José P. Zevallos, MD, MPH, FACS, The Eugene N. Myers MD Professor and Chair of the Department of Otolaryngology-Head & Neck Surgery at the University of Pittsburgh and Deputy Director of the UPMC Hillman Cancer Center was the sole presenter at the Eye & Ear Foundation’s November 17th webinar, “Improving Diagnosis and Outcomes for Head and Neck Cancer Using Liquid Biopsy.” He called the Department one of the founding centers for modern heck and neck surgery.
The Department of Otolaryngology-Head & Neck Surgery
The Department has 60 full-time faculty, 40 devoted to clinical and 20 to research. There are 7 clinical divisions, 25 residents, 7 fellows, 2 T32 training programs, SPORE in Head and Neck Cancer, and 8 clinical sites.
Head and neck cancer (HNC) is only treated effectively with a multidisciplinary approach. Pitt has a very strong multidisciplinary HNC team, which works together to ensure patients receive optimal care.
“We are at the forefront of research in many different aspects at Hillman Cancer Center,” Dr. Zevallos said, “and we are very proud to be one of the largest and leading groups in HNC both in clinical care and research.”
Circulating Tumor DNA (ctDNA)
Our cells are constantly producing, Dr. Zevallos said. As our cells replicate, we produce cell-free DNA, which is freely floating DNA in the blood that is shed naturally by all normal cells. Circulating tumor DNA is cell-free DNA that comes from cancer. These are tumor-derived fragments of cell-free DNA in circulation that are not associated with tumor cells, so they are the product of necrosis, apoptosis, or exosomal release.
When we talk about ctDNA, it is used most often for something called measuring molecular residual disease (MRD). When a person has their cancer treated, ctDNA can be measured in blood or plasma after their curative surgery as a marker of what is left behind residually.
“The idea is that we can use these fragments of DNA to predict a patient’s risk of recurrence or even in the setting of someone who is at risk of cancer but doesn’t have cancer yet,” Dr. Zevallos said.
There are two ways of doing this. In patients already diagnosed, a tumor-informed approach means taking a sample of the primary tumor, analyzing it, finding the cancer mutations, and using those to find the same mutations in the blood. The other way is tumor-naïve, in which the primary tumor is not needed. Mutations commonly found in this kind of cancer can be found in the blood as a marker. Both approaches work for different cancers.
Incorporating Data into Practice
Colon and lung cancers are the two places where using ctDNA has become standard of care. A very important paper published two years ago in the New England Journal of Medicine looked at the use of ctDNA for guiding adjuvant therapy after surgery for Stage 2 colon cancer. Patients were either selected to have surgery with some adjuvant treatment based on the pathology. The other group based the decision to have adjuvant treatment only on ctDNA. There was no difference in recurrence free survival in any of these patients. But half of the patients who were treated with standard of care received chemotherapy, compared to the ctDNA group. This means if decisions were made based on ctDNA, those patients were able to avoid chemotherapy. The outcomes were the same. This landmark study set the stage for using ctDNA to make adjuvant therapy decisions in colon cancer and has become an important part of the treatment paradigm for colon cancer.
Significant work has also been done for lung cancer. A 2017 study led by a close collaborator of Dr. Zevallos looked at early detection of MRD and lung cancer using ctDNA. The blood was checked at four months post treatment with chemotherapy and radiation and was able to determine whether it predicted recurrence. Patients who did not have any ctDNA detected had significantly improved outcomes. In other words, ctDNA MRD detection is highly prognostic after localized lung treatment.
Technology continues to evolve to incorporate ctDNA into workflows for patients with these cancers. But what about HNC? “Oftentimes because HNC is relatively rare compared to lung and colon cancer, we are often not at the forefront of industries as they start thinking about assays like this,” Dr. Zevallos said. But there has been work done specifically for HNC in the use of ctDNA. The best example is patients with HPV-associated oropharynx cancer.
HPV-Associated Oropharynx Cancer
A study from 2020 measured HPV DNA in patients with this cancer, looking at 115 patients who had treatment, of which there were 15 recurrences. The negative predictive value of HPV DNA in the blood was 100%, and the positive predictive value was 94%. So, this test is very good at identifying those patients at highest risk of recurrence after completing treatment. The median was 3.9 months prior to detecting cancer clinically. In one patient, HPV was found in their blood 14 months before it manifested as a clinical recurrence.
The challenge is patients can have molecular positivity up to 14 months before detectable cancer. “What do we do in that time period?” Dr. Zevallos asked. “Do we increase the surveillance? Do we incorporate more exams or imaging to find it? This could cause significant anxiety for the patient, but there’s nothing to do about it because we can’t see the cancer yet. This is where these technologies have to really be incorporated into how we treat our patients to ensure they have meaningful survival.”
First Prospective Study
All studies up until a study published earlier this year were retrospective, meaning they looked at patients who were treated in the past. This was the first prospective study using HPV DNA in patients treated with HNC. What they found was that the results were not quite as good as had been shown previously. The sensitivity was 73% for 115 patients and 15 recurrences. There were 40% early true positives, 33% “confirmatory” true positives, 27% false negatives, and 20% false positives. False negatives are patients who went on to have a recurrence, but the test missed it. False positives are patients who had HPV DNA but no recurrence.
These rates call into question the utility of this assay in clinical practice, making ctHPV DNA ineffective for adjuvant therapy decision making. Because the study failed to meet the primary endpoint, it was closed early. “We think the sensitivity of the test was not adequate enough to find those minute amounts of DNA predictive of recurrence,” Dr. Zevallos said.
Potential Utility of ctHPV in HPV (+) Oropharynx Cancer (OPSCC)
There are many potential usages, including early detection, evaluation of a lateral neck mass, response to induction immunotherapy/chemotherapy, measuring postoperative minimal residual disease, improving disparities and promoting equity.
Does measuring circulating HPV DNA save or extend lives? Is it cost effective? This test is now commercially available and ordered all the time in the community, but the reality is no data has shown this test leads to improvement in overall survival. A lot more work needs to be done before really incorporating HPV DNA into treatment.
Current Limitations of ctDNA Assays for HNC
- Prospective data for predicting recurrence is underwhelming
- No large prospective studies comparing ctHPV DNA to routine imaging/clinical examination
- No demonstrated survival advantage of earlier molecular detection
- No available NGS-based assay for variant detection
- Not actionable within the curative pathway
- Recent data on adjuvant therapy decision making is underwhelming
- What about HPV (-) HNC?
No tests are available for the last two items on the list. This led to the question: Can sensitivity for adjuvant therapy decision making be improved by sampling more proximally than blood?
Circulating Tumor DNA – Looking Beyond the Blood
“When I look at the challenges and where we can have the most impact in how liquid biopsy could be used, it comes down to incorporating liquid biopsy into the curative pathway, which means finding ways of bringing it into our treatment planning before recurrence,” Dr. Zevallos said. Although HPV+ cancers are critically important, survival is about 90%. HPV- cancer survival is around 50%, so there are more opportunities for precision treatments.
Instead of drawing blood from the arm and measuring it, is there a way of going directly to the source to measure biofluids in the head and neck? People have been looking at saliva for a long time with mixed results, and there are all kinds of studies looking at urine, stool, and other biofluids.
Plasma has inherent limitations, as it requires ultrasensitive detection given its very low variant allele fraction (VAF), and it does not provide localization of MRD. Proximal biofluids may be more enriched for tumor-associated variants and provide more sensitive detection of locoregional MRD.
Novel Approach Providing the Earliest Landmark Timepoint
Four years ago, at Dr. Zevallos’ lab at Washington University, they decided to look into surgical drain fluid. As a surgeon, Dr. Zevallos deals with it on a weekly basis. Prior to this work, no one had explored whether this fluid could be an accurate biomarker.
This fluid had the potential for prognostication as well as adjuvant therapy decision making, localizes the risk of recurrence, and is adoptable across multiple solid tumor cancers. Given the fluid comes directly from the tumor bed, it may have a stronger ability to find patients at risk of recurrence. Surgical drains are used very commonly in cancers, across the body, so it is a novel fluid platform.
The first paper was published in 2023, which looked at 120 patients with HPV OPSCC. Fluid was collected from their neck because they all had neck dissections in addition to tumor removal. This was compared to blood drawn at the same time. A PCR-based assay for HPV DNA was run and compared to next generation sequencing as well.
The amount of DNA in the neck – found in the fluid – correlated very well with the risk of the patient in respect to their pathology. In other words, surgical drain fluid (SDF) reflects nodal, rather than primary tumor features, and corresponds to adjuvant treatments. It also classifies patients according to their risk of recurrence. Patients that had HPV DNA in fluid measured 24 hours post-op had the highest risk of recurrence compared to those who were negative. Mutations also lined up with what would be expected to see, which gave the team confidence that they were dealing with cancer ctDNA rather than just HPV DNA alone.
A 2024 commentary written about this paper put the discovery in a general context. SDF does not replace blood or pathology but is one more data point used to provide precision care.
HPV- HNSCC
HPV- HNSCC is the deadliest. A similar study was published last week in Clinical Cancer Research, which had 120 patients in two different cohorts – one at Washington University and one at UPMC. The findings were similar. Twenty-four hours post-op, presence of tumor DNA in the surgical drain was significantly associated with survival. Looking at more specific subsets of HPV- HNC found that SDF was significantly more prognostic than blood.
The biggest difference between positives and negatives were the intermediate risk patients, where tough decisions must be made.
The test sensitivity was compared and found to outperform the majority of traditional pathologic features used to make decisions about radiation and chemotherapy post-op. Then they looked at other different kinds of patients with HPV- cancer – low risk patients. The test worked well. This patient population sometimes gets recurrence, and often they get surgery and no radiation. These patients in particular had a separation between those who were positive and negative. The numbers were very small, but the potential is there to use this test for this group.
A clinical trial could be designed to escalate therapy for patients positive in SDF. Based on pathology, patients would just get radiation. But if the fluid is positive, they would get chemoradiation. Therapy would not be taken away but would be added for those deemed highest risk based on the signal in the fluid.
Centers across the country are collaborating with Pitt to design trials and collect more fluid. “This is exciting because no one had ever done this work before,” Dr. Zevallos said. This fluid was considered waste fluid, but now it can be leveraged to improve patient care.
Pitt recently received a big grant from NIH to explore this in more detail. The focus will be on the pre-analytical side, to make sure the fluid is ready to be used for the sophisticated assays.
Dr. Zevallos anticipates in the future having three different bioanalytes that could be used immediately after surgery to predict risk of recurrence: surgical drain fluid and blood (nodal and distal burdens).
This fluid is also being used to explore immune cell composition, primarily lymphatic fluid. Immune cells that are found appear to be similar to the immune cells seen in the primary tumor, which means it could be potentially used as a window into the tumor microenvironment.
Smaller grants have been received, and a larger grant is being proposed to do a full characterization of the immune composition in this field. Dr. Zevallos is also working with Sunny Bao’s group at Pitt. She is an immunologist who can grow an organoid in a dish that has all the different characteristics of a tumor. The hope is to demonstrate that cells harvested can kill tumor cells and different types of therapies can be tested on these models to allow for more precision in patient care. One of the newest hires to the Department, Josh Smith, MD, is working with her.
Conclusions
- ctDNA has potential as a marker of molecular residual disease across multiple cancer types; utility extends from diagnosis to treatment selection, response, and surveillance
- Strength of the data in HNC is far behind other cancer types and should be interpreted w caution
- While widely utilized, prospective DNA using HPV ctDNA for surveillance is disappointing
- Locoregional MRD detection and postoperative detection remain challenges that may be overcome using more proximal biofluids
- ctDNA SDF, a novel proximal bioanalyte, may provide MRD data in HPV+ and – HNSCC, requires further exploration
Cohorts are forming in lung, bladder, and breast cancer across the country with cancer centers. “Hopefully we will have more data to share in the coming years,” Dr. Zevallos said.