NEW YORK – New findings presented at the American Society of Clinical Oncology in Chicago this week have offered a glimpse into a potential future for blood-based minimal residual disease tests in patients with sarcomas, a challenging application for liquid biopsy technologies due to a lack of recurring or predictable DNA biomarkers.
Data from the two studies — one led by investigators from Toronto's Princess Margaret Cancer Centre using Natera's Signatera MRD assays and the other by a team of hospitals in Guangzhou, China, employing an analogous, tumor-informed method —hints at how circulating tumor DNA (ctDNA) tests may prove clinically useful in sarcoma patient care, said Dana-Farber Cancer Institute oncologist and researcher Brian Crompton during a webcast discussion of the research.
Finding utility for circulating tumor DNA has been challenging in sarcomas thus far. "One of the things that everybody in this room is probably well aware of, which is so striking about sarcomas, is that they're very structurally driven," said Crompton, who was not involved in the studies. "It's rare to find recurrent hotspots or the same point mutations, and this makes it a much different challenge than carcinomas where you're finding activating tyrosine kinases over and over again."
Researchers continue to explore ctDNA as a prognostic tool, though, analyzing levels to estimate whether patients are likely to or have responded to therapy, as well as to sift for rare targetable alterations that could inform future treatment.
According to Crompton, multiple groups have shown, for example, that the presence of pretreatment ctDNA can be prognostic in Ewing's sarcoma, leiomyosarcoma, and rhabdomyosarcoma, among others. Liquid biopsies have also been used to identify activating tyrosine kinases in gastrointestinal stromal tumors (GIST), which "can be really useful for targeted therapeutics."
In Ewing's sarcoma, researchers have also attempted to use specific fusions detected in a patient's tumor to identify and track minimal residual disease, but it has been a challenge, Crompton said. "The upside is it's extremely specific; you only really see a fusion in the cell-free DNA if there's actually circulating tumor DNA there. The downside is it's one variant, so when ctDNA levels get very low, it gets hard to detect because there's only one variant to look for."
"The more shots on goal you have when ctDNA levels are extremely low, the better off you are, which I think is one of the strengths of today's presentations," he added.
Crompton described the development of ctDNA assays as "a balance" between depth and breadth of sequencing. "This is important, particularly, in sarcomas because there are so few recurrently mutated genes," he said. "If you were working in a tumor type where 20 genes are always mutated, you could easily create a panel that could sequence extremely deep … to find the presence of those variants. But when you don't know what you're looking for or what to expect in each patient, it's hard to develop an off-the-shelf assay that can give you that kind of depth [because] as you go broader, if you have fixed costs, it gets harder and harder to have the depth you need to find very, very small amounts of circulating tumor DNA," he said.
Both studies highlighted at the ASCO session employed what has come to be known as a tumor-informed, patient-personalized strategy, which eschews a universal or off-the-shelf panel in favor of developing personalized panels of DNA alterations based on an individual's tumor tissue sequencing. Natera has been offering this testing commercially for several years under the brand Signatera, and several competitors with similar platforms have also emerged over time.
Study results
In the first ASCO presentation, Princess Margaret investigator Abdulazeez Salawu reported results from a study in which he and his colleagues used Signatera to create MRD assays for patients with soft-tissue sarcomas (STS).
Despite standard treatment, which includes surgery and radiotherapy, up to 50 percent of patients with these cancers experience metastatic relapse, but routine use of adjuvant systemic therapy remains controversial, Salawu and his colleagues reported. The hope is that MRD might serve as a biomarker, identifying patients with higher risk of relapse who may benefit from added systemic treatment.
In the study, patients with localized, high-risk STS were enrolled prior to radiotherapy and surgery. The researchers collected blood samples for ctDNA analysis at diagnosis, post-RT, post-surgery, and every three months for up to two years. They also carried out whole-exome sequencing (WES) of archival tumor and matched normal tissue to identify patient-specific, somatic, single-nucleotide variants.
The team then used personalized and tumor-informed, multiplex PCR next-generation sequencing-based ctDNA (Signatera) assays to track ctDNA in serially collected plasma samples, with ctDNA levels expressed as mean tumor molecules per milliliter (MTM/ml) of plasma. Radiologic surveillance was performed every three months following surgery. The primary endpoint was a ctDNA detection rate of greater than 70 percent at diagnosis. Secondary endpoints included MRD detection after local therapy and correlation of ctDNA levels with disease relapse.
The team was able to collect samples from 20 patients who had adequate tissue quality for the WES necessary to create personalized Signatera assays. These assays were able to detect ctDNA in 80 percent, or 16 of the patients at initial diagnosis, of which 15 became ctDNA-negative when tested again immediately after surgery. Follow-up blood draws allowed detection of reemerging ctDNA, indicative of relapse, in four patients with a median lead time of 92 days before recurrence showed up on imaging.
Crompton cautioned that the analysis of pretreatment samples in the study was retrospective. Personalized MRD assays take time to create, and it could turn out that tumor-agnostic or off-the-shelf assays could suffice for that initial prognostication step.
The real strength of these patient-specific assays is in the follow-up setting, he said. "As you can see … there are several cases in this cohort where ctDNA was detected before it was known that the patient was progressing," he noted, making it a "harbinger of progression." According to Crompton, "this is the kind of study we need to start thinking about, where we can use therapeutics earlier on in patients who are eventually going to progress, hopefully to dramatically extend their life, if not help start effecting cures even in the context of relapse."
The authors wrote that these types of additional studies are already ongoing.
In the second ASCO study, investigators led by Junqiang Yin, an oncologist at the First Affiliated Hospital of Sun Yat-Sen University, created similar personalized MRD assays based on whole-exome tumor tissue sequencing for 84 sarcoma patients.
Crompton highlighted two standouts from the data. First, that 64 percent of the variants identified in the WES data were seen in only a single patient. "It goes back again to the genomics of sarcoma and how you can't rely on recurrent variants and certain genes to be sure that you have an assay that can detect MRD for that patient," he said.
Secondly, it was notable that there were 18 patients for which the team couldn't create a personalized assay with sufficient content. Crompton said that this shouldn't necessarily be looked at as a failure, but as more evidence of the "reality of what we're dealing with in terms of sarcoma."
"I think it's actually quite remarkable that they were able to develop an assay for so many knowing how structurally driven these tumors tend to be," he added.
Like the Princess Margaret group, Yin and his colleagues were also able to show that detection of ctDNA was often a harbinger of relapse, and, although the absence of MRD was not perfectly predictive of sustained disease-free survival, there was an association.
The team performed post-surgery MRD detection on 27 patients, of which eight had a positive result. Three of these developed a recurrence confirmed by imaging at 12, 30, and 101 days after detection of positive MRD, respectively. According to the investigators, no more imaging recurrence has occurred in the other patients thus far, including the five remaining patients with positive tests and the 19 patients with negative post-op MRD. Follow-up is still ongoing.
"I think the overall conclusion here is that tumor-specific tests are feasible in sarcoma," something that the field has not really explored before, Crompton said. Of course, much more data will be necessary to prove that this testing has true clinical utility.
"These studies are really just emerging, but I think seeing the success here is really exciting and encouraging," Crompton said.
"Now we need to think of larger studies to define specificity and sensitivity, and make sure that every time a blood draw is happening it coincides with imaging, so we know just how much lead time … these assays are actually providing," he added.
Finally, he recommended that future efforts also assess variation in utility in the context of different subtypes of sarcoma. "Some of these tumors [as evidenced in the second study] may have so few SNVs that it may actually not be feasible to develop patient-specific assays, whereas others may be much more successful," he said.