Recurrence rates in head and neck squamous cell carcinoma (HNSCC) remain substantial, and standard post-treatment surveillance using clinical exam and imaging has known limitations. Circulating tumor DNA (ctDNA) has demonstrated utility in HPV-positive HNSCC, but its role in non-virally associated HNSCC is less well-defined. This retrospective study evaluated a methylation-based ctDNA assay to detect recurrence and to assess associations with survival outcomes during post-treatment surveillance.
This was a retrospective single-institution cohort including patients with non-metastatic, non-virally associated HNSCC who received definitive therapy and underwent at least one post-treatment ctDNA test between January 2024 and October 2025. Fifty patients met inclusion criteria. Most primaries were in the larynx (38%) or oral cavity (36%). The cohort was weighted toward advanced disease: AJCC stage IVA–IVB in 72%, T3–T4 tumors in 69%, and N2–N3 nodal classification in 48%. Definitive treatments were chemoradiation for 54% and surgery for 46% of patients.
Plasma cell-free DNA was analyzed using Guardant360’s methylation-based approach that detects tumor-specific DNA methylation patterns via next-generation sequencing. A predefined threshold defined a positive ctDNA result as a methylation signal of 0.05% or greater. The analysis focused on diagnostic accuracy metrics and survival associations during follow-up surveillance.
Primary diagnostic measures reported included sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and area under the receiver operating characteristic curve (AUC). Survival outcomes were progression-free survival (PFS) and overall survival (OS), estimated with the Kaplan–Meier method. The study additionally reported the interval by which ctDNA detection preceded clinical recurrence detection and performed an exploratory analysis of tumor-associated mutations related to recurrence.
At a median follow-up of 18.0 months, 13 of the 50 patients experienced disease recurrence. Twelve patients had positive ctDNA results during surveillance. Diagnostic performance of the methylation-based assay was: sensitivity 76.9%, specificity 94.6%, PPV 83.3%, and NPV 92.1%. The AUC was 0.858 (95% CI, 0.733–0.982), indicating good discriminative ability in this cohort.
In patients with both recurrence and positive ctDNA (n = 10 reported for this comparison), ctDNA detection preceded clinical detection in 6 patients (60%), with a median lead time of 131 days. These findings suggest the assay can identify molecular recurrence before it becomes clinically or radiographically apparent in a substantial fraction of cases.
Positive post-treatment ctDNA was associated with markedly inferior survival outcomes. One-year progression-free survival was 25.0% among ctDNA-positive patients versus 96.4% among ctDNA-negative patients. One-year overall survival was 75.0% for ctDNA-positive versus 100% for ctDNA-negative patients. Both PFS and OS comparisons reached P < .001, indicating strong statistical associations between a positive methylation-based ctDNA result and worse prognosis in this cohort.
An exploratory analysis of tumor mutations identified associations between recurrence and alterations in TP53, SETD2, and ROBO2. The report presents these findings as exploratory and does not provide effect sizes or multivariable adjustment details in the abstract; further details were not reported in the source summary.
In this retrospective single-center cohort, a methylation-based ctDNA assay demonstrated high sensitivity and specificity for detection of disease recurrence in non-virally associated HNSCC and showed prognostic association with both PFS and OS. ctDNA frequently preceded clinical detection of recurrence by several months, supporting the potential for earlier intervention or intensified surveillance when molecular recurrence is identified. The authors conclude these results support integrating methylation-based ctDNA into surveillance strategies for non-viral HNSCC but emphasize the need for prospective validation.
This study is a retrospective analysis from a single institution with a modest sample size (n = 50) and a median follow-up of 18.0 months. The data derive from a preprint that has not undergone peer review. Detailed methods, timing and frequency of surveillance testing, criteria for clinical recurrence determination, and multivariable analyses were not reported in the available abstract; those details would be necessary to fully assess potential biases and generalizability. The authors explicitly call for prospective validation before routine clinical implementation.
The Albert Einstein College of Medicine Institutional Review Board approved the study. The authors declared no competing interests. De-identified data supporting the findings are available from the corresponding author upon reasonable request. The manuscript is a preprint and has not been certified by peer review, and the authors note this limitation in interpretation for clinical practice.