This story has been updated from a previous version to note that the drop in Natera's stock price Wednesday was likely attributable to a negative report about the company from short sellers Hindenburg Research, not solely the City of Hope research paper as originally reported.
NEW YORK – Researchers at City of Hope have published a new study calling into question the reliability and utility of Natera's Signatera assays for recurrence monitoring in colorectal cancer patients, highlighting the results as evidence that genomic minimal residual disease (MRD) testing may be at best unhelpful and at worst unreliable in detecting recurrences earlier than standard-of-care imaging and blood protein tests.
In the study, published Tuesday in JAMA Network Open, Signatera caught the first sign of recurrence in eight of the 15 patients who recurred in the study, although some cases were also simultaneously, or contemporaneously, identified by imaging.
In contrast, imaging plus proteomic testing identified 11 cancer recurrences in patients without being preempted by circulating tumor DNA (ctDNA) assay-positivity at an earlier time point.
Natera's stock was down around 31 percent at $37.66 in Wednesday afternoon trading on the Nasdaq, likely reflecting investors' reaction to the study and a negative report about Natera's business published by short sellers Hindenburg Research.
The study is a small one, but it provides evidence both that ctDNA can preclude radiologic detection of recurrence, and that, in the context of US-based screening recommendations, it often may not.
Investigators analyzed data from 48 individuals who had had their colorectal cancer resected between September 2019 and November 2021. Of these, 31 patients had stage II-III cancer and another 17 had stage IV tumors.
The researchers compared Signatera ctDNA results, most obtained after adjuvant therapy, to the standard of care recommended by the National Comprehensive Cancer Network, which includes CT scans and carcinoembryonic antigen (CEA) testing.
In contrast to some previous studies of Natera's approach, which compared post-adjuvant Signatera surveillance every three months to imaging at one and three years (a standard of practice in some ex-US areas), patients in the City of Hope cohort were imaged at the NCCN-recommended initial intervals of either three months, six months, or one year, depending on stage.
In their report, the investigators wrote that there were 15 recurrences, overall, in the 48-patient cohort. Positive ctDNA findings were among the first indication of disease recurrence in eight patients, with imaging leading recurrence diagnosis in nine patients, and elevated CEA levels evident at recurrence in three patients. A combination of imaging plus CEA levels, meanwhile, was the lead indicator in 11 patients.
Based on this, investigators calculated that the combination of imaging and CEA had better sensitivity (72 percent) than ctDNA (53 percent) in identifying disease recurrence in the cohort. They concluded overall that the results suggest that ctDNA assay provides "no definitive advantage" over the NCCN recommended standard of care.
Although there were cases where ctDNA positivity significantly predated radiologic recurrence, the study authors also calculated that there was no overall significant difference between the timing of ctDNA recurrence detection versus imaging or imaging plus CEA on average, with all modalities reporting positives at a median of about 15 months.
Marwan Fakih, codirector of City of Hope's gastrointestinal cancer program and lead author of the new study, said in a statement that imaging remains the most important surveillance tool in the follow-up of resected colorectal cancers.
"Clinicians should continue to abide by NCCN guidelines [and] for those who decide to implement ctDNA into their surveillance algorithm, they should be aware of the limitations associated with this assay," he said. "For now, ctDNA can be considered, if any at all, as a complement to the standard approach recommended by the NCCN."
In an email, a Natera spokesperson highlighted that the company "has always supported the use of Signatera in combination with imaging, to inform treatment decisions and detect recurrence as early as possible in patients with a history of cancer."
Notably, some recurring patients in the cohort were either not biopsied or did not have biopsy results described in the study and supplemental materials. Although the authors reported other indicators of true recurrence, such as response to therapy, the lack of biopsy analysis maintains a question of whether those cases represent Signatera false negatives or standard-of-care false positives.
The Natera spokesperson wrote that while it welcomes independent clinical research, "small clinical experience studies are subject to limitations and variability."
Alan Venook of the University of California, San Francisco's Helen Diller Family Comprehensive Cancer Center wrote a commentary accompanying the study report in JAMA, saying that the results highlight the fact that MRD has different implications in different cancers and circumstances.
It's not that there is no evidence of utility for Signatera, he argued. Prospective studies of the approach have showed that MRD positivity at a 30-day post-surgery time point can prompt the administration of adjuvant therapy for CRC patients when it would otherwise not have been used, with the subsequent disappearance of ctDNA correlating with long-term disease-free survival.
But he called that 30-day MRD testing a measure of disease "persistence" rather than disease recurrence.
In terms of recurrence detection, the hope for ctDNA assays has been that they can sound the alarm early enough that physicians can take some sort of similar action that also improves patient outcomes.
In this context, the fact that ctDNA positivity did preclude detection of recurrence on imaging in several cases in the City of Hope cohort would suggest at least some potential complementary value.
Notably, while the study calculated recurrence detection sensitivity for CEA plus imaging, it did not report the same for a combination of ctDNA and imaging, which, if calculated, would represent the highest performance in the cohort — identifying all but one recurrence ahead of CEA.
Impact on outcomes
The study authors also argued that regardless of earlier recurrence identification by ctDNA in some cases, it is unlikely that testing would have changed the treatment of the individuals in question.
For example, one patient with an initial ctDNA positive result was found to have multiple lung metastases that were deemed inoperable. Two other patients underwent curative intent surgery but would have "arguably experienced a similar intervention if followed up by standard surveillance," Fakih and his coauthors wrote. One had an elevated CEA level soon after the positive Signatera test, which they presume would have led to the same PET follow-up, and the other was already scheduled for an MRI that would have taken place just three months after the positive ctDNA result and would have "likely identified the recurrent lesion without compromising his clinical outcome."
The study was not randomized, however, which raises the possibility that — because doctors and radiologists had Signatera information in-hand — their subsequent imaging-based diagnosis of patients might have been influenced by those positive results.
The City of Hope team also highlighted that for two patients who had persistent ctDNA positivity for more than a year before an imaging-confirmed recurrence, the reemergence of cancer eventually came in the form of diffuse retroperitoneal disease and the patients "therefore were not candidates for curative-intent interventions."
Citing anxiety produced in such untreatable cases, Fakih said in an email that "caution should be exerted around widely using a test that is not proven to change outcomes." Despite this, he said he sees Signatera being adopted by oncologists as if that positive impact were a fact.
In his commentary, Venook further pointed out that "discovering MRD via ctDNA in a patient with colorectal cancer before the site of recurrence can be localized does not open the window of opportunity. Instead, it gives advanced notice that a patient has cancer recurrence when the only practical intervention is repeated imaging. In other words, patients receive bad news that is not actionable."
He further argued that patients similar to those in the study might be harmed rather than helped by the earliest discovery of recurrence, "learning of inevitable cancer recurrence with nothing to do but wait," he added.
This landscape is rapidly changing however, with clinical trials now recruiting ctDNA-positive post-adjuvant patients for additional therapy — for example the National Surgical Adjuvant Breast and Bowel Project's Phase II trial using Novartis' gevokizumab.
Fakih said in his email that absent definitive proof from a prospective study showing that ctDNA identifies recurrence earlier than US guideline-recommended surveillance, the assumption should be that liquid biopsy tests aren't an improvement over the current standard.
He added that he and his colleagues hope their findings will help highlight the need for larger studies to try to more definitively answer both this question and questions about resulting survival or outcome advantages.
Efforts to supply this data are already ongoing in studies like the Natera-sponsored BESPOKE trial.