David Polsky, MD, PhD, discusses the development and clinical utility of biomarkers in melanoma.
David Polsky, MD, PhD
As the search for blood- and tissue-based biomarkers continues in localized and metastatic melanoma, circulating tumor DNA (ctDNA) has emerged as a potential candidate for patients with BRAF-mutated disease, said David Polsky, MD, PhD.
"Many years ago, we heard critiques from grant reviewers who said, 'You don't have any good treatments, so what do you need a biomarker for?’” said Polsky. “Now we have a lot of treatments, combinations, and different sequencing [strategies]. Biomarkers may help us guide those treatments as well as new treatments coming down the pike."
In an attempt to identify a blood-based biomarker in melanoma, a large-scale, plasma sample analysis evaluated a pooled population of patients with unresectable or metastatic BRAF V600E/K—positive melanoma from the phase III COMBI-d trial. Investigators measured patients’ BRAF V600E/K ctDNA at baseline and 4 weeks after starting treatment with either the combination of dabrafenib (Tafinlar) and trametinib (Mekinist) or dabrafenib alone.
Results showed that baseline ctDNA was detectable in 93% of 345 patients and in 60% of week 4 samples. After 4 weeks of therapy, the median progression-free survival (PFS) for the 40% of patients who converted to ctDNA-negative was 12.94 months versus 7.06 months for the remaining 60% of patients who remained ctDNA positive (HR, 0.55; 95% CI, 0.39-0.76; P = .00028). Likewise, the median overall survival (OS) was 28.19 months versus 14.55 months, respectively (HR, 0.56; 95% CI, 0.40-0.79; P = .00086).
Importantly, ctDNA monitoring may be helpful for patients with high lactate dehydrogenase (LDH), as a greater PFS and OS improvement was seen in this subgroup.
In an interview during the 2020 OncLive® State of the Science Summit™ on Melanoma, Polsky, the Alfred W. Kopf, MD, professor of Dermatologic Oncology, Ronald O. Perelman Department of Dermatology Professor, Department of Pathology, and director of the Pigmented Lesion Service, NYU Langone Health, discussed the development and clinical utility of biomarkers in melanoma.
OncLive: How can biomarkers improve patient outcomes? Is biomarker development similar to drug development?
Polsky: Validated biomarkers that can [guide] a patient's management by follow-up or drug choice based on biomarker results and show clinical utility can improve outcomes.
There are examples of this being done successfully in breast cancer. In melanoma, there are [also] some examples, but we still have a lot of work to do.
In breast cancer, the results of Oncotype DX testing can spare patients with localized breast cancer [from] adjuvant chemotherapy. In melanoma, biomarkers are really under development rather than [utilized] in clinical practice to adjust therapy.
Could you explain the utility of BRAF testing?
We do tissue-based BRAF testing in melanoma. BRAF is not thought of as a biomarker because it is an open-and-shut case. You would never give a BRAF inhibitor to a patient unless they had a BRAF mutation.
In essence though, it is a biomarker because it is something we can measure in the tumor sample to determine treatment choice.
Could you discuss some of the exploratory tissue- and blood-based biomarkers?
Many groups are analyzing tissue samples to try to identify measurable [biomarkers] like proteins, RNA, or DNA mutations that are prognostic and may be associated with poorer outcomes. The hope then is to move them into the predictive space, where they may be able to tell us how a patient will do on follow-up therapy.
[At NYU Langone Health], we particularly work in blood, and we find ctDNA very exciting. I am obviously biased, but multiple labs around the world are doing this work as well. It has an exciting future and is being used in other cancers, too.
What is known about ctDNA?
ctDNA is the fraction of DNA in the plasma that is related to the tumor. ctDNA is found by identifying mutations from the tumor.
For example, BRAF mutations are very common in melanoma and we can find measurable BRAF-mutated DNA in the plasma. Some studies presented at the 2019 ASCO Annual Meeting showed that the levels of BRAF-mutated ctDNA prior to BRAF/MEK treatment with dabrafenib and trametinib were associated with survival. Patients with low ctDNA had better survival versus if they had high ctDNA.
At 4 weeks on treatment, LDH assessment showed that, for patients with a high tumor burden, undetectable DNA was associated with a doubling of their PFS and OS.
Other labs are showing interesting associations with ctDNA, survival, immunotherapy, and predicting relapse in resected stage III disease.
Do other promising biomarkers exist beyond ctDNA?
We are not there yet with tissue-based markers. PD-L1 was explored via immunohistochemistry, but it was not sufficiently predictive of benefit in patients who either expressed PD-L1, or who did not express it and then were given anti—PD-1 therapy. That is not going to move on as a predictive biomarker.
What is needed to bring ctDNA into clinical practice?
Ideally, what's needed is a large prospective study measuring ctDNA, and then adjusting patient's treatment based on the result. That clinical utility is the last phase of biomarker development.
Are there any plans being made to measure it in a larger study?
Many groups are planning [similar] studies, and actually, a couple of them are currently enrolling patients.
In a sense, LDH is a biomarker because patients with a higher level of LDH at the start of treatment are known to [have a worse prognosis]. A study from The Netherlands is treating patients with vemurafenib (Zelboraf) plus cobimetinib (Cotellic) to bring down the patient's LDH. Then the patient starts ipilimumab (Yervoy)/nivolumab (Opdivo). This is compared with patients who start on ipilimumab/nivolumab straight away. The goal is to see if reducing LDH with BRAF/MEK inhibition has a benefit.
How is gene profiling currently being used?
Gene expression profiling is being used in patients who have localized melanoma, to try to identify those who [will] have a worse-than-expected outcome. The test has had a lot of development using retrospective studies, but you need prospective data to validate these things.
To date, I am only aware of 1 fairly small prospective study. While the results were interesting, they are not developed [enough] to endorse [gene profiling] at this time. The National Comprehensive Cancer Network [recommends] that this test be used in a clinical trial setting rather than in a patient management setting.
It still has promise for those who will benefit, but we are awaiting additional studies.
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 15; abstr 9510). doi: 10.1200/JCO.2019.37.15_suppl.9510.