On Leveraging AI to Overcome Radiology Challenges - Women's Health Series with Élise Mekkaoui

Episode Transcript

[00:00:11] Claire: Is the health care system failing women? What do you think?

Welcome to a brand new episode of Narratives of Purpose. This is a five episode special series which is dedicated to female founders advancing women's health, featuring my conversations with the Voices for Change panellists at the first AI for Life Summit, which took place in Geneva, Switzerland in December 2023.

The Importance of Gender in Personalised Medicine

[00:00:39] Élise: Yeah, I think we have to bring awareness for women's health to investors mainly to pharmaceutical companies for example and to say that we have to invest in it and that personalised medicine, for example, just bring one parameter, which is gender. It's super important because it changes everything in medicine.

There is other parameters, and maybe we don't know some parameters, but we will find out maybe one day. But if we don't start with that, it's going to be a huge issue in society. Which has a huge cost.

[00:01:15] Claire: In this episode, I speak with Élise Mekkaoui, founder and CPO of Matricis.ai, a startup whose mission is to revolutionise the diagnosis of female pelvic pathologies, starting with endometriosis. Matricis brings cutting edge machine learning algorithms to endometriosis research in order to overcome the challenges in radiology and democratise MRI expert care.

Introducing the Host and the Podcast

[00:01:44] Claire: If you're listening to Narratives of Purpose for the first time, let me introduce myself. My name is Claire Murigande. I am your host on this podcast, which is all about amplifying a social impact. I bring you unique stories of change makers, stories of people who are contributing to make a difference in society. By showcasing these individual journeys, I would like to inspire you to take action.

Please take a moment to rate and review our show. This will help other listeners find narratives of purpose and further amplify the stories of impact that we share.

For now, listen to Élise's journey and her endeavours to improve endometriosis diagnosis.

The Purpose of Matricis.ai

[00:02:25] Claire: Élise Mekkaoui, welcome to the podcast. Thank you for joining me today. So we are here at the AI for Life conference and you are part of the panel on women's health and patient-led innovation. So first of all, can you briefly introduce yourself and tell us about Matricis.ai that you co-founded?

[00:02:48] Élise: So, I'm Élise. Thank you very much for inviting me, actually, I'm Élise Mekkaoui. I'm the Chief Project Officer and Co-founder at Matricis. We do AI for women's conditions and more specifically for Um, gynaecological diseases such as endometriosis, fibroma cysts, or things like this. And the first product that we wanted to develop is an AI companion for radiologists for endometriosis diagnosis. It is the first one that we are currently doing research on and we hope that it will be on the market in 2027. We hope so. After clinical trials and everything.

We began, uh, two years ago, but it's very, very difficult to have data. And yeah, we took, I think, one year to have access to a small data set. And we are currently trying to expand it in more long term things. We want to expand it to other types of data. for example, biological exams, um, medical interviews and things like this to have all the patient journey of, uh, the, the patients. And I think it will, I hope so leapfrog women's health to another level and, uh, we revolutionise the patient journey. We hope.

[00:04:05] Claire: Sounds amazing.

The Role of AI in Advancing Pelvic Pathologies' Diagnosis

[00:04:06] Claire: So when you speak about data, so you're talking about, first of all, diagnosis, right? Imaging. So what type of imaging? And we're talking about AI. How is that going to help advance that?

[00:04:16] Élise: It's very interesting because medicine starts with just a medical examination, after you have a diagnosis, and after you have a prognosis and surveillance. It is the path for everyone who's going to enter the health system. And we decided first to focus on diagnosis because we think it's the first brick who's going to, um, unlock all the bricks that there is after, for example, prognosis. So it is the reason why we wanted to focus on diagnosis.

So for women with pelvic conditions. Diagnosis could be several things. It could be medical examination by physicians, by gynaecologists, obviously, and it could be, um, biology tests, and it could be also imaging. And we have a background of computer scientists, so we decided to start with, uh, the imaging stuff. So, for women’s pelvic conditions, imaging could be ultrasound or MRI. It is both. Uh, and we decided to focus on MRI in the first time because it's very helpful for the prognosis. And it's also not invasive at all, contrary to, at the opposite of ultrasound, which is a little bit invasive and surgery, which is very, very invasive. And it's also, it's not, um, operator dependent, which is, uh, the case of ultrasound. And, uh, the ultrasound would be a very, very difficult challenge in technology because we have to do AI in real time, and, uh, yeah, it's very, very difficult. So we want to do ultrasound at one point to make a massive diagnosis, for example, in countries where MRI doesn't exist or it's not normally used, but first we decided to focus on MRI. And after we want to expand to other types of data, other types of imaging and to have a complete view of a woman's pelvic conditions diagnosis and after prognosis. It's our goal. So the first step, the first step, it takes us. I think it will take us five years, just for the first step, so it's a life commitment, you know, and we'll see where we're going, actually.

[00:06:34] Claire: And so basically the AI will help you analyse all these images and improve the diagnosis.

[00:06:40] Élise: Yes, sorry, I didn't explain. Um, the aim is to help the radiologists to read the images. So, uh, it's not good. We're not going to replace radiologists. We're going to give him a companion. So it means that we're going to plug in our AI algorithm into their workflow, which is the normal workflow of radiologists. He's going to read his MRI images and after just, to the right, for example, he's going to have more images with annotation.

For example, imagine that, uh, the patient has an endometrioma. It is a blood cyst in the ovary, for example. It could have many impacts on fertility, mainly, and, uh, the radiologist is going to see that on the screen, and he is going to see also an image where it is going to be written there is an endometrioma. So we have three types of, uh, analysis : we're sure there is an endometrioma, maybe there is an endometrioma and there's no endometrioma. And that's the way we want to help radiologists in order to improve the diagnosis.

And we also want to reduce time for radiologists to read these images because, for example, for pelvic MRI, it can take, uh, nearly 30 minutes to read an MRI because there is so much, so many things to look at. For example, the uterus and after ovaries, so many measures too. And we want to automate, uh, all these kinds of things, the diagnosis and the measures. And yeah, it will take a long time and we need a lot of data. And currently we have, uh, a great database actually, but we want to expand it because there’s always need for more data.

Challenges and Collaborations in Data Collection

[00:08:24] Claire: And so who do you collaborate with to find this data? Because as you were saying, you are the startup and you're going to offer this, this companion, as you said, or this support to radiologists. But obviously this is done in hospitals, in private clinics.

[00:08:38] Élise: For now, we collaborate mainly with hospitals. And public hospitals in France, actually, because we're a French company based in Paris. Uh, we are a spin off from INRIA, which is a research centre in machine learning in France. And, uh, APHP, which is the largest European network hospital in Europe, actually, uh, just after in the world, I think, uh, just after the Red Army in China, fun fact. So, yeah, we work closely with APHP. We also work with, uh, Hôpital Saint Joseph and maybe with other hospitals. And we are looking to work with American hospitals and with Swiss hospitals if it’s possible.

[00:09:20] Claire: And hopefully it will also help the time for diagnosis for the patients, because that's another issue, right? Especially when you're speaking about pelvic conditions like endometriosis, PCOS.

[00:09:29] Élise: Yes, exactly. For endometriosis, it's nearly seven years of medical errancy. Our goal is to reduce this medical errancy. But this is a huge clinical trial to make the point and show it's true. But, um. I think that there is a problem in all the value chain of the patient journey for endometriosis. We just decided to tackle one problem, but there is a problem of awareness for the symptoms. There is a problem for doctors who don't know what is endometriosis. And after there is a problem with imaging, in France, it's the main, uh, diagnostic of certitude. It's not in the United States, for example, we still cut. Women too, it was what was said, uh, in the panel and, uh, after there is a prognosis and treatments.

And we think, yeah, that's first it's not our specialty to tackle awareness. We have an expertise in computer science and data, so we decided to focus on it. And I think, yeah, diagnosis is a key and it's also a good fit with AI and machine learning because it's basically just crunching data and find correlations between the data set. So yeah, I think it's where we can make a difference and awareness is a subject that is treated by other amazing people and they're doing a great job about it.

Addressing the Gender Gap in AI and Women's Health Innovation

[00:10:48] Claire: And I'd like to come back to today's summit and the conference on the panel. What were your objectives coming in, you know, being part of this panel and especially what was the message you wanted to share with the audience?

[00:10 :58] Élise: First, I'm really, really happy to meet these incredible women. I didn't know them except for Fatoumata, who invited me and I really thank her for that.

Um, the second point I think we have to bring awareness for women's health to investors mainly to pharmaceutical companies, for example, and to say that we have to invest in it and that personalised medicine, for example, just bring one parameter which is gender. It's super important because it changes everything in medicine. There is other parameters and maybe we don't know some parameters, but we will find out maybe one day. But if we don't start with that, who is going to be a huge issue in society, which has a huge cost. For example, for endometriosis, it's nearly 100 billion of cost only in the United States for society because of lack of treatment, medical errancy, and also cost of society, because you just don't show up at work because you have too much pain, for example, or infertility. I don't know, there is many, many reasons, and I wanted that the audience just listened to that and heard this message. And I think, yeah, there were many investors in that scene. And I hope that they heard because, uh, they're very, very interested in AI and they are very interested in women's health. And after you present your business model and after you're going to have a reflection, which is. Women's health is a niche. How you're going to expand yourself. It's like a niche, a niche, not that much.

So yes, I'm really happy that, uh, we shared this message today.

The Future of Women-Led Innovation in the Age of AI

[00:12:37] Claire: And my final question is, you know, if you look in the future, how do you see this space of, you know, patient-led innovation or women-led innovation within this age of ai? How do you see this evolving?

[00:12:49] Élise: There are many barriers. I think the first one is that, uh, for example, the best master in machine learning in France, there is, I think, only 10 percent of women.

So, you see in women's health that it's always patients who are at the beginning of a company. So, if there are no women in AI, there's not going to be AI company. So, Yeah, first there is that, but because I see it in my company, we have both a foot in AI and a foot in women's health. And it's very, very different world because in AI there is almost nearly only men and in women's health there's almost only women. And this word doesn't meet and it's very strange because I feel sometimes quite schizophrenic. So yeah, it's very, very strange and very different mentalities. I like both, actually, but, uh, yeah, it's very strange.

And, uh, yeah, the second barrier, I think, is mainly investors and because it's a very, very difficult market because it's not clear who is going to buy the solutions. It's going to be insurance, radiologists, gynaecologists, I don't know, hospitals. And uh, the value proposition is hard to define. In health in general, in women's health in particular, because there is no data, it's very hard to find prevalence data who are correct. So you're always going to have the comments that you are a niche, and maybe I think there is a link between these two.

And, um, yeah, health is a very, very difficult market, and especially women. And I think it's going to be hard for investors to invest in it for a while, I think. And even as we are B2B, um, type of company and, uh, even for us, it's difficult to find a business model, even if there is companies that, uh, that does exactly the same thing as us, but for the type of conditions, for example, for prostate, there is many companies who are funded and for us, it's kind of difficult to, um, not to have the “niche“ comment.

[00:14:51] Claire: Really?

[00:14:52] Élise: Yeah, yeah, it's kind of funny actually. But maybe it is only the way we took the question. So I'm not sure it's only that.

And, uh, I think sometimes I'm very young, actually, I'm 26 and my partners are also 27 and 42. And we always get a comment that, uh, we are very young. And so they don't say women, but they ask us specific questions, for example, how can I be sure that you're going to be a good manager, for example. And we have this question all the time. And when we speak with male founders who have exactly the same age and exactly the same experiences, which is none or a PhD, they don't have this question, never. We spoke to several entrepreneurs and, uh, yeah, and we even have the question, how can I know that you are an entrepreneur? And I was like, because I am, it's my job. I tried something. So at this point I am.

[00:15:52] Claire: Well, at least I am hopeful because there's more and more women entrepreneurs. So that's a good sign. And I think if also on the investor side, this is also evolving, on two things : probably, I guess the mindset will change, but that would take a lot of time. But perhaps the second thing is also having more women investors.

[00:16:07] Élise: Yeah, I totally agree with you.

[00:16:09] Claire: Thank you so much.

[00:16:10] Élise: Thanks a lot.

Closing Remarks and Further Resources

[00:16:25] Claire: The healthcare system has indeed been failing women. Did you know that historically medical research has often excluded women? They entered clinical trials only in 1993, 80 percent of the drugs removed from the U.S. market between 1997 and 2000 were withdrawn because of side effects that occurred mainly or exclusively in women. Women's health has been considered a niche area. Mind you, women account for roughly 50 percent of the world's population.

Now that being said, things are starting to change. More and more women are advocating for their own health needs. They are creating impactful solutions with proactive and inclusive approaches. I truly believe that patients' lived experiences should be at the core of healthcare innovation. And let's not forget that there is a huge need to fund startups led by women. In fact, according to the research by the European investment bank in the U.S., just 2 percent of investments in startups are for women led ventures, although 38 percent of startup founders are women. Staggering numbers, right?

[00:17:31] Claire: If you are curious to learn more from some of our previous guests who addressed these issues, check out episodes number 22, 38, 42 and 63 Which focused on women's health. You can also check out episodes number one, 11, 33, and 60, which discussed the gender funding gap and impact investing.

Thank you so much for tuning in today. I appreciate you taking the time. That was my conversation with Élise McAwee at the first AI for life summit. You'll find more information about Matricis on their website at www.matricis.ai. That's M A T R I C I S dot A I. Be sure to listen to the other episodes of this special series featuring my interviews with the four panellists of « voices for change, women's health and patient experiences ».

If you enjoyed this episode, make sure to Please share it with a friend, with a colleague, or even with a family member. And don't forget to leave us a review on our website using this short link  www.bit.ly/narrativesofpurpose. You can also connect with us anytime and follow our activities through our linkedin account and Instagram handle at narratives of purpose podcast. Until the next episode, take care of yourselves. Stay well and stay inspired.

[00:19:19] Claire:  This episode was created in collaboration with Stephanie Trang at Artefact Open Innovation, Switzerland. This podcast series was written and hosted by Claire Murigande. This podcast was edited and produced by Tom Evan Hughes at Rustic Studios.