The approach, which was spearheaded by researchers at the University of North Carolina and is being commercialized by startup Naveris, would potentially allow doctors to more accurately and more easily identify which patients are free from cancer after radiation treatment, and which might be poised to recur.
This would either offer a way to pick out a subset of patients with a positive DNA result, for more expensive imaging surveillance, or potentially subsume the need for repeated imaging all together, by providing a new tool for recurrence monitoring.
Unlike liquid biopsy methods that search for non-viral tumor DNA mutations among the nucleic acid fragments circulating in the blood, the oral cancer assay that the UNC team described this week at the Annual Meeting of the American Society for Radiation Oncology relies on the fact that viral DNA — in this case from human papillomavirus, or HPV — is integrated in the genome of cancerous cells.
Because the target, this viral DNA, is the same in all patients, a test doesn’t require broad next-generation sequencing or the development of biomarker panels for each individual patient, as is being explored for the monitoring of other solid tumors like breast and colorectal cancer and being commercialized by firms like Natera.
The potential of viral DNA to allow blood-based detection of certain cancer types is something other groups have also hit on, most notably a team led by Chinese University of Hong Kong researcher Dennis Lo, which is now commercializing a test for early detection of nasopharyngeal cancers associated with Epstein Barr virus in partnership with US firm Grail.
Bhisham Chera, a co-investigator on the study on the HPV DNA test and Naveris board chair with his UNC colleague Gaorav Gupta, said that the principle at work in his team’s work with HPV cancers is similar to Grail’s EBV example, though he couldn’t speak to the details of the molecular detection strategies.
For HPV cancers, Gaorav, Chera and colleagues have developed a strategy that uses Bio-Rad Laboratories’ droplet digital PCR platform to detect and quantify circulating tumor HPV DNA representing several HPV types.
Chera said that the team has filed for IP protection on the method and can’t discuss it in detail, but it involves a strategy that uses multiple targets to make it possible to distinguish fragmented HPV from bits that are from an intact genome.
The group has also stated that the methodology distinguishes viral DNA coming from a tumor from non-tumor-derived HPV that might be present in the blood, and authors described the assay as providing redundant detection of the same DNA fragments.
Importantly, the team has already been able to collect evidence for the method in a prospective cohort, something that many similar applications of liquid biopsy are still working on.
The prospective data presented this week showed a very strong negative predictive value and promising sensitivity — every patient who developed a recurrence scored positively using the test, and not a single individual who tested negative has yet recurred.
The team’s prospective study involved 89 patients with HPV-associated oral and throat cancers whose disease had not metastasized. All patients received definitive chemoradiation therapy and had some combination of PET/CT and chest X-ray scans beginning three months after treatment completion.
Patients also had clinical exams every two to four months for the first two years of follow-up, and then every six months thereafter, and doctors took a blood sample at each visit to be tested for tumor-derived HPV DNA.
According to the authors, 70 of the 89 patients had undetectable levels of ctHPVDNA at every follow-up visit following the three-month, post-treatment scans. None of these 70 patients have so-far showed any sign of cancer recurrence.
The remaining 19 patients developed a positive ctHPVDNA test result at some point between 7.8 months and 30.4 months after treatment. Eight of these were diagnosed with cancer recurrence. The other 11 have not yet shown evidence of disease and are being monitored with repeat blood tests and imaging.
Researchers reported at the meeting that four of the individuals who had an initial positive test have now spontaneously cleared their HPV DNA signal, suggesting that they may have become free of disease, albeit later than others. This could represent immunosurveillance, Chera said.
Either way, the method is not designed to offer a single-timepoint prediction of recurrence, and so the investigators expect that the positive predictive value should increase as these patients continue to be watched and either the detectable signal clears or it becomes apparent that a cancer has returned.
The ability to see a change in DNA levels over time and to resolve these unclear cases is a crucial aspect of the technology, a boon provided by the team’s choice to use digital PCR, Chera added.
Moving forward the team plans to continue to use the method in research studies while Naveris works out a commercial timeline and strategy.
According to Chera, clinical practice varies in terms of how doctors surveil patients with these cancers. Some clinicians order quarterly imaging tests to try to catch a recurrence while others use longer intervals. The team’s liquid biopsy approach could help standardize that — offering a cheaper tool for close surveillance across the board. Or it could help rule out those who can avoid imaging by identifying patients who clear their DNA soon after treatment.
Naveris has not yet detailed its commercial plans or whether it expects to have to validate the findings in a larger cohort before it launches a clinical assay, but Chera said that the results from the data the team collected in their current prospective study has been positive enough that clinicians are already expressing interest in providing it to their patients.
HPV-associated cancers also occur in sites in the body other than the mouth and throat, suggesting that the same approach could also aid in the monitoring or detection of cervical and anal cancers that arise from the same virus.
The group hasn’t studied this yet, but according to Chera, there is good reason to think it would work. “We think that it will be translatable to those patients. It’s the same virus, so it should perform” similarly, he said.