Tumor-Informed ctDNA MRD Assay May Predict Recurrence and Survival Outcomes in HPV-Independent Head and Neck Cancer

Key Clinical Summary:

• Design/Population: A single-institution study evaluated a tumor-informed ctDNA minimal residual disease assay (PredicineBEACON, whole-exome–based) in 40 patients with locally advanced head and neck squamous cell carcinoma, predominantly HPV-independent (95%), treated with surgery ± risk-adjusted adjuvant therapy; MRD was assessed post-treatment and during surveillance.
• Key Outcomes: A positive MRD result at treatment completion and during surveillance was significantly associated with worse recurrence-free survival and overall survival. During surveillance, the assay demonstrated 67% sensitivity, 90% specificity, 83% positive predictive value, and 78% negative predictive value for recurrence. On multivariable analysis, MRD positivity was more strongly prognostic than traditional high-risk features (eg, extranodal extension, positive margins).
• Clinical Relevance: Tumor-informed ctDNA MRD testing provides meaningful prognostic stratification in HPV-independent HNSCC, a population with limited molecular surveillance data, and may inform future strategies for risk-adapted treatment escalation or de-escalation and earlier detection of recurrence.

Thomas Roberts, MD, MBA, Mass General Brigham Cancer Institute, Boston, Massachusetts, presented data evaluating a tumor-informed ctDNA-based MRD assay in primarily HPV-independent, locally advanced head and neck squamous cell carcinoma at the 2026 Multidisciplinary Head and Neck Cancers Symposium in in Palm Desert, California.

In 40 patients treated with surgery and risk-adapted adjuvant therapy, a positive MRD result at treatment completion or during surveillance was significantly associated with worse recurrence-free and overall survival, with strong specificity and positive predictive value for recurrence. 

Transcript:

Hi, my name's Tom Roberts. I'm a medical oncologist at Mass General Brigham Cancer Institute. I'm here today to talk about a study that we presented at the recent Multidisciplinary Head and Neck Cancer Symposium in Palm Desert, where we reported on the prognostic value of a tumor-informed ctDNA MRD assay, particularly focused on patients with HPV-independent, locally advanced head and neck squamous cell carcinomas.

As many of you know, there's been lots of work within oncology looking at the utility and the performance of various MRD assays across different tumor types. In the world of head and neck oncology, there's been a couple studies that have shown that these studies perform quite well in HPV-associated head and neck squamous cell carcinomas, but the data in HPV-independent squamous cell carcinomas is a little bit more limited.

Here we worked with a company called Predicine to use their MRD assay, PredicineBEACON, which is a whole exome sequencing assay. It looks at about 20,000 genes, and they make a tumor-specific panel for each patient. This is using tissue to make the tumor-specific MRD panel. That includes 50 personalized somatic variants, as well as a fixed panel of 500, more or less 500 actionable and hotspot variations. Then for the MRD testing, the applied ultra deep sequencing with the threshold for positive test being to detect 2 or more of the specified alterations for a specific patient.

We used that assay and we followed 40 patients who were treated initially with surgery, followed by risk-adjusted adjuvant treatment at Mass Eye and Ear infirmary here in Boston. Of these patients, 38 of them, so 95% of them, had HPV-independent disease, 88% of them had stage 3 or stage 4 non-metastatic disease, about 2/3 of them were new diagnoses, and 35% of them were recurrent disease, and most of them, 63% of them were current or former smokers.

We assessed the performance of the MRD assay at 3 points in each of their course. We did it during what we called the early phase, immediately after patients finished treatment, during the treatment completion so in a several week period after patients completed treatment, whether that being surgery alone or postoperative radiation with or without chemotherapy, and then during the surveillance period later on.

We found that particularly during the treatment completion period and during the surveillance period, after patients had finished their completion, a positive MRD result was associated with worse recurrence-free survival and worse overall survival that was in both the surveillance and the treatment completion period.

When we looked at the test characteristics of this assay during the surveillance period, we found that the sensitivity was 67%. The specificity was 90%. The positive predictive value was 83% and the negative predictive value was 78%. These values are all for recurrence. Overall, these values are consistent with previously reported performance for these assays.

Also when we did multivariate Cox proportional hazard models to look at how a positive MRD test compared to these traditional high risk factors such as extranotal extension or positive margins, we found that in this model, a positive MRD test was more strongly associated with worse overall survival and worse recurrence-free survival, and both of those findings were statistically significant.

Overall, we found that this tumor-informed ctDNA-based MRD assay was statistically significantly associated with worse recurrence-free survival and worse overall survival in patients primarily with HPV-independent head and neck squamous cell carcinoma.

The next steps for this work is really to continue some of the efforts that are going on to figure out how exactly we incorporate these assays into clinical care, whether or not they can be used to either escalate or de- escalate treatment to reduce the toxicity of treatment and/or improve outcomes for patients, and whether or not there's a role for them in treatment for patients with recurrent metastatic head and neck squamous cell carcinoma.

Source:

Roberts TJ, Ruiz-Torres D, Mendel J, et al. Prognostic value of tumor-informed ctDNA-based MRD assay in HPV-independent locally-advanced head and neck squamous cell carcinoma. Presented at Multidisciplinary Head and Neck Cancers Symposium. February 19 - 21, 2026. Palm Desert, California. 2034.