How AI Changes Breast Cancer Diagnostics, Recognizing Breast Density and Tumors on Mammography Fast and Reliable: Prof. Cristina Rossi, b-rayz

Prof. Cristina Rossi, b-rayz: 0:00

Our solution provides feedback within 30 seconds from the acquisition of the exam.

Alex Jani, X-Health.show: 0:05

[INTRO] Hi, I'm Alex and welcome to the X-Health.show where I talk to visionaries behind the latest innovations in healthcare. For the eXtra health of the future. We're meeting today at the University Hospital of Zürich, building in Stettbach, where hospitals IT is located. My guest today is Professor Cristina Rossi, the CEO of b-rayz, a physicist passionate about image data analyses. Her scientific career moved her from Rome, Italy via Tübingen, Germany to Zürich, Switzerland, where she embraced entrepreneurship. B-rayz is a Swiss startup that developed b-box, AI based platform supporting early detection of breast cancer. Born at the radiology department of the University Hospital of Zürich, for radiologists. You'll hear about mammography and mammography alternatives to screen for breast cancer, breast density tissue, AI algorithm training, the impact of a patient's stress, and more.[INTRO ENDS] I bet we all know someone who had the breast cancer. I myself remember my uncle's partner and this particular story, well, didn't end well. It basically ended way too quickly. We hear these stories everywhere. How prevalent it is, Cristina?

Prof. Cristina Rossi, b-rayz: 1:29

Well, breast cancer is the cancer type, which has the highest incidence and the highest mortality in the global female population. Statistics say that one in eight women will develop breast cancer.

Alex Jani, X-Health.show: 1:42

One in eight?

Prof. Cristina Rossi, b-rayz: 1:43

Exactly. During our lifetime. And nowadays, we still have over 600,000 women that die every year because of breast cancer. And I'm sure that these statistics are not comprehensive because they do not take into account regions where the healthcare system is missing the possibilities of tracking the statistics.

Alex Jani, X-Health.show: 2:11

How can we detect it? So what are the methods to screen for breast cancer?

Prof. Cristina Rossi, b-rayz: 2:17

Well, the breast cancer secondary prevention is the one that uses diagnostic modalities to detect the cancer at a very early stage. And one of the main modalities used in this context is mammography. Mammography is an X-ray based modality that makes images of breasts. Generally, there are two images per breast taken in different projection, so compressing the breasts in two different ways.

Alex Jani, X-Health.show: 2:48

You are mentioning that and I'm just cringing. Yeah, I remember, it's not a pleasant experience, is it?

Prof. Cristina Rossi, b-rayz: 2:53

It is not. There are also other

Alex Jani, X-Health.show: 2:53

No doubt about that. possibilities to get the examination without the compression. There is a lot going on in the topic of modalities for breast imaging. I personally had my third mammography. I didn't experience it as too bad, I have to say, I

Prof. Cristina Rossi, b-rayz: 3:07

Absolutely.

Alex Jani, X-Health.show: 3:08

I am thinking also about the know it's necessary. It's not the most pleasant thing in the world, but I know that this for a good reason and we need to do it. self-examination. We will not go deep into that, neither of us is a doctor, but I mean, you should do it too, right, plus...

Prof. Cristina Rossi, b-rayz: 3:35

...plus mammography, exactly. The limitation of the self-examination is that the self-examination can get a lump when the size is big enough to be felt by a person, even a person who is not well trained as a patient to detect some abnormalities. And mammography can get already the very early sign of breast cancer.

Alex Jani, X-Health.show: 3:59

What is then this very early sign? What is the difference – you can't touch it, right? You can't feel it?

Prof. Cristina Rossi, b-rayz: 4:07

You can touch it or feel – it depends very much on the position in which the cancer is developing. But there are also earlier signs of breast cancer, like the presence of calcifications, that indicates that something is going on in this special area of the breast. And for sure mammography has some limitations as a diagnostic modality, there is no diagnostic modality which is perfect in terms of sensitivity. So it gets everything that is there, or specificity that gives you an alert when it's necessary to get an alert. And we always need to have a compromise. But still, as of now, it is one of the best modalities that we can use for the early diagnosis of breast cancer. In breast cancer, indeed, getting the cancer early has a strong impact on the outcomes of the therapies. So women that get the therapy early enough, who have the possibility of choosing between different therapeutic options have better chances to survive. So early diagnosis is really key in the management of breast cancer.

Alex Jani, X-Health.show: 4:12

What can we do at this very early stage when, you were mentioning, there might not even be these big enough lumps to feel them?

Prof. Cristina Rossi, b-rayz: 5:32

Well, here the options are several and the breast radiologist together with a team of specialists decides on the individual therapeutical options for the single woman. And this is done generally in what is called tumor board, in which experts from different fields sit together and plan together the optimal therapy for the single patient. What we provide from our side is a support to the work of the radiologist in order to define, from the radiological point of view, what is the best strategy for the patient.

Alex Jani, X-Health.show: 6:11

And what are the current recommendations for how often should we do mammography?

Prof. Cristina Rossi, b-rayz: 6:17

This depends very much on the local healthcare system. There are some surveillance protocols, which are generally formulated by the different institutions, this could be the Krebs Liga, for example. And these recommendations are provided to the radiologist or to the oncologist and gynecologist, and they provide the guidance, how to manage the single case. This depends on the age of the patient, on the characteristics of the patient, for example, on the breast density, on the history of the patient. And this provides guidance in a very clear and structured way.

Alex Jani, X-Health.show: 7:01

I actually wanted to ask about, say, healthy population. Women at what age... I'm not sure but probably the age also was lowered in recommendations. And how often? A couple of years, or?

Prof. Cristina Rossi, b-rayz: 7:16

Yes, it depends very much on the country's possibilities to offer the services. Generally, screening programs, organized breast cancer screening programs, which are based on mammography, have a mammography every two years for women above 50 and till 69, I think in the most established breast cancer screening programs. Now there has been a change in this direction. So there are some countries that warn about the need of calling women earlier for breast cancer diagnostics and mammography. So there are some countries which have started calling women already at the age of 40, and again every two years. And here is where we believe AI can bring a substantial support in offering this service to a larger population of women because healthcare professionals specialize in breast cancer diagnostic and available on the job market is not enough to support this big volume of examinations, which are performed every year. Already now the numbers estimates for the number of mammographies performed per year globally is something like 270 million.

Alex Jani, X-Health.show: 7:28

Mammography is obviously the gold standard. Are there any alternatives or good alternatives?

Prof. Cristina Rossi, b-rayz: 8:45

To mammography? Yes, I have to say that especially for the last 10 years a lot of technological advances has taken place in breast cancer diagnostics. So for many years, there were no big alternatives to mammography. Nowadays, there are dedicated ultrasound devices for breast imaging. There are also ultrasound devices which use robotic arms to reduce the bias that may be introduced by the operator by pushing the ultrasound machine in a different way or in the interpretation or in the volume which is scanned during the ultrasound. There are also alternatives. One of the emerging ones is the breast CT which is a CT scan dedicated to the breast which is able to generate 3D volumes of the breast while the mammography is a projection technique. So we have a 2D image that projects the whole volume on a plane and there is also, especially in the USA quite popular, the use of what is called pseudo-3D modalities. So this is the tomosynthesis, for example, and the tomosynthesis provides nearly a 3D volume of the breast using the same acquisition device than the classical 2D mammography dedicated devices, which have been upgraded to provide this pseudo-3D acquisitions. So there are alternatives to mammography, it's important to discuss with the patient, what is the right modality and the right set of modalities? We may require different modalities for the single patient to get the necessary information.

Alex Jani, X-Health.show: 10:55

And what does it look like, the mammography? I'm also asking about what problems can like radiologists have with this? And actually we are slowly moving towards your solution. But first, I'd like us to identify what issues nowadays may radiologists face?

Prof. Cristina Rossi, b-rayz: 11:22

Yes, that's a good question. The mammography examination is aquired standardized examination. So let's talk about the 2D classical mammography. You have two projections per side – you take two pictures of the breasts in one picture, the breast is compressed from the head to the feet. This we call Craniocaudal. And in the other projection, the breast is compressed laterally, taking a part of the pectoralis muscle. And with these two images you want to cover most or ideally the whole breast tissue. What happens is that the performance and diagnostic quality of the image depends strictly on how the breast is positioned within the mammography device. So there are very strict guidelines on how mammography needs to be acquired.

Alex Jani, X-Health.show: 12:21

But I'm also thinking, there are so many breast sizes and shapes. How to standardize this?

Prof. Cristina Rossi, b-rayz: 12:29

This is a challenge and the quality criteria provide a framework within you can work as a professional. There is not only the breast size, there is also the chest anatomy, there is the cooperation of the patient – if she is sensitive to compression or if the patient is sitting on a wheelchair, so the mobility of the patient – the characteristics of the device, which could be more flexible or less flexible. There are many variables on the topic but there is a framework that provides guidelines to the professional. Once I assume the image is is acquired perfectly, then we need to look at the characteristic of the patient. As we said mammography may not be the sole modality you want to use for breast cancer early detection. In some women mammography needs to be supplemented with other modalities. And one of the characteristic of the single woman that affects the decision if you need a supplement diagnostic or not is the density of the breast. The breast density is a key factor that may triage women to get supplemented diagnostics or not. Why's that? Because having a breast dense of glandular tissue reduces the sensitivity of mammography.

Alex Jani, X-Health.show: 14:03

What does it mean, actually? This is what I wanted to ask, what does breast density mean for a woman and for a radiologist? Because these are two things also, right?

Prof. Cristina Rossi, b-rayz: 14:13

Yes, yes. I mean, having a dense breast doesn't mean anything bad, it is not a pathological sign to have a dense breast. But in terms of breast cancer diagnostic this has a fundamental effect on two points. On one point, having a dense breast makes the detection of breast cancer more difficult in mammography. Glandular tissue looks in mammography white – so, hyperintense with say. And breast cancer may also look bright. So if you have a very dense breast it's like looking for a snowball in a snowstorm. It is extremely difficult to get. And for this reason, radiologists often supplement mammography with breast ultrasound, for example, in case in which the density of the breast is high enough.

Alex Jani, X-Health.show: 15:09

They can actually see the snowball on the snow?

Prof. Cristina Rossi, b-rayz: 15:12

If you're trained enough – yes. And if you have a multiple examinations, for example, longitudinal examination of the patient, so you know she came already in the past for screening, then you can compare the two images. But sometimes even for very trained radiologists it is difficult to see a pathological change in the breast. And for this reason they recommend a breast ultrasound. Now, in order to get this recommendation most standardized the American College of Radiology has introduced four classes for the classification of the breast density. They are divided in A, B, C and D. A means that the breast has nearly zero glandular tissue, let's say, and D is a very dense breast, dense of glandular tissue. And the American College of Radiology has provided some recommendations in the handling of patients based on the breast density. They recommend for dense breasts, to supplement mammography with an ultrasound examination.

Alex Jani, X-Health.show: 16:20

How do we know that one has this dense breast? I mean, obviously, after mammography, but before mammography, can we detect that?

Prof. Cristina Rossi, b-rayz: 16:29

Well, there are different opinions on that. Experts, sonologists say that they can feel it from the physical examination of the breast. But for sure, they are not able, from the physical examination, to divide the breast density into four categories. So in order to be able to have a standardized classification, then you need to have a diagnostic method.

Alex Jani, X-Health.show: 16:54

And you know, I was actually googling that dense breasts and I found people asking, like, Can we do anything about breast density? Are there supplements for breast density? How to reduce breast density? Is this something anyone should worry about or should do anything about?

Prof. Cristina Rossi, b-rayz: 17:17

I think that it is important that a woman gets informed about breast density. As I said, having a high density of the breast doesn't mean having a pathological sign. It's a part of the natural aspect that the breast can have. But it has an influence on the screening strategy that you want to follow for breast cancer secondary prevention and it has also an association with breast cancer risks. So a dense breast has higher risk of developing breast cancer. And for this reason, also, in this case, the frequency of the screening may be adapted for woman which has a high risk of developing breast cancer.

Alex Jani, X-Health.show: 18:07

I can imagine, also for the radiologists, it's more difficult to find the snowball in the snow. So they can miss some tumors, smaller lesions. Are other any statistics about these missed tumors? I had a radiologist actually speaking about brain some episodes ago and he said that this is something really worrying. He also developed a platform to support radiologists – in finding these brain lesions. Yeah, what's it like with mammography

Prof. Cristina Rossi, b-rayz: 18:48

in breast cancer diagnostics? I think it's fair two separate two settings. The setting of the breast cancer screening, which is done in organized program is performed by healthcare professionals which are completely dedicated to breast imaging and by radiologists, which have high specialization training. In this case, the number of cancers which are missed is of course lower than in a different setting in which breast cancer diagnostic is performed maybe by radiologists, which are not fully dedicated to breast imaging – by a radiographer which works one day in mammography the other day in MRI, is not fully specialized and certified for breast imaging. Generally, there are some publications. The ones we refer to are the ones that claim that one over four detectable breast cancer lesions are actually missed in a clinical setting.

Alex Jani, X-Health.show: 19:50

One in four?

Prof. Cristina Rossi, b-rayz: 19:51

Yes. I'm a big fan of the screening program and I will say that this statistic does not apply to the screening program. We have a much lower rate in this case.

Alex Jani, X-Health.show: 20:06

I'll now ask about your solution to that. But before we get there, let's go to the radiology department of the University Hospital Zurich, when you met your co-founders before you became co-founders of b-rayz. So Andreas Boss and Alexander Ciritsis.

Prof. Cristina Rossi, b-rayz: 20:28

Well, the story is a special one. Indeed, we met at the radiology department in around 2017, I think when Alex Ciritsis moved to Zurich. I met Andreas already before. Indeed, Andreas says my husband, so we met already 2005. And I'm a physicist, Andreas is a radiologist so there were always some touching points among the things we were doing. But in 2017, I have to say that Andreas had the intuition of understanding that this was the right time and the right technology to create a digital platform dedicated for breast cancer diagnostics. And this intuition, which was realized through our work, work of everyone of the founders, and then the team, the first team at the beginning, came out really to be a great success commercially. And this allows us then to start the journey of b-rayz.

Alex Jani, X-Health.show: 21:37

Why it was the time? What was so special about the time to introduce the platform?

Prof. Cristina Rossi, b-rayz: 21:41

There were several points, actually. On one side, this revolutionary technology. For the first time a technology which was able to recognize patterns within

Alex Jani, X-Health.show: 21:53

And there are less and less of them. medical images in the same way as a radiologist would do. In the past, we always attempted to introduce a standardization or a support in the interpretation of the medical images. But technically, this was done in different ways. So we were formulating mathematical models and then assuming that the images would follow this mathematical model, and then we came out with some projections or some biomarkers, which were computed pixel per pixel within the image. And this was interesting from an academic point of view, but there is the big disadvantage that people are not looking at the image pixel per pixel. People are looking at the image per pattern and this technology was creating for the first time this capability of doing that and doing that with high performance, with time which are compatible with the clinical needs and the clinical requirements. So, technologically, this was the mature time to start. And from the point of view of the clinical needs, this was also the right time point, a time point in which medicine is becoming more and more controled, quality controled, structured, in which there is a dramatic need for highly specialized professionals, which are extremely expensive on the job market, which require a long time for education. Or there's more need for them.

Prof. Cristina Rossi, b-rayz: 23:33

Unfortunately, yes. So technologically, we are able to offer diagnostic opportunities to a large spectrum of patients, but we are limited by the human resources from the healthcare side. So there is a high demand with decreasing capability of offering this service.

Alex Jani, X-Health.show: 23:56

Let's just maybe say that machines are not going to replace humans because already there are humans missing in the field of radiology.

Prof. Cristina Rossi, b-rayz: 24:07

Absolutely. I understand the fear of professionals that think that machine may replace them. But actually, this is not the way in which these devices or these digital solutions, are engineered. No one wants to replace the human where they are, well, the human is not replaceable,

Alex Jani, X-Health.show: 24:28

Why a human is not replaceable?

Prof. Cristina Rossi, b-rayz: 24:30

Because history and the journey of the single patient is so complex that requires the coordination of this journey through a human person, a person which is able to understand the fear of the patient, that is able to understand the capability of the patient of complying with the therapy, of complying with indications which are performed. The machines are an instrument in the hands of these healthcare professionals that allow them to focus their strengths down where they are needed, and not fulfilling forms of performing repetitive tasks which can be automated by the machine.

Alex Jani, X-Health.show: 25:32

Are we speaking, Machine Learning or Artificial Intelligence, or both?

Prof. Cristina Rossi, b-rayz: 25:37

I think that the idea should be to use whatever technology is needed to deliver the value you want to deliver to your customers. Computer vision, machine learning, and natural language processing, generative AI – these are a big pool of tools you have that should be used in a meaningful way to deliver value to your customer.

Alex Jani, X-Health.show: 26:02

How did you start working on the idea?

Prof. Cristina Rossi, b-rayz: 26:05

As we say, we started with the intuition of Andreas Boss. Then we sat together with Alex and me, and we decided how we could approach this, this big ambitious project and what are the strengths within the team.

Alex Jani, X-Health.show: 26:25

You had a doctor, a data analyst – that's you, and?

Prof. Cristina Rossi, b-rayz: 26:30

Alex is also a data analysts and a software developer, AI specialists. So we decided that Andreas will be the medical adviser of the project, at the beginning it was a project, Alex, who would take care of the technical implementation, and I was the one which was responsible for finding money, basically, and at the beginning also supporting with the certification of the tool.

Alex Jani, X-Health.show: 26:59

All of you worked already at the hospital, you were looking to solve already existing problems in the hospital?

Prof. Cristina Rossi, b-rayz: 27:08

Yes, the idea was born directly from the clinical front. So what we have recognized was that there was a need of creating a system, which was a supporting part of the patient journey with the possibility of having a real digital platform built around this technology that would support the entire journey of the patient. So we decided to enter the journey of the patient from a special door, which was the door of the automation of the quality assurance of the positioning of the breast in the mammography device directly after the acquisition of the mammography. We did it because we understood that there are some steps which are performed during this diagnostic journey and this step to serve dedicated AI solutions that allow to reduce human error, to reduce the number of recalls of the patient, because there was a mistake in the image, that allow a different healthcare professional to work independently. So to reduce the number of consultations, but still guaranteeing high quality performance. And so this is the way we decided to enter this space and it turned out that actually this was not only the need of the University Hospital of Zurich but it was a general need, actually, a global need. And now we have customers in Europe that use our solutions, which are dedicated to specific aspects in the diagnostic journey of a woman.

Alex Jani, X-Health.show: 28:54

So you just started with what you just described and how did the idea develop? Because I also read that you screened 35,000 patients or images, and had eight clinical studies dedicated to your solution. So could you tell now about how the software or your solution developed? That's called B-box, right?

Prof. Cristina Rossi, b-rayz: 29:21

Exactly. I think meanwhile, we have for sure, more patients supported in clinical settings with our technology and a growing number of clinical studies published. We have been very blessed with good collaboration and with the trust of the first customers. How does it work? It works taking into account different aspects. First of all, the customer needs – you cannot build anything if this is answering a problem that no one has. So the first rule, ask the right questions to the customer. The second one, especially in healthcare, is the one of engineering a solution which is easy to be integrated. There are always several stakeholders, it is always difficult to change something in a setting which is extremely complex. It's difficult to change something if you are offering just a point solution. So people are ready to invest in your solution if they know that the solution is going to grow. And not that the solution is just solving a little piece of a long chain of tasks you need to perform during the day. And with the ultimate goal of delivering value, diagnostic value, quality of performance, standardization of decision, growing efficiency, reducing waiting time for the patient, reducing human horrors in the generation of these images, and providing personalized screening option for the single patient. And this is how the customers have seen also our solution. Now we were able also to present the solution to the first customers and this won the trust of a big large institution at the beginning and this was the good start for our fantastic story.

Alex Jani, X-Health.show: 31:24

[MID-ROLL] This episode is brought to you by the X-Health.show and me. So if you like this podcast, be generous hit Follow, leave a review. That'll help me invite more such amazing guests for you. Thank you. [MID-ROLL ENDS] This is actually surprising to me, because I was thinking introducing in structures like hospitals, huge and not movable too much, I was thinking that probably introducing something small for the beginning would work better. But you're saying they are actually looking for some more comprehensive solutions?

Prof. Cristina Rossi, b-rayz: 32:04

Yes, I think, actually, in this industry, there is a lot of movement going on. At the beginning, maybe five, six years ago, it was important to show that there were point solutions that were performing as humans are performing, nearly as good as humans. And this was the first, let's say, proof of concept, of embracement of this technology. But now that the trust has been won, people are looking for more and more comprehensive solutions. They want a dedicated a solution for a specific area in radiology at least but they want to have a solution that covers the complexity of the process. This is the feedback that we collect from our customers. And this matches perfectly with the ambition of b-rayz, which is the one of building a digital workspace completely dedicated to the needs of breast cancer diagnostics.

Alex Jani, X-Health.show: 33:08

Maybe you could say now, how does it work? How the b-box works and what does it mean for radiologists to implement it into their workflows? How difficult it is?

Prof. Cristina Rossi, b-rayz: 33:24

We are proud of having a plug and play solution. We took as a one of the requirements not to break the flow of data within the radiology department. We are a plugin that can be added and bring the diagnostic department in breast imaging to the next level. How it is done? It is done through a solution that is able to retrieve automatically the data from the different modality, elaborate the data and then reach the user there where the user needs to be reached. And having dedicated channels for different profiles of professionals in breast diagnostic. For example, in the radiology department, you have at least three players which are extremely important. One is the radiographer which is the tech person that is taking the picture – the first contact with the patient. Then you have the radiologists who is looking at the images and formulating a diagnostic report. And then you have the person who is often also a radiologist, who has the task or the responsibility of coordinating operations within the breast unit or among different breast units as it could be in the case of big radiological chains or in the case of breast cancer screening program. So the strength of b-rayz is the one of using this revolutionary technology to support the daily task of these professionals and to be smoothly integrated into their workflow. So to reach them, they're where they are and not to ask them to change anything in the setting in order to get the information from b-rayz.

Alex Jani, X-Health.show: 35:18

I'm just curious and I'm personally curious, I'm quite sure there will be some people in the audience also wondering how your AI or the machine can recognize better the lesions or changes in structure. So, basically be better in seeing through this breast densitym if it is on that high side.

Prof. Cristina Rossi, b-rayz: 35:45

We have a huge number of extremely well annotated data, which is fed into the algorithm. And we have also an architecture of the algorithm that take into account different tasks. So, generally, our AI is not only one algorithm. On one side, we have a huge amount of data which are extremely well annotated by specialists. When possible, we also refer to the medical report which has been reviewed by two radiologists. And the second one is to create AI algorithms, which are extremely specialized. So our solution is not relying on one AI algorithm. But there are different AI algorithms that work together and which are highly specialized for a single task. This allows us to control the performance of the software but also to achieve extremely high performance.

Alex Jani, X-Health.show: 36:45

So this is the current solution. What's the vision? Because you don't stop here?

Prof. Cristina Rossi, b-rayz: 36:51

Absolutely not. I think at the moment, we are just scratching the surface of what we can deliver with AI solution and embrace diagnostics. The first step as I said, and we have been extremely convinced since the beginning is the one of looking at breast cancer diagnostic as a complex process. A complex process that needs to fulfill different tasks that needs to involve different professional profiles. So the complexity of the platform, the idea of being the comprehensive platform. The technology is never stopping, the needs are growing. Every time we solve a problem, we create new possibilities, new opportunities coming from the radiologists, coming from the customers that need to be embraced in order to be in the system in the future. So a lot is to come. And we are proud of being an idea factory in this field and never stop the development of our solutions.

Alex Jani, X-Health.show: 38:03

I just realized we didn't speak about the results, the clinical test results. How good is AI in detecting lesions or tumors, or the breast density? Or actually assigning the classifications as it does it, too?

Prof. Cristina Rossi, b-rayz: 38:24

Yes, so there are several tasks. So each single task needs to be evaluated. At b-rayz, we have high standard for the evaluation of the clinical performance. This is based on the collaboration with clinics, which are partners of b-rayz, that allow us to challenge the performances of the AI algorithms.

Alex Jani, X-Health.show: 38:49

Can you name the clinics?

Prof. Cristina Rossi, b-rayz: 38:52

Yes, sure. We have a long collaboration, of course, with the University Hospital of Zurich, where we come from, but also with the University Hospital of Basel. We also collaborate with the Insel Spital in Bern. We have a very close collaboration with Cantonal Hospital of Arau and the Cantonal Hospital of Luzern, just to mention the Swiss sites. We have just started with a screening program with Kribs Liga Basel. So, we have the possibility of, just limiting now for the for the Swiss market, of challenging the use of our solution in different settings. As I said the screening setting but also the diagnostics setting, the university hospital or the cantonal hospital. So these are extremely complex settings, change very much in terms of population of patients which is screened or scanned, and availability of resources. And we want to add a solution which is flexible, which respond and adapt to the different needs.

Alex Jani, X-Health.show: 40:05

Is it as good as a radiologist?

Prof. Cristina Rossi, b-rayz: 40:07

b-rayz solution performed as good as a consensus of radiologists, which is by far better than a single radiologist.

Alex Jani, X-Health.show: 40:14

What's a consensus? Is it a couple of radiologists?

Prof. Cristina Rossi, b-rayz: 40:17

A couple of radiologists that are looking at the same images and that specialize in breast imaging. Because it's easy to outperform a human but it's not so easy to outperform specialists.

Alex Jani, X-Health.show: 40:28

And how long does it take for b-box to perform the task of identifying all the breast density categories and lesions, if there are any?

Prof. Cristina Rossi, b-rayz: 40:42

That's a good question. We wanted to have a solution, which is fully integrable into the clinical settings. And in order to be there, you need to be nearly real time. So our solution provides feedback within 30 seconds from the acquisition of the exam.

Alex Jani, X-Health.show: 40:59

That's amazing.

Prof. Cristina Rossi, b-rayz: 41:00

Yes, it is. We are super proud about that.

Alex Jani, X-Health.show: 41:04

So basically, it can be ready before the patient puts her wardrobe back on.

Prof. Cristina Rossi, b-rayz: 41:10

Everything needs to be ready when the patient is still there. I mean, if you need to retake an examination or a projection, because there was a mistake in the positioning of the breast, you want to do it when the patient is seated there. Because you want to avoid to recall her, to stress her with a thought that maybe there's something they don't want to tell me why I need to come back. Or with concerns to the breast density classification, you want to triage immediately the patient so that she doesn't need to come two times in order to provide all the diagnostic information which is needed for the radiologist to formulate a medical report.

Alex Jani, X-Health.show: 41:50

What are... I don't know if a mistake is a good word here, I'll use it anyway. What are mistakes to avoid in breast radiology? Or something that, I don't know, was used in the past and shouldn't be done? Or something that can be overlooked? Maybe overlooked is a better word here.

Prof. Cristina Rossi, b-rayz: 42:11

Well, I think for for a comprehensive answer, you should ask a breast radiologist.

Alex Jani, X-Health.show: 42:16

Fair enough.

Prof. Cristina Rossi, b-rayz: 42:17

I'm not a breast radiologist. But we see that there are some patterns which are recurring. For example, mistakes in the acquisition of the images – not the whole breast is depicted. Mistakes in the diagnostic strategy of the woman – dense breasts are not supplemented with further diagnostics. And, of course, detection – misclassification of lesion could be a problem.

Alex Jani, X-Health.show: 42:51

What I found on your blog, and which I really love that you also mentioned, how important communication with the patient is. And I mean, you mentioned that if a patient is anxious, which I didn't really thought about before, actually, the image quality may drop.

Prof. Cristina Rossi, b-rayz: 43:16

Yes.

Alex Jani, X-Health.show: 43:17

Why is that?

Prof. Cristina Rossi, b-rayz: 43:21

Everyone has expectations when you are entering a diagnostic room. Maybe you are coming already scared because you have a history of breast cancer in your family or because your friend told you that was extremely painful, this kind of examination. So these have an influence on your attitude while you're being subjected to these examination. And, of course, this has an effect on the outcome of the examination. It was funny, actually, because we were looking at the data of the quality of the examination in one center in Switzerland, and we were talking with the radiograph, who was responsible for the acquisition. I said, Look for the same kind of examination which was the craniocaudal projection, the one you compress the breasts from the head to the feet, I said You have worse performance for the right side than for the left side, how can it be? I would expect the same kind of mistakes. And she said, Oh, clearly this is the first side that we examine when the woman comes, she's still completely rigid. And every time we have the problem that this side gets worse.

Alex Jani, X-Health.show: 44:42

So she's what, trembling? Or what

Prof. Cristina Rossi, b-rayz: 44:45

Or maybe she's refusing more is it? compression, is complaining more, she is more rigid during the examination. This has an effect on the quality of the images. This was really extremely interesting actually, and should be taken into account in the definition of the protocol, on how these images are acquired.

Alex Jani, X-Health.show: 45:09

I can imagine five deep breaths before actually compressing the breasts.

Prof. Cristina Rossi, b-rayz: 45:14

Absolutely. But this will mean also to allocate more time for the patient to get prepared to the examination. And nowadays in a system which is already overloaded by requests and where you need, really, to make a lot of examination one after the other, is missing the time for this single person, for the approach to the single person, and this is a pitty.

Alex Jani, X-Health.show: 45:39

So actually, there's hope in AI.

Prof. Cristina Rossi, b-rayz: 45:41

Absolutely. Absolutely.

Alex Jani, X-Health.show: 45:43

I wanted to ask you now about the business side of b-rayz and when did you notice that things are accelerating? Because I can see you are hiring now and you've been doing that for a few months now. So I can expect things are accelerating, aren't they?

Prof. Cristina Rossi, b-rayz: 46:05

Yes. Well, first of all, we are super proud of the team we have at b-rayz. It is a fantastic team. But there are some figures who are missing and some aspects of the commercial activities which are growing, and require more support. This is a good sign, it means we are growing and we are growing in the right direction. It's a challenge from the management point of view, from the business development point of view, because at the beginning, you think, okay, all aspects, or all parts of this company will grow parallely once we scale activity. But actually, it's not completely true. So you in a startup, you will have pillars, which are growing faster, which need more resources than other pillars, and then the magic is a bit to find a compromise. Where do you start growing, or accelerating the growth of one pillar? When the time is right to it? And not when it is too late?

Alex Jani, X-Health.show: 47:14

Was there any moment when you noticed, like, Whoa, this is really... the rocket is starting?

Prof. Cristina Rossi, b-rayz: 47:23

I think every day I think that. When I walk to work and I think I'm really doing that. We are really doing that. Yes, I think, definitely, there is a change when you are a group of five, six persons, when you start to

Alex Jani, X-Health.show: 47:42

You closed your series A over a year being 10-15 persons, and then even bigger. Because the structure needs to change. Because the processes need to change. Because you have more expectations of people that come with the luggage of expectations and talent into this group, want to do great things, and you want to take them up on board so that they can deliver and bring you also to the next level. And I think this was in the last month, an extremely intense time for us, a time of scaling. And it's it's extremely exciting. ago and it was oversubscribed. How did you do that?

Prof. Cristina Rossi, b-rayz: 48:42

Well, we were lucky, we found the investors that understood the potential of the idea. And I think we were also able to communicate the potential of this idea. So they saw the big idea and the big ambition of the company. It was the time which was less challenging that nowadays. We know that now raising funds is extremely difficult for a startup. And I would like to encourage startups not to give up, do it, great idea will always get funded. But I have to say it's always a mix of good preparation, big luck and try to find the right opportunities, meet investors, go outside, talk about your idea and never give up. Because a door that closes can open bigger doors. So don't give up.

Alex Jani, X-Health.show: 49:46

This is what I wanted to acknowledge: the importance of luck, yes, but I can imagine it doesn't come without an effort. Well, maybe sometimes for some people does. Did it for you?

Prof. Cristina Rossi, b-rayz: 49:59

No, no, I think you need a good preparation. You need absolutely evidence that there is a commercial potential behind your idea. And I think that especially in digital health that is something that has changed as compared to eight years ago. I think eight to ten years ago, at that time, it was possible to raise funds just with an idea in your pocket and nothing else. Now investors are becoming more demanding. They want to have a CE marking even for your solution ready, when you raise money or they want to have already first commercial installations, which is quite a high level for a startup before you have certain amount of money.

Alex Jani, X-Health.show: 50:53

Definitely, you already need to have some investments to apply for CE marking and FDA also we're speaking here in the US, right.

Prof. Cristina Rossi, b-rayz: 51:01

Yes, yes.

Alex Jani, X-Health.show: 51:02

So these are costs or investments, well, whatever way you want to put it, but you need budget for it.

Prof. Cristina Rossi, b-rayz: 51:09

One possibility is to bootstrap the investment. And I think Switzerland is a good space for that. There are many excellent foundations, also Innosuisse – they support startups in this incubation phase. And then this needs to be a preparation to get to the professional investment, if this fits to the strategy of the startup.

Alex Jani, X-Health.show: 51:42

Was there any moment that you were considering quitting or anything real difficult?

Prof. Cristina Rossi, b-rayz: 51:50

I have to say, in the life of a startup, they are always ups and downs. You need to understand that this is a part of the game.

Alex Jani, X-Health.show: 51:59

How difficult was it or wasn't it difficult at all to understand that?

Prof. Cristina Rossi, b-rayz: 52:03

Ah, it's difficult because this is a work you do on your own and it depends very much on your personality. This was not a part of my personality, I have to say.

Alex Jani, X-Health.show: 52:14

You were a scientist, so you were hunting for grants. And then working on research, achieving higher and higher acknowledgement from the scientific community. And this is something different, right?

Prof. Cristina Rossi, b-rayz: 52:29

Yes, and I have to say talking to scientific audience was much more comfortable for me than going for a pitch in front of the investors. You need to leave your comfort zone. I think for us, it was good that there are so different personalities among the founders. I'm not the most optimistic one, I think Alex is more optimistic and he always sees something positive, even in the disasters. But I think it's good because sometimes you need a person who is realistic, sometimes you need that person who is optimistic. This journey of the entrepreneurial ship is changing you daily and you need to find your way along the journey as well. So understanding that this is a game, which has a lot of risks involved. You cannot control all these risks. You can do a great job, you need a bit of luck, also, but this is the part of the game and be happy with that.

Alex Jani, X-Health.show: 53:42

What's on this part that you can't control? I can imagine the investor environment now, for example, it's something you can't predict, right.

Prof. Cristina Rossi, b-rayz: 53:52

You can't predict that, you can't predict all the actions of the competitive landscape, you can't predict if in complex negotiations with an important client something changes internally within the team setting and a person who is a key person decides to leave the company, for example, for some things that are not related to you. This is this complexity in a process which depends on many variables that you cannot control. You can do great things, you can have a risk management plan, you can try to mitigate risk, to look for alternatives, but you need to also push forward and this means accepting that there is a residual risk that you will take with you in order to bring the company to the next level.

Alex Jani, X-Health.show: 54:51

I love it like you mentioned, you are in it alone. Many CEOs mentioned that actually, that you need to deal with that by yourself or well, you might have some ways to deal with that. Do you have anything special that helps you, I don't know, get up – to be resilient basically.

Prof. Cristina Rossi, b-rayz: 55:15

Well, I have to say I don't feel that alone. I think we try to have management which is taking into account that a single person cannot have all talents which are needed for bringing the company to success. So it's important to have complementary team in terms of personality, talents, life, time in which a person is feeling at the moment. But I think that what is extremely important is also to find a space where you can take a little bit of distance from this huge important project which is the startup. So it's really important also to recharge the batteries, to have a time in which you don't look at emails, that you have had an interest that allows you to recover and to keep your mind fresh enough to have the overview of the activities and the business development.

Alex Jani, X-Health.show: 56:16

And what do you do to recharge?

Prof. Cristina Rossi, b-rayz: 56:19

I try to spend time with my daughters, with my family – we go hiking, we cook something together.

Alex Jani, X-Health.show: 56:28

Here, it's like – I love it here, you just go to the mountains, and there is immediate reset.

Prof. Cristina Rossi, b-rayz: 56:34

Yes, that's true, that's fantastic.

Alex Jani, X-Health.show: 56:36

What's in the immediate horizon for b-rayz? You just said about the big vision but what's in the future in a few months from now?

Prof. Cristina Rossi, b-rayz: 56:50

First of all, the portfolio of our solution is constantly growing. We are now expecting the CE marking for the next generation of the b-box. And this will be a revolutionary technology. I cannot disclose more at this time point, so maybe we will have a second talk in a few months.

Alex Jani, X-Health.show: 57:06

Oh yeah, looking forward to that.

Prof. Cristina Rossi, b-rayz: 57:08

And we are expanding our commercial activities. We started with two key markets, the Swiss market and Norwegian one, and now we are expanding through Europe, which is a big exciting challenge for b-rayz.

Alex Jani, X-Health.show: 57:22

So where are you available and where theb-rayz or the b-box can be expected?

Prof. Cristina Rossi, b-rayz: 57:28

We are available in Germany, Spain, Italy, Netherlands, Norway, Switzerland, and more countries to come.

Alex Jani, X-Health.show: 57:38

Okay. So thank you very much, Cristina for this conversation.

Prof. Cristina Rossi, b-rayz: 57:44

Thank you, Alex.

Alex Jani, X-Health.show: 57:45

And you guys listening, if you want to, you know, follow b-rayz, please go to b-rayz.ch. Thanks very much.[POST-ROLL] I'm totally impressed by the

Prof. Cristina Rossi, b-rayz: 57:59

Thank you. audacity of researchers turned startup founders, doctors turned entrepreneurs or ordinary parents turned healthcare innovators. People battling the battles that no one fought before. For the eXtra health of the future. So if you see a startup posting on LinkedIn, show them some love, hit Like, comment, That's fabulous. If you have a couple of drops more of that altruism, follow the X-Health.show, leave a review here. I'll be able to bring more of these visionaries to you. So a big Thank you. You're awesome. See you next week.

SPEAKER: 58:32

The information in this podcast is for informational purposes only and should not be considered medical advice. If you have any medical questions, please consult your healthcare practitioner. The opinions on the show are Alex's or her guests. The podcast does not make any responsibility or warranties about guests statements or credibility. While the podcast makes every effort to ensure that the information shared is accurate, please let us know if you have any comments, suggestions or corrections.