0:02
So what happens when the cells are not happy or when you don't see that the cells are not happy?
0:09
Because if you measure offline the point here and you don't measure inline, you may miss one point.
0:17
And maybe in this.
0:18
Where the cells were not happy, they were not producing.
0:21
So you risk missing some titre.
0:24
And thanks to the fact that we provide sensors to basically the entire biopharma industry, we have the chance to get in touch with a lot of them.
0:33
Biopharma processes that can have a variation of 65% productivity.
0:39
If you compare that with any other industry, it's brutal.
0:44
I mean 60% less product really.
0:47
But in biopharma is accepted because we produce complex things with living cells and sometimes cell therapy.
0:53
The living cells are the product.
0:56
So make sure to monitor in real time.
1:01
Their viability can help us to understand if we have to do something.
1:04
Maybe we need to feed them and not wait once a day when we do the offline measurement to wait to understand that the cells were starving.
1:13
So to do that, first of all we need to know if the cells are happy, not happy.
1:20
And this is why we have introduced 2 technologies.
1:23
One is the Incyte - Impedance Spectroscopy, which measures the total cells, the viable cell density and then the density which is the total cell density.
1:32
If you put them together, you can calculate the viability.
1:36
So this one is our IR technology, and this one uses sensors that combine to measuring principles in transmissions and in reflectance.
1:45
So that means that independently from the sample thickness, the sensors function automatically and you can have the same accuracy within the entire calibration range.
1:53
And here are some examples of research.
1:55
This is coming from the University of Bielefeld in Germany where basically they were testing how well they were working against an offline method.
2:03
I hope that you see this lined dot and those technologies work seamlessly in fed batch as well as in perfusion processes.
2:12
So one technology that delivers real time measurements. Now we’ve just spoken about cell culture, but what about microcarriers?
2:23
We got someone was asking us, does it work on microcarriers?
2:26
Yes, it does.
2:28
So it's not if they are adherent or suspended cells, it's not a problem.
2:32
The viable cell density can be measured by impedance.
2:35
It's possible to use those technologies also in perfusion processes.
2:40
So I may ask now, how many of you are using perfusion?
2:47
That's much more than last time where it was just one person.
2:50
So it's going.
2:52
So it's good. We need to have technologies that works for that.
2:57
As you can imagine, having a sensor that works for one week is not like having a sensor that works for one month without drifting.
3:04
But we try.
3:05
It's part of the challenges that we solve with our technologies.
3:09
So that was the viable cell density.
3:11
Let's move to another critical process parameter.
3:13
I'm not talking about pH and oxygen because I think it's assumed that they must be measured.
3:19
But what about CO2 or dissolved CO2?
3:23
This is often overseen in cell culture and when it is overseen, here is research that comes from Roche.
3:34
They tested what happens when they don't control CO2 at all, when they control at the standard 5% and when they find out that for a specific cell strain it was better to have CO2 at 15%.
3:46
Seems strange, but for that cell that was the basically finding out what was the optimal CO2 target point caused the productivity to double.
3:58
The point was it had to be controlled in line and CO2 is rarely controlled in line.
4:06
Well, until a few years ago, there was not the right technology to do that and that's why we introduced that.
4:11
So we have introduced an optical sensor technology to measure very accurately CO2 in line.
4:19
And by doing that it's possible to achieve several things.
4:22
So it's possible for example, to avoid trouble starting the scale up.
4:29
How do you do the scale up even with the super-duper new technologies from Eppendorf or from Sartorius?
4:36
Probably they are going to work on the small scale, and they are going to just work on the KLA of the oxygen and scale it up.
4:44
But CO2 is not oxygen.
4:48
CO2 is 32 times heavier and stays in the bioreactor.
4:53
The problem is in a 2 litre bioreactor you can strip CO2 quite easily.
4:57
In a 12,000 or in 4000 litre bioreactors you can't.
5:01
CO2 accumulates and then you have to compensate with base.
5:06
And if you put more base than what was prevented into the small-scale bioreactor, the osmolality of your bioprocess goes up, osmolality goes up, viability goes down and productivity goes down.
5:19
And that's because you have not calculated the KLA of CO2 together with the one of oxygen.
5:26
And maybe you need even need to have a double sparge strategy to do that.
5:29
So I'm just dropping hints.
5:32
But if you do that, you can find out interesting things.
5:36
So like this example here, from the University of Northwest in Switzerland where they were really tried to see how it was important to monitor in plasmid and viral vector production, the accumulation of CO2. Especially in perfusion bioreactor where the CO2 level is going to be even higher than it was before.
5:59
So it has to do with scalability, it has to do with the better control of pH, CO2 is part of the pH controlling the cell culture process.
6:08
So at Hamilton, we like to believe we produce the best pH sensors, but the customer says can you have a pH sensor that can go up to 0.05 of pH because it's 0.5 more or less in the same process can help us drive our process.
6:27
Normally it's because you want to drive the internal intracellular pH and not the extracellular pH.
6:32
But with a normal pH sensor you can do that.
6:35
What you want to control is the CO2 that goes into the cells and changes the intracellular pH, and that you do with the CO2 sensor, not with the pH sensors.
6:45
Things that help you understand how to drive the process.
6:48
So more parameters, more information helps to better drive the process, especially if they are in real time.
6:57
Adding more information is one part of it.
6:59
You need to be capable of using that and sometimes having too much information is too much if you can't really use it.
7:06
So progressing digitalization.
7:08
Again, I'm just dropping some hints here.
7:12
Has anyone here heard about soft sensors and digital twins?
7:25
One that's good.
7:26
Again, it was 0 last year, so that's already, that's quite already good.
7:30
But the point is sometimes you cannot have a sensor for everything.
7:33
You want to combine signals to get this golden batch, and this is something that can be done. You need to have good information to get a good Golden Batch prediction. This prediction measure requires already having a sensor.
7:58
And if you have real time measure, maybe you can also do something else.
8:01
You can also do asset management, which I'm going to discuss a little bit in a few minutes, which is going to be good.
8:07
Now, soft sensor and digital twins, special digital twins or digital shadows depending if you use them to control your bioprocess, yes or no, they can be very important.
8:19
They are not always supplied.
8:23
It's because it's complex and they need to be recalibrated, especially if you want to go from R&D to production.
8:28
You need to do the entire validation of the GMP and for a soft sensor you need to revalidate everything.
8:34
If you just change a small part of it, the FDA probably is going to ask you to revalidate everything from scratch.
8:42
So maybe it's better to start with something easier.
8:46
We talked before about VCD, so what about instead of having just the permittivity signal, it's possible to have the viable cell count.
8:54
This is much easier than just developing a digital twin.
8:57
And this is something that we can provide and help you.
9:00
I mean with the permittivity impedance signal, you can also count the cells.
9:04
You can have even more information like the dimension of the cells, which is as important as the number to predict the productivity of the process.
9:14
This is something we can help you with and it is something that can be used into the sensor.
9:19
And now we’re almost at the end of this short presentation.
9:24
So more hard sensors, more soft sensors, this can all help to drive the process that we said before we go in direction of smart PAT.
9:34
But before I was saying something like assuming that the process, that the sensors and the models are working correctly, otherwise it's engineering bad data in or bad input in, bad data, bad, bad output or bad control out.
9:54
And again, I come back to the bio forum operation groups, the same one who delivered the metrics I showed before.
10:01
It's just a given.
10:02
But let's repeat it.
10:04
Ideally, sensors should have the capability to monitor their performances and predict when they will fail.
10:10
You need to trust your eyes and your ears if you want to it to work, right?
10:16
How do you do that?
10:17
Well, we thought instead of just providing the measurement in real time, we can provide in real time also, how well is the sensor measuring?
10:24
So how to what extent can you trust the measurement and everything that you can deliver with that?
10:30
So this is exactly what we were saying before.
10:33
So that's why we have developed the ARC technology, having a microtransmitter implemented in every sensor makes it much easier to validate it.
10:43
You have you occupy less space.
10:45
You can calibrate in the laboratory.
10:46
You don't need to calibrate that with the transmitter in your production process, not this smaller one.
10:52
I mean the bigger one where you have to hook up because you need to calibrate there.
10:57
This can be done directly in the laboratories and if you want you can have the diagnostic done in real time even through Bluetooth.
11:06
That's an add on.
11:08
It's not needed to run the sensor.
11:09
It's just to make the wireless maintenance and automated documentation which helps the with the valuation.
11:17
All of this helps us to go toward the direction of asset management.
11:22
Again, you don't just want to have real time measurement, you also want to know how well the sensor is measuring real time.
11:29
Because fed batch as well as perfusion processes you measure for one month.
11:34
Is the sensor drifting, you want to know that in real time and with current technology it is possible.
11:40
Actually we do our work with companies like Siemens to integrate our ARC technologies so that it's possible to do sensor fleet management.
11:52
Again, it’s possible to measure well and measure in real time and also know in real time how well are we measuring.
12:00
This also helps avoid planning downtime and it’s going to be faster.
12:05
So and by that we close.
12:08
So with the two pillars, so innovating real time measurement and progressing digitalisations, we come to the end.
12:15
We can deliver the metrics mentioned before.
12:19
And of course you have seen some examples.
12:21
We are working to bring even more parameter measured in real time to give more control and to give it in ARC format so that this is not just a guess, and this can become the reality.
