Longitudinal Circulating Tumor DNA Profiling as a Biomarker for Response and Resistance in Platinum-Refractory Head and Neck Squamous Cell Carcinoma

Wonyoung Choi; Jong-Ho Lee; Junsun Ryu; Sung Weon Choi; Yuh-Seog Jung; Joo-Yong Park; Chang Hwan Ryu; Sung Yong Choi; Weon Seo Park; Sun-Young Kong; Tak Yun

doi:10.4143/crt.2025.1089

Articles

Page Path: HOME > J Korean Cancer Assoc > Ahead-of print articles > Article

Original Article Longitudinal Circulating Tumor DNA Profiling as a Biomarker for Response and Resistance in Platinum-Refractory Head and Neck Squamous Cell Carcinoma: Wonyoung Choi^1,2, Jong-Ho Lee^1,3, Junsun Ryu^1,4, Sung Weon Choi^1,3, Yuh-Seog Jung^1,4, Joo-Yong Park^1,3, Chang Hwan Ryu^1,4, Sung Yong Choi^1,4, Weon Seo Park⁵, Sun-Young Kong⁶, Tak Yun^1,2; DOI: https://doi.org/10.4143/crt.2025.1089
Published online: November 4, 2025

¹Center for Rare Cancers, National Cancer Center, Goyang, Korea

²Division of Hematology-Oncology, Department of Internal Medicine, Goyang, Korea

³Oral Oncology Clinic, Goyang, Korea

⁴Department of Otorhinolaryngology-Head and Neck Surgery, National Cancer Center, Goyang, Korea

⁵Department of Pathology, National Cancer Center, Goyang, Korea

⁶Department of Laboratory Medicine, National Cancer Center, Goyang, Korea

Correspondence: Wonyoung Choi, Center for Rare Cancers, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang 10408, Korea
Tel: 82-31-920-0896 E-mail: wonyoungchoi@ncc.re.kr

Co-correspondence: Tak Yun, Center for Rare Cancers, National Cancer Center, 323 Ilsan-ro, Ilsandong-gu, Goyang 10408, Korea
Tel: 82-31-920-1621 E-mail: hmotakyun@ncc.re.kr

• Received: October 5, 2025 • Accepted: November 3, 2025

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

1,446 Views
95 Download

Full Article

Download PDF

Abstract

Purpose
Immune checkpoint inhibitors (ICIs) are the standard treatment for platinum-refractory head and neck squamous cell carcinoma (HNSCC). This study aimed to characterize genomic alterations, monitor dynamic changes in circulating tumor DNA (ctDNA) associated with treatment response, and identify novel alterations linked to resistance through serial ctDNA profiling.
Materials and Methods
Patients with platinum-refractory HNSCC receiving nivolumab were enrolled. ctDNA was analyzed using FoundationOne Liquid CDx at baseline, 6 weeks after treatment, and at disease progression.
Results
A total of 36 patients were enrolled, and 33 were evaluable for ctDNA. The most frequent baseline alterations were TP53 (75.8%), followed by TERT (27.3%), NOTCH1 (24.2%), CDKN2A (12.1%), and CCND1 (12.1%). Of 27 patients with measurable lesions, five achieved a partial response with a response rate of 18.5%. Dynamic ctDNA profiling revealed all responders had marked reductions in both the sum of variant allele frequencies (sumVAF) and the maximum variant allele frequencies (maxVAF) across detected mutations. A decrease in sumVAF or maxVAF correlated with longer treatment durations, and patients with a > 50% decrease in either sumVAF or maxVAF showed significantly longer progression-free survival. Additionally, serial profiling identified de novo mutations in 17 patients, detected during stable disease or progression.
Conclusion
Serial ctDNA analysis provides a noninvasive tool for monitoring treatment response and detecting emerging resistance in HNSCC treated with ICIs. A significant reduction in ctDNA at 6 weeks was associated with improved clinical outcomes and warrants validation in larger, prospective studies to define its role as a predictive biomarker in HNSCC.
Key words: Head and neck neoplasms, Immune checkpoint inhibitors, Circulating tumor DNA, Liquid biopsy, Next-generation sequencing

Introduction

Platinum-based chemotherapy remains the cornerstone of systemic treatment for advanced head and neck squamous cell carcinoma (HNSCC). In locally advanced disease, platinum agents are widely used both as radiosensitizers and cytotoxic agents during concurrent chemoradiation therapy, in adjuvant or definitive treatment settings [1-3]. In the metastatic setting, platinum compounds are frequently combined with fluoropyrimidines, with or without anti–programmed death-1 (PD-1) antibodies, as first-line therapy [4]. Despite their widespread use, resistance to platinum-based regimens is common and associated with poor clinical outcomes [5].

Patients who experience recurrence within 6 months of completing platinum-based chemotherapy or who exhibit disease progression (PD) while receiving such therapy are generally classified as platinum-refractory [6]. For these patients, nivolumab, an anti–PD-1 antibody, is the standard of care, provided they have not previously received PD-1/programmed death-ligand 1 (PD-L1)–targeted therapies [7]. The pivotal CheckMate-141 trial established nivolumab as an effective second-line or later therapy for platinum-refractory HNSCC, demonstrating a significant improvement in overall survival (OS) compared with the investigator’s choice of chemotherapy, although progression-free survival (PFS) remained similar between groups. In the CheckMate-141 trial, tumor PD-L1 expression, defined by a tumor proportion score (TPS) ≥ 1%, was associated with longer OS. However, PD-L1 status alone does not adequately predict therapeutic benefit, as it fails to reliably stratify patients by PFS [7]. Moreover, the molecular mechanisms underlying response and resistance to anti–PD-1 therapy in this disease remain poorly understood.

Circulating cell-free tumor DNA (ctDNA) analysis offers a minimally invasive approach for molecular profiling and disease monitoring in patients with cancer [8]. Liquid biopsy enables testing in cases where tumor tissue is difficult to obtain, captures intratumoral and intertumoral heterogeneity, and facilitates serial sampling to track dynamic molecular changes during treatment. Given these advantages, we conducted a prospective cohort study of patients with platinum-refractory HNSCC undergoing nivolumab therapy. Through ctDNA sequencing, we aimed to characterize the molecular profiles of platinum-refractory HNSCC, monitor dynamic molecular changes associated with treatment response, and identify novel alterations linked to therapeutic resistance.

Materials and Methods

1. Patients

Patients with locally advanced, unresectable, or metastatic HNSCC treated with nivolumab were recruited between September 2022 and April 2025 at the National Cancer Center (Goyang, Korea). All patients exhibited resistance to platinum-based therapy, defined as PD or recurrence within 6 months following the last dose of platinum-containing chemotherapy, whether administered as part of adjuvant treatment or for primary or recurrent disease management. Nivolumab was administered at a dose of 3 mg/kg every 2 weeks. Tumor PD-L1 immunohistochemical staining was assessed with a rabbit antihuman PD-L1 antibody (clone 28-8, Dako) and evaluated based on TPS, with a cutoff value ≥ 1% considered positive. Tumor response was evaluated using computed tomography or magnetic resonance imaging according to Response Evaluation Criteria in Solid Tumor (RECIST) ver. 1.1. Baseline imaging was performed within 4 weeks prior to the first dose of nivolumab, and follow-up assessments were scheduled every 6 weeks (after three cycles of nivolumab) until week 12, followed by evaluations 6 to 8 weeks thereafter.

2. Molecular profiling with ctDNA

Molecular profiling was conducted using the FoundationOne LiquidCDx platform (Roche) [9]. In brief, 20mL of whole blood was collected in specialized cell-free DNA (cfDNA) collection tubes for the assay. Samples were stored at room temperature and shipped to Foundation Medicine for analysis according to the manufacturer’s recommended protocol. Peripheral blood samples were collected at baseline (prior to the first dose of nivolumab), at week 6 following the initial response assessment, and at the time of PD in patients who demonstrated a partial response (PR) or stable disease (SD) at the initial restaging. ctDNA levels were quantified based on the variant allele frequency (VAF) of each mutation identified in the test report. Variants potentially associated with clonal hematopoiesis were classified according to the interpretations provided in the analysis report.

3. Statistical analysis

Statistical analyses were performed using R ver. 4.4.1 (R Foundation for Statistical Computing). Descriptive statistics are reported as numbers and percentages. Continuous variables were compared using the t test, and associations between continuous variables were assessed using Pearson’s correlation. Kaplan-Meier survival curves were generated to estimate PFS and OS, and differences between groups were evaluated using the log-rank test. PFS and OS were calculated from the first day of treatment initiation until the date of disease progression, death from any cause, or last follow-up, whichever occurred first. Patients without an event at the time of analysis were censored at the date of last disease assessment. A two-sided p-value < 0.05 was considered statistically significant.

Results

1. Patient characteristics

A total of 36 patients were enrolled in the study, and 33 patients were eligible for baseline ctDNA analysis. One patient was lost to follow-up after receiving two cycles of nivolumab, leaving 32 patients available for serial ctDNA assessment at the first restaging (Fig. 1A).

The baseline characteristics of the patients are summarized in Table 1. Most patients were male (84.8%), and had a history of smoking (78.8%). The median age was 69 years. The most common primary site was the oral cavity (57.6%), followed by the oropharynx (15.2%), hypopharynx (9.1%), larynx (9.1%), and maxilla (9.1%). Nivolumab was administered as second-line therapy in 30 patients (90.9%), third-line in one patient (3.0%), and fourth-line in two patients (6.1%). Most patients had an Eastern Cooperative Oncology Group (ECOG) performance score of 0 or 1; however, six patients (18.2%) had an ECOG score of 2. PD-L1 expression was evaluable in all patients, and 27 (81.8%) had a TPS of 1% or higher. Among patients assessed at baseline, ctDNA was detectable in 29, corresponding to a detection rate of 87.9% in our cohort (Fig. 1B).

2. Mutation profiles of platinum-refractory HNSCC

The landscape of mutation profiles in platinum-refractory HNSCC based on ctDNA analysis is shown in Fig. 1C. The most frequently altered gene was TP53 (75.8%), followed by TERT (27.3%), NOTCH1 (24.2%), CDKN2A (12.1%), and CCND1 (12.1%). Comparison with the HNSCC cohort from The Cancer Genome Atlas (TCGA) revealed that TP53 alteration was also the most commonly observed mutation, with a frequency similar to that of our cohort (S1 Fig.) [10]. Additionally, the mutation frequencies of NOTCH1 and CDKN2A were among the top hits in TCGA. Mutations in the TERT promoter regions were the second most frequent alterations in our ctDNA analysis; however, these mutations were not reported in TCGA. Blood tumor mutation burden (bTMB) levels were mostly low, with all patients having a bTMB ≤ 10 (Fig. 1C).

Clonal hematopoiesis of indeterminate potential (CHIP), part of the aging process involving the accumulation of somatic mutations in hematopoietic stem cells, can be detected in plasma cfDNA [11,12]. CHIP mutations were detected at baseline analysis in 23 patients (69.7%). ATM (33.3%) was the most frequently detected gene associated with CHIP, followed by DNMT3A (24.2%), ASXL1 (18.2%), TET2 (12.1%), and CHEK2 (12.1%) (S2A Fig.). Notably, the number of CHIP mutations was significantly correlated with patient age (S2B Fig.).

3. Clinical efficacy of nivolumab

The efficacy of nivolumab was assessed in 32 patients, with 27 patients having measurable lesions according to the RECIST ver. 1.1. The response rate was 18.5%, and the clinical benefit rate was 37.0%, with five patients achieving PR and five exhibiting SD (Fig. 2A). The median PFS was 1.3 months (95% confidence interval [CI], 1.2 to 3.5), and the median OS was 8.1 months (95% CI, 5.8 to not available [NA]) (Fig. 2B, S3A and B Fig.).

4. Dynamic changes of ctDNA correlate with response and survival

We evaluated the dynamic changes in ctDNA mutations in relation to clinical responses to nivolumab. The VAF of each detected genetic alteration was used as a quantitative index. For comparison, we analyzed either the sum of VAFs across all genetic alterations (sumVAF) or the VAF of the alteration with the highest value among the detected alterations (maxVAF). Variants classified as CHIP-related were excluded from these analyses, as they are unlikely to reflect tumor-derived signals. When stratified by best response, all patients with PR demonstrated a significant reduction in the sumVAFs, whereas those with SD or PD showed smaller reductions or even increases in the sumVAF levels (Fig. 2C). Similarly, when stratified by clinical benefit (defined as PR or SD lasting > 12 weeks), patients with clinical benefit exhibited significantly greater reductions in the sumVAFs than those without benefit (Fig. 2D). In addition, we examined changes in maxVAF levels using the same stratification criteria. Consistently, all patients with PR showed significant decreases in the maxVAFs relative to baseline, whereas those with SD or PD exhibited more variable changes (Fig. 2E). When stratified by clinical benefit, a trend toward lower maxVAF levels was observed in patients with clinical benefit; however, this difference was not statistically significant (Fig. 2F).

Subsequently, we assessed whether changes in VAF levels correlated with treatment duration and survival outcomes. Treatment duration on nivolumab was negatively correlated with changes in both sumVAFs and maxVAF levels measured at 6 weeks post-treatment, indicating that patients with greater reductions in their ctDNA levels tended to experience longer clinical benefit (Fig. 3A and B). Given that all patients with PR exhibited > 50% reductions in sumVAF and maxVAF from baseline, we stratified the cohort into two groups based on their dynamic changes. When stratified by sumVAF, patients with substantial ctDNA reduction had a median PFS of 4.0 months (95% CI, 1.3 to NA) compared to 1.2 months (95% CI, 1.2 to 1.5) in those without significant reduction (p < 0.01, log-rank test) (Fig. 3C). Similarly, when stratified by maxVAF, patients with significant ctDNA decline had a median PFS of 5.6 months (95% CI, 1.3 to NA), whereas those without significant reduction had a median PFS of 1.3 months (95% CI, 1.2 to 1.5) (p=0.020, log-rank test) (Fig. 3D). To account for potential confounding by baseline clinical characteristics, we next performed multivariate Cox proportional hazards analyses including age (≥ 65 vs. < 65 years), ECOG performance status (0-1 vs. ≥ 2), PD-L1 expression (≥ 1% vs. < 1%), and ctDNA reduction (≥ 50% decrease in sumVAF or maxVAF). In this model, ctDNA reduction remained independently associated with improved PFS (sumVAF: hazard ratio [HR], 0.22; 95% CI, 0.07 to 0.74; p=0.015; maxVAF: HR, 0.20; 95% CI, 0.05 to 0.79; p=0.021), while ECOG performance status showed a borderline association with shorter PFS. These findings suggest that early ctDNA dynamics serve as an independent predictor of nivolumab efficacy in platinum-refractory HNSCC (S4 and 5 Tables).

Furthermore, OS was assessed using the same stratification criteria. However, the median OS was comparable between the two groups when analyzed based on sumVAF and max VAF. When stratified by sumVAF, patients with significant reductions in ctDNA had a median OS of 8.4 months (95% CI, 8.2 to NA), compared with 6.1 months (95% CI, 4.7 to NA) in those without reductions (p=0.392, log-rank test) (S3C Fig.). When stratified by maxVAF, patients with reductions had a median OS of 8.4 months (95% CI, 8.2 to NA), whereas those without reductions had a median OS of 7.0 months (95% CI, 4.7 to NA) (p=0.541, log-rank test) (S3D Fig.).

5. De novo mutations detected in serial ctDNA assessments at SD or PD

In our analysis of dynamic changes in ctDNA, we investigated whether de novo mutations could be detected in liquid biopsy among patients who did not respond to nivolumab. Specifically, we examined subsequent ctDNA assays performed during periods of SD or PD to identify mutations that were absent at baseline but emerged during treatment. Among the 32 patients with available follow-up assays, de novo mutations were detected in 17 patients (Fig. 4). The most frequently observed genes were TP53 (23.5%) and TERT (23.5%), which were among the most commonly mutated genes at baseline. In addition, several less frequent mutations not detected at baseline were identified, including alterations in CUL3, SETD2, EMSY, RB1, APC, IRF2, SMARCA4, NOTCH2, FUBP1, FBXW7, CTCF, NFE2L2, and TBX3 (Fig. 4).

Two representative cases illustrate the detection of de novo mutations during periods of SD or PD. In the first case (NCC020), the baseline ctDNA analysis revealed mutations in TP53, NOTCH1, and PALB2. The patient initially achieved a PR, and follow-up ctDNA analysis at 6 weeks post-nivolumab showed a marked decrease in the VAFs of these baseline mutations. However, at the time of disease progression, ctDNA profiling demonstrated a re-elevation of VAFs for the baseline mutations, along with the emergence of de novo alterations, including an IRF2 intron rearrangement, and amplifications in FGF9, FGF4, FGF3, and CCND1 (Fig. 5A). In the second case (NCC021), baseline ctDNA analysis detected no genetic alterations. However, at the 6-week follow-up, when the patient exhibited SD, de novo mutations were identified in TP53 and the TERT promoter region. At disease progression, a further increase was observed in the VAF of the TERT promoter mutation, accompanied by additional de novo mutations in TP53 and SMAD4 (Fig. 5B).

Discussion

In this study, we demonstrated that ctDNA-based genomic profiling is feasible in patients with metastatic HNSCC and that dynamic changes in VAF are associated with clinical response and survival in patients treated with nivolumab. Importantly, our findings suggest that using a targeted ctDNA panel encompassing genes frequently mutated in HNSCC is not only practical for detecting genomic alterations but also enables serial monitoring of tumor dynamics during treatment. We utilized the FoundationOne LiquidCDx (Roche) assay, the U.S. Food and Drug Administration–approved companion diagnostic capable of analyzing more than 300 cancer-related genes [9]. This assay also reports bTMB and microsatellite instability status, both of which are recognized as potential predictive biomarkers for response to anti–PD-1 therapy [13,14]. Notably, no patient in our cohort exhibited high bTMB, and all were microsatellite stable based on ctDNA analysis.

Our study findings demonstrate that patients with a > 50% reduction in either sumVAF or maxVAF showed a significantly longer PFS compared with those without such reductions. The median PFS in patients with significant reductions was 4.0 months for sumVAF and 5.6 months for maxVAF, both of which were not only significantly longer than those observed in patients without significant reductions but also numerically longer than the median PFS of nivolumab-treated patients in the CheckMate-141 trial (2.0 months). However, these improvements in PFS did not translate into a significantly longer OS in our cohort. One potential explanation is that our study was not sufficiently powered to detect OS differences, as a larger sample size may be required. Additionally, a high proportion of patients in our cohort (81.8%) exhibited positive PD-L1 expression (28-8 assay, TPS ≥ 1%). In the CheckMate-141, tumor PD-L1 positivity was associated with longer OS (8.7 months vs. 4.6 months for PD-L1 positive vs. negative patients), although PFS remained comparable between nivolumab and cytotoxic [7]. These findings suggest that nivolumab may influence the response to post-progression therapies in patients with positive PD-L1 expression [15]. Given that most patients without a significant reduction in ctDNA in our study were also PD-L1 positive, we speculate that this may have contributed to the similar median OS observed between patient subgroups despite differences in PFS.

Similar associations between ctDNA kinetics and immune-checkpoint inhibitor response have also been reported in other solid tumors such as non–small cell lung cancer, melanoma, and urothelial carcinoma, supporting the broader biological relevance of ctDNA dynamics across cancer types [16-19]. Recent work by Ruiz-Torres et al. [20] also evaluated ctDNA dynamics in a small cohort of patients with metastatic HNSCC (n=16) treated with pembrolizumab with or without chemotherapy. Using individualized tumor-informed ctDNA assays, they demonstrated that undetectable ctDNA levels were associated with improved OS and PFS. Compared with that study, our analysis differs in two main aspects: the use of a tumor-agnostic, clinically approved assay (FoundationOne Liquid CDx) and a limited number of serial samples per patient. Tumor-informed assays typically achieve higher analytical sensitivity and specificity by tracking patient-specific variants identified from matched tumor tissue while excluding irrelevant loci [21]. In contrast, tumor-agnostic assays provide broader genomic coverage and are more readily applicable in real-world, as they do not require prior tissue sequencing or customized panel design. However, they are more susceptible to both false-positive signals from clonal hematopoiesis and false-negative results due to lower sensitivity. Thus, the two approaches should be viewed as complementary: tumor-informed assays offer precision at the cost of scalability, whereas tumor-agnostic assays such as ours enable feasible, standardized longitudinal monitoring across heterogeneous patient populations. Our findings extend the existing evidence by demonstrating that even with a tumor-agnostic platform, ctDNA dynamics can reflect treatment response in platinum-refractory HNSCC receiving nivolumab.

The genomic landscape of our platinum-refractory HNSCC cohort largely overlapped with previously reported profiles from TCGA, with TP53 representing the most frequently mutated gene, followed by alterations in CDKN2A and NOTCH family genes. However, one notable difference was the high prevalence of TERT promoter mutations, which emerged as the second most common mutation in our ctDNA analysis but were not reported in TCGA dataset. This discrepancy is likely attributable to differences in sequencing methodologies. TCGA utilized whole-exome sequencing (WES) to profile somatic mutations; however, the TERT promoter region is known to have poor coverage in WES-based approaches, potentially leading to an underestimation of its mutation frequency [22]. In contrast, the targeted ctDNA panel used in our study includes the TERT promoter region, allowing for more accurate detection of these alterations.

CHIP is known to be detectable in plasma cfDNA, and the frequency of CHIP-associated mutations has been shown to increase with age [23,24]. Because these mutations are also captured during the molecular profiling of liquid biopsy, careful interpretation of ctDNA results is essential to avoid misclassification and ensure accurate guidance for clinical decision-making [12]. Mutations in DNMT3A, ASXL1, TET2, and CHEK2 have been consistently reported as the most common CHIP-related alterations in previous literature. Although ATM is not typically among the most common genes classified as CHIP-associated, it represented the most frequently detected CHIP-designated alterations in our cohort. Whether this observation reflects a disease-specific phenomenon in HNSCC or represents a broader pattern across multiple tumor types warrants further investigation.

The de novo mutations detected in our cohort during serial assays may provide important insights into the molecular mechanisms of resistance to nivolumab. Most of these alterations involved genes that were also frequently mutated at baseline, such as TP53 and TERT promoter mutations. However, several mutations absent in the initial molecular profiling emerged during treatment, potentially indicating the outgrowth of resistant tumor clones or the activation of alternative resistant pathways [8]. While tissue re-biopsy at the time of progression remains the gold standard for identifying resistance mechanisms, it is often limited by tumor accessibility, procedural risk, and the availability of sufficient material for molecular profiling. In contrast, ctDNA analysis provides a minimally invasive, repeatable approach that can dynamically reflect genomic evolution across multiple tumor sites. Although functional validation is required, our findings suggest that longitudinal ctDNA monitoring may help elucidate mechanisms of acquired resistance and identify potential therapeutic targets. Nonetheless, the detection of de novo variants should be interpreted with caution, as some may arise from assay noise, CHIP-related alterations, or other non-tumor-derived sources. Therefore, integration of data with orthogonal validation methods will be essential to accurately determine the biological relevance of these mutations.

Furthermore, the use of nivolumab in relatively rare subtypes of HNSCC is noteworthy. In the CheckMate-141 trial, patients with primary tumors from the oral cavity, pharynx, or larynx were eligible for enrollment, whereas other primary sites (e.g., maxilla or external auditory canal) were excluded. However, in real practice, treatment strategies for these rare subtypes of squamous cell cancers are extrapolated from landmark clinical trials, and nivolumab is used as a standard-of-care, as in typical HNSCC [25,26]. In our cohort, three patients had maxillary cancer, all with a positive PD-L1 expression, and two achieved clinical benefit (one patient with PR, and one with SD ≥ 12 weeks). Although drawing a definite conclusion is difficult owing to the small sample size, these data suggest that nivolumab may be an effective option for maxillary cancers that are refractory to platinum-based therapy.

In summary, this study demonstrates that ctDNA-based genomic profiling is a feasible and informative approach for patients with platinum-refractory metastatic HNSCC treated with nivolumab. Additionally, our findings revealed a high prevalence of TERT promoter mutations and identified de novo alterations emerging during therapy, providing novel insights into potential mechanisms of acquired resistance. Our study is limited by its modest sample size and single-institution design, which may restrict the statistical power to detect survival differences. Because the majority of patients in our cohort were PD-L1 positive and only a small number were PD-L1 negative, we were unable to perform robust subgroup analyses to determine whether the prognostic impact of ctDNA reduction differed by PD-L1 status. Future studies with larger, more balanced cohorts will be necessary to clarify potential interactions between PD-L1 expression and ctDNA dynamics. Accordingly, the findings should be interpreted as exploratory and hypothesis generating, requiring further validation in larger, prospective cohorts. Nonetheless, these results support the utility of ctDNA analysis for molecular profiling and guiding precision strategies in metastatic HNSCC.

Electronic Supplementary Material

Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).

crt-2025-1089_S1_Fig.pdf

crt-2025-1089_S2_Fig.pdf

crt-2025-1089_S3_Fig.pdf

crt-2025-1089_S4_Table.pdf

crt-2025-1089_S5_Table.pdf

NOTES

Ethical Statement

The study protocol was reviewed and approved by the Institutional Review Board of the National Cancer Center (IRB number: NCC2022-0181) and registered at the Clinical Research Information Service (registration number: KCT0007598). This study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice. All patients provided written informed consent.

Author Contributions

Conceived and designed the analysis: Choi W, Kong SY, Yun T.

Collected the data: Choi W, Lee JH, Ryu J, Choi SW, Jung YS, Park JY, Ryu CH, Choi SY.

Contributed data or analysis tools: Park WS.

Performed the analysis: Choi W, Park WS.

Wrote the paper: Choi W, Yun T.

Review & Editing the paper: Kong SY, Yun T.

Supervision: Yun T.

Conflicts of Interest

Wonyoung Choi received consulting or advisory fees from Daiichi Sankyo Inc, Eisai Inc, and Boryung Pharmaceutical Co Ltd.; and received honoraria from Dong-A ST, and Bayer AG. Other authors declare no competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Funding

This work was supported by a National Cancer Center Research Grant (No. NCC-2310310-1 to W.C.). The funding source had no role in the study design, data collection, analysis, interpretation, or manuscript preparation.

Acknowledgments

We thank Hana Choi, Hyunjung Park, and Lina Kwon for their support in study coordination.

Fig. 1.

Genomic profiles of circulating tumor DNA (ctDNA) in platinum-refractory head and neck squamous cell carcinoma. (A) Flowchart of the study. (B) Pie chart of ctDNA detection at baseline analysis. (C) Oncoplot for genomic alterations assessed at baseline analysis. bTMB, blood tumor mutational burden; NA, not available; PD, disease progression; PD-L1, programmed death-ligand 1; PR, partial response; SD, stable disease; TPS, tumor proportion score.

Fig. 2.

Dynamic changes in circulating tumor DNA correlate with response to nivolumab. (A) Waterfall plot for nivolumab in patients with measurable lesions (n=27). (B) Swimmer plot for nivolumab in patients with available follow-up data (n=32). (C) Changes in the sum of variant allele frequencies (sumVAF) at 6 weeks post-treatment stratified by treatment responses. (D) Changes in the sumVAF at 6 weeks post-treatment stratified by clinical benefit. (E) Changes in the maximum variant allele frequencies (maxVAF) at 6 weeks post-treatment stratified by treatment responses. (F) Changes in the maxVAF at 6 weeks post-treatment stratified by clinical benefit. PD, disease progression; PD-L1, programmed death-ligand 1; PR, partial response; SD, stable disease; TPS, tumor proportion score.

Fig. 3.

Significant reduction in circulating tumor DNA levels is associated with improved survival outcomes. (A) Scatter plot of the change in sum of variant allele frequencies (sumVAF) at 6 weeks and duration of nivolumab treatment. (B) Scatter plot of the change in maximum variant allele frequencies (maxVAF) at 6 weeks and treatment duration of nivolumab. (C) Kaplan-Meier plot for progression-free survival (PFS) stratified by sumVAF reduction (< –50% vs. ≥ –50%). (D) Kaplan-Meier plot for PFS stratified by maxVAF reduction (< –50% vs. ≥ –50%). CI, confidence interval; mPFS, median PFS; NA, not available; PD, disease progression; PR, partial response; SD, stable disease.

Fig. 4.

De novo mutations detected in patients without responses via serial analyses. Oncoplot for de novo mutations detected at follow-up circulating tumor DNA analyses during periods of stable disease (SD) or progressive disease (PD). NA, not available; PD-L1, programmed death-ligand 1; PR, partial response; TPS, tumor proportion score.

Fig. 5.

Representative cases of circulating tumor DNA (ctDNA) dynamics during the treatment courses. (A) Serial ctDNA analysis of NCC020. ctDNA levels decreased when the patient achieved partial response, but increased along with de novo mutations at the time of disease progression. (B) Serial ctDNA analysis of NCC021. De novo mutations were detected when the patient achieved stable disease, and the allele frequency increased along with additional de novo mutations at the time of disease progression.

Table 1.

Patient characteristics

Characteristic	Patients for baseline ctDNA analysis (n=33)
Age (yr), median (range)	69 (32-80)
Sex
Male	28 (84.8)
Female	5 (15.2)
Primary tumor site
Oral cavity	19 (57.6)
Oropharynx	5 (15.2)
Hypopharynx	3 (9.1)
Larynx	3 (9.1)
Maxilla	3 (9.1)
Smoking history
Never	7 (21.2)
Former or current	26 (78.8)
ECOG PS
0 or 1	27 (81.8)
2	6 (18.2)
Line of nivolumab therapy
Second line	30 (90.9)
Third line	1 (3.0)
Fourth line	2 (6.1)
Prior surgery
Yes	19 (57.6)
No	14 (42.4)
Prior radiation therapy
Yes	28 (84.8)
No	5 (15.2)
PD-L1 expression (TPS score)
≥ 1%	27 (81.8)
0%	6 (18.2)

Values are presented as number (%) unless otherwise indicated. ctDNA, circulating tumor DNA; ECOG PS, Eastern Cooperative Oncology Group performance status; PD-L1, programmed death-ligand 1; TPS, tumor proportion score.

REFERENCES

1. Adelstein DJ, Li Y, Adams GL, Wagner H Jr, Kish JA, Ensley JF, et al. An intergroup phase III comparison of standard radiation therapy and two schedules of concurrent chemoradiotherapy in patients with unresectable squamous cell head and neck cancer. J Clin Oncol. 2003;21:92–8. Article PubMed
2. Forastiere AA, Goepfert H, Maor M, Pajak TF, Weber R, Morrison W, et al. Concurrent chemotherapy and radiotherapy for organ preservation in advanced laryngeal cancer. N Engl J Med. 2003;349:2091–8. Article PubMed
3. Cooper JS, Pajak TF, Forastiere AA, Jacobs J, Campbell BH, Saxman SB, et al. Postoperative concurrent radiotherapy and chemotherapy for high-risk squamous-cell carcinoma of the head and neck. N Engl J Med. 2004;350:1937–44. Article PubMed
4. Burtness B, Harrington KJ, Greil R, Soulieres D, Tahara M, de Castro G Jr, et al. Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or metastatic squamous cell carcinoma of the head and neck (KEYNOTE-048): a randomised, open-label, phase 3 study. Lancet. 2019;394:1915–28. PubMed
5. Wakasugi T, Nguyen TN, Takeuchi S, Ohkubo JI, Suzuki H. Pattern of recurrence after platinum-containing definitive therapy and efficacy of salvage treatment for recurrence in patients with squamous cell carcinoma of the head and neck. Front Oncol. 2022;12:876193.Article PubMed PMC
6. Horichi Y, Matsui H, Yamamura Y, Iwae S. Platinum resistance and sensitivity in recurrent/metastatic head and neck squamous cell carcinoma. Auris Nasus Larynx. 2024;51:132–7. Article PubMed
7. Ferris RL, Blumenschein G Jr, Fayette J, Guigay J, Colevas AD, Licitra L, et al. Nivolumab for recurrent squamous-cell carcinoma of the head and neck. N Engl J Med. 2016;375:1856–67. Article PubMed PMC
8. Bartolomucci A, Nobrega M, Ferrier T, Dickinson K, Kaorey N, Nadeau A, et al. Circulating tumor DNA to monitor treatment response in solid tumors and advance precision oncology. NPJ Precis Oncol. 2025;9:84.Article PubMed PMC PDF
9. Woodhouse R, Li M, Hughes J, Delfosse D, Skoletsky J, Ma P, et al. Clinical and analytical validation of FoundationOne Liquid CDx, a novel 324-Gene cfDNA-based comprehensive genomic profiling assay for cancers of solid tumor origin. PLoS One. 2020;15:e0237802Article PubMed PMC
10. Cancer Genome Atlas Network. Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature. 2015;517:576–82. Article PubMed PMC PDF
11. Abbosh C, Swanton C, Birkbak NJ. Clonal haematopoiesis: a source of biological noise in cell-free DNA analyses. Ann Oncol. 2019;30:358–9. Article PubMed PMC
12. Fairchild L, Whalen J, D’Aco K, Wu J, Gustafson CB, Solovieff N, et al. Clonal hematopoiesis detection in patients with cancer using cell-free DNA sequencing. Sci Transl Med. 2023;15:eabm8729Article PubMed
13. Le DT, Durham JN, Smith KN, Wang H, Bartlett BR, Aulakh LK, et al. Mismatch repair deficiency predicts response of solid tumors to PD-1 blockade. Science. 2017;357:409–13. PubMed PMC
14. Wildsmith S, Li W, Wu S, Stewart R, Morsli N, Raja R, et al. Tumor mutational burden as a predictor of survival with durvalumab and/or tremelimumab treatment in recurrent or metastatic head and neck squamous cell carcinoma. Clin Cancer Res. 2023;29:2066–74. Article PubMed PMC PDF
15. Saleh K, Daste A, Martin N, Pons-Tostivint E, Auperin A, Herrera-Gomez RG, et al. Response to salvage chemotherapy after progression on immune checkpoint inhibitors in patients with recurrent and/or metastatic squamous cell carcinoma of the head and neck. Eur J Cancer. 2019;121:123–9. Article PubMed
16. Zou W, Yaung SJ, Fuhlbruck F, Ballinger M, Peters E, Palma JF, et al. ctDNA predicts overall survival in patients with NSCLC treated with PD-L1 blockade or with chemotherapy. JCO Precis Oncol. 2021;5:827–38. Article PubMed
17. Murray JC, Sivapalan L, Hummelink K, Balan A, White JR, Niknafs N, et al. Elucidating the heterogeneity of immunotherapy response and immune-related toxicities by longitudinal ctDNA and immune cell compartment tracking in lung cancer. Clin Cancer Res. 2024;30:389–403. Article PubMed PMC PDF
18. Seremet T, Jansen Y, Planken S, Njimi H, Delaunoy M, El Housni H, et al. Undetectable circulating tumor DNA (ctDNA) levels correlate with favorable outcome in metastatic melanoma patients treated with anti-PD1 therapy. J Transl Med. 2019;17:303.Article PubMed PMC PDF
19. Tolmeijer SH, van Wilpe S, Geerlings MJ, von Rhein D, Smilde TJ, Kloots IS, et al. Early on-treatment circulating tumor DNA measurements and response to immune checkpoint inhibitors in advanced urothelial cancer. Eur Urol Oncol. 2024;7:282–91. Article PubMed
20. Ruiz-Torres DA, Merkin RD, Bryan ME, Mendel J, Efthymiou V, Roberts T, et al. Personalized circulating tumor DNA dynamics inform survival and response to immune checkpoint blockade in recurrent/metastatic head and neck cancer. NPJ Precis Oncol. 2025;9:298.Article PubMed PMC PDF
21. de Abreu AR, Wyninckx A, Vandamme T, Op de Beeck K, Van Camp G, Peeters M, et al. Circulating tumor DNA detection in cancer: a comprehensive overview of current detection methods and prospects. Oncologist. 2025;30:oyaf204.PubMed PMC
22. Zehir A, Benayed R, Shah RH, Syed A, Middha S, Kim HR, et al. Mutational landscape of metastatic cancer revealed from prospective clinical sequencing of 10,000 patients. Nat Med. 2017;23:703–13. PubMed PMC
23. Genovese G, Kahler AK, Handsaker RE, Lindberg J, Rose SA, Bakhoum SF, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med. 2014;371:2477–87. Article PubMed PMC
24. Jaiswal S, Fontanillas P, Flannick J, Manning A, Grauman PV, Mar BG, et al. Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014;371:2488–98. PubMed PMC
25. Sato Y, Fukuda N, Wang X, Urasaki T, Ohmoto A, Nakano K, et al. Efficacy of nivolumab for head and neck cancer patients with primary sites and histological subtypes excluded from the CheckMate-141 trial. Cancer Manag Res. 2020;12:4161–8. PubMed PMC
26. Hanai N, Shimizu Y, Kariya S, Yasumatsu R, Yokota T, Fujii T, et al. Effectiveness and safety of nivolumab in patients with head and neck cancer in Japanese real-world clinical practice: a multicenter retrospective clinical study. Int J Clin Oncol. 2021;26:494–506. Article PubMed PDF

Figure & Data

REFERENCES

Citations

Citations to this article as recorded by

PubReader
ePub Link

Cite

CITE: export Copy Download Format; Close

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

RIS — For EndNote, ProCite, RefWorks, and most other reference management software
BibTeX — For JabRef, BibDesk, and other BibTeX-specific software

Include:

Citation for the content below
Citation and abstract for the content below

Longitudinal Circulating Tumor DNA Profiling as a Biomarker for Response and Resistance in Platinum-Refractory Head and Neck Squamous Cell Carcinoma

DOI: https://doi.org/10.4143/crt.2025.1089

XML Download

Abstract
Introduction
Materials and Methods
Results
Discussion
Electronic Supplementary Material
NOTES
REFERENCES

Longitudinal Circulating Tumor DNA Profiling as a Biomarker for Response and Resistance in Platinum-Refractory Head and Neck Squamous Cell Carcinoma

Fig. 1. Genomic profiles of circulating tumor DNA (ctDNA) in platinum-refractory head and neck squamous cell carcinoma. (A) Flowchart of the study. (B) Pie chart of ctDNA detection at baseline analysis. (C) Oncoplot for genomic alterations assessed at baseline analysis. bTMB, blood tumor mutational burden; NA, not available; PD, disease progression; PD-L1, programmed death-ligand 1; PR, partial response; SD, stable disease; TPS, tumor proportion score.

Fig. 2. Dynamic changes in circulating tumor DNA correlate with response to nivolumab. (A) Waterfall plot for nivolumab in patients with measurable lesions (n=27). (B) Swimmer plot for nivolumab in patients with available follow-up data (n=32). (C) Changes in the sum of variant allele frequencies (sumVAF) at 6 weeks post-treatment stratified by treatment responses. (D) Changes in the sumVAF at 6 weeks post-treatment stratified by clinical benefit. (E) Changes in the maximum variant allele frequencies (maxVAF) at 6 weeks post-treatment stratified by treatment responses. (F) Changes in the maxVAF at 6 weeks post-treatment stratified by clinical benefit. PD, disease progression; PD-L1, programmed death-ligand 1; PR, partial response; SD, stable disease; TPS, tumor proportion score.

Fig. 3. Significant reduction in circulating tumor DNA levels is associated with improved survival outcomes. (A) Scatter plot of the change in sum of variant allele frequencies (sumVAF) at 6 weeks and duration of nivolumab treatment. (B) Scatter plot of the change in maximum variant allele frequencies (maxVAF) at 6 weeks and treatment duration of nivolumab. (C) Kaplan-Meier plot for progression-free survival (PFS) stratified by sumVAF reduction (< –50% vs. ≥ –50%). (D) Kaplan-Meier plot for PFS stratified by maxVAF reduction (< –50% vs. ≥ –50%). CI, confidence interval; mPFS, median PFS; NA, not available; PD, disease progression; PR, partial response; SD, stable disease.

Fig. 4. De novo mutations detected in patients without responses via serial analyses. Oncoplot for de novo mutations detected at follow-up circulating tumor DNA analyses during periods of stable disease (SD) or progressive disease (PD). NA, not available; PD-L1, programmed death-ligand 1; PR, partial response; TPS, tumor proportion score.

Fig. 5. Representative cases of circulating tumor DNA (ctDNA) dynamics during the treatment courses. (A) Serial ctDNA analysis of NCC020. ctDNA levels decreased when the patient achieved partial response, but increased along with de novo mutations at the time of disease progression. (B) Serial ctDNA analysis of NCC021. De novo mutations were detected when the patient achieved stable disease, and the allele frequency increased along with additional de novo mutations at the time of disease progression.

Fig. 1.

Fig. 2.

Fig. 3.

Fig. 4.

Fig. 5.

Longitudinal Circulating Tumor DNA Profiling as a Biomarker for Response and Resistance in Platinum-Refractory Head and Neck Squamous Cell Carcinoma

Characteristic	Patients for baseline ctDNA analysis (n=33)
Age (yr), median (range)	69 (32-80)
Sex
Male	28 (84.8)
Female	5 (15.2)
Primary tumor site
Oral cavity	19 (57.6)
Oropharynx	5 (15.2)
Hypopharynx	3 (9.1)
Larynx	3 (9.1)
Maxilla	3 (9.1)
Smoking history
Never	7 (21.2)
Former or current	26 (78.8)
ECOG PS
0 or 1	27 (81.8)
2	6 (18.2)
Line of nivolumab therapy
Second line	30 (90.9)
Third line	1 (3.0)
Fourth line	2 (6.1)
Prior surgery
Yes	19 (57.6)
No	14 (42.4)
Prior radiation therapy
Yes	28 (84.8)
No	5 (15.2)
PD-L1 expression (TPS score)
≥ 1%	27 (81.8)
0%	6 (18.2)

Table 1. Patient characteristics