NEW YORK – Princess Margaret Cancer Centre researchers have developed a next-generation sequencing approach for detecting very low levels of human papillomavirus (HPV)-related circulating tumor DNA.
The method, called HPV-Seq, detected ctDNA with high sensitivity and specificity, at levels approximately twentyfold lower than with digital PCR. As such, it could be used to help hone precision treatments of HPV-related cancers, and may eventually be applied to detecting other viruses that cause cancer, such as certain stomach cancers and lymphomas, according to the researchers.
HPV can cause cancers in the throat, mouth, and genitals, but tests to detect it in the bloodstream have shown limited sensitivity. The presence of viral sequences in ctDNA helps to distinguish tumor DNA from other sources of cell-free DNA.
Because these cancers typically present with locally or regionally confined disease, they can be treated with intent to cure. Treatment often consists of chemotherapy with radiation (chemoradiation) or surgery, followed by adjuvant therapy.
A more accurate test of minimal residual disease could enable more personalized adjuvant care by identifying those patients without MRD, who would therefore not benefit from further treatment.
In order to improve the accuracy of detection, the Princess Margaret team developed a dual-strand hybridization capture protocol to more completely cover HPV sequences found in patients.
In general, hybrid capture-based sequencing consists of fishing out fragmented target nucleic acid sequences with DNA or RNA-based probes, or "bait."
"There are over a dozen HPV types that cause cancer," Scott Bratman, the study's principal investigator, explained via email. "Because their genomes are so divergent, PCR-based methods cannot easily detect them all at the same time; this is precisely what hybrid capture sequencing is designed to do. HPV-Seq incorporates hundreds of baits that recognize whole genomes from multiple HPV types."
A single full HPV genome copy can produce roughly 50 cell-free DNA fragments, Bratman further explained, each of which harbors distinct HPV sequences. By using optimal concentrations of baits targeting both HPV DNA strands — and through further optimization of other hybrid capture reaction aspects — Bratman estimates that HPV-Seq can potentially capture all 50 fragments.
"We found that dual-strand hybrid capture boosted recovery of HPV fragments and could be incorporated into standard workflows," he said.
The team retrospectively tested their method, using patient samples from a cohort of 17 locally advanced cervical cancer patients and 13 oropharynx cancer patients taking part in a prospective multicenter clinical study (NCT02388698).
The team used nine HPV panels testing for over 38 HPV types. These panels covered a range of baiting schemes, from targeting the positive strands of the E6 and E7 genes of 38 HPV types, to focusing on both positive and negative strand baits of a narrower range.
This suggests that HPV-Seq could potentially be used to screen for a wider number of HPV types than is currently possible with PCR panels marketed by companies such as Abbott, Hologic, Qiagen, and Roche.
Overall, HPV-Seq detected approximately 0.3 copies/mL of HPV ctDNA, with 100 percent sensitivity and 67 percent specificity. The results of both assays were highly correlated and HPV-Seq also identified patients with reads mapping to alternate HPV types, indicating the likelihood of harboring multiple infections.
The ability to accurately and reproducibly detect multiple HPV strains at under one copy per mL could address issues related to variance between assays. Past studies have found that results can vary between commercially available tests, raising concerns about false positive and false negative readings.
The clinical utility of HPV-Seq — such as how therapeutically meaningful it is to detect more HPV types than commercially available assays — remains to be validated in larger study sizes.
Future studies will also investigate the method's generalizability to other HPV-related cancers and its usefulness in monitoring their related treatment responses.
The HPV-Seq protocol also potentially allows for sequences related to other viruses to be used as bait, further expanding the technique's utility.
In the near-term however, Bratman and colleagues are investigating ways to improve the assay's specificity while maintaining its sensitivity by expanding the number of timepoints at which patient samples are tested.
Bratman pointed out that chemoradiation can take weeks to reach its full effect, making it possible for a highly sensitive assay like HPV-Seq to find signs of residual cancer in the bloodstream even in patients who will ultimately be cured.
"We are now comparing different timepoints after completion of treatment to see if specificity can be improved while maintaining high sensitivity," he said. "We are also looking into whether the HPV ctDNA fragmentomic profiles can help distinguish between true and false positives."
The Princess Margaret team has filed a patent application for various aspects of the HPV-Seq technology through the University Health Network. It is available for licensing.