ctDNA Monitoring Carries Clinical Validity in Advanced BRAF-Mutant Melanoma

David Polsky, MD, PhD

Circulating tumor DNA (ctDNA) monitoring is clinically valid in predicting whether patients with unresectable, metastatic BRAF-mutant melanoma will benefit from dabafenib (Tafinlar) alone or in combination with trametinib (Mekinist), according to findings of a recent analysis that were presented at the 2019 ASCO Annual Meeting.

At baseline, the presence of ctDNA is linked with poor prognosis in patients with BRAF-mutant melanoma who are receiving therapy with BRAF inhibitors. In the analysis, investigators analyzed 345 patient samples from the COMBI-d trial (NCT01584648) at baseline and at week 4 of treatment. All patients harbored a BRAFV 600E or V600K mutation and were randomized to receive either dabrafenib and trametinib or dabrafenib alone.

Mutation-specific droplet digital polymerase chain reaction (PCR) assays were utilized, and investigators categorized ctDNA data as either positive or negative (±0.25 copies/mL) in this analysis. Moreover, progression-free survival (PFS) and overall survival (OS) were analyzed in all patients. Baseline lactate dehydrogenase (LDH) levels were evaluated as a prognostic factor, although LDH levels are not typically used to guide treatment decisions.

Results showed that baseline ctDNA was detected in 92.7% of patients but was not linked with survival; 201 patients had paired baseline and week 4 ctDNA samples available. Forty percent of these patients became negative for ctDNA at week 4 and experienced a prolonged PFS and OS compared with 60% of patients whose ctDNA remained positive.

An association was discovered between undetectable ctDNA at week 4 and a survival benefit. The prolonged survival was particularly noted in patients with high LDH levels at baseline. For patients with high LDH at baseline, ctDNA monitoring can be useful in predicting treatment outcomes with dabrafenib or dabrafenib plus trametinib.

In an interview with OncLive, lead study author David Polsky, MD, PhD, professor of dermatology at NYU Langone Health, discussed the findings from this analysis in patients with BRAF-mutant, unresectable, metastatic melanoma, as well as the next steps for validating these findings and potentially using ctDNA to help inform treatment decisions.

OncLive: How did you conduct this analysis on ctDNA as a biomarker in unresectable metastatic melanoma?

Polsky: This study is investigating ctDNA as a potential biomarker in melanoma; it’s being developed in many other cancer types, but we are particularly interested in how it can be used in monitoring disease activity in patients. We partnered with Novartis to study patients from their dabrafenib/trametinib versus dabrafenib trial, called COMBI-d, which led to the approval for dabrafenib/trametinib combination therapy for patients with BRAF V600E—mutated melanoma. The trial enrolled around 400 patients, equally randomized. We had samples available from 345 patients at baseline and about another 200 patients available at week 4. We measured, using a droplet digital PCR, ctDNA for the BRAF V600E/K-mutant samples. Those were predetermined by the study because it was an inclusion criteria for the study; everybody had to have a BRAF V600 mutation.

What were your findings from this analysis?

Polsky: What we found was that at baseline, the amount of ctDNA was predictive of survival, and this was in contrast to some other studies earlier that had shown, just plus or minus, whether it was present or absent. We had a very high detection rate of 93% of patients had detectable BRAF V600E in their plasma using the platform, which is the highest rate so far to date for a similar type of study. Because of the droplet digital technology, we are able to quantify the amount very precisely. When we analyze the levels as opposed to just the absence or presence, we found that the levels correlated with outcome in a Cox multivariant analysis. That was the first time that’s been done.

We then looked at week 4 samples since that was what was available to us, and we found that the level at week 4 was also predictive of survival benefit. If patients had no detectable ctDNA at week 4, they had an extended PFS and OS, roughly double [that of] the patients who had persistent ctDNA detected at week 4. We think that with additional studies and additional timepoints, this may become a monitoring tool for patients and doctors to help more fine-tune their treatment.

For example, patients who have a persistently positive ctDNA after starting treatment, doctors may want to add an additional treatment, which may have an additional risk of toxicity but also additional benefit. Whereas if patients clear their ctDNA during treatment, they may just stick with what they are on and wait and see if they progress, then add another therapy [as needed]. This is what we are thinking could be the future, but of course additional studies are needed. I think this is a big first step.

Were there any results that really surprised you from this analysis?

Polsky: We were very pleasantly surprised with the detection rate of 93% because previously, detection rates had really varied from 50% in some settings to the high 80s, but rarely over 90%. We were surprised on the upside that we had a 93% detection rate. Otherwise, I would say it panned out the way we were hoping it would pan out.

What additional studies would you like to conduct to further validate these data?

Polsky: We would like to look at additional timepoints. We only had week 4 samples available to us, but certainly samples from week 8, week 12, and etc. [would be useful] to use as a monitoring test. Right now, there isn’t a monitoring test available for patients; LDH is part of the AJCC staging criteria, and we know even in this group that it was predictive of survival outcomes. The ctDNA was independent of LDH with the baseline samples. We are doing some further analysis to try and develop maybe a survival model based on that, but we think it will be even more predictive as we have additional samples available.

Eventually, if it all still looks promising, we would like to do a clinical trial where we can randomize patients to have a treatment decision made, for example, a pre-prescribed second-line therapy built into the trial, then at progression patients switch to another agent and then they can be randomized to [therapy] based on a rise in their ctDNA versus radiographic progression, which is the standard of care. That would be an exciting trial for us to do to see if the marker has clinical utility, not just clinical validity, which is what we have demonstrated so far.

How does this tool fit in with others that are currently available for determining patient outcomes?

Polsky: This was all unresectable metastatic melanoma, so these patients are getting treated and monitored with CAT scans, basically. That’s the standard of care. Another potential application could be using this in conjunction with imaging, so often times at tumor board conferences, we have these indeterminate radiographic findings, and there is a paper out there from the John Hopkins group suggesting that if you add ctDNA into the information available, it can help people better resolve indeterminant findings. That was a very small study, but that’s another potential application that we can try down the road.

Do you have any next steps currently planned for this research?

Right now, this is research-use only. It’s not available for clinical use, but we’ve done a lot of work and have funding from NIH that we use to analytically evaluate these assays. We’re putting out protocols on how to run them, so technically other commercial companies could be running these assays, but you have to go from analytical validation to clinical validation, which is what does it mean when you have 20 copies/mL of mutant DNA? That’s what we are working on now. Clinical utility, of course, is whether or not it changes your management. This is sort of a biomarker development pipeline that’s similar to the drug development pipeline, so we are somewhere in between phase II and phase III, I suppose, with a phase III or phase IV being real clinical utility when we think about the different steps that are needed.

Do you see this technique being applied in any other tumor types?

Polsky: There are other cancers where this is being used, not with droplet digital PCR, but with other technologies. For example, there are clinical trials in colon cancer, stage II colon cancer, where after resection, patients have no other measurable disease but they take a ctDNA measurement of the mutations that are found in the patient’s tumor, and they look specifically for them. If they are still present, these patients who ordinarily would not receive adjuvant therapy are being randomized to either adjuvant therapy or not. Those studies are ongoing, and it’ll be very exciting to see the results, to see if there is a benefit in that population based on the presence of mutations in the ctDNA after surgery.

Syeda MM, Wiggins JM, Corless B, et al. Circulating tumor DNA (ctDNA) kinetics to predict survival in patients (pts) with unresectable or metastatic melanoma treated with dabrafenib (D) or D + trametinib (T). J Clin Oncol. 2019;37(suppl; abstr 9510).