NEW YORK (GenomeWeb) – Data from a new study of melanoma patients treated in an early trial of tumor infiltrating lymphocyte therapy has demonstrated that measuring DNA mutations in patients' blood could help predict treatment efficacy long before other clinical signs of recurrence.
The results, published last week in Clinical Cancer Research, suggest that there is a distinct pattern of ctDNA changes, at least in melanoma patients, associated with a much higher likelihood of having a durable response.
The study was part of an ongoing collaboration between the National Cancer Institute's laboratory pathology molecular diagnostics section and the surgery branch's tumor immunology section.
Mark Raffeld, head of the MDx section and a lead author of the study told GenomeWeb in an email that with data accumulating showing that circulating tumor DNA can be detected in the blood or other body fluids of cancer patients with potential impact on therapeutic decision-making, his team became interested in exploring its use as a way to monitor experimental therapies at the NCI, and started looking for a collaborator.
A pioneer in TIL therapy, Rosenberg had led a series of trials of TIL transfer in melanoma patients in which about 20 percent developed complete sustained responses — remaining cancer free for years after the treatment.
"When we started this collaboration, our goal was to simply assess whether we could monitor patients for responses … using ctDNA and hopefully also identify recurrences," Raffeld said in his email.
"Melanoma was ideal to pilot our exploration of ctDNA because with a single targeted assay for V600E we could monitor 50 percent of all melanoma patients," he added.
For their analyses, Raffeld and colleagues used Thermo Fisher Scientific's competitive allele specific TaqMan PCR (castPCR,) which he said his team brought in the lab as a high sensitivity test to detect BRAF V600E mutations in clinical samples.
"It was really a matter of convenience and availability that we chose castPCR for this study. Of course we like the assay, and it is very sensitive and specific. However, he added, castPCR is also expensive and as such the group has now converted to BioRad Laboratories' Droplet Digital PCR for studies targeting a few known mutations.
"Both technologies have excellent sensitivity and perform well in our experience," Raffeld said. "But if you have multiple potential targets, then you really have to go to a NGS system."
Using the assembled retrospective samples from three TIL trials conducted between 2000 and 2007, the group was able to identify a subset of 39 patients who had BRAF V600E-positive stored serum samples from before they began treatment and at least two samples in the first month following.
"Because we had numerous early time points in so many of these patients we were able to observe the immediate effects of the TIL infusion in most of the patients on the trials … From the clinical point of view this allowed us to identify patients likely to respond and patients unlikely to respond," Raffeld explained.
To do this, the researchers used the patterns of BRAF V600E DNA in the study samples to define three subgroups of patients: one in which there was an initial peak of ctDNA which then dropped to an undetectable level, one in which there was an initial peak, but where ctDNA did not then dissipate, and one in which there was no initial peak with or without eventual clearing of the mutation.
Overall, the authors wrote, patients that showed an early ctDNA peak and then cleared their serum of V600E were much more likely to achieve a complete response over the next one to two years than those with either of the two other patterns.
According to Raffeld, although many questions remain to be answered by future validation studies, the group's initial data supports a basic conclusion that patients with stage IV melanoma who do not develop a ctDNA peak within 10 days of their TIL infusion, or patients who develop a peak but do not clear their ctDNA, are highly likely to fail their therapy.
It remains unclear how this predictive ability might impact clinical practice, Raffeld added. "Even in in our own trial setting we haven't decided yet how to implement these findings," he wrote.
However, the results do hint that there might need to be a different clinical approach to the two different non-responsive ctDNA patterns. "The total absence of a peak may indicate that the TIL are completely ineffective and a change in therapy may be warranted," he said.
In contrast, a peak without subsequent clearing looks more like it might be a sign of the TIL not having enough potency to lead to total eradication of a patient's tumor, meaning that some form of dose intensification might be effective for patients with this pattern.
Since the earlier NIH studies there have been significant advances in cell-based immunotherapy, both by addition of other drugs to TIL transfer, and through genetic engineering of immune cells to target specific antigens.
Raffeld and his team's PCR-based approach focused only on a single mutation, BRAF V600E, which luckily is present in about 50 percent of melanomas. But he said that moving forward, his group is interested in taking a broader genomic approach that could allow ctDNA-based monitoring or prediction of TIL therapy response in patients without this one alteration.
Raffeld said that Rosenberg has initiated a trial at NCI in epithelial cancers, which is looking at an approach targeting immunogenic neoantigens identified by whole-exome sequencing.
"We have begun to use ctDNA to follow these patients," Raffeld said in his email. "For these cases we are using ddPCR. As you can imagine, this is a fairly large undertaking as we have to design personalized assays to the unique mutations in each patient's tumor."
A broader, individual-patient-based genomic strategy may be necessary if the approach of using ctDNA to predict TIL therapy response were to be expanded to other cancer types where there may not be a single mutation like V600E that occurs in a large proportion of cases.
In a recent study, a Johns Hopkins-led team used a similar broad genetic profiling approach but with targeted next-gen sequencing to identify a personalized mutation that could be used to monitor response to colorectal cancer adjuvant treatment in individual patients.
Raffeld wrote that his group is also looking into NGS.
Meanwhile, he said, "I think it is very important for the study to be confirmed by additional studies either by us or others … Of course a prospective study with well-planned collection dates would be best. The problem is that we have work with the current ongoing studies."
The team is also currently looking into expanding its retrospective melanoma study to a second larger cohort of patients, potentially targeting additional mutations besides BRAF V600E.