NEW YORK (GenomeWeb) – A Johns Hopkins University-led team has demonstrated the feasibility of finding early-stage, late-stage, and recurrent head and neck cancers by profiling DNA in blood and saliva samples.
As they reported today in Science Translational Medicine, the researchers tested blood and/or saliva samples from more than 90 individuals with squamous head and neck squamous cell carcinomas (HNSCC) to look for somatic mutations or human papillomavirus sequences corresponding to those in each individual's tumor. In the subset of HNSCC patients who provided both blood and saliva samples, they tracked down bodily fluid-based tumor markers in some 96 percent of cases.
Results from the team's proof-of-principle analyses hint that saliva may be a particularly useful source of tumor DNA for HNSCCs that arise in the oral cavity, while circulating tumor DNA in blood samples tended to be slightly more informative when searching for tumors originating in the larynx, oropharynx, or hypopharynx.
"Based on this study, the test is definitely better right now when both plasma and saliva are analyzed," co-corresponding author Nishant Agrawal, a genetics and otolaryngology-head and neck surgery researcher with Johns Hopkins University, told GenomeWeb. "Further studies will define which bodily fluid is indicated depending on the circumstances."
He and his colleagues believe the ability to detect tumor-related mutations in saliva and blood could eventually prove useful for tracking treatment response, watching for recurrence, or even performing primary screening — either in the general population or in individuals deemed at particularly high risk of certain cancers.
While the current study focused on finding mutations present in already-diagnosed tumors, Agrawal noted that it should be possible to find new cancer cases by sequencing DNA in blood, saliva, or other bodily fluids to look for mutations in panels of genes linked to various cancer types.
"This study … was based on identifying HPV or somatic mutations in the primary tumor," he said. "But we have developed a panel that includes HPV, TP53, and a few other genes, which should cover more than 95 percent of head and neck cancers. So that would preclude the need to sequence the primary tumor."
The US has seen a dip in oral cavity cases that's believed to reflect the declining popularity of cigarette smoking, the team noted. During the same time frame, though, the prevalence of HPV-positive tumors originating at other head and neck sites, particularly tumors at the base of the tongue, has been climbing.
For the current study, the researchers set out to look at how well DNA from such tumors could be picked up in blood and saliva — the latest in a series of studies that have searched for and scrutinized circulating tumor DNA in various bodily fluids.
To do this, they started by searching for marker mutations or HPV sequences in tumor samples from 93 individuals with HNSCC, including 73 individuals with advanced disease. Just under half of the cases originated in the oral cavity, while 34 individuals had HSCC tumors of the oropharynx. Ten individuals had larynx tumors and three had tumors arising from the hypopharynx.
Between HPV status and targeted somatic mutation testing, Agrawal explained, it was relatively straightforward to track down at least one tumor-specific marker to search for in the blood or saliva.
Just five cases required low-coverage whole-genome tumor sequencing. Another 30 tumors contained HPV16 sequences that were easily identified by PCR and another 58 tumors carried somatic mutations found by targeted, multiplex sequencing on HNSCC-related genes such as TP53, NRAS, and PIK3CA.
With these tumor markers in hand, the researchers used either digital PCR to find telltale HPV sequences in saliva samples from all 93 patients, or another PCR-based approach called the Safe-Sequencing System or Safe-SeqS to look for tumor-specific somatic mutations. Their search uncovered tumor DNA in 76 percent of the saliva samples.
A slightly higher proportion of blood samples yielded tumor DNA. The team found tumor DNA in 87 percent of blood samples from 47 HNSCC patients who provided plasma.
But when both blood and saliva were considered for the same 47 individuals, the researchers' ability to find circulating tumor DNA jumped to 96 percent.
The saliva samples served up DNA containing characteristic tumor markers for all 46 cases involving oral cavity tumors, but allowed detection of fewer than 70 percent of tumors arising from other sites. For individuals with HNSCC tumors originating in the larynx, oropharynx, or hypopharynx, blood tended to be a more reliable source of circulating tumor DNA.
Agrawal noted that these results might reflect methods used to collect saliva samples for the study. It remains to be seen whether DNA from tumors at other sites might be better detected by asking individuals to gargle when giving a spit sample, for instance, rather than simply rinsing their mouths.
He and his colleagues are exploring this and other strategies to notch up the sensitivity of their tests.
The team also plans to validate the approach in larger groups of individuals with or without head and neck cancers to further assess its performance, define indications for the test, and tease out factors that might lead to false-positives. For example, Agrawal said, researchers still need to figure out what effect HPV carriage in healthy individuals might have on this screening approach, if any.