Researchers at The University of Texas MD Anderson Cancer Center recently discovered a novel biomarker that re-categorizes human papillomavirus (HPV)-related oropharyngeal cancer into three—not two—subgroups (JCI Insight 2018; doi:10.1172/jci.insight.124762). The findings pave the way for a much-needed update to the treatment protocol, especially considering HPV-related head and neck cancer has become the most prevalent HPV-related cancer in the U.S., accounting for approximately 70 percent of all cases of oropharyngeal cancer.
Currently, HPV is categorized as either HPV-positive or HPV-negative, each with distinct treatment outcomes. While a combination of chemotherapy and radiation provides a 5-year survival rate of approximately 85 percent for patients with HPV-positive oropharyngeal cancer, patients with the HPV-negative form only have a survival rate of 55 percent.
“Most patients with HPV-positive head and neck cancer have good survival outcomes, but many suffer from long-term morbidity due to the intense radiation-based therapy that is usually given,” stated the study's corresponding author Curtis Pickering, PhD, Assistant Professor in the Department of Head and Neck Surgery-Research at the University of Texas MD Anderson Cancer Center. “There is also a subset of patients whose tumors don't respond to the therapy, and they have poor long-term survival.”
Although the intense treatment side effects have led to the development of de-escalation protocols to help reduce morbidity, no methods currently exist to identify good candidates for such reduced treatment plans. And, considering the 15 percent fail rate for those patients with HPV-positive head and neck cancer, de-escalation for all HPV-positive patients isn't an option.
“Since HPV function is believed to drive the good response to therapy, we wanted to understand if there were variations in HPV function that might be responsible for the variable survival outcomes among patients,” Pickering said.
Previous research has led to a better understanding of the mechanisms involved in HPV carcinogenesis, but the current understanding of HPV function remains incomplete (Nat Rev Cancer 2010; doi:10.1038/nrc2886). Pickering and his team didn't even have a consensus definition of HPV function to apply during their analysis, forcing them to use panels of genes linked to HPV, they noted in the study. While an imprecise measure, it provided significant findings.
New HPV Classification
The resulting study, recently published in JCI Insight, used data from 80 oropharyngeal cancers in The Cancer Genome Atlas. The researchers analyzed the tumors to assess both their level of HPV function and response to treatment. They found a biomarker that effectively classifies HPV-positive tumors into a high-risk and a low-risk group based on gene expression. The high-risk group has an intermediate gene expression profile between HPV-negative and HPV-positive tumors—with survival rates similar to HPV-negative cancers.
“This biomarker and its related biology are also associated with response to radiation treatment in cell lines, suggesting why the high-risk patients don't respond to therapy,” Pickering explained.
The newfound association between the high-risk HPV-positive group and treatment response could eventually lead to much-needed treatment advances, allowing for safer implementation of de-escalation treatment protocols. It could also help facilitate testing of new treatment approaches for patients with unresponsive tumors.
Since HPV-negative oropharyngeal cancers are largely tobacco driven, the researchers explored other variables that could account for the different gene expressions. They found that, while there were a higher number of nonsmokers in the low-risk HPV-positive group, neither the amount of smoking nor the specific risk score or clinical stage was significantly associated with survival. Categorizing patients based on the new signature was far more prognostic.
To further assess the validity of the gene signature, the team evaluated the biomarker in two other independent patient groups. First, they tested it among 47 HPV-positive oropharyngeal cancer patients from Johns Hopkins University. The biomarker again revealed two distinct patterns of gene expression with significantly different 5-year survival rates. Next, they tested the gene signature independent of tissue site with a cohort of HPV-related cervical cancer patients finding similar results.
The validation data confirmed that the gene expression remains consistent across cohorts and is associated with a poorer prognosis phenotype in some HPV-positive cancers. It also showed that the biomarker performed better than current clinical risk factors such as smoking status and tumor size. While the biomarker's prognostic capabilities across tissue sites may lead the way for a broader clinical application, the researchers remain focused on oropharyngeal cancers for now.
Bringing It to the Clinic
In an effort to bring the research one step closer to a clinically useful risk classification system, the team then identified a panel of 582 genes that distinguished high- and low-risk HPV-positive and HPV-negative groups. They further narrowed the panel to just 38 genes—and the researchers hope to one day end up with just one gene with the same prognostic ability. They already know at least three of the HPV-correlated genes have prognostic power when assessed alone.
“We hope this work will lead to a clinical assay that can be used to identify low-risk patients for reduced intensity treatment and new treatment options for high-risk patients,” Pickering stated. “It could provide a risk score that would then be used to select treatment options. Low-risk patients could be candidates for reduced intensity treatment to minimize long-term morbidity but still have good survival outcomes. High-risk patients could be selected for more intense therapy or clinical trials of new agents.”
Pickering and his team at MD Anderson are already forging ahead with efforts to translate their findings into a biomarker assay that could be used on routine clinical specimens. But they have a long road ahead of them with several hurdles.
The current study was limited by a small sample size, as well as its reliance on RNA sequencing data. Clinical specimens are often small, and the collection method doesn't necessarily preserve the RNA necessary to identify these specific genes, limiting the utility of this diagnostic approach. Despite this, Pickering is hopeful.
“It should still be feasible with currently available technology,” he noted. “Once the assay is optimized, we will test it on additional retrospective cohorts of samples and eventually in prospective samples, ideally as part of ongoing clinical trials.
Rebecca Hepp is a contributing writer.