A new study this month has added to the so-far scant data available on the ability of Guardant Health’s Guardant360 assay to predict responsiveness to immunotherapy using a blood-based signal to infer how highly mutated a patient’s cancer is.
Published in Nature Medicine on July 15, the study focused on a group of 61 patients with advanced gastric cancer, treated with pembrolizumab (Merck’s Keytruda) in a phase II trial. Investigators studied a variety of factors and their association with immunotherapy response including microsatellite instability (MSI) and PD-L1 expression, but also tumor mutational burden — more specifically, whether Guardant’s 360 assay could recapitulate the same response prediction as tissue-based measurements of TMB.
The idea that mutational burden, or load, predisposes cancers to respond to immunotherapy, has rocketed from a research hypothesis, to a mainstay of clinical trials, to the cusp of widespread clinical implementation in barely a few years.
Because TMB is based on a straightforward quantification of tumor mutations, multiple companies are pursuing assay commercialization. Clinical tests for TMB that use tumor tissue samples are already available, while close on their heels several companies have said that they intend to commercialize competing methods that can be performed on a blood sample.
Guardant has so far shared little about how it is developing methodology for gleaning a TMB readout from its Guardant360 liquid biopsy assay, and even less about how it would hope to distinguish its own approach from those of its numerous possible competitors.
The company declined to comment on its internal commercial progress regarding TMB, or on the new gastric cancer study this month, which was led by researchers from Samsung University in Korea. Academic authors from the study also declined to discuss the results after initially agreeing to answer questions.
In addition to evaluating known immunotherapy response biomarkers like MSI and PD-L1, the report included data from a subset of patients who had both blood and tissue sequencing results, which showed that the Guardant360 panel could accurately reproduce the predictive ability of TMB calculated from tumor tissue exome sequencing.
“ctDNA mutational load score was well correlated with radiographic response and, in a post hoc exploratory analysis, appeared to predict PFS at least as well as tissue mutational load,” the Samsung authors wrote.
Based on this, for patients unable or unwilling to undergo invasive tissue biopsy, “ctDNA mutational load can be a feasible option or at least a complementary testing tool,” they concluded.
University of California, San Diego researcher Razelle Kurzrock, who was not involved in the Samsung study, has also explored the use of Guardant360 results to provide a surrogate for tissue TMB. She and her colleagues published a study in Clinical Cancer Research last October in which they retrospectively studied liquid biopsy results from 69 patients treated with immunooncology drugs.
According to Kurzrock, the Samsung team’s strategy of stratifying pretreatment mutational load into tertiles is akin to what her team explored using a specific binary cutoff.
Although she cautioned that both studies were performed in small groups of patients, Kurzrock said in an email this week that growing evidence from such efforts has built confidence among researchers in the field that gleaning mutational load in ctDNA is “definitely possible and relevant.”
However, she wrote, “it may be more complex than initially thought.”
“There may need to be more than 73 genes sequenced in order to optimize the blood TMB calculation” she said. “[The] bottom line is that this is a very promising area of research for predicting immunotherapy response but will need some standardization and optimization.”
Luckily, this need for standardization has begun to be addressed — in at least two independent initiatives, one led by Friends of Cancer Research in the US, and the other by the Quality Assurance Initiative Pathology (QuIP) in Germany.
The Friends of Cancer Research effort has disclosed that it is examining TMB assays from Foundation Medicine, Personal Genome Diagnostics, Thermo Fisher Scientific, Illumina, Guardant Health, NeoGenomics Laboratories, Qiagen, and Memorial Sloan Kettering Cancer Center, with other interested parties still welcome to join.
Foundation Medicine already offers clinical TMB assessment using tumor tissue samples as part of the readout from its FoundationOne comprehensive NGS assay. But the company has said that it is also advancing a blood-based version of its TMB test, initially as a companion to Roche/Genentech’s immunotherapy drug Tecentriq (atezolizumab) in first-line treatment of non-small cell lung cancer patients.
Johns Hopkins and Memorial Sloan Kettering, meanwhile, were some of the first academic institutions to publish on the predictive nature of TMB in relation to immunotherapy response. And Personal Genome Diagnostics, originally a Hopkins spinout, announced in February that it has reached an IP agreement with MSKCC regarding commercialization of both tissue- and blood-based TMB products.
At the time, PGDx executives did not comment on whether its consolidation of IP in this area might reflect a possibility for future legal challenges as competing assays enter the market, but CEO Doug Ward said in a statement that the firm was “confident” in the protection of its “extensive investment and development plans for TMB-enabling tests.”
As efforts move forward to continue to validate TMB as a predictive marker and to harmonize existing assays, clinical researchers and oncologists are also continuing to explore what they expect will be a continuum of different tools that may all complement each other in determining which patients can benefit from revolutionary immunooncology drugs.
In the Samsung study this month, researchers didn’t only explore the use of pre-treatment TMB to predict immunotherapy response. They also took a look at whether serial measurements of circulating tumor DNA can provide information about whether a patient is going to respond to treatment or not.
Guardant and other liquid biopsy companies have already begun to collect evidence for the validity and utility of this type of serial testing in the context of targeted drugs. Kurzrock and colleagues, for example, have preciously reported that serial ctDNA could be used to predict response to EGFR inhibitors in the first week of therapy.
The Samsung team’s results offer a step forward for a similar technique in immunotherapy, albeit on a longer timeline. In the study, ctDNA changes were predictive for pembrolizumab response at a longer six-week timepoint.
According to Kurzrock, she and her colleagues have similar results they also hope to publish using a different liquid biopsy technology.