utDNA Could Provide Noninvasive Tool for MRD Detection Prior to Repeat TURBT in High-Risk NMIBC

Joshua Linscott, MD, PhD

Evaluating urinary cell-free tumor DNA (utDNA) alterations and copy number burden (CNB) prior to repeat transurethral resection of bladder tumors (rTURBT) was shown to accurately detect minimal residual disease (MRD) in patients with high-risk non–muscle-invasive bladder cancer (NMIBC), according to Joshua Linscott, MD, PhD, who highlighted the potential for this noninvasive biomarker to provide a reliable indicator of benefit with maximal resection.

Findings from a prospective study evaluating the predictive value of utDNA for MRD in this patient population were presented at the 2024 ASCO Annual Meeting. Data showed that mutations detected in utDNA were concordant with the tumors of individual patients. Tumor fraction and copy number burden (CNB) detected alterations such as single point mutations and chromosomal loss or gain.

Detected tumor fraction and CNB were higher in patients with MRD. The tumor fraction from urine samples collected prior to rTURBT showed an area under the curve (AUC) of 0.81 with a threshold of 0.039%; the AUC for CNB was 0.85 with a CNB threshold of 6.14. Notably, combining CNB and tumor fraction in a stepwise model improved the performance of the assay, achieving sensitivity, specificity, and overall accuracy rates of 92.3%, 85.7%, and 90.0%, respectively, in the validation cohort.

“[Cell-free DNA assays (cfDNA)] are going to be the future,” Linscott, a urologic oncology fellow at Moffitt Cancer Center, in Tampa, Florida, explained during an interview with OncLive^®. “Sometimes we dive into projects as scientists and clinicians and realize there might be a ceiling to what you can do. This technology is early on [in development] and relatively untapped, but I think it’s going to change things for the better for our patients.”

In the interview, Linscott discussed the rationale for using utDNA as a noninvasive biomarker for MRD detection in high-risk NMIBC, the results demonstrating the overall accuracy of MRD detection when combining tumor fraction and CNB assays of utDNA, and potential applications of this approach during the surveillance period and for other bladder cancer subtypes.

OncLive: What was the rationale for evaluating the accuracy of using tumor fraction and CNB from utDNA as biomarkers for detecting MRD?

Linscott: Most patients with high-risk NMIBC are going to experience disease recurrence at some point. Most of them won’t progress, but a small number will. The follow-up [for patients with high-risk NMIBC] is currently fairly invasive. We look in their bladder with cystoscopy, usually every 3 or 6 months, until they do [experience] recurrence.

There are a couple of major things that this patient population needs. One is a noninvasive test that could accurately detect whether they have disease. The other question is whether we can find a way to determine who’s going to be at risk of disease progression and who is going to benefit from which treatment and when. Those are still important questions in the field that we’re just starting to address.

What prior research regarding the predictive utility of utDNA supported this prospective analysis?

cfDNA is an exciting field right now. Most of us are aware of how it’s revolutionizing cancer care across the spectrum. There have been studies from our group [at Moffitt Cancer Center], as well as others, showing that cfDNA is released into the urine by these bladder tumors. In the localized disease setting, it seems to be the best way to look for things. We know that in early-stage disease, patients do not have the same level of cfDNA reaching the plasma or circulating tumor DNA, which we usually talk about. However, urine is constantly bathing these bladder tumors, and as they release these genomic contents it becomes detectable within the urine.

Why was utDNA evaluated in this particular patient population?

If you are looking for a patient population that does not have a high tumor burden, where they no longer have that original tumor that is going to be high volume and potentially releasing a lot of cfDNA, you’re now looking in the MRD setting. This is similar to an early recurrence, where the patient doesn’t have a lot of tumor present yet.

The rationale for this study was that if you take this population of high-risk patients who have an original tumor resection and are already going to undergo a repeat resection as part of the standard of care, you have, in essence, a population that should have MRD. [Accordingly,] you can evaluate whether the utDNA test works. You can then potentially extrapolate—[provided] that you perform the [appropriate] studies—that this approach will also work [for patients experiencing] early recurrence. Here we had a population where we knew a number of patients would have residual disease, so it would be a good way to demonstrate the power of the test and define how well it works.

What should be noted about the study design?

This was a prospective study. We had a cohort of patients who were in need of either their initial resection or, oftentimes, a repeat resection, with us being a referral center. We collected samples prospectively as patients received treatment to look at the MRD setting and likely the surveillance setting.

What key results from the study were presented at the 2024 ASCO Annual Meeting?

A couple important findings came from this study. Maybe the most important is that utDNA seems to be a good tool for detecting MRD within this patient population. There are multiple [genetic contents] you can look at within the urine. There are large structural changes to DNA, [such as] gain of an entire gene or loss of a chromosomal arm. You can detect those with relatively low-pass whole-genome sequencing. As bladder cancer has some of the highest number of copy number alterations, it’s a good marker for residual disease. However, not everyone’s going to have that, so we also looked at some of the individual point mutations and single nucleotide variations. We did this with a personalized assay based on patients’ baseline tumors. These bespoke assays, through which we’re looking for key mutations within the patient, provide us with a tumor fraction.

We found that tumor fraction does a good jo detecting disease when it is positive, but it’s not the most sensitive. Therefore, we combined it with the CNB score. Together, the 2 scores ended up leading to a 90% overall accuracy [score] for detecting whether these patients had disease or were disease free.

What are the potential implications for and the applications of utDNA in bladder cancer based on these results?

There is so much we can still do with this. [ Although this study was conducted] in the MRD setting, there are huge implications for surveillance. We will be looking at how well this does in the surveillance setting. [Additionally], I think this will work well for patients with muscle-invasive bladder cancer and patients undergoing neoadjuvant treatment.

We’ll also be looking at prognostic factors and signatures that predict who will respond to what treatments, especially as we expand our repertoire within this space. We would also like to expand [its use] to patients who maybe have more advanced MIBC and undergo neoadjuvant chemotherapy to determine who might be a better candidate to spare their bladder because they’ve achieved a complete response.

Reference

Linscott J, Miyagi H, Gould B, et al. Tumor fraction and copy number burden from urinary cell-free tumor DNA (utDNA) to predict minimal residual disease prior to repeat-transurethral resection in high-risk non-muscle invasive bladder cancer (HR-NMIBC). J Clin Oncol. 2024;42(suppl 16):4600. doi:10.1200/JCO.2024.42.16_suppl.4600