Validation of Keratin 17 as a Sensitive Marker for Upper Tract Urothelial Cancer - Woodson Smelser

January 27, 2025

Sam Chang discusses with Woodson Smelser research validating keratin 17 (K17) as a biomarker for both bladder and upper tract urothelial carcinoma. Dr. Smelser presents findings from multiple studies showing K17's high sensitivity (96%) and negative predictive value (98%) in bladder cancer, with new validation work demonstrating similar effectiveness in upper tract disease. The biomarker's ability to detect both low and high-grade tumors through differential expression patterns makes it particularly valuable as a potential screening and surveillance tool. The discussion explores K17's potential applications in reducing invasive procedures and radiation exposure, particularly in hematuria workup and surveillance of low-risk patients. Future directions include prospective studies examining K17's role in risk-adapted surveillance strategies and its potential to optimize screening protocols, especially in microscopic hematuria evaluation.

Biographies:

Woodson Smelser, MD, Assistant Professor of Surgery, Department of Surgery, Division of Urologic Surgery, Washington University in St. Louis, St. Louis, MS

Sam S. Chang, MD, MBA, Urologist, Patricia and Rodes Hart Professor of Urologic Surgery, Vanderbilt University Medical Center, Chief Surgical Officer, Vanderbilt-Ingram Cancer Center Nashville, TN


Read the Full Video Transcript

Sam Chang: My name is Sam Chang. I'm a urologist in Nashville, Tennessee. And we have really a great privilege and honor—I personally have a great privilege and honor—to introduce Dr. Woodson Smelser. Woody is actually an assistant professor at Washington University in Saint Louis in their Department of Urology, and he is a former fellow at Vanderbilt. One of our best Fellows that we've helped to work along and train. And he has done some outstanding work looking at a urinary marker for bladder cancer called keratin 17. And so, Woody, we look forward to your presentation. And obviously, the great things that you've done at WashU. And so I think you've got a presentation based on a recent manuscript looking at validation of K17 as a marker.

Woodson Smelser: Yes. And thank you, Dr. Chang, and thank you for the folks at UroToday for allowing us to showcase our work. So as Dr. Chang alluded, this is work that we did in upper tract urothelial carcinoma with a known and kind of validated biomarker in the bladder, called keratin 17. And keratin 17 is a biomarker that we now believe is reproducible in urothelial cancer, not only in the lower urinary tract, but also in the upper urinary tract.

And this is an oncoprotein that is produced early in the tumorigenesis of urothelial cancers. But that expression of keratin 17 is actually maintained throughout tumorigenesis both in low and high grade cancers. And so it makes it a somewhat ideal biomarker for detection because it can be detected at various stages and grades of urothelial cancer. There are now six published, independent prospective studies with more than 3,000 urine specimens that have looked at keratin 17 in the bladder, and they've shown a sensitivity of about 96%. And importantly, a negative predictive value of 98%.

And since this is helpful at detecting both low and high grade tumors, it is now available as a urinary marker. And it even has the ability to detect very low grade lesions like PUNLMPs.

So this is a summary slide that shows the kind of aggregate data and the average calculated sensitivities and specificities looking at keratin 17 in the bladder. And the important thing here is these were mixed samples. These were patients who either had an established diagnosis of recurrent urothelial carcinoma or people who were undergoing initial diagnostic workup for gross hematuria. And you can see, sensitivities have ranged anywhere from 100% down to about 92% with an average sensitivity of about 96%, and a negative predictive value of 98%. So obviously, a lot of the predictive features that we would look for in a good urine biomarker.

And so, at present, urologists who are using this—and this is a commercially available assay—use it as a rule-out test. And I think the potential there is to rule out unnecessary cystoscopy. But in performing these prior studies, there were several findings that indicated maybe patients had false-positive K17 urine results, but then when those results were scrutinized further and people did more workup using things like ultrasound or CT urogram or even ureteroscopy, what they actually found is oftentimes the test was correct and that there was a urothelial cancer present. It just wasn't in the bladder.

And so I think we all have experience as urologists in this realm, particularly with cytology, where we see a positive cytology, positive for high grade cancer, we don't see a lesion anywhere. We may even do things like enhanced cystoscopy in the bladder and still can't find anything. And obviously then we have to do things like prostatic urethra, urethral biopsy, or mapping biopsies or, importantly, upper tract urothelial carcinoma evaluation. And that's what really led to our hypothesis and study protocol to evaluate this in the upper tract, as these sort of earlier instances that were thought to be false positives.

And so our hypothesis was that, as in the bladder, K17 would also be a sensitive biomarker in the upper tract. And so we have a recent publication, as you alluded, in Human Pathology, looking at our validation study of this in confirmed cases of upper tract urothelial carcinoma versus normal controls. And it's important to note that there are a lot of collaborators in this project across disciplines. And I would be remiss if I didn't acknowledge our excellent expert GU pathologist, Dr. Giovanna Giannico, who really helped with not only the conceptualization of this project, but also the actual immunohistochemistry and examination of the stains.

But what we did is we took 139 cases of known upper tract urothelial carcinoma from a tissue biobank. And then we evaluated those using the keratin 17 assay. And we looked to define sensitivity and specificity characteristics across a spectrum of grade and stage, and compared these to normal controls of normal urothelium from kidney specimens that did not have upper tract cancer. And you can see from our results there, this test remained quite sensitive and specific in the upper tract setting.

So this is just an example of the immunohistochemistry. If you look at slide A and B here, you can see this is actually normal urothelium. And on the right with the immunohistochemistry stain, you see there's a basal pattern of expression of keratin 17. So this is present normally. But then you can see in slide D that with a higher grade upper tract tumor there's sort of pan-urothelial expression, not only in the basal layer, but in the deeper layers as well.

When we looked at how this sort of distributed across different stages, you can see that normal urothelium had more of a basilar pattern, and abnormal urothelium with upper tract carcinoma across stages Ta, Tis, and even T1 and T2 had a greater percentage that were a non-basal pattern.

And if you look at the receiver operating curves here, you can see this top line demonstrates that when you look at expression on a basal versus non-basal pattern, K17 performed very well with many more having a non-basal pattern. And we also used what's called an H score, which is another pathologic measure that performs similarly. And you can see the ROC on both of these is 0.83 and 0.81, respectively. So it performed well from most traditional measures of a biomarker in that regard.

So I think our key takeaways: K17, like in the bladder, is a sensitive and specific marker of urothelial neoplasia. And really, the pattern of expression—whether basal or non-basal—is more important than simply positive versus negative. And that's where you get the most discrimination in terms of performance. And I think this really indicates that in the future, this could be another tool for us to use to rule in or rule out the presence of upper tract cancer in patients in whom we see a normal bladder and could potentially decrease invasive tests like ureteroscopy.

Sam Chang: Woody, that was great. At this point we have so many different markers. Let's focus first on keratin 17, especially with some of the IHC that you showed. The ability to be a urinary marker is great. We clearly, though, are looking at other expression markers that have now become targets for therapy. Is there a chance that K17 could become another one of these targets—an antibody-drug conjugate–type target for therapies, especially with this preferential expression?

Woodson Smelser: I think potentially. I think the biggest limitation, though, is that, though it's an oncoprotein, it is expressed in normal urothelium as well, which is what some of those immunohistochemistry slides show. But if we could demonstrate that keratin 17 coexists with, say, another oncoprotein, that could be a more specific target. And I do see potential therapeutic applications in the future. But I really think if you talk to patients, what they're most concerned about is decreasing the invasiveness of treatment for both bladder and upper tract cancer.

And particularly, when you think about upper tract cancer, with a very much older and oftentimes more frail population—and a population that's skewed toward patients in whom surgical interventions or even surgical diagnostic tests like ureteroscopy may be high risk—I think that there's a significant value add in being able to do noninvasive diagnostic testing with overall high fidelity.

Sam Chang: So that leads to my next question. We have so many different tumor markers out there. We still—we always talk about cytology as being the gold standard, with all the caveats of the fact that it really isn't that sensitive. It is specific. Tell me, as you think about the different markers that are out there currently available, how do you fit in keratin 17 or K17 in your kind of diagnostic algorithm?

Woodson Smelser: Well, I would say it's very early obviously, since these are newly published results. But I think the way in which I would utilize it moving forward, especially as we plan further validation in prospective trials, would be in patients who are low-risk for recurrence but have a history of established upper tract cancer. I think it could simplify the surveillance and reduce the number of ureteroscopy evaluations, or even cross-sectional imaging tests that need to be used.

I mean, if you look at our recently published AUA guidelines on upper tract urothelial carcinoma, we have a very nice risk-adapted surveillance stratification with a very nice table that I know you and other authors contributed to, but it's still a very intense regimen. And even for patients who've had low-grade upper tract disease diagnosed with maybe even a small lesion that you just lasered or biopsied and removed, or maybe you treated them with Jelmyto and they've had a period of remission. It's a lot of cystoscopy. It's a lot of cross-sectional imaging.

And I really think if you look at value-based care and the ability to reduce cost, this is where we're going to make inroads. Because I just think for decades now, we've said cytology is the gold standard, but it does not have a high enough discriminatory capacity or capability to avoid those more invasive interventions and give patients the reassurance that they need that they haven't had a recurrence.

Sam Chang: Now, I think you're right in terms of avoiding not only the invasiveness of ureteroscopy, but the cost and radiation exposure, perhaps, of CTU—that combination of, hey, what do we use in replacement, perhaps delaying? Obviously, you're not saying nor is anybody saying that, hey, we can avoid any future ureteroscopy, we can avoid any cross-sectional imaging.

No, but instead of that, being able to just, as you say, put off one or two of those examinations, avoid the radiation exposure, avoid the trips to the operating room, especially as you take into account the risk-adapted strategies—it definitely makes sense. And to be able to delineate both low grade and high grade disease, I think really kind of helps separate keratin 17 as a marker for upper tract disease. So how do you use it, Woody, for bladder cancer patients at this point?

Woodson Smelser: Yeah, so I think it's very useful in patients who have a history of bladder cancer, and maybe they're outside the—and this is my personal bias again—but maybe they're outside the early intense screening regimen, and maybe they have discomfort with cystoscopy. And we all have these patients in our practice who maybe they're older or frail and they say, "Look, I did cystoscopy for 10 years and this never came back. But I don't want to just trust that everything's OK." And oftentimes, we'll do urinalysis with microscopy or cytology. I think this is a great patient in whom to consider that, because of the high negative predictive value of the test.

And I do think that an area where really this should be explored further is in hematuria workup, where I think all discriminatory urothelial cancer biomarkers have offered some promise. And I think we as urologic oncologists and then also our colleagues sort of in public health have not figured out the magic algorithm to perform cystoscopy or more invasive testing in the highest-risk patients, but avoid it in your common hematuria patient, particularly microhematuria, who may not need as extensive of a workup. And so I think that's really where there's additional promise because of the discriminatory characteristics of the test, which, as the table showed, has been validated in multiple studies at this point.

Sam Chang: Yeah, I think that's a really good point. Deescalating our evaluation of microscopic hematuria—I think the AUA guidelines have made great inroads in deescalating the evaluation, what needs to be done, when it needs to be done. Kudos to Dr. Boorjian and Dr. Barocas who helped lead those efforts. But then, like you say, if we can have a urinary marker that then really puts off or avoids the invasive procedures or even the imaging, I think it really makes sense.

Especially since, as the vast majority of individuals really don't need that evaluation. So I think really important points. Woody, last question: Where do you think K17 needs to go next in terms of studies? Are you doing anything specifically, or where do you think next steps would be in terms of helping validate or further show that this tool is something that would be very helpful for our practicing physicians?

Woodson Smelser: So I sort of envisioned two studies—both would be prospective. I think the first would be randomizing people to standard-of-care surveillance for low-grade or lower-risk upper tract using the AUA-validated stratification scheme, but substituting keratin 17 at certain evaluation points and looking at rates of recurrence, progression, et cetera, or just the discriminatory ability between direct visualization with ureteroscopy—which is what normally we would do for a patient, say, with low-grade papillary tumor in the renal pelvis—versus using keratin 17 at some time points.

And then I think the second area that I would look at would be in microscopic hematuria, especially now that we know this is valid in upper tract urothelium. And that's really the piece that was missing before this published study: we knew the biomarker was valid in the bladder, but we didn't know if there were differences in the upper urinary tract that maybe decreased its ability to detect true-positive cases. And now that we—at least through one study—see that upper tract urothelium presence of keratin 17 can be predictive, I think that allows K17 to potentially be a more catch-all test in the future. But that's of course going to have to be validated in a screening population.

Sam Chang: I think great points, and I think clearly to piggyback on those ideas would be that we have now therapeutic options for these patients, and this would be for not only advanced disease but for noninvasive disease, non-metastatic disease. But it would be great for any future therapeutic trial to measure those levels prior to treatment and then to measure afterwards. So really exciting points.

Dr. Smelser, young Woody, thank you so much. It is always a pleasure, and I can't hide my pride in how much you've done and how much you've accomplished, and look forward to future publications and future studies. And we wish you well. And we look forward to actually other times together on UroToday.

Woodson Smelser: All right. Thank you very much for having me as a guest.