as you’ve seen, has been demonstrated both in a primary brain tumour and also in metastatic disease.
So conclusions. Entrectinib is safe and well tolerated. I didn’t talk specifically about the safety data, you’ll have to come to the session at 4.15 today to see those results. It is an active targeted therapy for patients with NTRK, Ros1 and Alk rearranged cancers, specifically patients who are TKI treatment naïve. Most importantly, entrectinib is highly CNS penetrant which is a valuable characteristic that oncologists look for in terms of systemic therapy. There were durable responses in both primary brain tumours and metastatic disease and you saw the complete response in the CNS. Now the drug continues to be explored on the STARTRK-2 trial, so it’s an open label multicentre global phase II basket study of the drug for patients in that exact same population, so TKI naïve, NTRK1, 2, 3, Ros1 or Alk gene rearrangements. Thank you very much.
Thank you for that very interesting presentation. The take-home point to me here is that these gene fusions which are drivers of the malignant properties in certain cancers are really good targets for therapy. So again going back to that fragile strength analogy, these are the things that make these cancers bad but they are also the potential targets that can be disabled in order to render the cancers susceptible to therapies. I want to emphasise, this is the product of many decades of very sophisticated research that was conducted to understand the genetic abnormalities that drive human cancers. What we’re seeing here is this work was basic, it wasn’t clearly going to lead anywhere when it started, it was persistent, it was expensive, it was hard to get done and is exactly the kind of work that would have been presented at the AACR over the years. And here we are, witnessing what I would consider to be a bit of a glory moment, showing that we were able to benefit, all of us benefit, from the patient research that was done leading to exciting clinical results such as the ones you just heard.
I think one other point that is important here is that you’re seeing a series of clinical trials that are described that aren’t necessarily targeting people with a particular cancer but rather people who have cancers characterised by particular molecular abnormalities. It’s not always going to be possible for all patients with all cancers because there aren’t always these kinds of drivers. However, where you have these drivers the proper thing to do is not to worry about whether it works in a given disease but rather whether it works for people with that particular set of abnormalities.
So I have one question, just to kick things off, which is that the durability of the benefit seems to be reasonably promising with this treatment but there are people who are relapsing. Do you have any idea of what’s happening in the tumours molecularly when there is a relapse? Are they continuing to have these fusions or are they lost or are they selected out? Are there other pathways being utilised? It seems to be important to understand that.
Yes, absolutely. So we know the answer to that in a few patients; we’ve sequenced their baseline tumours and also tumours at progression. So, talking about NTRK, like we’ve seen for Alk and Ros1 rearranged cancers, for these fusions in general there’s this emerging notion that the solvent-front mutations, different from the gate-keeper mutations which you might expect in CML or EGFR sensitising mutations, which are actually the main mediators of acquired resistance. So we published a case of a MASC patient who progressed on therapy after ten months and subsequently acquired NTRK3-G623R solvent-front mutation which was a paralogue of some of the other solvent fronts seen in the other genes. But to address one of the points you had brought up, the fusion is maintained so it turns out these are truncal events that are carried forward despite the acquisition of a new mutation.
OK, thank you.
We have time for two questions. Go ahead.
Alex from MedScape. Just a question – you indicated that this was the total experience of 119 patients. We have data for 24 patients who have the best response. Do we not take into account the other 95 patients? They don’t show a response? I’m just a little… I just need some clarification on that.
I think experiences like this are very informative for doctors in the community. Essentially we have to learn where these targeted therapies are most likely to work and we’ve been taught by this experience that if you have a mutation in any of these five genes or amplification that the likelihood of response is low, in fact there were no responses that were seen on this trial, granted people were treated at different dose levels. So this has really taught us to pick out the patients who are very well poised to derive benefit from a targeted therapeutic. So we’re not minimising those patients, in fact they also teach us about the safety of the drug. But the reason we chose to take forward this population in the phase II trial is for the reasons I mentioned.