Liquid biopsy for the detection of non-viral head and neck squamous cell carcinoma

Liquid biopsy is a diagnostic method that detects diseases through bodily fluids such as blood and saliva. It targets materials including circulating tumor DNA (ctDNA), circulating tumor cells, exosomes, and microRNAs (1). Liquid biopsy using ctDNA has been developed for detecting minimal residual disease (MRD) after cancer treatment and for early cancer detection. Studies on ctDNA in head and neck squamous cell carcinoma (HNSCC) have primarily focused on virus-associated HNSCC, namely Epstein-Barr virus (EBV)-associated nasopharyngeal carcinoma (NPC) (2) and human papillomavirus (HPV)-associated oropharyngeal squamous cell carcinoma (HPV-OPSCC) (3). However, studies on ctDNA in patients with non-viral HNSCC have been scarce. Notably, patients with non-viral HNSCC may benefit more from MRD detection via ctDNA than their viral counterparts due to poorer prognoses and a higher likelihood of requiring adjuvant therapy. Hanna et al. conducted the largest analysis to date on the clinical utility of ctDNA assays in HNSCC, specifically examining the association between post-treatment ctDNA positivity and survival (4). They found that ctDNA was detectable in 86% of patients, post-treatment ctDNA positivity indicated persistent or recurrent disease, and its detection was associated with poor survival outcomes. The authors should be commended for providing real-world data on ctDNA assays in a cohort that includes non-viral HNSCC. However, this study has several limitations. First, the viral status of tumors was incompletely assessed. Although the study aimed to evaluate ctDNA detection in non-viral HNSCC, HPV status was investigated in only 56% of patients, and patients with NPC were included. To ensure a homogeneous study population, the authors should have excluded HPV-OPSCC and NPC or provided subgroup analyses limited to non-viral HNSCC. Second, the success rate of variant detection was presented according to treatment modality. From a clinical applicability standpoint, however, it would be more informative to stratify detection rates based on the type of tissue specimen used for sequencing. In practice, tumor tissue availability ranges from fine-needle aspiration samples to biopsy fragments or surgically resected specimens. If variant detection is feasible with minimally invasive samples, ctDNA testing could become more accessible, particularly for inoperable or recurrent cases. Conversely, if high-quality surgical specimens are required, ctDNA testing may be limited to surgical candidates. Therefore, evaluating ctDNA detection across specimen types is essential for its clinical translation. Third, the inclusion of early-stage disease limits the study’s generalizability. Stage I and II patients made up 24% of the cohort. While undetectable post-treatment ctDNA correlated with no recurrence in these patients, early-stage tumors often have favorable prognoses regardless of ctDNA status. Thus, in early-stage cases, ctDNA may function more as a confirmatory rather than a predictive tool. In contrast, in advanced-stage disease, ctDNA detection could significantly inform surveillance and treatment planning. Therefore, conclusions should be stratified by disease stage to avoid overgeneralization. Fourth, the short follow-up period (median 5.1 months) restricts the ability to assess long-term prognostic value. While short-term data provides early insight into the utility of ctDNA as a warning signal, longer follow-up is necessary to validate its predictive value. For instance, the LIONESS study with extended follow-up showed ctDNA detection in 96% of recurrence cases prior to clinical progression (5). Similarly, Flach et al. reported ctDNA positivity preceding radiologic recurrence in all five relapsed patients in a smaller cohort (6). These studies underscore ctDNA’s potential for early detection, particularly when integrated into longitudinal monitoring.

Prior to the above-discussed study, several similar studies had been published. However, these studies involved very small cohort sizes. For instance, Flach et al. investigated postsurgical ctDNA in 17 patients with p16-negative HNSCC (6). Among these patients, five experienced recurrences, and in all five cases, ctDNA was detected before clinical signs or imaging-revealed recurrence. In a subgroup analysis of the IMSTAR-HN trial, 19 patients with HNSCC, of whom 79% had non-oropharyngeal primary, were subjected to ctDNA testing during treatment (7). ctDNA testing has been shown to be a potential early indicator of recurrence. Janke et al. longitudinally quantified ctDNA in 16 patients with head and neck cancer treated with re-radiotherapy (8). They indicated that ctDNA dynamics during irradiation reflected response to radiation therapy. Collectively, all previous studies analyzed a small number of patients; therefore, the conclusions drawn from these studies could not be conclusive. In last year’s American Society of Clinical Oncology meeting, Flach et al. presented the results of the LIONESS study, where the ctDNA of 76 patients with surgically treated HNSCC was analyzed (5). They showed that in 96% of cases with clinical recurrence, ctDNA was detected prior to clinical progression. However, this large-scale prospective study on ctDNA in non-viral HNSCC patients has not yet been published, so the results are eagerly awaited.

The most common method for ctDNA quantification is measuring variant allele frequency (1), which requires sufficiently high-quality tumor tissue. In the Hanna et al. study, 16 out of 100 patients could not undergo ctDNA analysis due to insufficient tissue (4). This challenge is particularly pronounced in non-viral HNSCC, where tumor-specific somatic mutations must be identified for assay design. Alternative techniques, such as copy number variation profiling, bypass the need for tumor tissue but currently offer lower sensitivity (8). Non-invasive sample types like saliva or oral rinse also show promise for ctDNA detection (9,10), especially given the anatomical proximity of primary tumors. These approaches could broaden access to ctDNA testing, particularly for patients with inoperable or tissue-limited disease. Further research should evaluate the efficacy of these non-invasive or tumor-uninformed methods to expand real-world applicability. By contrast, ctDNA detection in virus-associated HNSCC is well established due to virus-specific biomarkers. EBV DNA has been used since the 1990s in NPC (11), and subsequent studies have firmly established its value in screening and prognostication (12). Similarly, HPV-OPSCC can be monitored via HPV E6 and E7 oncogenes, offering high sensitivity and specificity through digital droplet polymerase chain reaction (PCR) or next-generation sequencing (3). These virus-derived DNA fragments serve as consistent and shared biomarkers that facilitate reliable ctDNA measurement. In contrast, non-viral HNSCC lacks such universally present targets, making ctDNA detection more technically demanding, less sensitive, and harder to generalize. The absence of virus-specific biomarkers thus underscores a fundamental technological gap that must be addressed to improve ctDNA-based monitoring in non-viral HNSCC.

In conclusion, liquid biopsy for non-viral HNSCC requires more robust clinical evidence and continued technical refinement before it can be adopted into standard practice. Future research should prioritize several key areas: (I) the development of standardized protocols for ctDNA detection using a range of specimen types, including minimally invasive sources such as saliva or oral rinse; (II) the advancement of tumor-uninformed ctDNA assays to overcome limitations posed by insufficient tumor tissue; and (III) multicenter prospective trials to establish the prognostic value of ctDNA and to evaluate treatment strategies guided by ctDNA dynamics. Crucially, these studies must differentiate between early- and advanced-stage disease to clarify the contexts in which ctDNA monitoring provides the greatest clinical benefit. In parallel, the development of expert consensus and clinical guidelines is urgently needed to inform treatment decisions based on ctDNA results—particularly in scenarios where molecular recurrence is detected in the absence of radiologic or clinical confirmation. Currently, there are no established guidelines on how to interpret or act upon ctDNA positivity in HNSCC. While ctDNA may signal minimal residual or recurrent disease, its implications for management decisions—such as initiating or intensifying treatment—remain uncertain. For instance, it is unclear whether ctDNA-detected recurrence without radiographic evidence should prompt immediate systemic therapy or continued surveillance. This uncertainty represents a significant barrier to clinical adoption. To address these challenges, large-scale, multicenter studies are essential to generate high-level evidence and to develop standardized clinical pathways. Only through such coordinated efforts can ctDNA-based surveillance evolve from a promising research tool into a reliable instrument for personalized oncology. Moreover, ctDNA detection does not provide information about tumor localization. If the site of recurrence were known, local therapies such as surgery or re-irradiation could be considered. Without this information, treatment options are generally limited to systemic therapies such as immune checkpoint inhibitors, targeted agents, or cytotoxic drugs. Another unresolved issue is how to manage cases where ctDNA clearance suggests a poor response to ongoing systemic therapy. Should additional agents be introduced or treatment regimens changed? These questions further underscore the need for clinical trials to determine actionable strategies based on ctDNA dynamics. Many challenges remain, but with focused investigation, ctDNA has the potential to significantly enhance precision care in non-viral HNSCC.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.