Eroglu Examines the Clinical Utility of ctDNA as an Emerging Biomarker for Melanoma Treatment

Serial plasma circulating tumor DNA (ctDNA) dynamics appear to correlate with response to anti–PD-1 therapy or disease relapse in patients with melanoma, underscoring its potential as a noninvasive biomarker, according to Zeynep Eroglu, MD, who added that remaining questions include defining clinically meaningful lead times, optimal sampling schedules, and cost.

In a presentation delivered during the 2021 Society Melanoma Research Congress, Eroglu detailed research exploring the utility of the personalized and tumor-informed ctDNA assay, Signatera, in monitoring for minimal residual disease (MRD), response, and recurrence in different subsets of patients with advanced melanoma.

“From our research and many other groups, the serial plasma ctDNA dynamics do seem to correlate with response to immunotherapy, and it may potentially be an early marker of treatment response to immunotherapy and other systemic therapies, or a marker of eventual disease progression,” Eroglu said. “For our patients in the adjuvant setting, the serial ctDNA dynamics may correlate with eventual disease relapse, and there may be this lead time where the ctDNA in the plasma becomes detectable before clinical radiographic relapse. However, we need to clinically validate these assays.”

In an interview with OncLive^®, Eroglu, medical oncologist, the Department of Cutaneous Oncology, Moffitt Cancer Center, assistant professor, the Department of Oncologic Sciences, the University of South Florida, Morsani College of Medicine, discusses the potential of ctDNA as a noninvasive biomarker in melanoma, outlines different assays under development, and details potential opportunities to leverage ctDNA in treatment decision making.

OncLive^®: How might ctDNA represent an emerging noninvasive biomarker in melanoma?

Eroglu: There is a lot of interest in ctDNA, as detected in the plasma. One of the challenges we [face with our] patients with melanoma is that we do not have a great noninvasive biomarker the way other solid tumors may have, such as prostate-specific antigen and prostate cancer. We must use imaging and clinical exam to monitor our patients for disease recurrence or disease progression while they are on treatment or after surgery. However, in terms of how frequently we can do either CT or PET CT scans, [we] may be limited to every 2 to 3 months.

With ctDNA, [we want to explore] whether it is possible [to use this] to detect relapse earlier than clinical or radiographic relapse, or to give us an idea of how a patient may be responding to a drug, such as immunotherapy, earlier than the time point in which we would do scans. [ctDNA] may provide, particularly in melanoma, a minimally invasive opportunity to be able to monitor and characterize the disease status in our patients.

You recently presented a 3-cohort pilot study that leveraged the personalized, tumor-informed Signatera ctDNA assay for MRD detection and treatment response monitoring in patients with advanced disease. What inspired this study and what were the key objectives?

[We have seen] a lot of great research that has evaluated ctDNA assays in several solid tumors, including in those with melanoma, either in the adjuvant setting or [in those] undergoing immunotherapy or targeted therapy. These studies, including those in melanoma, have looked at a variety of different plasma ctDNA assays. More commonly, they have used a digital droplet PCR assay, which generally tracks 1 specific mutation, such as BRAF V600E ctDNA. At Moffitt Cancer Center, with our patients, we have also looked at this digital droplet BRAF V600E assay.

The Signatera assay had been looked at in various other solid tumors, such as [those in the] colon, lung, and breast in the adjuvant setting, and it appeared to have very high levels of sensitivity in detecting relapse. [As such, we considered whether] we could use this multiplex PCR assay that could look at multiple mutations in the blood in our patients with melanoma. Specifically, [we wanted to evaluate its use in] the adjuvant setting where we monitor our patients for relapse, in the metastatic setting where patients are receiving treatment, and in the post-immunotherapy setting where patients have completed their treatment.

How is this specific test utilized, and how does it differ from other tests in development?

There are 2 main categories of ctDNA assays. There are tumor-informed assays, which track mutations that are specific to a patient's tumor and you would expect to be present in all cancer cells. [Examples include] the digital droplet PCR or multiplex PCR [assays]. The 1 disadvantage [with these assays] is that you must have tumor tissue to be able to figure out the mutations that you are going to be following.

There is also a category of tumor-naïve assays, and these are all commercially available and can detect ctDNA using this pre-selected panel of actionable mutations in advanced cancers. However, there are limitations that come with using this fixed gene panel approach with regard to sensitivity and specificity, as opposed to the mutations you already know are in the patient's cancer. The advantage is that you do not necessarily need tumor tissue with this ctDNA assay.

With regard to the tumor-informed assays, there are various methodologies, but those that have been most evaluated in melanoma have been the digital droplet PCR [assays]. These look at 1 specific mutation, generally BRAF V600E, which we can find in about 40% to 45% of our patients with advanced melanoma; that is an appealing mutation to be able to follow. Other mutations to follow include NRAS Q61K, but generally, we track 1 mutation.

One of the downsides with digital droplet PCR is that when you are tracking 1 mutation, there are many patients with advanced melanoma who do not have BRAF or NRAS mutations. That makes it difficult to be able to follow their cancer with this methodology. There are also high levels of tumor heterogeneity, so the [question] is: If you are able to follow several tumor-specific mutations, will that give you a more robust representation of the patient's tumor dynamics and burden?

The [assay] used in this project was a multiplex PCR ctDNA assay, specifically Signatera. Available tumor tissue is sequenced, and then there is a personalized PCR assay that is designed for that patient that targets the top 16 clonal mutations in the tumor. Any serial plasma samples that are obtained from the patient can then be tested with this personalized ctDNA assay to look for the presence of ctDNA, and this can be quantified. There is a number that can be given as tumor molecules per mL, or it can be reported as undetectable.

What was the design of the study, and what were the methods utilized?

We were interested in [looking at] this multiplex PCR ctDNA assay in 3 specific cohorts. The first included patients with stage III melanoma who had recently undergone surgery and were about to start adjuvant systemic therapy, specifically anti–PD-1 therapy. [The goal was to] look at the prevalence of ctDNA detection at baseline, post-surgery, and with serial monitoring, and [determining] how that may correlate to disease relapse. We obtained the plasma samples every 4 weeks from the patients.

In the metastatic setting, we were looking at patients receiving immunotherapy, either anti–PD-1 monotherapy or combination therapy, such as ipilimumab [Yervoy] and nivolumab [Opdivo], in the first-line setting. [We examined] how well serial ctDNA dynamics correlated with tumor response, disease progression, and [we explored] its relation to progression-free survival [PFS] and how this could help with pseudoprogression.

Our third cohort was comprised of patients who completed immunotherapy. [Maybe at approximately] 2 years, we decided they were done with treatment, and they were just put on surveillance. [We studied] how starting ctDNA monitoring at that time and moving forward could help with potentially detecting relapses earlier than CT or PET scans.

What were the findings from the study? Were there any results that you found to be particularly interesting?

Our findings mirrored some of the research that has already looked at ctDNA in the adjuvant setting in solid tumors. We were able to look at 28 patients with stage III melanoma post-surgery. Twenty-four percent had detectable ctDNA at baseline. These are post-surgery patients who have been rendered disease free, both clinically and radiographically. For patients who ultimately had disease relapse, we were able to detect 75% with the ctDNA assay, prior to radiographic or clinical evidence of relapse. The average lead time before radiographic relapse, where the ctDNA assay became detectable, was about 3 months. [It was] a small number of patients, but some interesting data with regard to the potential to detect relapse before we start to see it on scans or clinical exam.

The median follow-up in these patients was about 15 months, so we need to follow them longer. In terms of looking at relapse-free survival [RFS], we found that compared with the baseline post-surgery time point, the longitudinal monitoring, looking at ctDNA values every 4 weeks over time, was more informative.

We found that the median RFS was about 6.7 months for patients whose plasma ctDNA was undetectable [and] became positive, or if it was positive in the beginning and remained positive, as they never cleared it. That cohort patients had a poor median RFS and relapsed fairly quickly vs the median RFS that has not been reached in the other group of patients with undetectable ctDNA, or who were able to clear it.

In the metastatic cohort, the patients who were receiving immunotherapy in the first-line setting, 92% [had] ctDNA detected at baseline, [which is not] surprising since these are patients with stage IV melanoma starting their treatment. We did find that the ctDNA dynamics—again, following serial measurements—after 6 weeks of treatment were significantly associated with PFS. At 6 weeks from the start of immunotherapy, the ctDNA values were decreasing from baseline. For that group of patients, the median PFS has not been reached. We have, at this point, about 1 year of follow-up for those patients.

If the ctDNA was increasing at about 6 weeks from start of immunotherapy, the PFS was very poor, at less than 3 months for those patients, and they ultimately all had disease progression on immunotherapy. We also had 1 patient with an interesting case of pseudoprogression, where there was concern that the patient may be progressing with a new lesion that appeared on clinical exam, but the ctDNA had decreased and became undetectable. Eventually, that lesion also decreased, and the patient went on to have a complete response.

Lastly, [in the] post-immunotherapy cohort, which had a small number of patients, we did find that ctDNA was remaining positive post-completion of immunotherapy in a patient who subsequently had disease relapse after 6 months from finish of immunotherapy, and [ctDNA] then became detectable in another who relapsed 3 months after completion of immunotherapy. As such, there may be a role for [ctDNA] in post-immunotherapy surveillance.

What are limitations exist with these assays, in terms of viability of use in clinic?

In terms of limitations, for a tumor-informed assay, we must have a tumor [sample] to be able to do the sequencing to find the mutations to follow. We did have a few patients in these cohorts who did not have the minimum tissue required to be able to do the sequencing. We also had a couple patients with a very small volume of disease. The patient [in whom] we were not able to pick up on the relapse prior to the radiographic relapse was someone who had a very small volume, basically a dermal satellite, that appeared on the skin. For [those with] small volume [disease], there is this question of, [will the] ctDNA [be] able to pick up a relapse in those patients?

Also, [the use of ctDNA in] patients with melanoma who have brain-only disease, is another interesting area of research. How effective is plasma ctDNA going to be in detecting central nervous system–only relapses? We still have many questions that need to be answered.

What are the next steps necessary to understand the clinical utility of ctDNA, within melanoma and beyond?

We have some questions on how to define a clinically meaningful lead time. How many months in advance should ctDNA become positive before radiographic relapse—is it 3 months or 6 months, etc.? How often should you be collecting samples? There are various types of ctDNA assays, different methodologies. Ultimately, as the field moves forward, looking at more standardized ctDNA assessments, there are also questions about the cost of these assays—especially the multiplex PCR assays that are more expensive than digital droplet PCR.

The next step is going to be with the clinical trials, which are [either ongoing] or in development and are using the ctDNA assessments in interventional trials to demonstrate the clinical utility of ctDNA as a biomarker for treatment.

There are some small studies in melanoma that are incorporating ctDNA levels in treatment decision making, [including] whether to use the numbers to make treatment switches, for example, between targeted therapy and immunotherapy, or whether to add targeted therapy to immunotherapy. In other cancers and solid tumors, ctDNA measurement [is being incorporated] to make treatment decisions on the type of adjuvant therapy that should be used, or whether to give them adjuvant therapy at all.

With these ongoing studies that are now starting to incorporate ctDNA into treatment decision making, and as we start to get data from them and future efforts, we are going to have a much better idea of the clinical utility and validity of ctDNA as a biomarker and whether it eventually will become part of clinical practice and the treatment of [patients with] melanoma.

Are there any specific trials that are in early development or that are ongoing that are further exploring ctDNA?

The phase 2 CAcTUS trial [NCT03808441] is looking at using ctDNA levels on patients who receive BRAF-targeted therapy to see when they should be switched to immunotherapy.

At the Moffitt Cancer Center, we have a small clinical trial, where we have 2 cohorts of patients. One [is] getting triplet, anti–PD-1, BRAF-targeted therapy continuously, and in the second arm of the study, all patients start with anti–PD-1, but then we start targeted therapy or withdraw it based on the plasma ctDNA level. We are using the results of the assay to make that decision on whether to start or hold the BRAF-targeted therapy. We are trying to see whether we can make triplet therapies in melanoma more effective and less toxic if we do it based on ctDNA levels.

Reference

1. Eroglu Z, Krinshpun S, Martin J, et al. Circulating tumor DNA based molecular residual disease detection for treatment monitoring in advanced melanoma patients. Presented at: 2021 Society Melanoma Research Congress; October 28-30, 2021; virtual.