The Uromigos Episode 124: Legends in GU Cancer Series—Dean Bajorin

Dr. Dean Bajorin discusses developments in bladder and collaboration in testis cancer.

Episode Transcript

Tom:
We’re joined here today by Dean Bajorin, who is going to introduce himself in a second. I’ve known Dean since the beginning of my career in urothelial cancer and he’s seen as the godfather of many things that we do.

He’s even got a prognostic scoring system named after himself, which again, Brian and I haven’t managed that. We’ve tried very hard, but Dean, listen, thanks so much for joining us firstly. If you’d like to introduce yourself, that would be great. And then we might ask you a couple of questions if that’s okay.

Dean Bajorin:
Sure. To the audience, hi. I’m Dean Bajorin. I am an attending at Sloan Kettering and a professor of medicine at Weill-Cornell medical college. Been in the business quite a while, actually, and started working at MSK in 1986 as a faculty member previously there for three years in training as a fellow.

Tom:
Dean [crosstalk 00:01:25].

Brian:
Talk about …

Tom:
You go, Brian.

Brian:
In respect to the early days, talk about your background. I know you did a lot of cisplatin work early on, so let’s go back to the beginning and talk a little bit about that.

Dean Bajorin:
Sure. So, when I was a fellow actually, it’s really interesting. These are the days before the internet. These are the days at the time entry-level computers, et cetera. And these were times when it was very difficult to understand biologically what was happening to drugs in patients.

And so, one of the first studies I did was actually looking at cisplatin. And it was really a lot of work behind the scenes.

Actually, you had to identify the non-protein bound cisplatin because active drug is unbound drug. And then we used to ultra-filtrate the plasma, identify the platinum, measure the platinum, et cetera over time. And then you had to plot the curves out and identify the area under the curve.

What people don’t really know then is actually we had to go to the patient’s bedside, because all patients were hospitalized for cisplatin back then. There were no anti-emetics, you collected the specimens. You had to run them down to the unit, to ultra-filtrate them because you couldn’t let time pass. Because they got more protein bound over time. And then you had to plot the curves.

And so interestingly, back in that time, the entry-level computers were actually microprocessors stuff. They had very little capacity. And so, my background is actually in physics and mathematics.

I remember being told nobody in medical school needs somebody who’s a physics and mathematics background back then, you won’t survive. You need biology and chemistry. But we were able to finally get computers and actually demonstrated mono exponential decay areas under the curve.

And back then with testis cancer, we knew that 80 to 100 per meter at that time, per course was satisfactory and cured patients. So, we actually went up to giving 200 per meter per course with hypertonic saline, et cetera. And I can tell you, those were really very challenging studies.

And at the same time, because I could do pharmacokinetics and I could model computers, et cetera. Then I was asked, well, listen, we have these patients getting monoclonal antibiotics. Can you do that as well? And so that was my background in terms of my original interest.

Tom:
And Dean, before we jumped on, you told us a little bit about how some of this data, you presented some of this data where originally, and you talked also a little bit about international collaboration and perhaps some of the first international collaboration in GU cancer.

Could you expand on that a little bit?

Dean Bajorin:
Yeah, sure, Tom. I think this is really a lesson from the past and for the future. Present and future actually. So, when I was a fellow, one of my side projects was actually looking at staging systems and prognostic systems for testis cancer.

It all started George Basel had a model, which is an exponential model, mathematical model to predict who should be good risk and who should be poor risk. And these are really entry level studies with trying to minimize the toxicity of the good risk and maximize cure and poor risk.

But at the time actually, the Institut Gustave-Roussy had their model, Indiana, Larry had his model, we had our model and then the NCI had their model. And so, one day in clinic, I said to George, I said, “Well, what makes your model the best?” And he goes, “That’s a good question.” And it took me two years to solve that.

A lesson for all of us, right?

So, I tabulated all of our patients and looked at our model, then recategorized them according to the [inaudible 00:05:34] model. So, I had to measure everything out. Ours was using markers, AFP and HCG and LDH. And then we looked at the Institut Gustave-Roussy model and the NCI model.

And every one of them, now, all these patients were treated at MSK, but when you recategorize them according to the different risk criteria, you had remarkably different curves for the good risk and poor risk. And so, it was misallocation, and it was not better therapy.

And so, I had to present that first at ASCO, which I did in front of Larry. I remember being petrified as a junior attending, and Larry came up and he goes, “Great study.” I wasn’t hung out to dry.

But then the next step was where do we go from here? And so, we contacted everybody. Now we had Germans on board, the UK on board, the French, two groups in the US. And so, we agreed that we would pool all the data.

There would be one repository for the data. We would provide two-thirds of the patients to any group in the world that wanted to look at their system to refine the system with two-thirds of the patients.

And then we reserved one-third of the patients for the validation set, and we prospectively agreed that the publication that came out of it would be unaltered and be just in the name of the group, the collaborative group.

Ben [Mead 00:07:12] From the UK was going to be the corresponding author for the study.

Okay. So, each one of us kept it to … Sally [Stenning 00:07:19] from the MRC got all the data and the data were set up for the investigational dataset and the validational dataset, everybody got their two-thirds. We spent months working our model and everybody else did the same thing.

And we agreed that we would meet in-person and discuss it. We didn’t have any money. It wasn’t a formal group. And so, Sally, because she worked for the MRC, said, “I think that we can get the University of Leeds to give us free housing and food for a meeting of basically four to five days. And if they’re willing, we could do that.”

So, we sent out all these invites to everybody, everybody interested in Europe, in the US, and we all flew over. And if anybody has ever been to the University of Leeds during the summer, when students aren’t there, the food is dreadful, the accommodations are dreadful.

The bathroom and showers are all at the end of the hallway. So, everybody had, remember college? You had to go to the end of the hallway to go find the bathrooms.

And I can tell you, people said to us at the meeting, “How could you do this to us?” Thing was, the great thing was, is that everybody was incredibly collaborative because when they wanted to get the hell out of there. And so, all the work got done in a very short period of time, everybody was highly collaborative.

We actually, because there were US interest, UK and French interest, we actually had a German chair of the meeting. So that it would be completely independent. And it would be arbitrarily decided whoever had the best system for allocation to good and poor risk would be the winner.

And it turned out actually it was good, intermediate, and poor risk. That was the first iteration of the international criteria. I was fortunate enough to be the chair of the AJCC and UICC at the time. And we would incorporate markers into the staging system if it looked like it really worked. And the publication was in the name of the group and not in the name of the individuals.

So, I think the take home point is if you really want to get something done, go to the University of Leeds during the summertime. But the second thing is everyone actually cast aside the egos, the institution and everything else, and really it was for the good of the disease to make progress. And I think that’s a lesson for eternity.

Brian:
Yeah. It’s really an amazing story. And you were saying before we started recording, the data was recorded on, was it floppy disk or even on paper?

Dean Bajorin:
So, yeah. So, I actually have mine on huge spreadsheets that I actually did because we didn’t have, we actually did not have databases and what we call relational databases at the time.

And so, we created them on spreadsheets and then we then created ours in very old formats with semi-colon delimiters and all that for Sally, who then put it into their mainframe at the MRC. And then she and I independently did the proofing for every single data point for every single patient. It took about a year.

I remember sending it to her on New Year’s Eve at about 10:00 saying, “Sally, these are our data for our independent work.” So, if you have a chance to interview Sally, she probably can give you a great perspective of the MRC from back then.

Tom:
Dean, if we switch to bladder cancer, there is the Bajorin score, which is used. How did that come about? And is that through a similar collaborative process?

Dean Bajorin:
It wasn’t quite as collaborative as it was in the germ cell community. So, part of it was that in the work that I had done in germ cell tumors at the time, my predecessors and mentors, which included [inaudible 00:11:31] and George Basel and actually Howard Scher who was just a couple of years ahead of me said, “Can you bring into bladder cancer what you did in testis cancer?”

And so, Howard was moving into prostate cancer completely at that time. And there really was a lull in bladder cancer at the time. And we started examining it the same way.

The first thing we did is we started looking at the number of cycles. Why are we treating patients way back indefinitely? It just really didn’t make sense. And we look at all the patients we had treated, and we identified that a maximum of six cycles is what you achieve for benefit, and anything beyond that was toxicity.

And then the second thing was to look at post-chemotherapy consolidation surgery, because we are treating patients for a long time with these radiographic residuals, like we had done previously for testis cancer. And we just asked a simple question. Could it be just radiographically evident residual and not cancer?

So that was the first step. And that is the post-chemotherapy consolidation data that really emanated from the prior charts to the data. And then came the issue of prognostic features.

Now, the reason the prognostic features came up is that we were developing new regimens. And my past experience was that, is it the regimen or is the allocation of patients? And so, I had done data looking, we compiled data of all the patients treated with MVAC.

We had developed a new regimen called ITP and, and we were trying to identify what were the individual subsets of patients, specific predictive factors. And was the mix of those factors affecting the predicted outcome?

And so that was another exercise. We compiled all the data. We actually had compiled over 150, what we believe potential predictive factors. And it was more than just the fact that you had visceral disease or that you had a poor performance status in the beginning, and included hemoglobin and albumin, et cetera, but the model was too complex to be used.

And so, we looked at what were the major factors, which included visceral disease and performance status. And then to validate it, we couldn’t externally validate it. So, we internally validated it by bootstrapping.

These were all the techniques that were we were learning at the time, and we found that it validated quite well. That’s what we published.

But the outgrowth of that was this new regimen that we thought would be great, ITP, median survival of 20 months. It turns out it was mostly patient-related factors and not necessarily the drug. It was a little contribution, but you could actually look at the mix of different patients, the zero, one, and two in your clinical trial, and you could actually predict based on that mix of patients, what your long-term outcome would be with therapy.

And so, I think that’s a take home point when we start looking at our treatments, we have to really make sure that when we have that brand new phase two study of a new regimen, we have to really make sure that when we preparing it to historical patient populations, that we have really controlled the study for predictive values and prognostic factors, rather than just say this is the median survival in clear to improvement.

And we go onto a phase three. And we did, our exercise was not only to provide stratification, but to eliminate wasteful phase three studies on regimens that were better only because of the patient mix and not because of the regimens.

Tom:
Dean, when did it become apparent to you that bladder cancer was very responsive to chemotherapy? And when did it become apparent to you that it wasn’t going to parallel the testis story and chemotherapy was not going to cure bladder cancer patients?

Dean Bajorin:
So that happened basically early nineties, ’94 to ’95, as we did prognostic modeling, it was really clear. Actually, if you go back to the original publications, from Cora Sternberg and Howard Scher was that patients who had CR, their median survival was basically over 40 months and patients with PR was around 19 months and if you were a non-responder, it was basically seven or eight months.

So, it was really clear that there might be an element of response in terms of survival. But the second thing is, when you started looking at it, most of the patients actually had nodal only disease. I was lucky I was actually treating melanoma at the time.

Back in the old days, we treated the head neck in melanoma, and I’ve got publications there, but it was clear that the only melanoma that we saw respond was disease in lymph node with all the therapy.

So that was the first indication. And then when we feathered it out for zero risk patients, good performance status, metastatic nodal only disease, that patient population in our hands had a 24% five-year survival and a little over 30% five-year survival and a 24% 10-year survival. And so, it was really clear that that population was curable.

But if you look at a patient with visceral disease and poor performance status, none of those patients survived. And so that really was the element that that was a much greater population of patients than we had seen with testis. And we clearly needed new therapy, and we really needed to understand the biology for new therapy.

Brian:
I want to go back to something you said a little bit ago, how you gave this great story about a collaborative effort in germ cell tumor. And then when you started talking about bladder, you said it wasn’t quite as collaborative.

What do you think the key elements are for a disease as a whole to be academically collaborative, which you said is really for the benefit of the patients? You’ve been in many different circles and diseases. So, can you identify key factors?

Dean Bajorin:
I can actually. In the beginning for bladder cancer, there was a constant tension between medical oncology and urology. And it wasn’t the same. In testis cancer, what happened was the urologist said, “This is your disease.”

Whereas in bladder cancer, when I first entered that, I got the sort of the pushback, “This is our disease.” And in fact, during that time period, most urologists in the world were giving their own chemotherapy. Because that’s what was happening with GYN surgery. And I remember being told you have your areas, we have ours, and this is not yours.

I remember being introduced by a famous urologist to give a talk at one point, and I was introduced that, “This is Dean Bajorin, and he’s going to throw everything but the kitchen sink at your patients.” And I remember being really frustrated at that point in time.

And that actually continued for well over a decade in terms of adjuvant therapy. In terms of neoadjuvant therapy, it was a constant uphill struggle during the early days. Eventually I think what happened is it was about a 10-year period of time, but there became a very close collaboration between urology and medical oncology that built up over time.

But that happened at our institution because actually all of our urologists and medical oncologists had this … We actually had the same offices together, shared staff facilities, actually some of my closest friends live close to me, they’re also surgeons in our neighborhoods.

And so those collaborations helped with us. And I think that model of the integration of urology and medical oncology slowly built momentum, and now it’s much more collaborative, and it’s much more collaborative across the board.

If you take a look at it in terms of urology, medical oncology, radiation oncology, we developed great relationships in terms of radiation with testis cancer and that evolved in bladder cancer.

And then lastly, the scientists. Each one of our groups started developing programs that involved the scientists. And now all of us are actually friends. But it took a while to get there because of sort of the provincial nature of who owned the disease at certain periods of time.

That’s all a thing of the past now. And it’s great now. Just take a look at the studies that we’ve now produced, neoadjuvant, adjuvant, marker development. Look at our meetings in terms of urology, medical oncology, radiation oncology, a completely different world than it used to be.

Tom:
Dean, what are the … So, you’ve explained the six cycles came initially from essentially a time when you felt it was right to stop. Is that a fair comment?

Dean Bajorin:
Yeah, actually it evolved from two different directions. Everyone in lung cancer was given therapy for as long as you could possibly do it. For us, we were doing the same thing. But what we did is we just started taking it, who gets six cycles, eight cycles, 10 cycles. I had one patient at 13 cycles. And we kept saying, it just looks like we’re gaining toxicity, scans aren’t changing, et cetera.

That was one aspect. The team of oncologists were smarter than we were. They did randomized trials of five versus 10 and six versus 12 cycles. They really were well powered studies at the time, but it really … what we learned is that they maximize therapy by five or six cycles. And so, Bob [Ossels 00:22:10] who’s at the MCR, I remember sitting down and talking to them and we’re saying, these are both epithelial tumors.

Why would they act any differently in terms of the number of cycles of therapy? And that’s how it evolved. And then once we determined that six cycles was the max, that’s when we started … we’re now able to do randomized trials of six cycles.

Tom:
When carboplatin came along, the Spanish did a randomized phase two study of 80 patients comparing the two and showed cisplatinum was better than carboplatin. It seems that’s to be, was that the study that generated the cisplatin push?

More recent data suggests more similarities than differences between those. What’s your feeling on the cisplatin carboplatin issue and where it currently stands?

Dean Bajorin:
So, my background is actually, I actually did the pharmacokinetics on carbo. It was called [inaudible 00:23:00] CBDCA a JM-6, came out of the UK. And so, actually I actually authored a paper on carboplatin etoposide versus cisplatin etoposide in good risk germ cell tumor patients and showed that carbo was inferior.

Wasn’t my best paper in terms of the findings, but it was an important study at the time. So having experience with carbo with testis, bladder, looking at all our data and other data, I’ve always felt that it is an inferior drug to cis based on the body of information. It has its role in patients who obviously can’t tolerate cisplatin, but I don’t think it is a major advance in disease treatment. I think it’s an intermediate step.

Tom:
There was a barren period in urothelial cancer. I remember Matt [Galfsky 00:24:12] getting up at ASCO one year and saying that there are no oral presentations and no poster presentations in bladder cancer this year. “Has bladder cancer disappeared in the US?” That was the question he asked me, which I remember it was one of my first interactions with Matt.

I thought, well, that’s a good question. When did it change? Because clearly there’s a huge amount of work in urothelial. At ESMO this year there’s more bladder than there is kidney, there’s probably as much … So, when did it change and why did it change?

Dean Bajorin:
So, I think there are two things. I think the turning point was 2007. And I say that because I had a background in immunology actually when I came on board because of modeling for monoclonal antibodies and I used monoclonal antibodies in therapy for melanoma. Again, multiple different diseases at the time for all of us.

In 2007, Pam [Scharmer 00:25:09] was actually my fellow. And so, we talked about that immunotherapy should work in bladder cancer. We just can’t figure out the reason why. And so, and I had had patients, for example, whose disease I would follow for years without treatment. And I remember saying, gee the tumors are infiltrated with lymphocytes. What is it?

And then I was lucky, Lloyd Old, who’s a grandfather of immunology actually. I got to know him at MSK and some of his teachings. And so, the fortunate thing was we were able to get monoclonal antibodies to start studying different subsets of cells.

And in fact, Pam’s study looking at patients who just underwent surgery, no other therapy, looking at long-term survivals, those patients whose tumors were infiltrated with CDA positive cells did much better than anybody else.

Okay. So, we knew there had to be some recognition, some immune response. And then the second thing that was helpful is the development of immune therapy and PD-1 and PD-L1 PD-1 pathway. That was helpful. A lot of that happened, Jim Allison at MSK. So, we could see where the field was going. We just weren’t there yet.

But 2007, these immune studies just interrogating the specimens was the first real attempt at the underlying biology. And I think that was the turning point. And then we could design the study. The agents could design the studies, et cetera.

Tom:
Dean, let’s get to 2021. Let’s go into the adjuvant setting. You led the adjuvant New England Journal publication. It was a positive trial. It’s in the perioperative space. As you said, it requires collaboration. We’ve had unsuccessful adjuvant trials, Cora Sternberg’s chemotherapy trial, which never really enrolled enough patients, but the data looked great.

And had we completed it, it probably would have been positive in my opinion. How did you make that trial successful, firstly? And secondly, what’s the current role and what do we do next in the perioperative space? Is this a practice changing study?

Dean Bajorin:
So, I think that the first thing is the design of this study and now you’re seeing it is a placebo-controlled study. It’s extraordinarily expensive. It’s a big gamble by a pharmaceutical company when you are randomizing versus placebo versus randomizing against standard of care surveillance. And so, the escalation in costs is really substantial.

I remember sitting down with the folks at BMS at the time, and this was their first foray into that space. And we talked about a number of different approaches. It was really clear that a placebo-controlled study would be the very best. Because we’ve been there before, where patients, you have patients, you tell them that you have the worst disease status that we know of, very high likelihood of recurrence. And we have this promising therapy, and you can get randomized to that or nothing at all.

And patients say okay, and then they get randomized to nothing at all. They started seeking out other therapies, et cetera. So, the dropout rate is high, number of events. Which actually, and so that can affect trial reporting and results. So, this placebo control was designed from the outset. We knew it would take much longer.

It was designed to have a cap for the upper tract because we knew the biology might be different, et cetera. And this is really a testament to all the investigators that it was extraordinarily rigorous. We’ll see where it goes in terms of survival, et cetera, and the FDA and EMA and all that. But I think it shows rigor.

And now you’re seeing that in other studies. You look at the pembro study that Tony just reported. Again, placebo controlled, same thing in esophageal. Placebo controlled. These kinds of studies are really important. So, I think that that’s the first space.

The second space is single agent therapy. We know it is an interim position. We’re hopeful that more combinations as we study them will be more effective as we dissect out more biology, which is terrific. I used to try to get people to love bladder cancer, now people come to us seeking out studies in bladder cancer.

In terms of the scientists, it’s so much different than it was before. And so, the old joke is you have to get them to love you to be able to collaborate. Now it’s really scientifically fascinating.

Brian:
So, Dean, we’re going to run out of time. So, we’re going to end with our speed round. We’ve done this with our previous legend series. I’m going to start. I tend to ask more serious questions, Tom less so. So, we’ll start just for a minute here. So, what was the year and topic of your first publication?

Dean Bajorin:
’84. I think the first one cancer researchers on pharmacokinetics, it might be ’84, ’85. And then I think ’85 was the JCO article on the prognostic factors. That might be ’86, but right in that ballpark.

Brian:
I have ’86 according to your CV, and a pretty impressive impact factor for publication number one. Go ahead, Tom.

Tom:
Dean, I’ve got artists, I’ve got Jackson Pollock or Pablo Picasso.

Dean Bajorin:
Pablo Picasso.

Tom:
Seems reasonable. I’d agree with that.

Brian:
What was your most embarrassing publication? The one you wish you could take back.

Dean Bajorin:
I’m not sure embarrassing. I think it was a study that was important but was not impactful. And that was the carboplatin etoposide versus platinum etoposide. It’s important to do studies, but I remember as the data came in and I realized it was failing, I was disappointed.

And I remember as we were sitting down and talking about, we said, “This is truly disappointing.” It’s almost devastating that you are planning for something to be better, and it turns out to be worse in a highly [crosstalk 00:31:51]

Brian:
Happens to Tom all the time.

Tom:
Story of my career.

Brian:
Throughout his entire career.

Tom:
I wouldn’t … yeah. Okay. This should be easy. York or New York? In the knowledge, there are similarities between York and Leeds.

Dean Bajorin:
So, New York got its name because it was the new home for the duke of York. And so, I’ll take the new home for the duke of York.

Tom:
That seems reasonable, although that’s actually Prince Andrew at the moment, which is quite complicated.

Brian:
Favorite location you’ve ever been invited for a talk?

Dean Bajorin:
I would say Southern France, Northern Italy. That stretch. I’ve given a couple talks up there. I really love both countries. My wife’s first language is French. She’s from Europe and I actually lived in Italy for a few years. And so that’s a favorite of ours.

Tom:
Dean, this has been amazing and terrific. I think that it’s clearly been a really … I’ve learned quite a lot about the origins of urothelial cancer today. I’ve got to say thanks so much for joining us. I know it’s really valuable time of yours.