Agent Orange and Bladder Cancer: Insights from a Study on 25 Million Veterans - Stephen Williams

May 30, 2023

Stephen Williams and Sam Chang discuss the relationship between Agent Orange and bladder cancer. They highlight a retrospective cohort study conducted on over 25 million veterans to determine the association between Agent Orange exposure and bladder cancer incidence. The study controlled for various factors and found a slight increase in bladder cancer risk among veterans exposed to Agent Orange. The study also employed a natural language processing model to assess muscle invasion in bladder cancer cases. The findings support the designation of bladder cancer as an Agent Orange-associated disease and emphasize the importance of providing resources and services to affected veterans. Future research will explore additional variables such as race, treatment modalities, and stage at diagnosis to improve outcomes and identify high-risk patients.


Stephen B. Williams, MD, MS, FACS, Chief, Division of Urology, Director of Urologic Oncology, Director of Urologic Research, Co-Director of Department of Surgery Clinical Outcomes Research Program, Medical Director of High-Value Care, UTMB Health System, Galveston, TX

Sam S. Chang, M.D., M.B.A. Patricia and Rodes Hart Endowed Chair of Urologic Surgery Professor Department of Urology at Vanderbilt University Medical Center

Read the Full Video Transcript

Sam Chang: Hello, my name is Sam Chang. I'm a urologist in Nashville, Tennessee, and I work at Vanderbilt University, and I have the honor and privilege today to spend some time with Stephen Williams. We were joking, before we started, about how many titles and roles that Stephen currently has, but he is the Professor and Chair of Urology at UTMB. He's also, I think, the Chief Medical Officer at UTMB, as well. His executive leadership roles have been only increasing, as he has been able to deliver on efficiency and healthcare quality. Those are other things that, hopefully, we'll be able to talk about in the future, but today you're going to focus on, actually the relationship between Agent Orange and bladder cancer, a really important abstract in a AUA 2023. So I look forward to hearing about the abstract, and thanks so much for being here.

Stephen Williams: Well, thank you so much, and thank you UroToday for inviting us to speak again. Sam, you're a good friend and I'm really excited about this work. It's not my work, but really as you know, it's a team. And what's beautiful about this is, actually I leveraged a team out in the VA. So Durham, actually in North Carolina, and Stephen Freedland, who has a large body of work actually in prostate cancer, and then really elucidating the association of prostate cancer and Agent Orange. So, I thought to myself over time, and you know that relationships are critical, is to leverage that and determine actually the association, if there's any, between Agent Orange and bladder cancer.

So what we ended up doing is, getting together our minds, and then really performing the largest retrospective cohort study that has been done to date, controlling for key factors, such as smoking, age, deployment, race, also to service entry, all these other factors that are critical. But what was really exciting is actually, also at that time to help fund this, the Department of Defense CDMRP, actually elicited an RFA, to perform, not necessarily this study... you know, you had to read between the lines, but the Institute of Medicine determined that there was no strong association to determine whether or not bladder cancer could be an AO-associated disease, AO being Agent Orange. So anyway, this was kind of a, I don't know if it's a calling, if you will, but really-

Sam Chang: Almost at that point it stimulated more research in terms of trying to help answer that question, yeah.

Stephen Williams: Exactly. And also in bladder cancer, in general. As you know, you're a longstanding expert, well recognized that finally we're getting the funding to support the investigation in this disease. So getting back to this study, is a little over 25 million veterans-

Sam Chang: 25 million.

Stephen Williams: Correct. That were eligible for inclusion, between the period, its age... or year 2001 to 2019. And the reason for that is, their VA EHR record, which is VINCI, started 2000. So that was really how we were able to put this cohort together. And then we matched in a criteria of 1:3, AO exposed, unexposed, by age, as well as looking at race, service entry, as well as the service branch, to then have some exclusion criterion. And at that time in the military, there weren't a large body of females, so we restricted the cohort to males, which is a caveat of this study, but it's important that we really try to control, not only for the compound, as we mentioned, such as smoking, but then look retrospectively on the incidence of bladder cancer. Of note as well, which is key, is prior to that year 2001, we excluded patients with a prior diagnosis of bladder cancer, as-

Sam Chang: So these are de novo, not exactly... got it.

Stephen Williams: As well as, also death, prior to that, and also lost a follow-up. So it's a real restricted cohort because, as you know, this has important implications for our service members and the VA itself. So getting to the meat of everything, we ended up finding an increased risk of bladder cancer. And also as a part of this cohort, the VA registry, they have a registry, but roughly two thirds of the staging is not present. So we had to develop a natural language processing model, which is a total separate topic.

Sam Chang: Yeah. Project... of how you did that, absolutely. Okay. All right.

Stephen Williams: So, using people much smarter than myself, data scientists, they were actually able to develop this NLP model to determine muscle invasion, yes or no, and has a 94% accuracy, which we published this data in the Journal of Clinic... JCO Clinical Informatics. So anyway, getting back to this one, we found the increased risk of bladder cancer, but it was slight. It was 3% increased risk. But then we also wanted to look at aggressiveness. And that gets into the NLP model. And we found actually a decreased risk of muscle invasive bladder cancer. Granted, this is indexed, and we're looking at the largest equal access system in the US. So in turn, those are the two major findings that we have.

There's other segues using this data and really leveraging this database that we're looking into. But very exciting results. Prior to this work, as I mentioned before, the VA did not designate bladder cancer as an Agent Orange-associated disease. And now they have, and I think this work really helps support that. And really helps support our service members, their families, to getting the benefits and the services they deserve. You know at Vanderbilt how critical that is to ensure our veterans have their due diligence, but also to our government. There's a large body of money that goes into that funding, that it's appropriately allocated.

Sam Chang: Right. And I think that combination of, they have to have the respect that they're due, number one, but then properly allocating the funds in terms of understanding that, hey, this is scientific data proving, showing this relationship, taking into account the other factors that you can take into account, is actually really important. I mean, because that's not been done before, correct, Stephen?

Stephen Williams: Correct. I mean, there's studies that have been done, however, they're either smaller numbers, not controlling for smoking for instance, or even Agent Orange exposure itself. But what's important is, this is an epidemiological study. And although we use the VA health system, which is the largest health system and largest cohort of veterans, that we can assemble, this designation doesn't come lightly by the VA, but it is a surrogate variable, and not biological. So we don't know concentration, dose response, all these caveats. But still, at the same time, it does highlight a moment of pause, given our other conflicts that we've had, the Iraq conflict as well, burn pits, I think. There's a number of initiatives, as you know, and we work alongside, partner, with BCAN, for our survivors, in trying to better understand other toxicology type risks.

Sam Chang: That may... you know, you start off with a relationship and then you start understanding the possible independent impact of these variables, I think is really important. I'll be honest, I think for a lot of the younger listeners or residents or Chinese, their familiarity with the exposure that Vietnam veterans had, I think is very minimal. So to me, I had no idea that in terms of tracking... so every veteran that was exposed in any way to Agent Orange, there's a designation or there's something on their record. Is that correct?

Stephen Williams: Correct, yes. In the VINCI system. So there is a variable that's there to extract, and it's yes or no. But what's critical to get that designation, this was very interesting for me, not only did they need to be in a conflict, particularly Vietnam in that era, they also need to have an AO-associated disease, or, also too, they have to have an AO-associated disease that occurred in roughly 10% of their occurrences and interactions with the VA, unless it's like an acne or some other, where it needed to occur within one year of the service date. So they're really strict.

Sam Chang: It's pretty strict, yeah.

Stephen Williams: And then after that, it has to go through a committee to designate whether that veteran, yes or no, because the implications are quite surmount... obviously providing the resources in our healthcare climate in general. So it's pretty important.

Sam Chang: So there are 25 million though, that met the criteria-

Stephen Williams: Correct.

Sam Chang: Of Agent Orange?

Stephen Williams: Well, met the criteria of inclusion, whether or not they were exposed or unexposed. And then restricting down was a little over 2.5 million.

Sam Chang: Got it.

Stephen Williams: That were actually included in our study.

Sam Chang: I got it.

Stephen Williams: That were the matched controls, 3:1, to those that were exposed versus not exposed.

Sam Chang: Right. So the impact on this is, I think, quite significant. So what's the next step? Stephen, what's next here?

Stephen Williams: Well, I think before the next step is understanding the limitations of this study... is a male only study, VA, some may contemplate, it's not generalizable, but it's a veteran's study. Also, too, it was tremendous with the millions of veterans that were included, were overpowered to detect differences. You combined the p-value. And so we actually did standardized differences. It was well to better understand that. So putting this all together, you know, you can't definitively hang your hat, but it's pretty certain that there is an increased risk of bladder cancer, which biologically and just common sense would-

Sam Chang: Makes sense.

Stephen Williams: Correct. Yeah, exactly. So the next steps really is, we're presenting a number of other abstracts because it's the largest equal access healthcare system. Race in bladder cancer's important to understand, treatment modalities, radical cystectomy, trimodal therapy, definitive therapies, also looking at stage at time of diagnosis.

What's interesting here, and supports our findings, there's an increased risk of bladder cancer among patients that smoked. But we found also a decreased incidence among African Americans, which corroborates our prior knowledge. Historically, Caucasians have a higher increased risk of bladder cancer, and then also African Americans decreased. So that was interesting. But also too, in this study, which is hypothesis generating, is those with muscle invasive disease... remember I mentioned Agent Orange and decreased of muscle invasive or the aggressiveness. But we found increased risk of aggressiveness among African Americans versus Caucasians, which also, too, is something that we've observed in other large population-based studies, general population-based studies.

Sam Chang: Because they seem to have worse disease. But it's interesting though, the incidence though, seems to be lower, but if you have the cancer, if you develop it, it tends to be more dangerous.

Stephen Williams: Yes, more aggressive in this study. So we're actually investigating that further, because that's very interesting in an equal access center, what we're observing, and hopefully help inform just really how we're able to hopefully identify the highest of high risk bladder cancer patients, and really improve our outcomes. And it also is another segue. It's one of the most costly cancers to treat. So with increased novel treatments, appropriate allocation, identification-

Sam Chang: Of those that are really high risk, and even those that are perhaps even lower risk-

Stephen Williams: Correct.

Sam Chang: Being able to actually then match up the treatments. Ah, that's fantastic. Well, I think... I mean, this accumulation of data will... I mean, the number of studies I know that your team already must be working on or thinking about, I can't even... it's going to... you have a lot of interviews, Stephen. That's all I got to say. There's a lot of things to discuss, but-

Stephen Williams: Well, hopefully not with me. As you know, you pay it forward. And you have one of your... returning.

Sam Chang: Yeah, absolutely.

Stephen Williams: Angel's coming back. Dan Joyce, just a remarkable gentleman. I've had the pleasure of working with, learning from them. Kelly Bree, she's at MD Anderson.

Sam Chang: That's right. Yeah.

Stephen Williams: Stayed on faculty there, and a number of others. So, I think one thing that I love about this is, sharing the opportunity for others to learn, but for me to learn, and really pay it forward and hopefully help others' careers. But also, too, important underpinnings and understanding of this disease at large.

Sam Chang: Stephen, thank you so much for all your efforts, your team's efforts, what you've accomplished, and what you've been able to actually educate in terms of all treating physicians and patients, as well. And thanks for spending some time with us.

Stephen Williams: Thank you very much.