0:01
Thank you so much for sticking with me.
0:03
I'm realise I'm the one between you and the drinks and I think this is the hottest room in the whole building.
0:09
So I'm pretty warm.
0:12
So the title of the talk is here.
0:16
You're all familiar with these numbers.
0:18
Usually they are on different slides or different presentations, but I think they do belong together.
0:23
There's a terrible lack of accuracy both in the clinical care of cancer and also in the drug development.
0:30
And I believe, or many believe that these things are related.
0:36
Coming to the end of the day, we can try to be a little philosophical.
0:40
Why is this the case?
0:43
And first of all, it's like the little choice, unhappy families.
0:49
All the cancers, they are very unique in its own way.
0:53
They present an enormous functional diversity represented with a heat map there where you see patients and drugs.
0:59
They, there's so much of variation.
1:03
We use very simplified model systems, cell lines and mice that are inbred, which makes it difficult to bring things forward.
1:14
And also maybe another problem is that the thing we are doing is inherently very difficult to predict.
1:22
There was this Netflix hit series, or this book called The Three Body Problem.
1:27
But the premise is that if you have this is physics, but if you have a planetary systems with just one sun and one plant orbiting, you can predict this system for the next 1 billion years, how things are going to evolve.
1:42
If you have three bodies orbiting each other, you can't really say anything about the next year or yeah, that's the premise of this book.
1:51
And also it's a familiar concept of physics.
1:56
So what's the solution to this?
1:58
If something is very hard to predict, maybe you can measure instead.
2:02
So what we do, and many others are starting to do is using more complex model systems.
2:09
So here there are cancer cells that are supposed to grow, but what you can see those of you that are familiar with cell lines, cell lines are usually very homogeneous when they grow with them form nice small layers.
2:25
While we here see some quite advanced structures that actually reminiscence the micro architecture of the originating tumour.
2:37
So we have 3D structures of cells organising cell types.
2:40
So they're all coming from one single stem cell, but they form the same kind of complexity that we see in tissues when we stain them for instance with antibodies that serve as markers for different kind of functionality.
2:54
[Unclear], it's also cytokeratin, which says something about the epithelial origin.
3:03
It's been mentioned here earlier today FDA is very keen on moving on to more advanced model systems through this FDA's modernization act.
3:15
But these kind of model systems, tumour rates are organised, they are precise, predictive, scalable and could be cost effective in terms of cancer drug development.
3:27
Our company, it has basically 2 legs to stand on.
3:30
1st is that we want to push this into clinical use.
3:33
So for each cancer patient coming into the clinic, we want to grow the cancer cells and come back to the clinician with a report on what kind of drugs should use.
3:41
But this is quite a long pathway for a startup.
3:44
So in the same time, we're pushing this, all the same technology, all the same tools for drug development, but we're not the only one.
3:53
Some cut out screenshots from the right here.
3:58
You see that Merck here is all on board.
3:59
You see that Roche is all on board.
4:02
And there are also small biotechs in Sweden that has put on us on board, helping them generate data to support their drug development projects.
4:12
The company on the far [right], Isofol, they raised €10 million on Friday based on part of the data that we have generated in this project that started one year ago.
4:25
So our technology platform we believe is clinically useful for drug testing, both for the individual patient but also for drug developers alike.
4:35
And we're going to document this with a clinical trial that start now in August where we're going to recruit 75 patients, not me, the clinicians, but we are going to do the testing, and the clinicians are going to treat selected drug based on this kind of technology.
4:50
We have a number of clinical studies ongoing, one that's completed just before Christmas where we are wrapping up the paper showing the kind of data we get out of these efforts.
5:00
But we also have pancreatic cancer that is ongoing from Sahlgrenska in Sweden and we have colorectal cancer in Oslo in Norway.
5:13
So this is a heat map.
5:15
You're all seen it's patients on the X axis, and it’s drugs on the Y axis.
5:19
And the colours show the heterogeneity and drug response.
5:23
For instance, here we have a block, two patients that are sensitive to a lot of different chemotherapies.
5:33
Illustrated here with epirubicin.
5:34
Epirubicin is not the standard drug in this clinical setting.
5:38
Maybe this patient could fare better if they get this drug.
5:42
It's a cheap common chemotherapy.
5:46
Everybody has worked with cell lines.
5:49
I worked with cell lines.
5:50
So this is the technology that is supposed to supplement or and actually maybe replace cell lines.
5:58
And what you see here is a 2D plot where we plot the independency, how dependent are the samples of having EGF when we grow them and how sensitive are they when we put MEK inhibitors in the medium.
6:12
And what we see is that there is a very nice correlation within this disease.
6:16
Most samples, they live along this axis, except these groups of samples here that behave like totally different animals, basically because they're wired very differently.
6:26
In some of these cases, we know why this is the case.
6:30
One of them, for instance, KM12.
6:33
It's driven by NTRK fusion.
6:37
The one on the far right there is driven by HER2 amplification.
6:43
So the thing is that you are able to pull a lot out, a lot of biology from these kind of assays.
6:51
All of these are published data that we hope to wrap up within this year.
6:54
From these patient derived models, it's very easy to measure things in culture.
7:02
The thing is how clinically relevant are these data?
7:05
And one thing we show here is a very strong correspondence between the RAS state status, whether we have KRAS or NRAS or BRAF mutation and the response to EGF or anti HFR antibodies.
7:19
You see that near all the mutations red they're above the dot horizontal dashed line.
7:26
That means that they don't really depend on the EGF in the media.
7:30
You have one single model which is red and that's KRAS mutated and actually need EGF and that's because it has a very different mutation.
7:39
So not all mutations are created equal, which adds to the complexity when you're deciding clinical trials or when you're going to select the patient for work, for models, for your testing.
7:53
So here you see that for instance, G12D, which is a very powerful KRAS mutant has a higher activity than the G13D.
8:06
The differences are not very big, but they are sufficient to that some believe are actually able to translate into clinical differences outcome.
8:14
There are people that or groups that suggest that KRAS G13D mutants should have anti EGF treatment therapy.
8:26
This is the new kid on the block.
8:29
When I went to conferences maybe 20 years ago, KRAS was Death Star.
8:33
No one could touch it.
8:34
It was like undruggable.
8:36
But now you're like have this long tail of different KRAS inhibitors entering the clinic.
8:41
And one of them is this really selective one. And you see that it's really specific to the mutant allele that is supposed to target.
8:57
What we want to find is basically therapies that work in patients, therapists that they wouldn't originally get, but that we can offer them based on the testing that we do in vitro.
9:08
And what you see here on the left is a dose response curves for 70 different models where you have one red sample indicated red, which is extremely sensitive to a proteasome inhibitor, which is approved for multiple myeloma.
9:25
The median IC50 is very low, 2 nanomolar.
9:31
No, the minimum I mean, is 0.5 nanomolar.
9:34
So it's a complete outlier when it comes in terms of sensitivity to this specific drug, which just suggests that these patients should probably get the chance to get this drug.
9:45
And this is the thing we're going to test in this clinical trial.
9:50
So to summarise, patient derived tumoroids, they are coming from patients and then we can generate data from them within weeks.
9:59
They recapitulate known pharmacogenomic associations.
10:03
Regulatory authorities and big pharma are convinced that this is actually going to improve the drug development process.
10:12
We are able to identify novel biomarker-less, biomarker-less vulnerabilities, but sensitivities that you are not able to actually, regardless of what kind of genomic data you have or kind of sequencing you do, you couldn't pull out this proteasome inhibitor, a sensitive sample.
10:34
What we have now is that the platform to actually screen drug candidates, we can do it timely.
10:40
And the data would be useful selecting disease candidate drugs, disease subgroups and patients subgroups.
10:50
So thank you for your attention and I think you will enjoy the refreshments.
10:55
But these are the things that we want to contribute to the drug development community.
11:01
So thank you.
11:02
Thank you so much.