The future of oncology

Show Notes

Daniel sits down with Dr. Kim Blackwell, former oncologist and Eli Lilly trialist who is now the Chief Medical Officer at Tempus, a technology company that has built the world's largest library of clinical and molecular data as well as an operating system to make that information accessible and useful for patients, physicians, and researchers. They discuss big data, genetic sequencing, and the tech that drives today’s transformative precision cancer treatments.

What is Healthy Conversations?

Healthy Conversations brings together leaders and innovators in health care to talk about the biggest issues facing patients and providers today. Every month, we explore new topics to help uncover the clinical insights and emerging technologies transforming health care in real time.

Kim Blackwell:
I lived in a space, both in academics and pharma, where there was real risk aversion. And a lot of times when there was a problem be solved, we would come up with a list of 20 reasons why we couldn't solve it and one reason why we could. I'm realizing that a tech approach in particular is, "Okay, we're going to solve this."

Daniel Kraft:
Welcome to Healthy Conversations. I'm Dr. Daniel Kraft and I'm in Healthy Conversation today with Dr. Kim Blackwell. So we're both oncologists. We're going to have a, I think, a really fascinating conversation about the cutting edge and future of oncology and beyond, but you introduce yourself to our listeners from being a clinical trialist to a practicing oncologist to embolden big pharma, and now as a CMO and senior vice president of clinical oncology for Tempus.

Kim Blackwell:
I was six months short of the quarter of a century in academics. I was a practicing oncologist at Duke University for close to 25 years. Loved, loved taking care of patients, mostly focused on young women facing breast cancer. But what got me up in the morning was developing new drugs. And luckily I was involved in some of the early successes drugs like Trastuzumab or Herceptin, and then the pill targeted, HER2, called Lapatinib and then ultimately binding chemotherapy to antibodies known as antibody drug conjugates, one known as TDM1. And it was really a great time because 20 years ago we had chemotherapy, we had chemotherapy, we had chemotherapy, and as I explained to patients for more than two decades, it basically was it kills every dividing cell in your body. I still remember the first patient I gave TDM1 to, and she was just crying hysterically because she wasn't going to have to lose her hair.
But I got up to a point in my career where I realized if we just continued the pace of drug development, even just in breast cancer, we were only going to see maybe three or four major breakthroughs in the way we take care of patients before I retired. And so I decided, well, I'm going to join big pharma. And so when this opportunity came along at Tempus to really think about how do you use big data in the context of, I call them tech solutions, you can call it AI, to not only identify new targets that exist in cancer, especially in rarer populations of molecularly defined cancers, but also speed up the process of drug development. And I can say even today, I still believe that big data with tech solutions is the way that we will accelerate getting new treatments to cancer patients and ultimately to patients facing many types of diseases.

Daniel Kraft:
The challenge is turning that data into insights and actual information and narrowing the gap to bring that to the point of care, whether it's in the oncology office or in public health.

Kim Blackwell:
My experience in talking to people, even before I joined Tempus, was people were working on problems they knew they could solve. People were collecting data that they knew they could collect if they were in the data space, and then the clinicians were trying to answer questions with just a tiny piece of the data because that's all we could really understand. At least that's how I was trained. You have one target, you have one drug. We now have the ability from a single piece of tumor to look at 15,000 little forms of the RNA known as the transcriptome. It was just too much data for even the practicing clinician or clinical researcher to understand. So it is a huge challenge, but again, I'm thinking about it in terms of how do all of those three analytics data and clinical come together to define what is the problem that we actually are trying to solve.

Daniel Kraft:
And so we're talking about precision medicine. Where is it now and where is it heading?

Kim Blackwell:
I mean, I'm pretty simple-minded about this. It's getting the right drug to the patient that will benefit from it. When I explain it to my teenage children, I say 20 years ago I had this vision that we put a piece of tumor in a machine and out would pop a pill. When I was working on Lapatinib, which was a pill targeted against this protein known as HER2, I always had that vision of the Jetsons, like how they cook their food in the kitchen, for those of you who are old enough to remember the Jetsons. But it's here now. We have at Tempus thousands of organoids that we already know what their DNA is, RNA, what their protein is. You can dump 2000 of those drugs on it and understand which of those organoids from patients actually respond to the drug. So we're close to having all the components of that precision medicine machine, sorry, to my pharma partners who are listening, I guess they're the fourth circle that we really have to partner with. And how do we expedite getting what we know in the tech big data company sequencing into the pharma companies?

Daniel Kraft:
Maybe run us through a sort of how that process works through Tempus.

Kim Blackwell:
Although we offer what we call panel testing of tumors, which is a very highly selected group of over 600 genes. We also offer something known as whole exome sequencing, which is basically sequencing every gene in the tumor. It generates a lot of data. For many years when I was a practicing breast cancer oncologist, we were making decisions off of three things, estrogen receptor, progesterone receptor, and HER2. I used to always say, "If triangles had gods, they'd have three sides" and now we have 15,000 RNA endpoints. We have the DNA, we have the DNA from the normal tissue as well. And that's kind of important because we realize now that a lot of cancers that we didn't think were genetically linked as their cause are now genetically linked up to 5% of lung cancers have a germline predisposing gene.
But the classic thinking was if you got lung cancer most likely you were a smoker, not always, and that caused it, so we didn't really need to worry about you or your family or your genetic makeup that led to you, and that's just not the case. And then the other fundamental goal of the company though was to take all of that and make sense of it for patients so they can have a discussion with their doctors for doctors so they can have a better discussion with their patient. And I think that's what most cancer patients want. It's just hard to get those two things connected. The data with how do you have a conversation with your doctor

Daniel Kraft:
And for the practicing oncologist out there, how do you boil all that thousands of data points into something that is actionable?

Kim Blackwell:
We offer a digestible form of all of this data, but we also believe that patients own their data. So anytime a patient wants their raw data, we give it to them and we give it to them in a usable format. We're working on this diagnostic algorithms but really applied to meaningful problems. So we're doing what we call tumor origin algorithms, and that's really interesting. That's another way of digesting large amounts of data. And so for the 25 years that I gave 5FU, there was that one out of 10 patient that really suffered a lot of side effects from these drugs, mostly diarrhea, a lot of hand foot syndrome, and I just wrote it off to, it's the luck of the draw, but there's actually a genetic test that you can click on our report this all in one, kind of find out everything you need to know about the patient and the tumor and have this tested.
And 90% of the time it's going to be completely normal and you don't have to do anything about it. But that 10% I know that I as a patient or I as a physician would say, "Hey, why don't we start a lower dose with this first cycle?" And even that is a separate test nowadays, if you don't work with companies like Tempus and the companies that are really trying to do alt one stop shopping, and so then the burden falls on the patient, right/ I got two weeks before I'm starting treatment or it falls on the physician, it falls on both. Wouldn't we want to have state-of-the-art knowledge about what disease we're facing in the clinic to help our patients even more, even if it's scary, even if it's stuff we don't understand. It is pretty scary, as a physician, I barely can deal with CT scans and bone scans and now you got a whole nother nine page report to deal with and having the ability to not only talk to other people in the genomics world, but talk to colleagues about how they would use the information in a virtual tumor board. The one thing the pandemic has taught us at Tempus is you can do a lot remotely and that brings together isolated cancer care.

Daniel Kraft:
So that sort of takes us maybe the future of the Virtual Tumor Board is what opportunity to now look at response to therapy and really predict how someone's going to do, particularly as things mutate and change through the course of therapy.

Kim Blackwell:
I would envision five years from now what we call methylation changes on circulating DNA, which doesn't require you know which genes you're looking for, but basically just look for pattern changes will be the standard of care that will predate any 1970 CT scans to let patients know that maybe their therapy is not necessarily going to work long term for them in what we call the metastatic or incurable setting or that they've done all of the treatment they thought they needed to do, but they're still at risk of their cancer coming back. It's going to be a blood form of the machine, which is you finish this test, you finish your treatment. I used to tell patients you've done everything you can within reason to prevent your cancer from coming back, which again, this is stage one through three cancers, which is the majority of cancers diagnosed luckily in the US today. And then I put your blood tube in the machine and it says, "You're cured." Like wouldn't that be great? But if it says you're not cured, it's going to say, "And this is what we think is important to your tumor and we're going to reanalyze your tumor." It might be three years from now the way things are going. And then out pops the pill again in the Jetson like machine of precision medicine.

Daniel Kraft:
How do you see this sort of new world where you have new massive data sets changing rapidly and new maybe off patent drugs leverage into the ability to learn and do trials when it's no longer large buckets of this is one form of cancer treated in a standard form.

Kim Blackwell:
So we clearly have to do things differently. My vision of doing things differently, I break it down very simply into if you have a rare population and a drug that works and doesn't hurt people, my perception is you can get drugs approved in that space. And we have examples now. We didn't, Lapatinib was one of a very few for about a decade because we had a driver alteration in breast cancer and we had a pill that we could target and then Osimertinib in lung cancer with EGFR mutations and erlotinib and gefitinib is another example. My opinion is we've plucked the fruit of those ones that occur in one out of five cancers. But for these rarer subtypes, there's two problems in getting the drugs developed. One is how do you identify the patient?
How many of those patients actually exist? Is it 0.1% across the spectrum? Is it 1%? Is it 5%? So helping drug companies gather that data and then identifying those patients that are, they're not all in the academic medical centers. They're everywhere. They're in your backyard. And then the second part of the problem, which I think is so multifactorial, once you find those patients, how do we get the patient on the trial that is just enough to get the drug approved for widespread use? So we have a program now available to over a thousand oncologist where if you identify a patient, we've already worked out all the contracts and the IRB and we open up these trials in less than 14 days. And when I joined Tempus a year and a half ago, I was like, "Yeah, right." I mean I seriously was the biggest cynic, and then I saw it happen. If you get the lawyers and the finance people out of the way and you focus on what do we really need to do to open up a very standard single agent pill form for that one patient that might derive benefit from it in Yuma, Arizona? You can actually get it done fairly quickly if you've laid the groundwork. It's really putting precision medicine into practice and then help doctors and patients get on these trials without having to travel for a hundred miles once a week.

Daniel Kraft:
What's the actual brass tacks of how do you get a patient enrolled or engaged, sort of the one-on-one of how to make that happen?

Kim Blackwell:
Yeah, it's a little bit of a scrappy process. You'd be amazed at all of the barriers that exist to get a patient. And I don't think this is just in cancer. My favorite analogy is just how do we staple the Staples Easy Button, you know that red button? I can get a jar of peanut butter to my house by 8:00 AM tomorrow morning. I would think that we could figure out how to get a patient on a clinical trial. I actually think the patients get better care when they're on a clinical trial because they're followed more closely. They get more frequent assessments. So I don't think that part of the system is broken.

Daniel Kraft:
Yeah, I think the takeaway here is sometimes it's not about to say technology, particularly in health care. It's the workflow. And I know Tempus is only what, six years or so old and many diseases wait for five year outcomes. Any taste of the impact you're making?

Kim Blackwell:
So lots of successes. One that comes to mind is a 40 something year old gentleman that was in Georgia who literally I think was going on hospice and his tumor was sent for sequencing and we opened up a trial and he went on and he responded, it's just amazing. And if you haven't taken care of cancer patients, I have, you just don't realize how magical and special it is when you're able to offer something that no one else can offer and extend someone's life in a meaningful way. And so we are thinking a lot about how they can advocate for themselves without being worried about upsetting their doctor. And I don't think doctors realize how scary we are. The most interesting component of our successes is how engaged patients are in it. So they see the report, they go to our website, they see that we have this network. They go in, they ask their doctor. Their doctor actually says, because we're so far reaching, "Oh yeah, we have that available." And then next thing you know, they need a new treatment and they're getting put on clinical trials. So I'd say that we're 70% of the way there.

Daniel Kraft:
All sorts of new forms of data, and we can imagine in the next decade we're going to see entire new convergences as well. The challenge is how do you connect those dots?

Kim Blackwell:
It's not going to be perfect. And I'm most excited about the intersection of all of the diseases and how they're interrelated. And so we have to start by not being afraid of how big of data set that would look like. Just combining cardiology and cancer on one platform is a major data handling effort, but they're all actually pretty intertwined. I see a day where there's a Google-like for the patient sitting in front of me and all of their data thinking about it like organ system. And I know you see it too. I mean, I think you saw it long before many other people did when I discussed it with you 10 years ago, which is I pop on my computer and I have all of your data in front of me in a meaningful way across the diseases. And that is the vision of Tempus.

Daniel Kraft:
Amen to that. We're still using low bod modems and fax machines to communicate a lot of our healthcare data. Even Eric, your founder, came from the consumer marketing world, et cetera, and I brought that mindset and tech ability and we're seeing many of the most interesting, exciting healthcare startups founded by folks outside.

Kim Blackwell:
Yeah, that's probably what I love the most about working at Tempus. I lived in a space both in academics and pharma where there was real risk aversion. And a lot of times when there was a problem to be solved, we would come up with a list of 20 reasons why we couldn't solve it and one reason why we could. I'm realizing that a tech approach in particular is, "Okay, we're going to solve this." And I'm amazed at how resistant I, as a physician and my colleagues are, to say, "Why don't we question why we're doing the things we're doing?" And I think some of it's just fear of change. I mean, change is hard. I think some of it's fear of it's so much data that it's hard to comprehend, which I think is really kind of a mindset that happens a lot. In particular in the cancer space.
"It's too hard." I hear that a lot. The biggest thought I'd like to leave the practicing doctors out there is to not be afraid of some of this new technology. Actually before you discount it, at least try it. And I know that everyone's busy, but I'm realizing that I never gave it as much of a chance when I was in day-to-day practice because I was so set in my ways. And now that I have had a chance to look outside, I think some of these things are incredibly helpful and would've been helpful to me and my patients had I just given them a chance. So that's my parting words for what it's worth.

Daniel Kraft:
Thank you for this amazing, Healthy Conversation. I think Tempest is a great example of the future coming faster than we might think. Hopefully accelerate this future of health and medicine and precision.