Predicting therapeutic responses in head and neck squamous cell carcinoma from TP53 mutation detected by cell-free DNA

Introduction

Head and neck squamous cell carcinoma (HNSCC) is one of the most common cancers in humans. More than 600,000 new cases of HNSCC worldwide are recorded each year, and of these cases, approximately 355,000 result in death (1). Smoking, human papillomavirus (HPV) infection, and alcohol consumption are the most significant risk factors for HNSCC (2). Tumor-suppressor TP53 mutation is the most frequently mutated gene in HNSCC, and it is present in approximately 70% of cases, although the frequency varies in different head and neck regions (3). Different TP53 mutations may have varying gain-of-function characteristics. The loss of p53 function would weaken the activation of cell cycle checkpoints and cell apoptosis, thereby resulting in the acquisition of additional mutations and gradual accumulation of a significant tumor mutation burden (4). TP53 mutations are closely associated with adverse outcomes and treatment options in HNSCC.

Cell-free DNA (cfDNA), as a biomarker of blood, has attracted considerable attention for its potential as a minimally invasive tool for cancer monitoring (5,6). Apoptosis, necrosis, and the release of viable cells with newly synthesized DNA are major sources of cfDNA (7,8). Quantification of nontumor-specific and tumor-derived cfDNA and the presence of genetic and epigenetic variants in cfDNA are potential biomarkers of cancer.

In this study, we investigated the association of TP53 mutation with the prognosis of patients with HNSCC from The Cancer Genome Atlas (TCGA) dataset. Furthermore, we investigated whether the TP53 mutation is associated with immunotherapy and chemotherapeutic responses. Ultimately, we found that the presence of TP53 mutations could be detected by cfDNA. It may be possible to predict the response to immunotherapy and chemotherapy for patients with HSNCC by detecting TP53 mutations in cfDNA. We present this article in accordance with the REMARK reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-878/rc).

Methods

Data acquisition and analysis

We downloaded information on HNSCC gene expression profiles, clinical profile, and TP53 mutations from TCGA database (http://xena.ucsc.edu/). Level 4 transcriptomic and reverse-phase protein array (RPPA) data of patients with cancer were obtained from TCGA. Correlation analysis was performed using the expression of genes extracted from the corresponding packages in R software (The R Foundation of Statistical Computing). The masked somatic mutation data were summarized and visualized with R software “Maftools”.

Estimating immune cell infiltration

To verify differences in immunity between the TP53 mutation and wild-type groups, the “estimate” package for R was used to calculate the stromal score, immune score, and ESTIMATE score. Meanwhile, the single-sample gene set enrichment analysis (ssGSEA) algorithm was employed to evaluate the immune cells in the HNSCC samples from gene expression data.

Immunotherapeutic and chemotherapeutic response prediction

The tumor immune dysfunction and exclusion (TIDE) can be used to evaluate the possibility of tumor immune escape in the gene expression profile of tumor samples (9). In this study, TIDE was used to calculate immune measures of HNSCC (http://tide.dfci.harvard.edu). The “pRRophetic” R package was used to predict the chemotherapy response as quantified by the median maximum inhibitory concentration (IC₅₀) for each patient with HNSCC (10,11).

Gene Ontology (GO) analysis, Kyoto Encyclopedia of Genes and Genomes (KEGG) database analysis, and GSEA

Differentially expressed genes (DEGs) between patients with and without TP53 mutations in the HNSCC cohort were obtained using the “limma” package in R. According to the following significance standard, 403 DEGs with | log fold change | >1 and false-discovery rate (FDR) <0.05 were selected. The “pheatmap” R package was used to generate a heatmap plot to visualize the DEGs. GO functional annotation and KEGG pathway enrichment analyses were conducted using R software packages “clusterProfiler”, “org.Hs.eg.db”, “enrichplot”, and “ggplot2”. GSEA was performed to identify the signaling pathways wherein DEGs were enriched between the 2 subgroups (12).

Detection of mutations in plasma DNA

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the independent ethics committee of Tianjin First Central Hospital, China (protocol no. 2020N115KY). Informed consent was obtained from all individual participants. Patients histopathologically diagnosed with HNSCC were recruited into the study. The clinical investigations, including of tumor size, lymph node involvement and any evidence of metastasis, and clinical tumor node metastasis (TNM) staging, were performed in accordance with the American Joint Committee on Cancer Manual of Head and Neck Cancer Staging, eighth edition (13). Patients who received any treatment or had a history of other malignancies were excluded. Before participating in the study, patients who enrolled in Tianjin First Central Hospital agreed on the use of their tissue, blood samples, and data. Collected tissue and plasma samples were stored in a −80 ℃ freezer for later examination.

Steps for mutation detection in cfDNA: (I) DNA sample testing: cfDNA was extracted from the patient’s plasma as a tumor sample, and the peripheral blood mononuclear cells (PBMCs) of the patient were selected as the normal control. A Qubit 2.0 fluorometer (Thermo Fisher Scientific, USA) was used to measure the sample concentration, and Agilent 2100 bioanalyzer (Agilent Technologies, Inc., USA) was used to detect the sample fragment size. cfDNA samples with DNA concentrations ≥20 ng/µL, total amounts of more than 20–30 ng, and no genomic pollution were used to build the database. (II) Library construction: As most cfDNA samples were small fragments of approximately 170 bp and there was a small number of fragments of 170-bp integer multiples, the biggest difference between cfDNA samples and nucleic acid samples was that the cfDNA samples did not need to be randomly broken into small fragments by a Covaris crusher. After end repair, phosphorylation, and A-tail addition, the ends of the fragments were connected to connectors to prepare a DNA library. (III) Library detection: After the construction of the library, the Qubit 2.0 fluorometer was first used for preliminary quantification, and then the Agilent 2100 was bioanalyzer was used to detect the insert size of the library. After the insert size was found to meet expectations 170 bp, quantitative polymerase chain reaction (qPCR) was applied to accurately quantify the effective concentration (3 nmol/L) of the library to ensure library quality. Given that cfDNA was a special sample, its fragment size was 170 bp, and its integer was multiple, the target fragment with an insert size of approximately 170 bp was selected for subsequent sequencing through purification. (IV) Computer sequencing: after the library inspection was qualified, sequencing with a NovaSeq system (Illumina, USA) with pair end 150 bp (PE150) was conducted according to the effective concentration of the library and the data output demand. In the constructed small fragment library, insert DNA, that was, insert fragments, was considered to be the unit of high-throughput and direct sequencing. Double-ended sequencing was used a method for sequencing the two ends of each inserted segment. The length distribution of inserted fragments was known so that the sequences at both ends of the fragments could be determined, and the length between the two fragments was also known during the double-ended sequencing process for subsequent comparison and analysis.

Study design

A retrospective analysis method was used to analyze the TP53 mutation status in patients with HSNCC, as well as the impact of TP53 mutations on patient survival, immunity, and chemotherapy response. In addition, gene mutations that could be detected in patients’ cfDNA and tissues were compared. We explored the feasibility of using cfDNA in blood as a noninvasive detection method to detect TP53 mutations in patients.

Statistical analysis

The statistical analyses in this study were performed with R software version 4.1.3. One-way analysis of variance (ANOVA) was used to evaluate the relationship of TP53 mutation and the clinical characteristics of HNSCC. In all analyses, statistical significance (α value) was set to 0.05, and all P values were two-sided.

Results

Somatic genomic mutations in HNSCC

We identified 30 frequently mutated genes using the HNSCC cohort in TCGA, and the 10 most frequently mutated genes were TP53 (69%), TTN (37%), FAT1 (21%), CDKN2A (20%), CSMD3 (18%), MUC16 (18%), NOTCH1 (18%), PIK3CA (17%), SYNE1 (16%), and LRP1B (15%) (Figure 1A). TP53 was the most frequently mutated gene in HNSCC. We explored concordance and exclusivity relationships between the mutant genes shown in Figure 1B. Figure 1C is a lollipop plot drawn by Maftools, which shows the mutation distribution of TP53 in HNSCC. The results indicated that the majority of TP53 mutations were missense mutations. The effect of TP53 mutation on overall survival (OS) prognosis was assessed according to the Kaplan-Meier survival curve, and TP53 mutation was found to be a poor prognostic factor for OS. Our result was consistent with previous reports (14,15) (Figure 1D). We then further analyzed TP53 mutations in subsequent analyses.

Figure 1 Somatic genome mutations in HNSCC, as well as TP53 mutations and their impact on survival. (A) Waterfall plot of commonly mutated genes in HNSCC in TCGA cohort; (B) the coincident and exclusive associations across mutated genes; (C) Lollipop map showing the location and type of TP53 mutations in HNSCC; (D) effect of Kaplan-Meier survival curve of TP53 mutation on survival. HNSCC, head and neck squamous cell carcinoma; TCGA, The Cancer Genome Atlas; TMB, tumor mutational burden; WT, wild type.

Correlation between the TP53 mutation and clinicopathologic characteristics

This study included 507 patients, 357 of whom had TP53 mutations, with the other 150 having the wild type. As shown in Table 1, the clinical characteristics of gender, clinical stage, and pathologic M (pM) showed no significant differences between the TP53-mutation and wild-type groups. Meanwhile, the clinical characteristics of age (χ²=4.279; P=0.041), pathologic N stage (χ²=4.550; P=0.037), pathologic T stage (χ²=12.448; P<0.001), and tumor grade (χ²=10.780; P=0.002) showed significant differences between the two groups.

Immune cell infiltration landscape of TP53 mutation in HNSCC

We performed an ESTIMATE analysis of the immune properties of TP53 mutations based on the expression of immune cell types. The results revealed significant differences in immune (Figure 2A) and ESTIMATE scores (Figure 2B) but not in stromal scores (Figure 2C) between the two groups. The ssGSEA method was applied to the transcriptome of HNSCC samples to assess the distribution of immune cell types. Immune infiltration in patients in the TP53-mutant and wild-type groups was then elucidated. As shown in Figure 2D, more than half of the immune cell types were downregulated in the TP53-mutant group compared with the wild-type group. We observed significantly reduced levels of activated B cells, activated CD4 T cells, activated CD8 T cells, and eosinophils in the TP53-mutant group. CD56 bright natural killer (NK) cells were upregulated in the TP53-mutant group.

Figure 2 Immune cell infiltration in HNSCC. (A-C) Analysis of the immune score, ESTIMATE score, and stromal score between the TP53-mutant group and the wild-type group based on TCGA database; (D) differences in the distribution of immune cells between the TP53-mutant group and wild-type group based on TCGA database. *, P<0.05; **, P<0.01; ***, P<0.005; ****, P<0.001; ns, P>0.05. HNSCC, head and neck squamous cell carcinoma; TCGA, The Cancer Genome Atlas.

The TP53 mutation was associated with the expression of immune checkpoints

The expression of programmed cell death-ligand 1 (PD-L1) is widely recognized as a reliable biomarker for guiding the administration of immune checkpoint inhibitors (ICIs). To explore the role of TP53 mutation in determining the response to ICIs, we performed several analyses to determine the association between TP53 mutation and PD-L1 expression. PD-L1 (CD274) messenger RNA (mRNA) expression in the TP53-mutant group was significantly lower than that in the wild-type TP53 group (P<0.01) (Figure 3A). This result was further confirmed at the PD-L1 protein level by RPPA data, and TP53 mutation was associated with a lower PD-L1 protein level than was wild-type TP53 (P<0.05) (Figure 3B). We also compared programmed cell death protein 1 (PD-1) and p53 expression levels in each TP53 group. Compared with the wild-type group, the expression of p53 protein in the TP53-mutant group increased (Figure 3C), but no significant difference in PD-1 levels between the groups was observed (Figure 3D).

Figure 3 The relationship between TP53 mutation and immune checkpoint expression level in HNSCC. (A,B) Correlation analysis of TP53 status with PD-L1 mRNA and PD-L1 RPPA levels based on TCGA database; (C,D) correlation of TP53 status with PD-1 and p53 RPPA levels; (E-G) correlation of TP53 status with miR-34 family levels. HNSCC, head and neck squamous cell carcinoma; PD-L1, programmed cell death-ligand 1; RPPA, reverse-phase protein array; WT, wild-type.

Previous studies have shown that TP53 mutation prevents HNSCC from downregulating PD-L1 expression via miR-34 (16,17). Therefore, the expression of miR-34 family members in the TP53-mutant group and the wild-type group was compared. The results showed that miR-34a was significantly decreased in the TP53-mutant group, and no significant difference was found for miR-34b or miR-34c (Figure 3E-3G), suggesting that the TP53 mutation may not reduce PD-L1 levels through miR-34 regulation in HNSCC.

We also explored the expression of other immune checkpoints in TP53 mutation, including cytotoxic t-lymphocyte associated protein 4 (CTLA-4), lymphocyte activation gene 3 (LAG3), galectin 9 (LGALS9), hepatitis a virus cellular receptor 2 (HAVCR2), programmed cell death 1 ligand 2 (PDCD1LG2), and T cell immunoreceptor with Ig and ITIM domains (TIGIT). We found that the expression levels of CTLA-4, LAG3, LGALS9, HAVCR2, and TIGIT were significantly reduced in TP53-mutant patients with HSNCC (Figure 4). These results suggested that immunotherapy that targets immune checkpoints might be less effective for patients with the TP53 mutation.

Figure 4 Differences in immune checkpoint–associated genes between TP53 mutation and wild-type patients in HNSCC. **, P<0.01; ***, P<0.005; ****, P<0.001; ns, P>0.05. HNSCC, head and neck squamous cell carcinoma; WT, wild-type.

Immunotherapeutic and chemotherapeutic responses of patients with TP53-mutant or TP53 wild-type HNSCC

To investigate the relationship between TP53 mutation and ICI response in HNSCC, the TIDE algorithm was used to estimate the potential clinical efficacy of immune checkpoint blockade (ICB) therapy. The TP53-mutant group had a higher TIDE score (P<0.01; Figure 5A,5B), composed of higher TIDE exclusion scores (P<0.01; Figure 5C) and lower TIDE dysfunction signatures (P<0.01; Figure 5D). Our findings suggested that anti-PD-1 therapy was less effective in most TP53-mutant patients with HNSCC.

Figure 5 Immunotherapy and chemotherapy response in patients with HNSCC with mutant and wild-type TP53 based on TCGA database. (A-D) Violin plot of different prediction scores of TIDE, together with the dysfunction score and exclusion score in the two groups. (E) Drugs with significantly different chemotherapy response (P<0.001) in the HNSCC test. WT, wild-type; TIDE, tumor immune dysfunction and exclusion; HNSCC, head and neck squamous cell carcinoma; TCGA, The Cancer Genome Atlas.

We tested other chemotherapeutic and targeted drugs, and the results showed a difference between the predicted value of IC₅₀ for patients with TP53-mutant HSNCC and those with TP53 wild-type HSNCC. TP53-mutant patients with HNSCC exhibited a higher sensitivity to camptothecin methotrexate, etoposide, bleomycin, lenalidomide, and rapamycin (P<0.001). Patients with HNSCC and TP53 mutations were less sensitive to methotrexate, etoposide, bleomycin, lenalidomide, and rapamycin (Figure 5E).

Enrichment pathway analysis of TP53 mutation

To further clarify the difference between the TP53-mutant and wild-type group, the analysis of 501 HNSCC samples identified 403 DEGs, including 253 upregulated and 150 downregulated genes. We generated a hierarchical clustering heatmap to present the top 40 DEGs of each group (Figure 6A). We also performed GO and KEGG enrichment analyses to identify the most common biological processes and pathways involved in these DEGs. GO enrichment analysis showed that these DEGs were mainly enriched in striated muscle tissue development, structural constituent of muscle, and endopeptidase activity (Figure 6B), while the results of KEGG enrichment analysis indicated that these DEGs were primarily enriched in neuroactive ligand-receptor interaction, phosphatidylinositide 3-kinases/protein kinase B (PI3K–AKT) signaling pathway, MAPK signaling pathway, and endocytosis (Figure 6C). Furthermore, GSEA was performed based on the TP53 mutation, and it showed that the TP53 mutation could upregulate the signaling pathways involved in epithelial-mesenchymal transition, Kirsten rat sarcoma viral oncogene homolog (KRAS) upregulation, and the glycolysis pathway and downregulate signaling pathways involved in interleukin 6–Janus kinase–signal transducer and activator of transcription 3 (IL6–JAK–STAT3) signaling, interferon α response, and E2F targets (Figure 7).

Figure 6 GO and KEGG enrichment analysis of TP53 mutation-related DEGs in HNSCC based on TCGA database. (A) Heatmap showing the top 40 DEGs. (B) GO enrichment analysis of the enriched BPs, CC, and MF. (C) KEGG enrichment analysis. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; DEGs, differentially expressed genes; HNSCC, head and neck squamous cell carcinoma; TCGA, The Cancer Genome Atlas; BPs, biological processes; CC, cell composition; MF, molecular function; WT, wild-type.

Figure 7 GSEA of TP53 mutation-related DEGs based on TCGA database in HNSCC. GSEA, gene set enrichment analysis; DEGs, differentially expressed genes; TCGA, The Cancer Genome Atlas; HNSCC, head and neck squamous cell carcinoma.

Concordance between TP53 mutation identified in plasma cfDNA and matched HNSCC tissue DNA

Figure 8 illustrates the specific mutation spectrum and clinical characteristics of each of the nine sample pairs with mutations. The results showed that the mutation frequency of TP53 was the highest among all mutated genes. In cfDNA samples, the mutation frequency of TP53 was 66.67%. In matched tumor biopsy, TP53 mutation was detected in all patients. Only 27.27% of the tissue tumor variants were detected outside of plasma when all TP53 mutations were considered. Table 2 depicts the detailed mutations of TP53 in paired samples of cfDNA and tumor tissue.

Figure 8 Somatic mutational landscape of nine cfDNA samples in tDNA and cfDNA. cfDNA, cell-free DNA; tDNA, tumor DNA.

Discussion

Certain studies indicate that TP53 is a tumor suppressor that can inhibit the occurrence and development of tumors by regulating tumor proliferation, apoptosis, angiogenesis, and DNA repair (18,19). TP53 mutations have been observed in various cancers and are correlated with reduced OS (20). In this study, TP53 was the most frequently mutated gene in HNSCC. TP53 mutation was also associated with HNSCC poor prognosis. Numerous studies have suggested that TP53 mutations play a crucial role in tumor recognition and antitumor immune surveillance via the immune system (21,22). In our study, the TP53-mutant group showed lower immune and ESTIMATE scores, thereby suggesting poor prognosis, which was consistent with the poorer survival in the TP53-mutant group. Immune cell infiltration is an important feature of the tumor microenvironment. The TP53-mutant group exhibited higher CD56 bright NK cell infiltration than did the TP53 wild-type group of patients with HSNCC. The levels of activated B cells, activated CD4 T cells, and eosinophils were markedly decreased in the TP53-mutant group. These results indicated that the TP53-mutant group was prone to immune escape.

PD-1 is an immune checkpoint that usually inhibits the antitumor immune responses of tumor cells by combining with PD-L1 (23). In our study, the level of PD-L1 in the TP53-mutant group was significantly lower than that in the TP53 wild-type group, and the result was consistent across multiple databases (RNA-sequencing and RPPA). Previous studies have reported that p53 can enhance PD-L1 expression by regulating the loss of function of miR-34 (24,25). However, we did not find a positive correlation between miR-34 and TP53 status in our study. We evaluated the potential clinical efficacy of immunotherapy in TP53 mutations using TIDE. The higher the TIDE prediction score is, the higher the likelihood of immune evasion. The results showed a higher TIDE score in the TP53-mutant group, indicating that ICB has poor efficacy in patients with TP53 mutations. We can easily predict the efficacy of ICB therapy in patients using TIDE, but the specific treatment effect and identification of biomarkers still need to be clarified by extensive clinical trials for verification.

Additionally, the sensitivity of the TP53-mutant group to chemotherapeutic and targeted drugs was explored. The data suggested that the use of camptothecin, methotrexate, etoposide, blemycin, lenalidomide, and rapamycin for those with TP53 mutations may be effective.

According to the DEGs between the TP53 mutation and wild-type groups, GSEA and GO and KEGG enrichment analysis were performed, and we found that the TP53 mutation was associated with the glycolysis pathway in patients with HNSCC. This metabolic phenotype is characterized by a preferential reliance on glycolysis (the process by which glucose is converted to pyruvate and then lactate) to produce energy anaerobically (26). Glycolysis plays an important role in oncogenic regulation. Therefore, overcoming drug resistance by inhibiting the glycolysis in cancer cells is a novel strategy (27).

In this study, TP53 mutations were detected in tissue and serum cfDNA derived from patients with HNSCC. As the cfDNA of tumor cells is released in the blood, the mutational status of cfDNA may reflect the genetic characteristics of the primary or metastatic lesions (28). In our study, only 27.27% of the tissue tumor variants were detected outside of cfDNA in a consideration of all TP53 mutations. These results suggest that the TP53 mutation in patients with HNSCC could be measured in serum DNA.

Conclusions

Our study further confirmed that TP53 mutation in patients with HSNCC was associated with poor prognosis. We also explored the response of patients with TP53-mutant patients with HNSCC to immunotherapies and chemotherapies. Furthermore, the detection of TP53 mutations in serum-derived cfDNA from patients with HSNCC was demonstrated to be feasible. In the future, we may predict the prognosis and treatment of patients with HNSCC by detecting TP53 mutations in cfDNA.

Acknowledgments

Funding: This study was supported by the Tianjin Key Medical Discipline Construction Project, the Tianjin Health Research Project (ZC20010); Tianjin Science and Technology Project (21JCQNJC01770, 21JCZDJC01200, 21JCQNJC01670, 21JCQNJC01780, and 21JCYBJC01570), and the Tianjin Health Research Project (TJSJMYXYC-D2-021).

Footnote

Reporting Checklist: The authors have completed the REMARK reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-878/rc

Data Sharing Statement: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-878/dss

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-878/prf

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-23-878/coif). The authors have no conflicts of interest to declare.

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. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and was approved by the independent ethics committee of Tianjin First Central Hospital (protocol no. 2020N115KY). Informed consent was obtained from all individual participants.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Johnson DE, Burtness B, Leemans CR, et al. Head and neck squamous cell carcinoma. Nat Rev Dis Primers 2020;6:92. [Crossref] [PubMed]
2. Sabatini ME, Chiocca S. Human papillomavirus as a driver of head and neck cancers. Br J Cancer 2020;122:306-14. [Crossref] [PubMed]
3. Mulder FJ, Pierssens DDCG, Baijens LWJ, et al. Evidence for different molecular parameters in head and neck squamous cell carcinoma of nonsmokers and nondrinkers: Systematic review and meta-analysis on HPV, p16, and TP53. Head Neck 2021;43:303-22. [Crossref] [PubMed]
4. Deneka AY, Baca Y, Serebriiskii IG, et al. Association of TP53 and CDKN2A Mutation Profile with Tumor Mutation Burden in Head and Neck Cancer. Clin Cancer Res 2022;28:1925-37. [Crossref] [PubMed]
5. Vitale SR, Sieuwerts AM, Beije N, et al. An Optimized Workflow to Evaluate Estrogen Receptor Gene Mutations in Small Amounts of Cell-Free DNA. J Mol Diagn 2019;21:123-37. [Crossref] [PubMed]
6. Nikanjam M, Kato S, Kurzrock R. Liquid biopsy: current technology and clinical applications. J Hematol Oncol 2022;15:131. [Crossref] [PubMed]
7. Lin D, Shen L, Luo M, et al. Circulating tumor cells: biology and clinical significance. Signal Transduct Target Ther 2021;6:404. [Crossref] [PubMed]
8. Lin D, Shen L, Luo M, et al. Circulating tumor cells: biology and clinical significance. Signal Transduct Target Ther 2021;6:404. [Crossref] [PubMed]
9. Jiang P, Gu S, Pan D, et al. Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nat Med 2018;24:1550-8. [Crossref] [PubMed]
10. Geeleher P, Cox N, Huang RS. pRRophetic: an R package for prediction of clinical chemotherapeutic response from tumor gene expression levels. PLoS One 2014;9:e107468. [Crossref] [PubMed]
11. Vanden Heuvel JP, Maddox E, Maalouf SW, et al. Replication Study: Systematic identification of genomic markers of drug sensitivity in cancer cells. Elife 2018;7:e29747. [Crossref] [PubMed]
12. de Jong A, Kuipers OP, Kok J. FUNAGE-Pro: comprehensive web server for gene set enrichment analysis of prokaryotes. Nucleic Acids Res 2022;50:W330-6. [Crossref] [PubMed]
13. Lydiatt WM, Patel SG, O'Sullivan B, et al. Head and Neck cancers-major changes in the American Joint Committee on cancer eighth edition cancer staging manual. CA Cancer J Clin 2017;67:122-37.
14. Shi C, Liu S, Tian X, et al. A TP53 mutation model for the prediction of prognosis and therapeutic responses in head and neck squamous cell carcinoma. BMC Cancer 2021;21:1035. [Crossref] [PubMed]
15. Chen Y, Li ZY, Zhou GQ, et al. An Immune-Related Gene Prognostic Index for Head and Neck Squamous Cell Carcinoma. Clin Cancer Res 2021;27:330-41. [Crossref] [PubMed]
16. Cortez MA, Ivan C, Valdecanas D, et al. PDL1 Regulation by p53 via miR-34. J Natl Cancer Inst 2015;108:djv303. [Crossref] [PubMed]
17. Song D, Lyu H, Feng Q, et al. Subtyping of head and neck squamous cell cancers based on immune signatures. Int Immunopharmacol 2021;99:108007. [Crossref] [PubMed]
18. Wang Z, Strasser A, Kelly GL. Should mutant TP53 be targeted for cancer therapy? Cell Death Differ 2022;29:911-20. [Crossref] [PubMed]
19. Tang Z, Zeng M, Wang X, et al. Synthetic lethality between TP53 and ENDOD1. Nat Commun 2022;13:2861. [Crossref] [PubMed]
20. Donehower LA, Soussi T, Korkut A, et al. Integrated Analysis of TP53 Gene and Pathway Alterations in The Cancer Genome Atlas. Cell Rep 2019;28:1370-1384.e5. [Crossref] [PubMed]
21. Nathan CA, Khandelwal AR, Wolf GT, et al. TP53 mutations in head and neck cancer. Mol Carcinog 2022;61:385-91. [Crossref] [PubMed]
22. Ganci F, Allegretti M, Manciocco V, et al. Two distinct TP53 mutations in HNSCC primary tumor: Only one circulates in the blood. Oral Oncol 2021;115:105096. [Crossref] [PubMed]
23. Yi M, Zheng X, Niu M, et al. Combination strategies with PD-1/PD-L1 blockade: current advances and future directions. Mol Cancer 2022;21:28. [Crossref] [PubMed]
24. Pan W, Chai B, Li L, et al. p53/MicroRNA-34 axis in cancer and beyond. Heliyon 2023;9:e15155. [Crossref] [PubMed]
25. Liu C, Rokavec M, Huang Z, et al. Curcumin activates a ROS/KEAP1/NRF2/miR-34a/b/c cascade to suppress colorectal cancer metastasis. Cell Death Differ 2023;30:1771-85. [Crossref] [PubMed]
26. Reinfeld BI, Rathmell WK, Kim TK, et al. The therapeutic implications of immunosuppressive tumor aerobic glycolysis. Cell Mol Immunol 2022;19:46-58. [Crossref] [PubMed]
27. Paul S, Ghosh S, Kumar S. Tumor glycolysis, an essential sweet tooth of tumor cells. Semin Cancer Biol 2022;86:1216-30. [Crossref] [PubMed]
28. van Dessel LF, Vitale SR, Helmijr JCA, et al. High-throughput isolation of circulating tumor DNA: a comparison of automated platforms. Mol Oncol 2019;13:392-402. [Crossref] [PubMed]