Immunotherapy Response Monitoring in Melanoma - Episode 6

Using ctDNA to Monitor Patients With Melanoma

An expert in melanoma cancer treatment examines monitoring patients with melanoma using ctDNA.

Matthew Fowler: How might ctDNA be used to monitor patients with melanoma?

John Kirkwood, MD, PhD: There are 2 arenas. First, the measurable active disease where we can make correlations much quicker because those patients have radiologically assessable or palpable disease. So, when we say it gets larger either on the x-ray, the CT, the PET [positron emission tomography] scan, or by physical exam, we can measure whether ctDNA has tracked with that. The area where ctDNA is the largest potential benefit is the adjuvant arena, where the patient doesn’t have any measurable disease. We historically give patients a whole year of treatment with interferon, the agent we developed with the ECOG trial 30 years ago.

Since 2017, PD-1 of 2 different types, dabrafenib-trametinib and MEK and BRAF inhibitor combinations, which are twice as effective as our earlier adjuvant therapies. We’re still shooting in the dark because we don’t have measured disease in those patients to follow, and we give them arbitrarily a year of therapy, which may waste many months of those patients’ time if the ctDNA evidence of disease relapse could be monitored and could be shown to indicate the need for additional alternative therapies in a subset of patients who are relapsing under our eyes. There’s nothing we can see to document this and to occasion a change in the treatment that they’re pursuing.

With the avalanche of new therapies that we have, the alternative options for treatment aren’t limited at this point. We stay with therapies for an arbitrary year because that’s what we did 35 years ago in the original adjuvant trial. There’s no solid reason to stay on that therapy for a whole year if there’s molecular evidence of progression, of failure of disease resurgence that should dictate alternative therapies. The flip side to this is the option for terminating therapy, of declaring victory, may exist in patients who have no evidence of disease clinically because all these patients had been resected free of disease but also have no molecular evidence of disease on treatment at some point in time. Is it 3 months? Is it 6 months? Is it 9 months? When we know these data—it will not be 12 months—we can stop earlier and spare these patients the cost, the morbidity, the time required for the arbitrary adjuvant 1 year of treatment that we had been giving these patients with all our agents and some of them for even more than that.

The ipilimumab adjuvant regimen approved in 2015 was a 3-year regimen. The PegIntron regimen approved in 2011 was a 5-year treatment. Were these necessary? I think not. Even 1 year, which has been our standard since E1684, the original interferon trial [in 1996], has no real solid justification. With sensitive molecular markers that are precise, it would be possible to intelligently, rationally approach this and limit the therapy we’re giving our patients.

Matthew Fowler: As you know, ctDNA testing is relatively new. What evidence would you want to see to incorporate a new test into routine clinical practice?

John Kirkwood, MD, PhD: Rigorous phase 3 correlatives are needed. We need evidence that both relapse-free survival in the adjuvant setting and overall survival correlate to the molecular parameters that we’re following. The FDA and some of our NCI [National Cancer Institute] and CTEP [Cancer Therapy Evaluation Program] statisticians have, for many years, said only survival makes a difference, and an impact that shows survival impact of this marker-guided therapy would be the absolute gold standard. These days, when almost none of our trials has a primary end point of survival anymore because it takes a decade to get to, or at least many years, we have increasingly thought about alternatives.

One of them, for instance, is the neoadjuvant trial where you give therapy, prior to surgery, for 1 month, 2 months, at the outside 3 months, and then use the pathological evidence at resection to give a pathological index of pathologic complete response, pathologic major response, or nonresponse. Some of the proponents of neoadjuvant trials, which we have done more of in Pittsburgh than any other place in the world, have said this should be already accepted by the FDA.

The FDA have stood their ground and rigorously insisted that we have more trials and more statistically solid data to argue that the neoadjuvant treatment is effective and that that pathologic readout is as valuable as all of us think it is. That data will come with larger numbers from multisite studies. In our cooperative groups in the United States, we’re doing the first multicenter neoadjuvant trials in the National Clinical Trials Network. Those will provide the rigor and multiinstitutional prospective validation that the FDA is properly asking for.

This transcript has been edited for clarity.