NEW YORK – Investigators of a recent study at Yale University were surprised when their analysis of circulating tumor DNA, which had been ubiquitously successful as a treatment response monitoring tool for them in the past, failed to correlate with outcomes in a new study of combination PARP and VEGFR inhibitor therapy.
As forays into different features of circulating cell-free DNA such as epigenomics and fragmentation patterns have been overcoming hurdles that stymied earlier mutation-based approaches, liquid biopsy has become increasingly applied not only in cancer genotyping but in new niches like response monitoring and cancer screening. But results like those of the Yale authors suggest that not all applications are guaranteed success.
Their trial, published in JCO Precision Oncology, set out to determine whether changes in ctDNA levels could predict clinical outcomes for a combined targeted therapy used as a last resort in patients who had failed multiple other treatments.
The study recruited patients with metastatic pancreatic ductal adenocarcinoma, triple-negative breast cancer, small cell lung cancer, and non-small cell lung cancer who had progressed on an average of three previous lines of therapy.
In the trial they received up to seven days of cediranib 30 mg orally once daily monotherapy as lead-in followed by the addition of olaparib 200 mg orally twice daily. Investigators collected plasma samples before and after cediranib monotherapy lead-in and the combination therapy at seven days, 28 days, and every 28 days thereafter.
The team used a multigene mutation-based assay to quantify circulating tumor DNA, hoping to see a match between changes in the amount of tumor DNA in the blood and response to the trial drugs. But in the 63 evaluable patients, there was no association between either baseline ctDNA levels or ctDNA decline and patient's radiographic response or survival.
Abhijit Patel, the study's senior author, said the result was a surprise to him and his team. "We were expecting it to be just another study confirming that ctDNA works," he said. "We had these samples that we had signed up to collect through our lab, and we tested them thinking that we would just see good correlation with response like everyone else tends to."
With the opposite result in hand, the group started wondering why and hypothesized that the heavily treated population might be responsible. Patients in the trial were particularly unlikely to respond to treatment after progressing on a number of other therapies, so Patel and colleagues guessed that in that situation where responses aren't particularly robust, there wouldn't be a high signal in terms of differences among responders and non-responders in circulating DNA.
Indeed, baseline ctDNA levels were much higher in the Yale study population than reported in previous studies of ctDNA-based treatment monitoring. "At such late stages of disease where tumor burden and ctDNA levels are very high, baseline ctDNA concentrations may be less effective at predicting relatively small differences in generally poor survival outcomes," the authors wrote.
"In the setting of such modest therapeutic responses, a clear association between longitudinal ctDNA change and treatment efficacy can be more difficult to discern," they added.
Patel said he's not aware of other studies in heavily pretreated cancers that reported similar failures, and he stressed that his team's findings shouldn't lead others to avoid pursuing ctDNA monitoring in future trials.
"I don't consider this to be a definitive case-closed kind of study because it could be something else, right? It could be just it's not the heavily treated population. That's just a hypothesis on our part. It's possible that it was just this particular therapy, so I think it's certainly worth others looking at other types of therapy in a similar advanced population."
It might not seem clear what the benefit is in monitoring ctDNA in patients with advanced cancers. Such patients, if they do benefit from treatment, only do so for a short time, and if they fail to respond, have few alternative options.
In earlier-stage cancer, trials have begun to prove out specific cases where ctDNA detection and monitoring can be used to make therapeutic decisions, such as whether to pursue adjuvant treatments or even to switch patients off of one drug and onto another in real time when a liquid biopsy shows they aren't benefiting.
Patel predicted that even more applications could emerge, for example, in settings where patients are now being treated with immunotherapy and chemotherapy in combination, there could be stepwise induction strategies that help spare individuals from dual treatment who are not likely benefiting from added immune drugs.
"We know with immunotherapy alone, there is a 20 percent to 40 percent response rate. Most of the time, people aren't even responding, so one interesting approach could be to give a round or two of immunotherapy first and then use ctDNA to determine if they are responding. Then only the people who are actually benefiting from it would continue with the combo," he said.
Even though the same utility likely isn't possible in the context of third- and fourth-line treatments in patients with refractory, advanced disease, Patel said that ctDNA monitoring still has usefulness as a quicker and more agile measure of response and progression than radiologic imaging.
"I think there's enough thought leaders buying into the prospect of using ctDNA in these ways, and it's worked well in multiple studies now. I think it's likely to be established soon in the appropriate clinical settings … so this paper is more just a cautionary note," Patel said.