Front Row Seat: We'll Never Cure Cancer Without AI, Say Steve Brown & Lisa Booth

What if a misdiagnosis nearly cost you your life—and AI became your second opinion? In this episode, we sit down with Steve Brown, CEO of CureWise, and Lisa Booth, both living with cancer and pioneering the use of AI to navigate their treatment journeys.

Steve, a tech veteran, was diagnosed with rare blood cancer in the emergency room after months of missed symptoms. When he fed his medical records to an AI agent he'd built, it immediately spotted what his doctors had overlooked. That AI became CureWise—where Steve is CEO and patient number one, and Lisa is patient number two.

Lisa has been living with metastatic breast cancer for nearly 11 years, successfully navigating seven lines of therapy and multiple clinical trials. She uses CureWise to identify trial options, understand complex treatment data, advocate with her oncologist, and even appeal insurance denials.

This conversation isn't just about technology—it's about empowerment. We explore:

• How AI agents trained in different medical specialties can analyze your records and offer multiple expert perspectives
• The gap between elite cancer care and what most patients receive
• Off-label treatments and why insurance companies often say no
• How to bring AI insights to your doctor without damaging the relationship
• HIPAA compliance, hallucinations, and data protection
• The future of patient advocacy in the age of AI

Steve believes that someday soon, doctors who don't run patients through an AI model will be considered negligent. He also thinks we'll never cure cancer without AI—and after hearing his and Lisa's stories, you might agree.

If you or someone you love has ever wondered, "Is there a better treatment option out there that my doctor doesn't know about?"—this episode is for you.

Guests: Steve Brown (CEO, CureWise) | Lisa Booth (Patient Advocate)

Hosted by: Victoria Goldberg with Jill Terabassi, Linda Weatherby, and Abigail Johnston

Learn more: CureWise

Thanks for listening. If you enjoyed the episode, subscribe and leave a review — it really helps. Follow us on social media @livefromstage4 and visit our website at www.livefromstage4.org for show notes and links.

 Your support helps us continue to share important stories and advocate for those living with metastatic breast cancer.

Until next time, take care and keep pushing for progress.

Transcript

SPEAKER_06 0:10

Could your foot cancer be closer than you think? Welcome to Life from Stage 4, where MBC takes center stage as we talk to experts, share inspiring stories, break down signs, and shine the spotlight on what matters most. Because when it comes down to it, the spot for us and by us is all about us. What happens when a misdiagnosis puts your life on the line and artificial intelligence becomes your second opinion? Welcome to your front row seated live from stage four. I'm Victoria Goldberg, joined by my podcast team members, my friends Jill Terabasi, Linda Weatherby, and Abigail Johnston. Today, Linda is in the driver's seat for a conversation that could change the way you think about cancer care forever. Our guest, Steve Brown, believes that someday soon, doctors who don't run their patients through an AI model will be considered negligent. He also thinks we'll never cure cancer without AI. And he might be right. Because he has lived it, Steve's own doctors missed his aggressive blood cancer for months, telling him to take gas X for symptoms that turned out to be something far more serious. After an ER diagnosis finally caught what others had missed, Steve, a tech veteran, built an AI agent named Haley. He fed it the same test results his original doctors had seen. The AI immediately spotted the warning signs they had overlooked and recommended the exact tests that should have been run much sooner. That AI became cure-wise, where Steve is CEO and patient number one. Lisa Booth became patient number two when she received her metastatic breast cancer diagnosis. She faced the same overwhelming question millions of patients grapple with. Is there a better treatment option out there that my doctor doesn't know about? She's been navigating her care with CureWise in ways she never thought possible. This episode isn't just about technology, it's about empowerment. It's about asking the hard questions in the exam room. And it's about what happens when patients arm themselves with information that goes beyond the standard of care. But we're not just celebrating wins. We're asking the tough questions. Are your data protected? How do you bring AI inside to doctor without damaging the relationship? What happens when insurance says no? Everything Steve is on was prescribed by doctors, but the AI coached him on how to talk about it. If you or someone you love has ever wondered, is there a better treatment option out there? This episode is for you. Let's get started.

SPEAKER_04 3:49

So, Steve, can you introduce yourself to us and tell us how this all came about that you came to start CureWise?

SPEAKER_01 3:56

Hi, Linda, it's great to be here. So I have been involved in technology and healthcare for much of my career, but it was always developing things for somebody else's problem. But exactly a year ago, today, I was in the hospital. I got diagnosed with a rare blood cancer related to multiple myeloma in the emergency room. And if you're getting diagnosed with cancer in the emergency room, chances are somebody missed some things before that because you're not supposed to get diagnosed in an emergency room. You're supposed to have figured things out before it got that urgent. I was in the hospital, and immediately I was diving into. And what the heck is going on here? Then I started getting these biopsies and things with pages and pages of stuff that looks like Greek and with a lot of Greek letters in it, probably. And what I found at this point, having a bunch of different doctors, there were different opinions about what was going on. And there wasn't one absolute right answer. There's never an absolute right answer. There's a lot of different opinions and a lot of hard choices to make. So how do you even navigate that? You have to educate yourself and become an expert in what's going on with you. So I was probably like most people in 2026, when you have questions, you go to AI. So you can go to AI and get educated on a lot of things. But what I needed was AI that looked at my own medical record and looked at it from many different points of view and helped me figure out what's really going on. So that's where I started, and that was one year ago. I wasn't trying to make this for the world or for anyone else. I was thinking about just how do I deal with this? As I started to figure things out and started to get on better treatment pathways and actually started to get better, a lot of people started saying, hey, how do I get this? How can I get this for my friends? Lisa said, I need this, how can I try it? So she became user number two. Look, there's a big unmet need here, obviously. This year, another 600,000 Americans are gonna die of cancer. This is a big unsolved problem. But there's also a lot of progress in the field. And there's a gap between what's possible and what most people are getting. And what most people are getting is what's in the guidelines. And what's in the guidelines might be great for a lot of people. Cancer isn't just one disease, everybody has their own unique version of it, and there are a lot of different possibilities, and it's really important to figure out what are the best possibilities for you. So that's the problem I was trying to tackle for my own situation. And turns out I'm not the only one with these issues. Everybody's dealing with this.

SPEAKER_04 6:44

So, Steve, maybe we should go back to your particular experience. What I read and heard in other interviews with you was that you had been working with a team near your home, and they were not coming up with any diagnosis. And after an event in an emergency room, you get your diagnosis, and then you started to employ these AI tools, which eventually, like you said, found the notation from the Quest Diagnostics report, which ended up sending you in the right direction.

SPEAKER_01 7:17

Well, actually, the ER doctors in Rancho Mirage, California figured it out. I had really experienced doctors, and then I had some young ER doctors. My question to the AI was why didn't they figure this out before? The real thing that benefited me most from the AI, because I've already been diagnosed at that point, the real benefits were in helping me understand what the treatment options were. Standard of care was not the right thing for me because I had specific mutations. So figuring that out was where the AI had the life-changing difference for me, because that helped me understand what to ask and what to advocate for. Hey, you know what? There's some other drugs out there that are really good if you have this mutation, happens to be off label for your disease because your disease is too rare. No one's gonna even have the money or time to do a clinical trial on it. But Mayo Clinic would do this for you, and the UCSF would do this for you, but your doctor's not gonna do it for you because it's not in the guidelines. And then it's off label, so then the next thing is, oh, your insurance company's not gonna pay for it because it's off label. So then you gotta go fight that battle. The AI was also super helpful in fighting that battle. Help me explain to the insurance company why this is actually gonna save them a lot of money to do the right thing rather than the wrong thing. So it's not just, hey, I found this idea, it's the whole process of how do I make this work over the long haul. One last example of where the AI was instrumental. A friend of mine told us multiple myeloma. My brother just died of multiple myeloma. And then he said, actually, he didn't die of multiple myeloma, he died of an infection. And I'm like, oh my gosh, I think I've got this all figured out, but I can still die of an infection because I'm immune compromised. What do I do about that? So I'm going back to the AI and saying, what do I do about that? And it turns out there's a whole bunch of stuff I could do about that. But I wasn't gonna get it unless I asked my doctor about it. Once I asked my doctor about it, the doctor said, Yeah, that's a good idea, I'll prescribe it. If I didn't ask, I wouldn't have gotten it.

SPEAKER_06 9:23

This is Victoria, and my first interruption. Steve just mentioned off-label treatment. And this is something every cancer patient should understand because it might be the key to your best treatment option. Here is what off-label means. When the FDA approves a drug, they approve it for a specific use. For example, they might approve drug XA for metastatic breast cancer with a HER2 mutation. But here is the thing cancer research moves fast. Doctors and researchers might discover that drug XA also works for rare blood cancer with similar genetics. Using drug X for those other cancers, that's off-label. The drug is FDA approved, but just not officially approved for that specific use yet. This happens all the time in oncology, especially at major cancer centers. In fact, about one in three cancer prescriptions are off-label. But here is another catch. Insurance companies often refuse to pay for off-label uses, even when the science is solid and top oncologists recommend it, even when it might save your life. That's what Steve is talking about, and why having data and evidence from AI analysis can be so powerful when you're fighting for coverage. If you are at a community cancer center, your doctor might not even suggest off-label options because they know insurance won't cover it. But at places like Mayo Clinic or UCSF, they do it routinely. This is one reason why educating yourself matters so much. Now back to Linda and our interview.

SPEAKER_04 11:03

So what I'd love to do now, Lisa, I'd love to have you explain how you found out about Steve's technology here and CureWise, and tell us your story and how it got you to something that was not standard of care that's working. And then we'll toss to you, Abigail, for your questions.

SPEAKER_03 11:23

So Steve and I have been friends since we were kids. Our dads practiced medicine together. The folks on the podcast, I was a de novo stage for oligometastatic cancer patient in April of 2015. And I'm heading into my 11-year anniversary. And I firmly believe that I am alive because I do clinical trials. And since 2015, I have done a lot of clinical trials. I'm currently in my fifth interventional drug trial. And a couple of years ago, when AI started coming out, I was working through a mild recurrence and I wanted to find a new trial. And it was so tedious to find clinical trials. And a friend of mine who is a biologist that does cancer development, we would sit at my kitchen table and we would look through the thousands of trials that came back from clinicaltrials.gov. I kept a spreadsheet. This one looked interesting, this one didn't look interesting. And so I pitched to Steve and his brother, who also does AI, that they needed to help me build this thing. And neither of them bit. They were not interested at all in that. So I went back to my spreadsheet and I typed some more. And then last year, Steve's sister reached out to me and I reached out to Steve, and then we started talking about his diagnosis and all of the types of things and why clinical trials are important. And then he started telling me about this AI he was building for himself. And I was like, I want in. I need this, and I need it for what's going on with me. I was in the very first clinical trial for the drug called inhair to. And then when it was approved by the FDA, I turned to standard of care in Seattle, still getting in hair too. And I was on it for a total of a seven and a third years. But I started having minor recurrences after it was approved. And so I needed to figure out what to do with this single solitary liver lesion. We're lucky in Seattle. We have a liver board at the University of Washington ran by a guy named Ray Young, who's a surgeon. And through that, I ended up talking to an interventional radiologist, a radiation oncologist, a surgeon, an oncologist. And we looked at my entire medical record and they said, you know what, we think you should stay on in hair too, and you should go get that one lesion cyber knife. And so Steve and I took that specific scenario with his multi-agents and said, okay, let's load up all of my medical records and let's ask a virtual radiation oncologist, a virtual interventional radiologist, a virtual surgeon, what would you do? And just like the tumor board itself, the surgeon said, cut it out. The radiation oncologist said, let's radiate it or cyber knife it. The interventional radiologist said I could ablate it. And the oncologist said, by all means, change systemic treatments, which is exactly what the tumor board said. And I was like, oh, that's pretty cool. And I was like, that was really helpful. And one of the big questions I had when I went through the tumor board myself was, what's the difference between cyber knife and radiation? And what is the difference and effectivity between these and which is the best for me? And I was able to use the tool that Steve had built to really answer all of those questions. And I spent hours figuring out data that was actually out there, but I didn't know how to get to it. And Steve's tool really helped me do that. And after that, I had secondary recurrence of cancer. And this time it wasn't just a single spot, it was multiple spots. And it was a case where I couldn't stay on in hair to any longer. But I was really torn about what to do. And so I used CureWise all summer long. I used it first to figure out what clinical trials are available. And then based upon my goals, which is I tend not to have a lot of side effects with drugs. So I care about overall survival and effectiveness and progression-free survival and complete responses. And so I asked it for all of the drugs in development for my activating HER2 mutation, which ones had the greatest chance of a complete response. And it force ranked them and it gave me all the data, which is something I'd never had before. And I got to sit down with my oncologist. And she and I looked at all the data and we said, yeah, we agree that I think the right trial for me is one that it's going to use a new ADC. So it links to my cancer cells, but it delivers a different payload than in hair two does. So it said that. And then it gave me the choices of which ones were out there. And my oncologist and I agreed that the right one to do was one that had to cat nib. Unfortunately, that trial was not opening up again with spots for two months. And so I was stuck. Okay. So do I wait or do I take an oral chemo? And if I took the oral chemo, I couldn't get on the clinical trial because I would have exceeded the lines of treatment. So I had to do a big decision like, what happens if I wait two months? And so once again, I reached out to the CureWise tool and I said, okay, help me understand it. It wrote a risk profile along the lines of these are your options. This is what the risks are. This is how we're going to monitor it. And it gave us a specific monitoring plan and it helped me understand that I had a moderate risk of progression. It was probably a good idea to wait. Once again, my oncologist and I sat down and we thought, this is pretty good. And we did it. We did one more treatment of inhair to and we upped the cadence by which my personalized signature tumor markers were checking so we could see if I was progressing and we needed to deviate or not. I knew the slot was going to open, which was going to be at the UCSF in San Diego. After seven and a third years on inhair to my platelets were low. And so I was like, oh, I can't get into this trial, not because of lines of treatment or I don't qualify, but this was the first time there was something in how my body had gotten tired of chemotherapy that it was going to prevent me. And it's an arbitrary number. In that particular study, your platelets had to be 100 and mine were at 82 or something like that. And so the next thing I did was I said, okay, cure wise, I've got a problem with my platelets. What are my options? And it gave me ideas, vitamin B, iron, whatever it is. I would have picked my nose and eaten it if it would have given me the trial. Whatever AI would have told me I would have done. But I tried all of the natural ways of doing it. And then I said, was there a drug way of doing this? And it said, yes, you can try steroids. And then there was another drug that I can't remember the name of that I could take. I could take both of those and I could give them to the clinical trial team and say, if I take these drugs to boost my platelets, will it invalidate me from the trial? And they had the time to go research it and they said, no. I had the time to talk to my oncologist about it. And the first thing was we did was the steroid treatment, and it boosted my platelets up so that I qualified for the trial. I have the world's greatest oncologist. I love her. We have a very good working relationship. But she sees 600 patients. She's absolutely slammed. She wouldn't have had time to research all of these types of things and talk to the clinical trial team.

SPEAKER_06 17:47

But CureWise empowered me as a patient to self-advocate because some AI terminology might sound like tech jargon, but once you understand it, the whole conversation becomes so much clearer. First, what's an I agent? And how is it different from a chatbot? Think of a basic chatbot like a customer service rep who can answer questions from a script. You ask, What are your hours? and it tells you. Simple back and forth. An AI agent is much more sophisticated. It's like having a specialist consultant who can actually do things for you, research, analyze, make recommendations, and even take action. An agent has a specific role, expertise, and purpose. Now, what does it mean to train an agent? Imagine you wanted to create a virtual oncologist, you'd see the AI thousands of medical textbooks, millions of research papers from PubMed, clinical trial data, treatment guidelines, and patient case studies. The AI learns patterns, what symptoms suggest which diagnoses, which treatments work for which mutations, how the side effects are typically presented. But here is the key you can train different agents with different specialties and perspectives. One agent is trained to think like a medical oncologist, another is trained to think like a radiation oncologist, another like a surgeon. Each has access to the same information, but they're taught to prioritize and evaluate it differently, just like real specialists do. So when Steve talks about his multi-agent system, he's describing something powerful. It's like having a virtual tumor board where multiple specialists, AI agents, each trained in a different discipline, all analyze your case independently, then compare their findings. One agent might say, based on this patient's mutations, I'd recommend drug A. Another might say, But look at their prior treatment history. Drug B might be more effective. A third might add, have you considered this clinical trial that just opened? They're not just chatbots speeding out information. They're trained specialists offering nuanced, evidence-based perspectives. And because these are multiple agents cross-checking each other, it's much harder for errors or hallucinations to slip through. This is what makes your wise different from just Googling your symptoms or asking ChatGPT a one-off question. It's a team of AI specialists that are working together and analyzing your specific medical data. All right, now we've got the foundation.

SPEAKER_02 20:56

That's incredible. How long have you been on this trial now? I've been on the decisive MVD and Tacatnib trial since August.

SPEAKER_03 21:04

And I officially hit a partial response. So in this particular trial, you have to have a 30% reduction in your tumor burden. And my January scan showed that 30% reduction. So that's the great news. The bad news is that this particular drug has a side effect of neuropathy. And this is the first time in my entire cancer treatment that I've actually had a side effect that makes me think quality versus quantity of life. And that's another way I use the cure-wise application, right? So I needed to understand what is this neuropathy? How important is it? Is it going to go away? Is it not going to go away? And what are my options? Because the trial has certain constraints. If you're above a certain grade, they're going to take you off the drug. But there's this kind of grain middle ground like what happens if I downdose? Is the drug going to be effective or not? And what data is out there? Because the drug, the stism of via detin, is actually approved in China. So there's a fair bit of data out there. But I could ask the CureWise AI what What does it mean? What does down dosing mean? What does it look like for me? What are the odds that with the level of neuropathy that I have that it's going to go away with the chemotherapy break? It helped me understand the different treatments that are available for neuropathy and really helped me make an educated decision. Durantologist has what about seven to 10 minutes with you to talk about something serious? It's a life and death decision for me. I am on my uh seventh line of therapy, which is amazing. There's other standard of care lines out there. There are some new ADCs that are coming in development that will help me with my particular diagnosis, but it's not like there's 25 options out there. So moving off a treatment that's working is a big deal for. And so I need time to think about it. I need to go to a trusted source. I need something that looks at my entire medical record and can see that I'm an exceptional uh survivor with respect to ADCs. I need something that understands that I have a long goal, I'm very healthy, and things like that. So I can have that conversation back and forth in a safe, secure environment to mentally prepare for my meeting with my oncologist, to send my oncologist notes ahead of that meeting so she has time to think about them and we can have a better conversation.

SPEAKER_04 23:14

Yeah. So this suggests that once you start to engage with CureWise, you are maintaining the updates from all of your medical records as you go along. And again, Steve, in the past, this has been such a daunting thing to share your medical records, but you're saying that once we give permission, CureWise can import the data. Is it straightforward for the patient? Can I go that far and say that?

SPEAKER_03 23:38

Yes, I can say that because I've been beta testing it with other users. It's beautiful. And I love the CARES Act of 2022. In the old days, we used to wait for our doctor to tell us our scan results. And now our scan results get released to us at the same point in time. That's the CARES Act of 2022, and that's what enables us to get to our own medical records electronically. So you go in through the CureWise app and you say, I want to connect to a provider. So I'm going to connect to the Fred Hutch Center where I have my medical records. It then takes me to my Epic portal login and I authenticate and do the two-factor authentication. I look at everything and they say, I want to release my data to CureWise, and I get to pick a day, three months, six months, a year indefinitely. It's my choice as a patient. I click the button and all of it gets sucked in automatically into CureWise. And it does take time for us to import all that data and process it. I have over 3,000 medical records. So it doesn't happen in a minute, it happens over an hour. We are in the process of upgrading the system so it'll automatically refresh when you log in so that the data is always current.

SPEAKER_02 24:46

That that's an achievement just in and of itself. So being able to ask questions about your medical record without trying to figure out what report it's in and things like that.

SPEAKER_03 24:54

I love Epic, don't get me wrong, but it is really hard to use as a patient.

SPEAKER_04 24:58

Yeah, for sure. Okay, watching our time, I feel like this is a good time to pass it over to Abigail, who is our resident legal expert and attorney. And Abigail, why don't you ask the pressing questions about legal concerns with AI?

SPEAKER_05 25:14

Sure. So sorry, I've had a little bit of a cold, so I brought my frog voice today. Since this is a very new and changing legal landscape, I do have a couple of questions that I wanted to ask you. And the first is about HIPAA compliance. How are you protecting the data that patients are giving you?

SPEAKER_01 25:34

So we're going through all that process. We're not fully done with the HIPAA compliance where we're doing everything in a HIPAA compliant way. I've been a digital health for much of my career, and we were the first ones to take remote patient monitoring through the FDA, and we did HIPAA, we did a whole bunch of other things. So it's a known thing, it's a known process, and we're going through it.

SPEAKER_05 25:54

So the people that are participating now with the beta testing, how are you protecting them if you're not yet HIPAA compliant?

SPEAKER_01 26:00

That's one of the reasons why it's a closed beta right now. First of all, if you give permission for your own medical record or upload your own medical record, you're outside of the kind of main intent of HIPAA, but it's a closed beta invitation only, and you need to understand that this is what status we're in. That's why we haven't opened it up to the public. People say, Why can't I get it? Why can't I just sign up? It's because we need to go through all these things. We need to ensure that we've met all those requirements and we need to do a lot of quality assurance. When we have all of those things finished, we can open it up to the public.

SPEAKER_05 26:33

So, what about the fact that there is a lot of data behind paywalls? How are you getting around the fact that, especially clinical trial data? You're talking about all these great solutions that are being offered, but what about all the information that's behind a paywall?

SPEAKER_01 26:48

You mean hub meds of medical research. You've probably seen in the last couple months all these announcements about Anthropic and OpenAI and the AI companies and Gemini really diving into medicine, which is extremely helpful to us on this front as being a part of the developer community of doing specific applications that leverage all of that. They've fought those battles. Anthropic has already trained their AI on all of PubMed and all of these medical resources, and we use that. There are certain things that are billion-dollar problems that a startup is not really in a position to solve. We are really lucky that the big tech industry is solving some of those problems for us. We're leveraging billions of dollars of investment that's going into training AI on all of the data, and it's not perfect and it's constantly evolving. But it's so much better than asking Google. It's so much better than not knowing it all. The reason why I got into this in the first place is that I had something quite rare, and that's like one in a million people. And I was just astonished at how good the medical knowledge was at great detail that was in there. That's what convinced me to probe that more deeply and say, hey, has this already been trained on medical knowledge? And then you've seen from the recent announcements that the big AI companies are going all in on training their models on medical knowledge.

SPEAKER_05 28:13

In terms of application, though, right? There's this data set. And the idea is that you train this AI or bot, whatever we're going to call it, on that knowledge. But I've got a mug in my kitchen that says, don't confuse your Google search with my law degree, because there are professionals, just like our doctors, who are trained in the application of that data, legitimately from a professional who had to synthesize all of that data. And it was my training and experience that made my recommendations better. How can we rely on the application from an AI versus a human being?

SPEAKER_01 28:56

First of all, just to take my own case as an example, every treatment I'm on was prescribed by my doctor or one of my doctors. My doctors are the ones that are doing both the diagnosis and treatment. The AI is helping educate me so I'm actually an informed participant in that conversation. Because if I knew nothing, and I suppose this concept in healthcare, shared decision making, and now I know what that means. It's like, hey, Steve, you can do this and you can do this, and it's totally different future depending on what you decide, but you decide. If you're not knowledgeable about what all this means, how are you gonna even pretend to make that decision? And you're not making it all by yourself, you're making it with your doctor, but you need to be a good counterparty in that conversation and be able to have an informed conversation about it. So the fact is, medicine is delivered by doctors, but doctors are not perfect, and doctors make mistakes and doctors miss things. So if there's two brains at the table, not just one, and they're both informed about it and they're collaborating on it and know more about it, you're gonna get to a better decision than just one who's overloaded and it's gonna default to the guidelines.

SPEAKER_05 30:08

So, what you're saying is you cure wise is a tool for education and knowledge for patients to be able to then take to their doctors, not something that is going to give the answer.

SPEAKER_01 30:21

It says at the top of the screen, and you can't even delete this thing. It says this is not for diagnosing or treating your disease. That's what your doctor is for. This is to help educate you on how to be a better counterpart in that conversation and how to ask better questions. Because, like I said for my own case, if I didn't ask the question, there's a lot of don't ask, don't tell in medicine. If I don't ask the question, that door doesn't even get open. When I do ask the question, it's oh, that door is open. I better think about it. And I have a dozen examples from my own treatment where if I didn't ask the right question and I wasn't involved in it, I would have been on the wrong path. Lisa's talking about her cancer. It's like a whole foreign language. She sounds like a microbiologist talking about her cancer. She got to become an expert in your disease. So I became an expert and all these specific things about plasma cells. And the more I understood it, the more I understood that it's not some arbitrary decision. Once I understood like why and how this stuff was working, I felt like I could be the smarter person at the table in making those decisions.

SPEAKER_04 31:27

Let's go to Jill Tirabasi, who is our resident provider and clinician and physician expert, also living with metastatic breast cancer.

SPEAKER_00 31:37

I am not an oncologist. So first off, I am primary care, sports medicine, preventive medicine, boarded physician. The cancer was not my MO. And so when I go talk to my own oncologist, I feel like I don't know enough. I still have that perspective even as a physician. And so I think about patients, and I think about the people here are very well educated and have, I would say, much more capacity and resources than typical patients. And then we think about patients that go to NCI institutes versus community centers, right? We're talking about differences in resources and whatnot. Assuming that you're going to see your oncologist who is an open, happy person to discuss your AI results with you. I'm curious what has seemed to be the best approaches when you're bringing this new next level data to them. And the other big piece, also two-parter, the insurance coverage piece and how we approach that. Because once we're getting off standard of care, I literally just got a denial letter the other day. You're not following standard of care. How do you approach those battles? Because that's what we're going to need to do to have the N of one experience, to have the individualized cancer experience that we need.

SPEAKER_01 32:44

Yeah, this is the biggies. It's like you learn something, but you now got to convince your doctor. I went through exactly those battles. But I don't go to my doctor and say, oh, here's my AI said, no, aren't I smart? I don't want to be a patient from hell that does that. But what I do go in is if I learn that there is some other treatment that is matched to the genetic mutation of the cancer, I say, it's like, hey, I saw in my record I have this mutation, and I think there's a drug that's matched to that. Can we talk about it? Now you can use the AI to make a cheat sheet for you to help you prep for your doctor's appointment. What I would actually do in talking to the AI agents is rehearse for my doctor's appointment. So I just would rehearse the conversation. I'd rehearse it like 10 times. Let's say you get 10 minutes. You don't want to spend that time on the wrong questions. You want to spend that time on a really targeted thing where you've already done your homework. So I look at it as I've done my homework, I've already rehearsed the conversation, so that I've worked through some things. So now I'm really focused. If I've got 10 minutes, how do I make every one of those 10 minutes count? And the fact is, my doctor was not going to prescribe this because it was off label. He says, I just don't do that here. I went and got a second opinion. I ended up getting a second and a third and a fourth opinion. I went to Mayo Clinic and I went to UCSD and I went to UCSF. And I basically told them, I said, look, just write your notes in the medical record so my doctor will see it. And my doctor might not believe me. If you write it in there and say, Yeah, this is a good idea, then he'll probably do it.

SPEAKER_00 34:20

Some peer pressure, I hear, yeah. Uh huh.

SPEAKER_01 34:22

Yeah, yeah. So not everyone's gonna feel comfortable doing that, but there's some basic things where, hey, I heard that every cancer is genetically unique, but is there any test I can do to sequence the cancer to see if there's anything in there that might lead to some other clinical trials or treatment ideas? Looking at my medical record, I don't see that I ever got any sequencing. Shouldn't I get that? That'd be a good question to ask your doctor, and your doctor might say, Yeah, okay, I'll do it for you. So that's one thing. And then the second question was, how do you do with your insurance company? I had to write appeals to the insurance company, the doctors had to write appeals to the insurance company. I got denied on standard of care stuff because it's expensive. That's a default thing. That's just their process. Expect to get the denial, be prepared for the appeal. That's just the way they work. So CareWise also helped me in that process as well.

SPEAKER_07 35:14

Lisa, you wanted to add something. Why don't you jump in?

SPEAKER_03 35:18

Yeah, so in terms of the insurance approval, so I have an activating HER2 mutation. And what that means is that my cancer behaves like a HER2-positive cancer, but the drugs aren't available to me. So I can't get a prescription for Herseptin or any of the other HER2-positive drugs. So while we were going through looking at clinical trials, my doctor prescribed the HER2 Klein protocol and it got rejected by insurance. And she would have definitely written their own appeal. But what I did in working with Dr. McCormick because I used CureWise and it actually did a brilliant job of describing why my activating mutation would work on it and things like that. And it forgot all of the articles. And I was able to give that back to my oncologist, and she used it with the insurance company, and they approved, I think it was passed on Hersepton, and I didn't get to Catney, but it came back and it approved everything. And so it just made it faster. And I was able to help my oncologist and do that piece of it and really frame it up. And I think that's an important piece. But I think that the other thing that we highlighted here, and I think it's an important piece of what is going to really differentiate cure-wise, is that while I know how to self-advocate, and I've been doing it for a decade, Steve knows how to self-advocate, and probably everybody on this call is self-advocators. And we are all going to tier one cancer centers or but the broader community doesn't necessarily have access to that based upon where they live, or they are used to having the provider say, This is what you do, and the provider decides everything. And so one of the things we're going to be building into the tool is that guide and that advocacy for the patient. How do I prepare for my oncology appointment? Or what would you suggest? How do I do that? What should I send my doctor in advance? How do I practice that conversation with my oncologist? Or how do I describe my symptoms? And so I think that those are the pieces of the puzzle. And I think somebody had asked, well, doesn't your provider have access to this? And the answer is, I firmly believe that there is so much investment going on in both the provider side of the house and the patient side of the house that I think our providers will have access to more AI-based tools. I think their hurdles are different than the patient's perspective because as a patient, I can make decisions to gather information and look at things. I don't have to worry about my liability insurance. I don't have to worry about my medical provider and those types of things. And so I think the barriers are much higher to entry for our provider community than they are for patient community. And so I think there's a real opportunity here that we as patients can go faster than our providers can right now. And as Steve talked about, that broader perspective of shared decision making means that you as a patient need to understand and be your own advocate. And so you need to be equally educated and have that opportunity to ask questions. So in that 10 minutes you have with your provider, you can make meaningful steps forward because we rely on our providers. And our providers are well-intended, educated, competent people. When people say my doctor, blah, blah, blah, blah, blah, and they're not competent. I was like, no, they're human. And there's a thousand things that could have been, and they need to work through that puzzle of figuring things out. And we as patients need to be partnering with our providers to make that happen. And so the best cancer outcomes comes with a tight partnership between the provider and the patient.

SPEAKER_04 38:28

I do want to say this has been an amazing conversation. I really appreciate the angles that you both have spoken to today. This is really good tools for our audience. When you do come to market, hopefully soon, what's that going to look like for patients and how does that work? What is the cost model for this? Right now, you're asking for patients to sign up for the wait list. I did that, by the way. I'm really interested in this.

SPEAKER_01 38:54

Yeah, so we we are making it so you can just pay out of pocket for it while we go to the insurance world and we go to the diagnostics world and the therapeutic world and find people who are aligned with this and try to make this more part of the practice of medicine. But we don't want to hold it back for that because there are people who want this now. While we go fight all those battles, we will make sure that people can just pay out of pocket in the meantime. Because a lot of people are not gonna want to wait.

SPEAKER_07 39:22

Are you comfortable in telling us how much out of pocket we're talking about?

SPEAKER_01 39:26

One of the things we're looking at is that the more AI analysis you do, it's fairly expensive. So we're trying to analyze that because we need to make sure that we're committed to being around for the duration. It's gonna be on the order of$50 a month,$500 a year. So our goal is to make sure that we make this accessible and also sustainable at the same time because we want to be around for a long time.

SPEAKER_04 39:52

We will uh provide the link to your website, of course, for our listeners in the how it works section. You walk through 10 steps, and I just wondered if you could speak to that process. And specifically, as I'm listening to you, you guys are skilled at asking great questions of the agents. What does a patient do who doesn't know what questions to ask? Will the agents give information regardless of how good your question is?

SPEAKER_01 40:20

I do this with AI all the time. Asking the right question is the key to unlocking the knowledge. But what do you do if you don't even know where to start? That actually is a good question. I don't even know where to start. So we're building some of that assistance in there so it gives you some questions to, hey, maybe here's a list of questions you might want to start with. And then as you go in the conversation, hey, here's some things you might want to ask next. So we do offer some suggestions that can kick off the conversation.

SPEAKER_04 40:48

Okay, that's really helpful. I think this might be intimidating to a lot of people, but my gosh, the benefits are tremendous.

SPEAKER_06 41:14

Let's pause for a moment. I'm about to ask Steve a question that may need some explanation. I want to make sure that everyone understands what we mean by AI hallucinations. Because it sounds actually scarier than it really is. An AI hallucination isn't like a visual hallucination. It's when an AI confidently generates information that sounds completely plausible, uses the right medical terminology, besides what seems like real studies, but it's actually making things happen. Imagine asking a friend about a restaurant. They confidently tell you, oh yeah, that Italian place on Main Street, they have the best carbonarian, they're open until midnight. Sounds specific and believable, right? But when you get there, you discover that the restaurant closed two years ago. Your friend's brain filled in gaps with false details. That's essentially what AI hallucinations do. Here is why your wise approach matters. By using multiple AI models of law, Chat, GPT, Gemini, and having them cross-track each other, hallucinations become much easier to spot. If one AI says something that the others don't agree with, that's a red flag. Plus, these systems are constantly improving. As Steve mentioned, what was pretty good a year ago is dramatically better today. And cure-wise, that's another layer. It's analyzing your actual medical records, not just making general statements. Still, this is why Steve and Lisa both emphasize the AI is for education and preparation. Your actual treatment decision still happened with your doctor. The AI helps you ask the right questions and understand your options. Okay, let's get back to my question. I agree, and I want to ask just a very quick question as a follow-up. There are a lot of people who are very suspicious of everything. And now with AI, there is so much talk about hallucinations and how you get garbage in, garbage out, and all of that stuff. So how do we, and I understand that your tool doesn't say this is how you have to do it, and you have to go to your doctor and talk to your doctor about it. But how do you make people more comfortable with the results that you provide?

SPEAKER_01 43:44

We're working on actually the measurement of that too, so that we can be more rigorous about it. But I can say that in general, even over the last year, since I've been doing this in medicine, it was really good a year ago. It's even better now, and it's getting better every Day. So it's not like we're the only ones trying to solve this problem of hallucinations. There's an entire multi-billion dollar industry working to solve that problem, and it's getting really good. We also have our own guardrails on top of that. And one of the initial methods of dealing with that is to have multiple agents address the same questions from multiple points of view. And when they agree and when there's convergence on something, that's usually a pretty good signal that the knowledge is more reliable than if it was like, hey, the 10 agents do this, and only one of them said this, and everything else said that. That's helpful too. So it's an ongoing process to improve the quality every day. It's kind of the self-driving car thing. It's 10 years ago, maybe they crashed a little bit, but now they're better than a human driver. They may not be perfect, but they're getting really good. It's the same thing. It's getting better every day, and it's so much better than not having it all or what it was a year ago. That yes, you should be skeptical. You should be skeptical of all this stuff, but don't let that hold you back from educating yourself. If you're skeptical, she asked that as a follow-up question. I'm skeptical. How do I know you're not hallucinating? Ask it. Because what you're doing is you're probing this black box of AI, which has compressed all this human knowledge into this black box. And like, how do you know whether or not you're getting the right stuff? What if there's garbage in there? How do you even know it's a black box? So you're probing it. And the more you probe it, the more you're getting to the nuggets and you're getting smarter in the process.

SPEAKER_03 45:32

That's wonderful. Hey, Victoria, I have something to add to that. And so one of the things that Steve spoke about, but I really think is something that is important, is when I use CureWise, I get to look at the anthropic model, which is Claude. I get to use Gemini, which is Google's. Whenever you start a new drug, the question that we ask as patients is, okay, I'm optimistic that it's going to work, but there's always that fear till you get to the first tumor marker or you get to the first scan. You know, what happens? And how I've managed my 11-year journey is that I always make sure I know before I get the scan, if the scan is bad, what am I going to do next? And so I had selected the clinical trial that I was in. And then I started asking curies, what happens if this drug doesn't work? What would I do next? And at the time we were using Lambda, Groc, Quad, all of the models. And I asked all of them. And they do, although they're trained on the same data, they do answer questions differently at times. And so it's using the multiple models. And to Steve's point, having them talk against each other means that we're talking with each other because we have this agent debate thing where they go off on each other. So you have that built in. Yeah, good. And so it's not that a model is right or wrong, it's how it's seeing the information. And I know hallucination is a problem with respect to AI. I know that from my personal experience, and my oncologist has looked at each one of every major treatment decision we've been doing with Curewise. My oncologist is in there with me. We're looking at the data, and we've both been really happy. I deferred to Steve and explained why Curewise works so well. I just know as a patient, it's been pretty phenomenal.

SPEAKER_01 47:09

It's wonderful. Wow. There's multiple models and multiple agents with multiple points of view. What that's also telling people is that there's not like one perfect God answer on this. There's a lot of different possibilities. There's a lot of possibilities, but you need to do your own investigation with it and explore that.

SPEAKER_06 47:30

To clarify, we all hear about all these big names coming into the space and doing work in this space, and ChatGPT is doing something, and my son, who works for Google, is saying they're working on something. So for you and for your tool, this is not competition. This is actually assistance, right? Once they come out with this stuff, you will be able to leverage what they produce to give us better results.

SPEAKER_07 47:59

Am I correct about that?

SPEAKER_01 48:01

We're taking in all three of those and more so that you're benefiting from all of them. I think that's one of the big benefits of going with us, is that we're saying that we care about this and we're doing this for ourselves. But we're committed to being on the frontier of what's possible. And right now, the best model might be this, and tomorrow it might be that. But we're always going to be on the frontier of what's possible. The year before the I was doing this, I was doing stuff in education, and I gave a keynote talk at a big conference about empathic AI. And the whole point was it's like we can train these agents to actually be very empathetic and very personal, and like little things like that are actually really big things.

SPEAKER_03 48:45

You, as a cancer patient, are faced with life and death decisions with imperfect information. It's overwhelming and it's scary. And so building an application that meets you where you are, that supports you and guides you. The knowledge is incredibly important, but we all know that knowledge is knowledge, but it's how do we absorb the knowledge, how do we interact with it, how can we get to it? How do we feel empowered as patients? And that's what we're building. And it needs to be specific to oncology. OpenAI and Claude are going to build medical navigators because what did somebody say 25% of queries to AI right now are medical things. So it's huge, but I am again of one, and I need to have this conversation in the context of my full medical record. I need to have it in a way that I can speak to it and it understands me and I understand it, and I can practice and consume things, and I'm guided uh with the specialty of oncology. I am guided towards clinical trials. I'm guided towards how to figure out how to get into them and not just providing information and not guiding to say that I'm the doctor, but guiding me as a patient to how to be my best advocate.

SPEAKER_07 49:58

I think it's very important.

SPEAKER_06 50:00

You know how they say there are no stupid questions, but there are so many people who are so intimidated by asking their doctors questions. And this is a completely different setup. You feel comfortable asking this new friend that you have and who doesn't think that you're asking a stupid question.

SPEAKER_01 50:20

For that alone, it's very helpful. There's only one stupid question, and that's the one that you're afraid to actually ask. That's true. One thing with the AI, what I learned pretty early on is that people are much more afraid of asking a real person what they think might be a stupid question, and even their doctor, where it really counts, but they're not afraid of asking the AI is not going to judge you.

SPEAKER_07 50:44

Absolutely. And I've been very actively using them, and I consider Chad GPT my personal friend. But it's true. It is absolutely true. The more you use it, the more comfortable you are with it, the more comfortable you are with asking questions.

SPEAKER_01 50:59

These human factors of feeling like you're supported, feeling like somebody cares and somebody's there to listen to you, that's a huge part of the benefit of what we're doing because it makes the knowledge accessible and it encourages you to ask the questions, it encourages you to get educated, and it just is there for you.

SPEAKER_06 51:19

As you're ramping it up, how many people do you foresee being able to get access to it at some point soon? And when will you go from the beta wait list to the actual subscription when people can start using it?

SPEAKER_01 51:34

Yeah, there are some things like PIPPA and things like that that we want to make sure are in place. And until everything is in place, call it a soft launch where we're opening it up to more and more people. But our goal is to have it certainly by the middle of this year. Anyone who wants it can get it. The stakes are really high, and we need to be very careful about this. We started doing this for ourselves, and now we're opening it up to more people. We want it to be available to everyone, but we really need to get it right.

SPEAKER_07 52:02

And you will, and we're so happy that we are here and we'll try to help.

unknown 52:07

Yeah.

SPEAKER_06 52:08

Thank you for being here, and thank you for spending so much time with us. Yeah. We're happy to do it. Thanks, you guys. Take care. What a powerful conversation with Steve Brown and Lisa Booth and my friends, Jill Terabasi, Linda Weatherby, and Abigail Johnston. You know, when Steve said the only stupid question is the one you're afraid to ask, that really captures what this is all about. For too long, cancer patients have been told to just trust the system. Trust that your doctor has time to research your bring your clinical trial. Trust that your standard of care is the best option for you specifically. Trust that if something important was missed, someone would have caught it. But if the historian listens a story, show us something different. They show us that being an informed, empowered patient isn't about challenging your doctor. It's about becoming a partner in your care. And now, for the first time in history, we have tools that can help us do that. Not to replace physicians, but to help us show up for those seven-minute appointments with the right questions already researched, the clinical trials already identified, the appeals already drafted. I think what drug memo was Lisa saying she uses cure-wise to practice conversations with her oncologist. Just think about that. She is literally rehearsing how to advocate for herself, testing different ways to ask questions, exploring what-if scenarios, getting comfortable with the language of her own disease. That is not medical education, that's empowerment. Now I know some of you are thinking, this sounds amazing, but I'm not tax-avvy, I'm not a programmer, I barely understand how to use ChatGPT. And here is what I want you to hear. You don't need to be at Lisa's not a technologist. Steve built this because he needed it as a patient, not as a Silicon Valley entrepreneur. The future of cancer isn't about replacing the human connection between doctor and patient. It's about making sure that when you have that connection, you're bringing your whole self to the table, educated, informed, and ready to advocate for what you need. You can join the wait list for beta access. And yes, they're working on making this accessible and affordable. For those of you who are thinking, I need this now, I hear you. Steve and Lisa hear you too. They're moving as fast as they responsibly can while making sure keeper compliance and quality assurance are rocket solid. Because this isn't just another app. This is potentially life-saving technology, and it needs to be done right. A few key takeaways to remember. First, AI agents are in chatbots, they're trained specialists that can analyze your specific medical records and offer multiple expert perspectives. Second, hallucinations are real, but using multiple AI models that cross-check each other reduces the risk. The cure wise is always improving as the technology evolves. And third, this is a tool for education and advocacy and not diagnosis or treatment. Your doctors are still your doctors. But now you get to be an informed partner in those decisions. One more thing. If you found this conversation valuable, please share it. Share it with your cancer community, share it with your support group, share it with that friend who is drowning in a medical jargon and doesn't know where to start. Because the futures even least are building, it's not just for tech savvy or the well-connected, it's for every single person who's ever sat at an oncologist's office wishing they understood more, could ask better questions, and knew what options they weren't being told about. Until next time, keep advocating for yourself, keep asking questions. And remember, knowledge isn't just power, it might just save your life. This has been Live from Stage 4. I'm Victoria Goldberg. I want to thank my friends Jill Terabasi, Linda Wetherby, and Abigail Johnston. And of course, our guests, Lisa Boots and Steve Brown, for joining me for this important conversation. Take care of yourselves and each other.