NEW YORK – New evidence has further confirmed that while levels of pre-treatment circulating tumor DNA can be used to predict melanoma patients that will have better responses to first-line immunotherapy, the same is not true in individuals receiving this treatment as a second-line option.
Published last week in Clinical Cancer Research, the study analyzed pre-treatment ctDNA in separate discovery and validation cohorts with both first- and second-line-treated patients, finding that baseline ctDNA only predicted response when patients were getting immunotherapy as their first treatment.
The results highlight the need to carefully consider disease context when implementing new or exploratory biomarkers, the study authors noted. Although the study featured small cohorts, it did replicate its finding across two independent sets of patients with the same clear discrepancy emerging between first-line and second-line outcome prediction.
The researchers said the findings also reiterate a need for more studies evaluating how the accuracy of ctDNA, as well as other immunotherapy biomarkers now entering the clinic, may shift when applied in different contexts.
According to Elin Gray, an associate professor at Australia's Edith Cowan University and senior author of the new report, immunotherapy has become an increasingly common treatment for patients with melanoma. For patients with BRAF mutations, it is commonly used in the second line after treatment with BRAF inhibitors.
As in other cancer types, melanoma oncologists are eager for biomarkers that can help predict which patients stand to benefit from immunotherapy, and which won't.
Durable response to immunotherapy is only seen in a minority of advanced melanoma patients, and debate continues about the optimal sequencing of immune versus targeted therapies, the utility of combination treatments, and the selection of the most effective first-line option.
Meanwhile, research has emerged showing that baseline ctDNA levels correspond to longer progression-free survival in patients with melanoma, though questions remain about how this might hold fast for different types and contexts of therapy.
In their study, Gray and colleagues paid specific attention to both type and line of treatment. The team began by analyzing pre-treatment ctDNA levels and patient outcomes for a cohort of 125 adults with metastatic melanoma treated at Sir Charles Gairdner Hospital and Fiona Stanley Hospital in Perth, Australia. Of these patients, 32 received either single or combination immune checkpoint inhibition in the first-line setting, and 27 in the second-line setting after either initial targeted therapy or initial single-agent ICI. The remaining 66 were undergoing first-line treatment with a molecularly targeted therapy.
Consistent with previous studies, the researchers found that patients with low baseline ctDNA levels (fewer than or equal to 20 copies/ml) who were treated with targeted therapy had longer progression-free survival (PFS) than those with high baseline ctDNA levels (greater than 20 copies/ml).
But looking at the immunotherapy-treated groups, there was a significant divergence between the first- and second-line subsets. Individuals receiving first-line immunotherapy who had low baseline ctDNA had an 80 percent longer PFS than those with high baseline levels. And this association remained significant even after controlling for other factors, such as sex, age, tumor stage, BRAF mutation status, and brain metastases.
In contrast, there was no significant association between PFS and baseline ctDNA in the second-line ICI group.
In an email this week, Gray said that two factors could be affecting this difference between first- and second-line immunotherapy.
"First, it is the fact that ctDNA does not reflect the presence and volume of brain metastases, which are common in patients relapsing after BRAF inhibition. Second, the tumor dynamics that regulate ctDNA shedding might be affected once [resistance to BRAF-targeted drugs] is developed," Gray wrote.
Unfortunately, the second mechanism of ctDNA shedding dynamics remains relatively poorly understood, she added. "We know that not all cancer types generate the same amount of ctDNA relative to tumor size, but what mechanisms regulate that are unclear."
To put the findings to the test even further, Gray and her team set out to replicate the same analysis in a separate validation cohort of 128 melanoma patients recruited at other Australian treatment centers. This group included 77 patients who received first-line immunotherapy and 51 who were treated in the second-line setting.
Similar to what they saw in the discovery cohort, baseline ctDNA levels in this group failed to significantly correlate with PFS in patients getting ICI treatment in the second line. But for first-line treatment, low baseline ctDNA levels were associated with a 58 percent longer PFS independent of other potentially confounding factors.
As a final experiment, the group combined the two cohorts to try to examine whether ctDNA was predictive specifically of response to combined ICI treatment. Although the results weren't statistically significant, they did see a trend in the data for high ctDNA levels corresponding to longer survival on combination therapy versus single agent anti-PD1 treatment.
According to the study authors, the selection of first-line immune checkpoint monotherapy versus combination is currently a complex decision.
Although the study findings have to be validated in future independent cohorts, Gray and her coauthors said that in this context, their results at least suggest that more aggressive treatment might be particularly beneficial to those patients with high ctDNA levels.
Whether it's possible to bring ctDNA into clinical practice as a predictive biomarker depends somewhat on whether it emerges as something crucial for the advancement of specific drugs or drug combinations, Gray said.
"Getting liquid biopsies into clinical practice is a challenge. In my view, the most successful [biomarkers] are those associated with qualifying a patient for a drug [as] pharma will be more likely to sponsor the required clinical studies and regulatory approvals," she added.
"In the case of ctDNA testing to escalate or deescalate immunotherapy, the ultimate stakeholder (in addition to the patients) is whoever is paying for the drugs … and managing the side effects." As a result, there is also a need to engage these players if this type of testing were to move forward.
Gray said that she and her colleagues are in discussion with an undisclosed pharma company now to access samples from previous or ongoing clinical trials, which will be an easier next step than starting a new trial to validate their findings. They are also hoping to further study how tumor biology and the tumor site affect the release of ctDNA and evaluate ctDNA as a biomarker of disease progression.
Gray added in her email that the group is well aware of the current interest in multi-biomarker strategies for immunotherapy response prediction and is also carrying out a study comparing ctDNA levels and tumor mutational burden.
"As our work is mainly on blood biomarkers, we are also trying to derive TMB from ctDNA [and] exploring other blood-derived biomarkers such as circulating tumor cells, exosomes, and autoantibody profiles," she said.
Another immunotherapy area where researchers feel ctDNA might hold promise is the early on-treatment setting, where changes in levels of mutated, or overall circulating DNA can provide an early indication that therapy is or is not working.
Although she said that greater value to clinicians may be in knowing if a therapy is suitable before it is given, Gray and colleagues have been exploring on-treatment monitoring in melanoma as well, publishing their findings earlier this year. This area of research hasn't yet addressed the question of first versus second line treatment, though, so that might be something to explore in the future, she added.