Synopsis: In this episode of the AI Leadership Insights podcast, Amanda Razani speaks with Tzvia Bader, CEO and co-founder of Leal Health, about an AI-powered treatment platform and a new GenAI feature that's helping patients and staff.

Amanda Razani: Hello, and welcome to the Techstrong AI Leadership Insights. I’m Amanda Razani, and I’m excited to be here today with Tzvia Bader. She is with Leal Health. How are you doing?

Tzvia Bader: I’m very well. Thank you for having me, Amanda.

Amanda Razani: Can you explain a little bit about Leal Health, and what services do you provide?

Tzvia Bader: Absolutely. We are an AI-based platform that help HCPs and cancer patient find treatment that are relevant for them. With the world of cancer changing so rapidly and new innovation, treatment, drugs, trials are being added daily, finding the right treatment and understanding your option become real challenging, both for the oncologist and for patients. That’s exactly where we come in.

Amanda Razani: Wonderful, and that’s such a blessing to so many, I imagine. Now, you recently provided an AI-powered treatment platform, I believe. Can you share a little bit more about that?

Tzvia Bader: Yeah. Absolutely. So what we’ve developed, basically, if you think about it, matching a patient to a treatment became an art. There are so many attributes that comes into play in order to find the right treatment for the right patient. It’s almost like matching on dating. Right? There’s a lot of attributes in patient profile when it comes to cancer, what type, subtype he has, stage, what treatment he had before, didn’t have before, how he responded, and most importantly these days, what we call the biomarkers, the NGS, the genetic mutation in his tumor. Because there’s so many advancement in terms of targeting those mutation, those biomarkers.
So how do we do it? What we developed is an AI-based engine that can read through treatment protocols, clinical trials protocol, FDA-approved drugs, and extract from it an understanding of all the attributes that this, or of the patient that we’ll feed this drugs, both inclusion and exclusion. Therefore, in a matter of minutes, a patient or a physician can build a medical profile and get pre-qualified for those treatments. So if you think about it, we cut the search of what’s out there. We cut the physician spending time reading through protocols of treatment. We cut him needing to remember what he heard in a conference or study and the time that patient spend on Doctor Google’s trying to understand what’s relevant or not.
All the information is being presented in a very patient-friendly, curated information. So not only we pre-qual you to a treatment in less than a minute or two minutes, but we also help that patient understand those treatment options. So when he has the discussion with his oncologist, it’s really about, “What’s right for me? How do I want to approach my treatment, more aggressive, less aggressive? What makes sense?” and not about, what’s the option out there? There isn’t such a gap in the understanding and the knowledge, and they can actually make the decision together, the best decision for the patient.

Amanda Razani: That’s wonderful, and I imagine for the patient, who’s already probably experiencing a lot of stress and worry, that this really helps streamline the process. Also, on the backend for the staff, it’s probably a lot easier and more efficient, as well.

Tzvia Bader: Of course. I mean, that’s the whole point. Physician cannot be in a position that they know it all. With such a rapidly-changing world, they’re also trying to catch tail with what’s going on, and we can’t expect them to know it and remember. So by reducing this burden of knowing it all by baking a tool that is smart enough to understand and bring them only the relevant information at the relevant time that the patient needs to make a treatment decision, we cut down not only the waste of time, a lot of the barriers in accessibility of care. It’s important. It’s important for patients who are being treated in the community. It’s important that this way, we can expedite diversity. We can reduce barriers in access to treatment for specific population and basically make sure that everybody gets the best treatment out there for their cancer.

Amanda Razani: Now, do I understand that you also unveiled a new generative AI tool this year?

Tzvia Bader: Absolutely. So what I just described of the art of matching patients to treatment, we actually based it on what now we can refer to, maybe, previous generation of an AI. But basically, there’s also a lot of tech and patent technology around what we call the unsupervised NLP that helps us read through this treatment protocol and get matched. What we realized throughout our journey, that patient and physician has hard time understanding and reading the genetic testing reports.
So like I’ve mentioned before, those genetic testing are extremely important, as they identify the biomarkers and mutation in your tumor that, now, drugs can address more effectively. But if you don’t know what you have, it’s hard to look for it or find it, and those testing reports are extremely complex. So what we’ve developed, leveraging a combination of LLM with our unsupervised NLT, and basically created a new capability that we call Precise Medical Modeling. We’re now able to read those NGS reports.
So think about it. A patient doesn’t need to get lost in trying to read it. A physician doesn’t need to get lost. They upload it into our platform. We, our system, our AI, our Precise Medical Modeling, can read through it and basically let them know which mutation and biomarkers they have, and then take it to the next step and actually say which treatment are available, whether in clinical trials or approved drugs, for those mutation. So we completely close the cycle, starting by telling you, helping you better understand your diagnosis, your cancer characteristic, all the way to finding you the best treatment that are available for this.

Amanda Razani: Wow. That’s incredible, and the fact that the treatment can be found so quickly, which I know that speed is of the essence when it comes to treating.

Tzvia Bader: Exactly, especially when it comes to cancer, and also staying on top of things. So you’re able to see all treatment that are relevant right now that are just became available, as treatment path is changing so quickly, as well.

Amanda Razani: Absolutely. So from your experience in the industry, what are some other use cases for generative AI when it comes to health care?

Tzvia Bader: I think health care is an industry that could really leverage gen AI, because if you think about it, it’s a matter of complexity of data, of tools and knowledge that requires a lot of expertise. So it starts from use cases that we saw with an AI and can now really be leveraged with gen AI, of reading on the diagnostic phase, diagnosing x-ray reports, PET scan. So the entire pathology world of the diagnostic can now be done much quicker, in a much more effective manner, and probably with more accuracy to a certain level, all the way to identifying and predicting drug, to identifying mutation that can be, then, drugged, develop drugs to target it.
So we’re talking about the drug discovery element of identifying and predicting how a mutation, how it will behave, and what will it responds to. Those are the drugs, and then all the way to what we do, which is matching the relevant patient to a drug and hopefully, also, then, predicting how a patient will responds to a treatment, and maybe diagnosing fairly quickly on if he is responding or not so you can amend treatment or change treatment course early enough in order to ensure success. So throughout the entire cycle from diagnostic, to treatment development, to prescribing a treatment, to monitoring a treatment, we can really see AI leveraging, and helping, and driving innovation into this industry that’s been waiting for innovation for a long, long time.

Amanda Razani: My thoughts on this, I know there’s still a lot of people who are really hesitant, even fearful when they hear the word AI. And so when you talk about all the ways in which AI can honestly save lives, I think it’s a new view and a new perspective that people can look at and embrace.

Tzvia Bader: I agree with you, and I also understand the hesitation. AI can be hallucinated at time, can be directed at the wrong places. It can be biases based on the data that it was trained on. So I think there is a lot of legit to the concern, but there is ways to overcome. So I don’t think AI can be treated as one, and it’s not a ChatGPT that I ask to summarize my article or to edit a video. When it comes to the healthcare, we have to develop tools and measurements to overcome hallucination, to ensure accuracy, to ensure that the data sets that we’re using are relevant one, and we control biases that can be created by them.
There is ways and measurement, like I’ve mentioned before. We’re combining LLM with unsupervised NLP on our existing data sets in order to reduce the hallucination and create accuracy and consistency. So there is ways to overcome the shortcoming of the AI. Also, remember, it’s not replacing the human, but it does help what needs human attention to get to the human quicker, because we reduce the workload.

Amanda Razani: It’s an augmentation tool, but that human element is still certainly important for sure.

Tzvia Bader: Absolutely. No doubt.

Amanda Razani: So from your experience when trying to implement and harness this AI technology, there’s a lot of work involved with that. What are some of the roadblocks or the barriers that you’ve seen, and what’s your advice when it comes to the process?

Tzvia Bader: So one, we just described. I think hallucination is a big, big problem, especially when we look at the worlds of the gen AI and, of course, data sets that are less trainable. So I think measurement in controlling hallucination in the healthcare profession are key in our industry. The other challenge is that there isn’t large data sets when it comes to the healthcare, or there are for some things. But in order to be really accurate and to support the trend, especially on cancer, but also in other disease of precision medicine, the data sets are really small.
So you also need to be able to find a way to leverage an AI capability on smaller data sets and not be dependent on large data sets. Looking at the data sets, again, we have to remember, a lot of the data sets we have has embedded biases into them, because they’re taking from a specific set of patient, which is not necessarily reflecting of the real world, and other barriers. We need to make sure that any tools or measurement that we leverage in the AI does reflect real world and we have access to real world data, or we mitigate the biases and understand the biases in advance in order to make sure that we don’t take those embedded biases and continue duplicating them in our conclusion and way of looking at it.
I think domain expertise are key to be part of this process. So those cannot be … As much as I represent the tech side of the house, I cannot exist, and I cannot develop the tools that I have if I hadn’t had, by my side, the domain expert, those that come from the world of the medicine, of the healthcare, that basically check what we do, qualified what we do, provide their guidance to the system in order to develop it to do a good job.

Amanda Razani: Wonderful. So if there’s one key takeaway you can leave with our audience today, what would that be?

Tzvia Bader: I think that we should not be afraid of technology. I think AI can help us finally be more in control of our diseases, have access to a better care, have better outcome out of it. So we shouldn’t be afraid of the tech, and we should know the tech does not replace human. It just accelerate our accessibility and accelerate better care for us. That’s what we need to remember, and I think for patient everywhere out there there, don’t be afraid that the fact that you’re not … you haven’t been trained in the medicine professor. For this, we now have tools that will help you understand and partner with your physician in order to have a better outcome, and I wish us all health in ’24.

Amanda Razani: Yes. Most definitely. Well, I want to thank you for coming on our show and sharing your insights today.

Tzvia Bader: Thank you so much for having me. It’s been a pleasure.

Amanda Razani: Look forward to speaking with you again soon.

Tzvia Bader: Thank you.