Purpose:

The purpose of this study was to better understand the complex molecular biomarkers and signatures of head and neck cancer (HNC) among Black patients and identify possible molecular changes associated with HNC disparities.

Experimental Design:

Molecular subtypes and genomic changes in HNC samples from patients of African and European ancestry in The Cancer Genome Atlas, Memorial Sloan Kettering Cancer Center, Broad Institute, MD Anderson Cancer Center, and John Hopkins University were identified. Molecular features (genomic, proteomic, transcriptomic) associated with race and genomic alterations associated with clinical outcomes were determined. An independent cohort of HNC tumor specimens was used to validate the primary findings using IHC.

Results:

Black patients were found to have a younger age at diagnosis, more aggressive tumor types, higher rates of metastasis, and worse survival compared with White patients. Black patients had fewer human papillomavirus–positive tumor types and higher frequencies of laryngeal subtype tumors. Higher frequencies of TP53, MYO18B, KMT2D, and UNC13C mutations and a lower frequency of PIK3CA mutations were observed in Black patients. Tumors of Black patients showed significant enrichment of c-MYC and RET-tyrosine signaling and amplifications. A significant increase in tumor expression of c-MYC in Black patients was observed and was associated with poor survival outcomes in the independent cohort.

Conclusions:

Novel genomic modifications and molecular signatures may be related to environmental, social, and behavioral factors associated with racial disparities in HNC. Unique tumor mutations and biological pathways have potential clinical utility in providing more targeted and individualized screening, diagnostic, and treatment modalities to improve health outcomes.

Translational Relevance

Evaluation of the molecular characteristics of patients with head and neck cancer (HNC) based on genomic ancestral assignment was performed. Black patients showed a higher prevalence of the laryngeal tumor site, higher rates of metastasis, and more advanced tumors compared with White patients. Higher TP53, MYO18B, KMT2D, and UNC13C and lower PIK3CA mutation frequencies were identified in Black patients. Significant enrichment of c-MYC and RET-tyrosine signaling and amplifications were found in tumors of Black patients. An independent cohort of patients with HNC showed significant increase of c-MYC expression in tumors of Black patients, which was associated with poor survival. These results suggest genomic differences may be related to existing racial disparities in HNC. This study advances our knowledge about the diverse and dynamic nature of HNC molecular signatures that may be associated with differential exposures and subsequent disparities on a group level. These findings have screening, diagnostic, and therapeutic potential.

Head and neck cancer (HNC; Kown, #2208) refers to several cancers affecting the oral cavity, nose, throat, and sinuses. HNC is the sixth most common malignancy worldwide and accounts for approximately 300,000 deaths per year (1). Metastasis and recurrence result in a 5-year survival rate of approximately 50% after surgery, which has not improved over the last decade (2). Epidemiologic data have identified disparities between racial groups in HNC screening, detection, treatment, and survival (3).

Patients of African ancestry (Black patients) are diagnosed with HNC at a younger age (4), in contrast to patients of European ancestry (White patients). This finding may be a composite outcome resulting from differences in group-level environmental exposure to carcinogens and socioeconomic factors. As evidenced in other diseases, these varying group-level exposures may activate carcinogenesis pathways that may be superimposed on group differences in gene polymorphisms (5, 6). However, the molecular modifications and genomic signatures in HNC that may be associated with, or a consequence of these exposures, have yet to be established. Although environmental and social factors may be drivers of the existing racial disparities in HNC, the molecular and biological changes at play alongside these factors warrant investigation to establish biomarkers, identify risk factors, and provide more effective therapeutics to reduce health care disparity (2).

Tumorigenesis and metastasis involve acquired alterations at the genomic, transcriptomic, and methylomic levels. These changes may include point mutations, copy-number alterations (CNA), differential gene expression, and methylation. For many cancer types, racial differences in pathways involving hormones, growth factors, and inflammatory signaling, as well as SNPs and CNAs resulting in differential gene expression, have been reported previously (7–9). For example, EGFR mutations are significantly more common in patients with lung cancer of Latino and Asian descent, directly impacting the clinical and therapeutic regimens for these cancers (10, 11). Triple-negative breast cancer (ER, PR, HER2), a well-established, more aggressive form of breast cancer, shows an increased incidence in those of African descent compared with other races, which correlates with a lower survival rate among African American patients (9). These genomic and molecular differences identified as more prevalent and unique to certain racial groups have led to more targeted and individualized diagnostic, monitoring, and treatment approaches. To date, comparative genomic analyses involving different racial groups are lacking for HNC.

As a result, there remain significant questions as to the association of race with risk factors and survival outcomes of HNC. A better understanding of HNC's complex genomic and molecular uniqueness among different racial groups is required to address the survival disparities observed between the groups. Identifying genetic and molecular modalities between diverse populations will not only improve targeted tumor therapies but may also lead to improved screening, risk assessment, and early cancer detection. This study aimed to identify the molecular uniqueness of head and neck cancer in patients of African ancestry to aid in better understanding the complete setting, both social and biological, contributing to the existing disparities.

Data source

The patient cohorts analyzed in this study are listed in Supplementary Table S1. For the Tumor Cancer Genome Atlas (TCGA) analysis, preprocessed datasets from the Broad Institute TCGA Firehose were used (https://gdac.broadinstitute.org/). Overall survival time was used as the clinical endpoint. Overall survival was chosen as the main endpoint because it reflects an objective and unambiguous event; it is considered the gold standard for oncology clinical trials and is widely available across different studies (12). Clinical data, copy-number variation, mutation, and DNA methylation were downloaded as preprocessed files from cBioportal. The following covariates were obtained for all tumors: biological sex, genomic race, smoking status, anatomic tumor location, human papillomavirus (HPV) status, tumor stage, surgical margin disease, lymph node status, and rate of metastasis.

For copy-number variation analysis, we used HG19-segmented somatic copy-number alterations (SCNA) corrected for germline SCNAs. The Genomic Identification of Significant Targets in Cancer (GISTIC 2.0) algorithm (13) identified recurrent chromosome arm amplifications and deletions within HNC tumors. The GISTIC module also calculates the q-values for aberrant regions, accounting for the FDR using the Benjamini–Hochberg method.

We generated survival curves of HNC cases in the combined cohort according to race, and Kaplan–Meier curves were plotted. The P values for the Kaplan–Meier curves were calculated on the basis of the log-rank test. Univariate and multivariable Cox proportional hazards regression were used to assess the association of race with overall survival, controlling for different covariates using the R survival package (https://cran.r-project.org/web/packages/survival/index.html) to compute Z scores and P values. For correlation analysis, Pearson correlation coefficient was calculated with a Bonferroni-corrected P value ≤0.05, which was considered statistically significant.

Determination of genetic ancestry

To avoid self-reported biases, the genetic ancestry assignment for each patient was determined using a computationally derived analysis previously described by Zhang and colleagues (14). The ancestry for each patient was assigned as the consensus between five independent methods based upon SNP array genotyping calls (Broad Institute, Washington University, and University of California San Francisco methods) and exome sequencing data (University of Trento and ExAC/Broad methods). Each of the five methods evaluated associations between ancestry and molecular data. Regression analyses were performed to categorize the patients by genetic ancestry independently. Consensus calls between the five methods were made based on the ancestral population that received the majority of assignments for each patient. We adopted the 2021 JAMA guidelines on reporting race and ethnicity; patients of African ancestry are termed Black patients, and those of European ancestry are termed White patients in our current article (15).

Validation cohort

Clinical samples were obtained from the tumor bank at the Columbia University Irving Medical Center. All samples were collected in accordance with Institutional Review Board and recognized ethical guidelines. Histologically confirmed HPV-negative primary head and neck squamous cell carcinoma samples limited to the oral cavity from a Columbia cohort were included in this study. We confirmed 80% tumor contact in each sample. Patients with clinical information and a minimum survival time of 12 months were included in this study. Samples of untreated patients were collected at the time of surgery. Subjects who demonstrated occult lymph node metastasis following initial surgery were excluded from the cohort. Thirty-six Black patients (self-reported) who met inclusion and exclusion criteria were identified, and 36 White patients (self-reported) were matched on the basis of gender, age at diagnosis (5-year interval), smoking history, and tumor–node–metastasis (TNM) stage. The clinical and demographic information was obtained from the electronic record. The cohort included 48 males and 24 females. The mean (SD) of age was 66 (5.8) years for the cohort. A total of 53% of patients had TNM stage I and II HNC, and 47% had TNM stage III and IV HNC. The progression-free survival for the cohort was defined as the time from treatment to locoregional relapse, distant recurrence, or death, whichever came first. For each participant, archived formalin-fixed paraffin-embedded tissue blocks were retrieved. The initial HNC surgical tissue sample was utilized for the analysis in case the subject had recurrent and/or second primary HNC.

TCGA methylation analysis

Expression profiles of gene-specific DNA methylation data of patients with HNC were downloaded from TCGA database (https://tcga-data.nci.nih.gov/tcga/). The Illumina Human Methylation 450 BeadChip (450 K array) was used to measure the DNA methylation data. In total, 482,421 CpG sites were assessed throughout the genome. Level 3 methylation data from TCGA data portal were downloaded (16). Gene methylation was compared between Black and White patient samples and was corrected for multiple comparisons with an FDR <0.05 by the Benjamini–Hochberg method.

DESeq2

Gene expression correlations with ancestry (Black patients vs. White patients) were analyzed using the DESeq2 algorithm (DESeq, RRID:SCR_000154). The results were ranked according to the generated scores. This adjusted P value was calculated from the Benjamini–Hochberg adjustment of the P value, with an FDR of 0 < 0.05. The “stat” category was used to sort by t test, determining the significance of the upregulation or downregulation of expression.

IHC and automated quantification

Paraffin-embedded human tissue microarrays were stained for c-MYC protein using the Leica Bond automated platform via a high pH pretreatment solution, followed by heating at 95°C for 40 minutes and detection using the Leica Refine DAB system. The c-MYC antibody (Abcam, catalog no. ab 32072, RRID:AB_731658) was purchased from Abcam (ab32072) and diluted 1:100. Following immunostaining, slides were scanned using a Leica Aperio AT2 scanner (Vista). The digitalized slides were then imported into QuPath open-source software, where positive cell detection was performed on five regions of interest (ROI) with an average area of 454,614 μm2. Within each ROI, a classification tool was used to discern tumor cells from stromal cells, followed by positive cell detection with a 1+ intensity threshold of 0.06. The results of the five ROI were then averaged for both cytoplasmic and nuclear positivity and compared between the samples.

Gene set enrichment analysis

Genes differentially expressed in Black patients versus White patients were ranked based on their fold changes. Using the preranked gene set enrichment analysis (GSEA) algorithm, the significantly enriched pathways in the gene list were identified. The recommended numerical values determined the cut-off values for the pathway rankings for significance (family-wise error rate < 0.2) and sorted them by the enrichment score. GSEA includes different categories of pathways, including C1—positional gene sets, C5BP—Gene Ontology biological processes, and H-hallmark annotated gene sets.

Protein interaction networks and molecular complex detection

For proteins that were overexpressed in the Black patient cohort, physical interactions in STRING (score > 0.132) and BioGrid were used to form networks of protein physical interactions (17). The molecular complex detection (MCODE) algorithm (ref. 18; RRID:SCR_003032) was applied to identify densely connected network components. Pathway and process enrichment analyses were applied to each MCODE component independently, and the three best-scoring terms by P value were retained as the functional description of the corresponding components. MCODE networks were visualized using Cytoscape (v.3.8.2; RRID:SCR_003032).

Statistical analysis

The figure legends or text report the number of subjects and technical and biological replicates for all human subject data and experiments. Parametric Student t test or Mann–Whitney U test was performed to compare the two groups based on the underlying distribution of data. Parametric one-way ANOVA or Kruskal–Wallis test was used to compare more than one group. Survival curves were calculated using the Kaplan–Meier method and compared using the log-rank test. Disease-free survival (DFS) was defined as the time from surgery to recurrence, second primary cancer, or death. Data are presented as mean ± SEM or mean ±  SD, as indicated. Two-sided P values less than 0.05 were considered statistically significant. In GSEA, P values were calculated using a permutation test. Data were analyzed using R (version 4.0.3) and GraphPad Prism version 8 (RRID:SCR_002798). In all cases, ****, P < 0.0001; ***, P < 0.001; **, P < 0.01; *, P < 0.05.

Data availability

The data analyzed in this study were obtained from cBioPortal datasets at https://www.cbioportal.org/datasets using squamous cell carcinoma of the head and neck cases, including the following datasets:

  • (i) TCGA dataset

  • (ii) Memorial Sloan Kettering Cancer Center (MSKCC) dataset

  • (iii) Broad Institute dataset

  • (iv) MD Anderson Cancer Center dataset

  • (v) John Hopkins University dataset

All relevant data are presented in the article or included as Supplementary Materials and Methods. Raw data generated or analyzed in this study are available upon reasonable request from the corresponding author.

Clinical and demographic

We analyzed data from 1,099 patients with HNC, including 95 Black and 1,004 White from seven cohort studies led by five institutions (TCGA, MSKCC, Broad Institute, MD Anderson Cancer Center, and John Hopkins University; Supplementary Tables S1 and S2). The ancestry assignment of race was based on genomic analysis. The mean age at diagnosis for Black and White patients were mean ± SD of 58 ± 8 and 61 ± 12 years, respectively (P = 0.0293). Compared with White patients, Black patients had worse overall survival [median of 25.94 months for Black patients and 60.43 months for White patients (P = 4.229e-4)] (Fig. 1A). Cox proportional hazards models indicated that lower survival of Black patients is independent of HPV status (P = 0.0412).

Figure 1.

Survival, clinical, tumor, and genomic data in patients with HNC differ by ancestral race. A, Overall survival of patients with HNC by ancestral race. Clinical data differ by ancestral race for primary tumor site (B), biological sex (C), HPV status (D), and smoking history (E). Tumor data by ancestral race differ in disease stage (F), metastasis stage (G; M0: no metastasis, M1: distant metastasis, MX: metastasis cannot be assessed), and disease of surgical margin (H). Genomic differences based on ancestral race in 6q status and 3p status (I), and tumor mutation burden (J). n = 95 for Black patients and n = 1,004 for White patients. *, P < 0.05; **, P < 0.01; ***, P < 0.001 after FDR adjustment.

Figure 1.

Survival, clinical, tumor, and genomic data in patients with HNC differ by ancestral race. A, Overall survival of patients with HNC by ancestral race. Clinical data differ by ancestral race for primary tumor site (B), biological sex (C), HPV status (D), and smoking history (E). Tumor data by ancestral race differ in disease stage (F), metastasis stage (G; M0: no metastasis, M1: distant metastasis, MX: metastasis cannot be assessed), and disease of surgical margin (H). Genomic differences based on ancestral race in 6q status and 3p status (I), and tumor mutation burden (J). n = 95 for Black patients and n = 1,004 for White patients. *, P < 0.05; **, P < 0.01; ***, P < 0.001 after FDR adjustment.

Close modal

The primary tumor site differed between the two groups, with Black patients showing a higher percentage of tumors in the larynx than White patients (P = 5.93E-04; Fig. 1B). Moreover, Black patients showed a lower percentage of tongue tumors (P = 5.931e-4; Fig. 1B). Male patients comprised a more significant portion of the Black patient pool (P = 4.21e-3; Fig. 1C). White patients were more likely to have HPV-positive tumors than Black patients (P = 2.19e-3; Fig. 1D). There was a higher prevalence of smoking history among Black patients (Fig. 1E). No significant difference in alcohol consumption was observed between the two groups (P = 0.369). Tumor samples from Black patients were at more advanced disease stages (stage III and IV) compared with samples from White patients (P = 1.97e-7; Fig. 1F). Black patients had a higher percentage of positive lymph node stages (P = 2.21e-4; Supplementary Fig. S1A). Tumors from Black patients showed a higher rate of metastasis (P = 8.47e-3; Fig. 1G) and were more likely to show a positive disease of the surgical margin than White patients (P = 8.09e-3; Fig. 1H). We found a difference in chromosomal 6q and 3p status; tumor samples of Black patients showed more losses at both loci than White patients (Fig. 1I). Deletion of chromosome 3p has been linked to both kidney and squamous cancers (19). Overall, tumors from Black patients showed higher mutation burdens than White patients (P = 2.274e-3; Fig. 1J).

Mutation profiles

Next, we examined the mutation profiles of both groups. Among the genes that showed the highest mutation frequencies overall, the genes with significantly higher mutation frequencies in the Black patient pool were TP53, TTN, KMT2D, MUC16, LRP1B, CSMD3, FAT1, CDKN2a, NOTCH, UNC13C, DNAH5, SYNE1, PKHD1L1, and COL11A1 (Fig. 2A). Interestingly, COL11A1, SYNE1, and LRP1B have been implicated in tumor progression and more aggressive tumor types in other squamous cancers (20–22). PIK3CA was the only gene more frequently mutated in the White patient pool (Fig. 2A). Gene mutation frequencies with the most statistically significant differences between Black and White patients were higher in the Black patient group. The top 15 genes were IGKV1D-42, DHX33, OR5L2, SDR9C7, RNF139, PPARD, OR10R2, GOLGA5, FGF21, CCDC142, ZNF451, MED25, FOXO1, YIPF7, and TLR6 (Fig. 2B). Many of these genes are involved in other cancer types; DDHX33 plays a critical role in cell proliferation and growth and is upregulated in colon cancer (23). PPARD expression is increased in many human cancers and may be a key regulator of metastasis (24). Deletion of FOXO1 in mice promotes tumor growth. Evidence that FOXO1 acts as a tumor suppressor are discussed in the context of a potential target for therapy in prostate cancer (25). Recurrent mutations in HNCs from the Black patient cohort in TCGA data are presented in Fig. 2C.

Figure 2.

Gene mutation and methylation differences in HNC tumors by ancestral race. Gene mutation frequencies in tumor samples by ancestral race overall (A) and lowest P value (B). C, Most frequently mutated genes in HNC tumors of Black patients. Pathway analysis of hypermethylated (D) and hypomethylated (E) genes in HNC tumors of Black patients. n = 95 for Black patients and n = 1,004 for White patients. *, P < 0.05; **, P < 0.01; ***, P < 0.001 after FDR adjustment.

Figure 2.

Gene mutation and methylation differences in HNC tumors by ancestral race. Gene mutation frequencies in tumor samples by ancestral race overall (A) and lowest P value (B). C, Most frequently mutated genes in HNC tumors of Black patients. Pathway analysis of hypermethylated (D) and hypomethylated (E) genes in HNC tumors of Black patients. n = 95 for Black patients and n = 1,004 for White patients. *, P < 0.05; **, P < 0.01; ***, P < 0.001 after FDR adjustment.

Close modal

Turning to CNAs, the most frequently amplified genes were SHANK2, ANO1, CTTN, FADD, PPFIA1, TPRG1, CCAT1, CASC8, FGF19, FGF4, and FGF3 (Supplementary Fig. S1B). The most frequently deleted genes were CDKN2A, CDKN2B, CDKN2A-DT, and CSMD1 (Supplementary Fig. S1B). All of the most commonly amplified and deleted genes showed higher frequencies in the Black patient cohort. CNA frequencies with the most significance between the two groups were also observed in the Black patient cohort. These included higher copy-number variations of CFAP300, BIRC2, TMEM133, PGR, ARHGAP42, NSD3, LETM2, LSM1, BACH1, USP25, NCAM2, MMP8, MMP27, PLPP5, and DDHD2 (Supplementary Fig. S1C). BIRC2 amplification has been studied in squamous carcinomas, and increased levels have been shown to impair anticancer immunity and immunotherapy efficacy in melanoma and breast cancer cells (26). In nasopharyngeal cancer, ARHGAP42 promotes cell migration and invasion (27).

Methylation profile

Next, we identified differentially methylated genes by ancestral race in the patients with HNC available in TCGA data (Supplementary Fig. S2; Supplementary Table S3). Our results on methylation activity closely coincided with our previously discussed mutation patterns. Genes found to be significantly hypermethylated in Black patients compared with White patients included MEDAG, RASIP1, RGPD5, HEBP2, SFRP1, C2CD4D, MFN2, TXNDC17, PLEKHN1, and MEI1. Genes found to be significantly hypomethylated in Black patients compared with White patients included SLO1A2, SYT1, CALCR, SERHL2, RGS7, SSPO, C1D, SNAP91, LINC00526, and MDGA1. Next, we performed an overrepresentation pathway analysis based on the Reactome pathway. Hypermethylated genes in Black patients are associated with TP53 regulation, PUMA activation, megakaryocyte development, platelet production, mitophagy, and PI3K-AKT signaling (Fig. 2D). Interestingly, PUMA has been shown to mediate the apoptotic response in many cancers, including colorectal cancer (28). Hypomethylated genes in Black patients were associated with pathways involved in PPARA activation, lipid metabolism, telomere maintenance, glucagon signaling, and insulin resistance (Fig. 2E). Telomere maintenance is a major event in tumor growth and survival. When hypomethylated, its activity may increase and play a role in the aggressiveness of HNC tumors in specific individuals.

Gene expression profile

Comparing the gene expression profiles of HNC in Black patients versus White patients, we found 150 differentially expressed genes with an FDR of less than 0.01 (Fig. 3A; Supplementary Table S4). The top 10 genes showing increased and decreased protein expression in the Black patients are illustrated in Fig. 3B. Next, we broadened our gene expression analysis to include pathway analysis. GSEA of differentially expressed genes showed positive enrichment for β-catenin signaling, negative enrichment for inflammatory response, negative enrichment for IFNα response, and negative enrichment for IFNγ response (Fig. 3C). These data indicate the presence of more immune-cold tumors in Black patients than in White patients. We then performed network analysis based on the transcriptomics of HNC tumors in the Black patient cohort. We applied the MCODE algorithm to identify densely connected network components. Pathway and process enrichment analyses were independently applied to each MCODE component. The best-scoring terms by P value were identified as neddylation pathways, HTAC acetylate histones/chromatin modification, and pyrimidine metabolism (Fig. 3D). Functional pathway analysis showed enrichment of the c-MYC pathway and RET tyrosine kinase–regulated signaling events as the top enriched pathways with the highest OR and smallest P values in Black patients (Fig. 3E). When we integrated data with a known transcription factor protein–protein interaction network, we found that c-MYC acts as a master regulator for HNC tumors in Black patients (Fig. 3F). Interestingly, of the 1,685 patients that were available in the combined cohort, all genomic changes in c-MYC (amplifications, gains) were significantly more common in Black patients than in White patients (76.84% in Black patients vs. 64.57% in White patients; P = 0.0392).

Figure 3.

Gene expression and pathway analysis of HNC tumor samples by ancestral race. A, Volcano plot depicting differentially expressed mRNAs in Black patients versus White patients [purple dots represent genes with significant P values (−log10P > 2), and gray dots represent genes with nonsignificant differential expression). B, Top genes with decreased or increased mRNA expression in Black patients compared with White patients based on mRNA expression in TCGA dataset. C, GSEA for β-catenin signaling, inflammatory response, alpha response, and gamma response in Black patients versus White patients. D, Transcriptomic-based network analysis showing connected network components in HNC tumors of Black patients. E, Functional pathway analysis of significantly overexpressed genes in Black patients with HNC depicts pathways significantly enriched in Black patients with the highest OR and lowest P value (corrected for FDR < 0.05). F, Functional transcription factor analysis demonstrates the binding site of c-MYC on many upregulated genes in HNC tumors of Black patients. n = 49 for Black patients and n = 449 for White patients.

Figure 3.

Gene expression and pathway analysis of HNC tumor samples by ancestral race. A, Volcano plot depicting differentially expressed mRNAs in Black patients versus White patients [purple dots represent genes with significant P values (−log10P > 2), and gray dots represent genes with nonsignificant differential expression). B, Top genes with decreased or increased mRNA expression in Black patients compared with White patients based on mRNA expression in TCGA dataset. C, GSEA for β-catenin signaling, inflammatory response, alpha response, and gamma response in Black patients versus White patients. D, Transcriptomic-based network analysis showing connected network components in HNC tumors of Black patients. E, Functional pathway analysis of significantly overexpressed genes in Black patients with HNC depicts pathways significantly enriched in Black patients with the highest OR and lowest P value (corrected for FDR < 0.05). F, Functional transcription factor analysis demonstrates the binding site of c-MYC on many upregulated genes in HNC tumors of Black patients. n = 49 for Black patients and n = 449 for White patients.

Close modal

Tumor location subtypes

As we observed differences in tumor location sites between Black and White patients, we conducted a stratified analysis based on the tumor location with sizable sample sizes between the two groups. We continued to observe different genomics and molecular changes in the three most abundant anatomic subtypes: laryngeal, oral cavity, and oropharyngeal.

Laryngeal subtype

In the laryngeal HNC subtype, samples from Black patients showed more aggressive lymph node stages (P = 4.71e-3) and more advanced stages (stage III and IV; P = 1.056e-4) at the time of diagnosis, similar to the pan-HNC cohort (Fig. 4A and B). Comparing the gene expression profiles of laryngeal HNC tumors in Black patients versus White patients, we found four differentially expressed genes with an FDR of less than 0.01: SDK1, FOXO1, LRAP1L, and PIK3CA (Fig. 4C). The genes most frequently mutated and the highest CNAs overall and with the lowest P value between Black patient and White patient samples are illustrated in Fig. 4D and E. Patients in the Black cohort had more mutations in TP53, KMT2D, MYO18B, CMYA5, and RUNX1T1 and showed higher CNA frequencies for CDKN2 genes. Unlike the pan-HNC cohort, mutations in the FRG1BP, SMG1, and MACF1 genes were present in the laryngeal cohort. These genes warrant further investigation for their uniqueness in the laryngeal subtype of HNC.

Figure 4.

Laryngeal HNC subtype clinical and genomic data by ancestral race. Laryngeal HNC in Black and White patients differs by lymph node stage (A) and disease stage (B). C, Volcano plot depicting differentially expressed mRNAs in laryngeal HNC in Black patients versus White patients [pink dots represent genes with significant P values (log10P >2); gray dots represent genes with nonsignificant differential expression]. Laryngeal subtype gene mutations (D) and CNAs (E) overall and by lowest P value. n = 20 for Black patients and n = 101 for White patients. *, P < 0.05; **, P < 0.01; ***, P < 0.001 after FDR adjustment.

Figure 4.

Laryngeal HNC subtype clinical and genomic data by ancestral race. Laryngeal HNC in Black and White patients differs by lymph node stage (A) and disease stage (B). C, Volcano plot depicting differentially expressed mRNAs in laryngeal HNC in Black patients versus White patients [pink dots represent genes with significant P values (log10P >2); gray dots represent genes with nonsignificant differential expression]. Laryngeal subtype gene mutations (D) and CNAs (E) overall and by lowest P value. n = 20 for Black patients and n = 101 for White patients. *, P < 0.05; **, P < 0.01; ***, P < 0.001 after FDR adjustment.

Close modal

Oropharyngeal subtype

In the oropharyngeal subtype, genes with the highest mutation and CNA frequencies and the most significant P values were all in patients of African ancestry (Supplementary Fig. S3A–S3D). Black patients with oropharyngeal HNC showed higher mutation frequencies of ABCF1, CSMD1, KLKB1, MED25, and RASGRP2. ABCF1 has been shown to promote epithelial–mesenchymal transition (EMT) and chemoresistance in hepatocellular carcinoma (29). Loss of CSDM1 is frequently found in cancer cells and may be associated with a poor prognosis (30). The CNA frequencies of LINC02912, a long noncoding RNA (ncRNA), and PVT1 were higher in Black patients versus White patients for the oropharyngeal HNC subtype. Long ncRNAs are involved in cancer progression; long ncRNA LINC02273 has been shown to drive metastasis in breast cancer (31).

Oral cavity subtype

In the oral cavity subtype of HNC, patients of African ancestry had higher lymph node stages at diagnosis (Fig. 5A). The neoplasm stage at diagnosis was also more advanced in Black patients versus White patients for this subtype (Fig. 5B). Differences in gene mutations and CNA frequencies between the two groups were observed (Fig. 5C and D). TP53, TTN, MUC16, KMT2D, and FAT1, among others, were more frequently mutated in the Black patient group (Fig. 5C). Patients of African ancestry had higher CNA frequencies for DKN2A-DT, TPRG1, and CDKN2A. (Fig. 5D). Pathway analysis of differentially expressed genes in the oral cavity subtype demonstrated activation of ncRNA metabolic processes, ncRNA processing, DNA metabolic processes, and alpha-amino acid metabolic processes in the Black patient cohort. In contrast, actin cytoskeleton and cytoskeletal protein binding were suppressed in Black patient tumors compared with White patients (Fig. 5E and F). Integration of data with transcription factor perturbation experiments and ENCODE and CHEA databases showed c-MYC as the master regulator of many genes that were upregulated in oral cancer (Fig. 5G).

Figure 5.

Oral cavity HNC subtype clinical and genomic data by ancestral race. Oral HNC in Black and White patients differs by disease stage (A) and lymph node stage (B). Oral HNC highest gene mutations (C) and CNAs (D). Pathway analysis based on differentially expressed genes in oral HNC indicates significantly enriched pathways (E) and significantly activated and repressed pathways (F). Pathway analysis showed high OR and a small P value of c-MYC enrichment in oral HNC (G). n = 22 for Black patients and n = 303 for White patients. *, P < 0.05; **, P < 0.01 after FDR adjustment.

Figure 5.

Oral cavity HNC subtype clinical and genomic data by ancestral race. Oral HNC in Black and White patients differs by disease stage (A) and lymph node stage (B). Oral HNC highest gene mutations (C) and CNAs (D). Pathway analysis based on differentially expressed genes in oral HNC indicates significantly enriched pathways (E) and significantly activated and repressed pathways (F). Pathway analysis showed high OR and a small P value of c-MYC enrichment in oral HNC (G). n = 22 for Black patients and n = 303 for White patients. *, P < 0.05; **, P < 0.01 after FDR adjustment.

Close modal

Interestingly, IHC staining of untreated HNC cases from patients of African ancestry showed a significant increase in c-MYC staining of tumor cells, tumor stroma, and dysplastic epithelium (P < 0.0001 for both comparisons; Fig. 6AE). No significant difference in staining was observed in the stroma of oral dysplastic lesions (Fig. 6F). HNC tumors from patients of African ancestry showed higher c-MYC expression and had significantly lower DFS (P = 0.0049; Fig. 6G).

Figure 6.

c-MYC is overexpressed in tumors of Black patients. A and B, representation of IHC images and image segmentation with high c-MYC expression and low c-MYC expression in HNC (A) and dysplasia (B). The levels of c-MYC staining were quantified in HNC tumor cells (C), HNC stroma (D), dysplastic epithelium (E), and dysplastic stroma (F). HNC samples were dichotomized to high c-MYC expression versus low c-MYC expression (n = 18 per group). DFS was compared (P = 0.0049; G). n = 36 for Black patients and n = 36 for White patients. Data are presented as mean and SD. ****, P < 0.0001 after FDR adjustment.

Figure 6.

c-MYC is overexpressed in tumors of Black patients. A and B, representation of IHC images and image segmentation with high c-MYC expression and low c-MYC expression in HNC (A) and dysplasia (B). The levels of c-MYC staining were quantified in HNC tumor cells (C), HNC stroma (D), dysplastic epithelium (E), and dysplastic stroma (F). HNC samples were dichotomized to high c-MYC expression versus low c-MYC expression (n = 18 per group). DFS was compared (P = 0.0049; G). n = 36 for Black patients and n = 36 for White patients. Data are presented as mean and SD. ****, P < 0.0001 after FDR adjustment.

Close modal

Molecular changes in advanced HNC

Because our initial results indicated more aggressive tumor types in Black patients, molecular changes in stages III and IV in Black patients versus White patients were compared. The genes with the highest mutation frequencies in Black and White patients are presented in Supplementary Fig. S4A. Interestingly, we found mutations that were significantly present in the advanced tumors of Black patients, including CD5L, PCDHL12, LIPI, ESM1, LCAM1, UGT2B11, RIPK4, IL6T, ISL1, UGT2B17, and CCDC142 (Supplementary Fig. S4B). There were 560 differentially expressed genes in stages III and IV HNC between the two groups (Supplementary Fig. S4C). Black patients showed decreased expression of PPIL3, LRRC37A2, IL22, USP32P1, LRSC37A, C6ORF52, RABL3, AMH, TMEM187, and GBP7. Reduced levels of and increased methylation of LRRCC37A2 are associated with poorly differentiated gastric cancer tumors and may have similar activity in HNC (32). Furthermore, Black patients had increased expression of PI4KAP2, TBL2, POLR2J2, ULK4, SERHL, SERPINC1, PYCR1, TOP1MT, CCL2LI, and HL2 (Supplementary Fig. S4D). SERHL is a long ncRNA shown to downregulate the expression of essential regulatory genes of the RAS and NFκB signaling pathways; increased SERHL expression is associated with worse survival in hepatocellular carcinoma (33). Pathway analysis showed activation of MYC pathways in stages III and IV tumors of Black patients (Supplementary Fig. S4E). The differential molecular changes observed in advanced HNC between Black and White patients indicate that the molecular changes observed overall in the pan-HNC cohorts are most likely not only based on a higher percentage of aggressive tumors in Black patients.

While race remains a social construct, many factors, including socioeconomic, environmental, and behavioral characteristics, may play a role in the differential health outcomes seen in complex diseases such as cancer. Current molecular findings and clinical outcomes associated with racial ancestry are inadequate for constructing the advanced variation observed in cancer and other diseases (34).

In this study, less favorable survival and more aggressive tumors were observed in Black patients. These findings at the molecular and clinical levels are likely not intrinsic to race per se but instead are a reflection of differences in cumulative life stressors. Several studies have shown that life stressors, such as socioeconomic status, contribute to DNA methylation levels which may in turn influence the regulation of proinflammatory stress responses (5, 6). Group-level differences in exposures and life stressors related to social disparities may result in group-level genomic and epigenetic modifications that mediate more aggressive tumor types and worse HNC outcomes through proinflammatory states.

Most of the current literature identifies and describes racial disparities in HNC based on epidemiologic data alone (3, 4, 35). Historically, cancer biomarkers and molecular signatures have primarily focused on samples from White patients. This study posed a unique opportunity to update and advance the recorded molecular data to improve representation of the true population. In recent years, several large studies have attempted to determine the effectiveness of targeted drugs matched with tumor molecular profiles (36, 37). Several studies have shown higher response rates in tumors treated with matched therapy and have confirmed its association with improved overall survival (36, 37).

In the context of HNC, many oncogenes have been explored as they relate to carcinogenesis. Upregulation of EGFR, c-MYC, BCL, and members of the RAS family are involved in the development and progression of HNC (38). Our current study showed that the upregulation of c-MYC and RAS pathways in HNC is more frequent in Black patients than in White patients. These overexpressed genes may be implicated in developing the more aggressive tumor types in specific individuals. c-MYC is a transcription factor involved in a wide variety of cellular processes, including cell proliferation, progression, differentiation, and apoptosis, and functions by binding to enhancer box (E-box) sequences (39). Specific genetic changes in c-MYC have been linked to certain cancers, such as its translocation in Burkitt's lymphoma and its amplification in colon cancer. In particular, c-MYC upregulation has been observed in gastric, colon, prostate, breast, and lung cancer (40). Furthermore, increased expression or copy-number gain of c-MYC is associated with a worse prognosis in HNC (41).

In this study, laryngeal and oral cavity HNC tumors of Black patients showed activation of c-MYC. Activation of c-MYC signaling was validated in an independent cohort of Black HNC tumor samples using IHC. Quantification of c-MYC staining in tumor cells, tumor stroma, dysplastic cells, and dysplastic stroma all showed overexpression of c-MYC. Interestingly, Black patients with high c-MYC expression had worse survival outcomes. This might be due to the fact that MYC as a master regulator of oncogenesis and plays an important role in various cellular biological processes, such as cell invasion, metabolism, differentiation, proliferation, drug resistance (42). On the basis of this finding, we suggest that further work is warranted to study c-MYC, specifically in the context of head and neck tumors of patients presenting with more aggressive tumor types. It should be noted that our validation cohort and TCGA data both included patients who had not received treatment. While this represents a potential limitation of this study, future investigations in the context of treatment modalities and tumor response between the two groups should be undertaken.

RET gain-of-function mutations are often observed in human cancers, and our study showed enrichment of RET-tyrosine signaling events in Black patients. RET regulates several intracellular signaling pathways involved in cell survival, differentiation, proliferation, and migration, such as the RAS/RAF and PI3K/AKT (43). Elevated RET levels have been previously observed in HNC and correlated with increased tumor size, advanced tumor stage, and decreased overall survival (44). Mechanistically, RET knockdown inhibited HNC cell proliferation and invasion, both in vitro and in vivo. The role of RET tyrosine–regulated events in the progression of more aggressive HNC tumor types is not well established. Our results showing increased RET mutation frequencies in tumors of Black patients warrant further investigation.

TP53 mutations are observed in more than two-thirds of adult solid tumors. Here, we found that TP53 mutations are more frequently found in HNC of Black patients, both in pan-HNC and within the three anatomic subtypes. Interestingly, methylation of genes involved in TP53 pathways was also observed at a higher frequency in Black patients’ HNC tumors. Previous studies have shown an association between TP53 mutations and decreased overall survival (45). Our data suggest that TP53 mutations and hypermethylation may correlate with tumor aggressiveness and be associated with survival.

Apart from differences in TP53 mutations, two other highly significant differential mutation frequencies were observed for KMT2D and UNC13C; Black patients had higher mutation frequencies for both genes. KMT2D is part of the histone lysine methyltransferase (HMT) family, and dysregulation of HMTs has been shown to have downstream effects promoting cancer progression; KMT2-mutant cancers show upregulation of cell cycle, metabolism, and IFNα/β response pathways (46). In a recent study, patients with gastric cancer with KMT2 mutations that were treated with immune checkpoint inhibitors were found to have a more prolonged overall survival rate when compared with KMT2-wild-type patients (47). UNC13C was recently identified as a tumor suppressor in oral cancer via its role as an essential regulator of EMT signaling pathways. Furthermore, UNC13C expression is inversely correlated with cancer stage in oral cancer samples (48). KMT2D and UNC13C are yet two additional translational avenues to explore to assess the clinical and treatment utility of screening for such mutations.

On the basis of the observed differences in tumor location sites between Black and White patients, a stratified analysis was conducted on the basis of tumor location. A greater MYO18B mutation frequency in the laryngeal subtype of HNC in Black patients was found to occur when compared with White patients. MYO18B has been previously identified as a tumor suppressor in other cancers but not in HNC (49). Further work is warranted to study the diagnostic, predictive, and clinical utility of MYO18B in HNC. Higher deletion frequencies of CDKN2A were also found in the Black patient pool. CDKN2A encodes the p16 tumor suppressor and is frequently inactivated via hypermethylation in human cancers; loss of CDKN2A has been associated with poor survival in certain HNCs (50). CDKN2A offers another potential diagnostic, prognostic, and clinical tool for patients presenting with particular molecular subtypes.

Using ancestry-informative markers to computationally assign ancestry in our HNC patient cohort, we found Black patients present with a younger age at diagnosis, more aggressive tumor types, higher rates of metastasis, and worse overall survival than White patients. Furthermore, Black patients had fewer HPV-positive tumor types and higher laryngeal tumor site rates. Pathway analysis of HNC tumors from patients of African ancestry identified c-MYC and RET tyrosine–regulated signaling events as master regulators of these tumors. This study revealed a differential genomic profile among the two ancestral cohorts that may be associated with the existing disparities in HNC. Given the recent attention to molecular profiling and targeted drug therapies in cancer, the differential mutation rates and pathway activity found in our study may have potential clinical utility in the screening, diagnosing, monitoring, and treatment of HNC.

A.M. Taylor reports grants from Ono Pharmaceutical outside the submitted work. No disclosures were reported by the other authors.

N. Mezghani: Conceptualization, data curation, formal analysis, investigation, methodology, writing–original draft. A. Yao: Formal analysis, investigation, methodology. D. Vasilyeva: Investigation, visualization, methodology. N. Kaplan: Data curation, methodology, writing–review and editing. A. Shackelford: Investigation, visualization. A. Yoon: Data curation, investigation, visualization. E. Phillipone: Investigation, visualization. S. Dubey: Investigation, visualization, methodology. G.K. Schwartz: Conceptualization, resources, investigation. A.M. Taylor: Data curation, formal analysis, investigation, visualization, writing–review and editing. F. Momen-Heravi: Conceptualization, resources, data curation, software, formal analysis, supervision, funding acquisition, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing.

This work was supported by the 2020 AACR-The Mark Foundation for Cancer Research “Science of the Patient” (SOP) grants, grant number 20-60-51-MOME (to F. Momen-Heravi), NIH/NIDCR (DE031112) grant to (to F. Momen-Heravi, A.M. Taylor), NIH/NCAT (KL2TR001874) to F. Momen-Heravi, Columbia CTSA (UL1 TR001873), Columbia University Herbert Irving Comprehensive Cancer Center NIH/NCI(P30CA013696).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

1.
Ferlay
J
,
Colombet
M
,
Soerjomataram
I
,
Mathers
C
,
Parkin
DM
,
Pineros
M
, et al
.
Estimating the global cancer incidence and mortality in 2018: GLOBOCAN sources and methods
.
Int J Cancer
2019
;
144
:
1941
53
.
2.
Keir
JA
,
Whiteside
OJ
,
Winter
SC
,
Maitra
S
,
Corbridge
RC
,
Cox
GJ
.
Outcomes in squamous cell carcinoma with advanced neck disease
.
Ann R Coll Surg Engl
2007
;
89
:
703
8
.
3.
DeSantis
CE
,
Siegel
RL
,
Sauer
AG
,
Miller
KD
,
Fedewa
SA
,
Alcaraz
KI
, et al
.
Cancer statistics for African Americans, 2016: progress and opportunities in reducing racial disparities
.
CA Cancer J Clin
2016
;
66
:
290
308
.
4.
Gourin
CG
,
Podolsky
RH
.
Racial disparities in patients with head and neck squamous cell carcinoma
.
Laryngoscope
2006
;
116
:
1093
106
.
5.
Hunter
RG
,
McCarthy
KJ
,
Milne
TA
,
Pfaff
DW
,
McEwen
BS
.
Regulation of hippocampal H3 histone methylation by acute and chronic stress
.
Proc Natl Acad Sci U S A
2009
;
106
:
20912
7
.
6.
Saban
KL
,
Mathews
HL
,
DeVon
HA
,
Janusek
LW
.
Epigenetics and social context: implications for disparity in cardiovascular disease
.
Aging Dis
2014
;
5
:
346
55
.
7.
Guda
K
,
Veigl
ML
,
Varadan
V
,
Nosrati
A
,
Ravi
L
,
Lutterbaugh
J
, et al
.
Novel recurrently mutated genes in African American colon cancers
.
Proc Natl Acad Sci U S A
2015
;
112
:
1149
54
.
8.
Krishnan
B
,
Rose
TL
,
Kardos
J
,
Milowsky
MI
,
Kim
WY
.
Intrinsic genomic differences between African American and White Patients with clear cell renal cell carcinoma
.
JAMA Oncol
2016
;
2
:
664
7
.
9.
Lindner
R
,
Sullivan
C
,
Offor
O
,
Lezon-Geyda
K
,
Halligan
K
,
Fischbach
N
, et al
.
Molecular phenotypes in triple negative breast cancer from African American patients suggest targets for therapy
.
PLoS One
2013
;
8
:
e71915
.
10.
Arrieta
O
,
Cardona
AF
,
Martin
C
,
Mas-Lopez
L
,
Corrales-Rodriguez
L
,
Bramuglia
G
, et al
.
Updated frequency of EGFR and KRAS mutations in nonsmall-cell lung cancer in Latin America: the Latin-American consortium for the investigation of lung cancer (CLICaP)
.
J Thorac Oncol
2015
;
10
:
838
43
.
11.
Couraud
S
,
Zalcman
G
,
Milleron
B
,
Morin
F
,
Souquet
PJ
.
Lung cancer in never smokers–a review
.
Eur J Cancer
2012
;
48
:
1299
311
.
12.
Driscoll
JJ
,
Rixe
O
.
Overall survival: still the gold standard: why overall survival remains the definitive end point in cancer clinical trials
.
Cancer J
2009
;
15
:
401
5
.
13.
Mermel
CH
,
Schumacher
SE
,
Hill
B
,
Meyerson
ML
,
Beroukhim
R
,
Getz
G
.
GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers
.
Genome Biol
2011
;
12
:
R41
.
14.
Carrot-Zhang
J
,
Chambwe
N
,
Damrauer
JS
,
Knijnenburg
TA
,
Robertson
AG
,
Yau
C
, et al
.
Comprehensive analysis of genetic ancestry and its molecular correlates in cancer
.
Cancer Cell
2020
;
37
:
639
54
.
15.
Flanagin
A
,
Frey
T
,
Christiansen
SL
;
AMA Manual of Style Committee
.
Updated guidance on the reporting of race and ethnicity in medical and science journals
.
JAMA
2021
;
326
:
621
7
.
16.
Zhu
Y
,
Qiu
P
,
Ji
Y
.
TCGA-assembler: open-source software for retrieving and processing TCGA data
.
Nat Methods
2014
;
11
:
599
600
.
17.
Szklarczyk
D
,
Gable
AL
,
Lyon
D
,
Junge
A
,
Wyder
S
,
Huerta-Cepas
J
, et al
.
STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets
.
Nucleic Acids Res
2019
;
47
:
D607
D13
.
18.
Bader
GD
,
Hogue
CW
.
An automated method for finding molecular complexes in large protein interaction networks
.
BMC Bioinformatics
2003
;
4
:
2
.
19.
Zabarovsky
ER
,
Lerman
MI
,
Minna
JD
.
Tumor suppressor genes on chromosome 3p involved in the pathogenesis of lung and other cancers
.
Oncogene
2002
;
21
:
6915
35
.
20.
Wu
YH
,
Chang
TH
,
Huang
YF
,
Huang
HD
,
Chou
CY
.
COL11A1 promotes tumor progression and predicts poor clinical outcome in ovarian cancer
.
Oncogene
2014
;
33
:
3432
40
.
21.
Faraj Shaglouf
LH
,
Ranjpour
M
,
Wajid
S
,
Jain
SK
.
Elevated expression of cellular SYNE1, MMP10, and GTPase1 and their regulatory role in hepatocellular carcinoma progression
.
Protoplasma
2020
;
257
:
157
67
.
22.
Cowin
PA
,
George
J
,
Fereday
S
,
Loehrer
E
,
Van Loo
P
,
Cullinane
C
, et al
.
LRP1B deletion in high-grade serous ovarian cancers is associated with acquired chemotherapy resistance to liposomal doxorubicin
.
Cancer Res
2012
;
72
:
4060
73
.
23.
Zhu
Y
,
Du
Y
,
Zhang
Y
.
DHX33 promotes colon cancer development downstream of Wnt signaling
.
Gene
2020
;
735
:
144402
.
24.
Liu
Y
,
Colby
JK
,
Zuo
X
,
Jaoude
J
,
Wei
D
,
Shureiqi
I
.
The Role of PPAR-delta in metabolism, inflammation, and cancer: many characters of a critical transcription factor
.
Int J Mol Sci
2018
;
19
:
3339
.
25.
Gheghiani
L
,
Shang
S
,
Fu
Z
.
Targeting the PLK1-FOXO1 pathway as a novel therapeutic approach for treating advanced prostate cancer
.
Sci Rep
2020
;
10
:
12327
.
26.
Samanta
D
,
Huang
TY
,
Shah
R
,
Yang
Y
,
Pan
F
,
Semenza
GL
.
BIRC2 expression impairs anti-cancer immunity and immunotherapy efficacy
.
Cell Rep
2020
;
32
:
108073
.
27.
Hu
Q
,
Lin
X
,
Ding
L
,
Zeng
Y
,
Pang
D
,
Ouyang
N
, et al
.
ARHGAP42 promotes cell migration and invasion involving PI3K/Akt signaling pathway in nasopharyngeal carcinoma
.
Cancer Med
2018
;
7
:
3862
74
.
28.
Ming
L
,
Wang
P
,
Bank
A
,
Yu
J
,
Zhang
L
.
PUMA Dissociates Bax and Bcl-X(L) to induce apoptosis in colon cancer cells
.
J Biol Chem
2006
;
281
:
16034
42
.
29.
Fung
SW
,
Cheung
PF
,
Yip
CW
,
Ng
LW
,
Cheung
TT
,
Chong
CC
, et al
.
The ATP-binding cassette transporter ABCF1 is a hepatic oncofetal protein that promotes chemoresistance, EMT and cancer stemness in hepatocellular carcinoma
.
Cancer Lett
2019
;
457
:
98
109
.
30.
Kamal
M
,
Shaaban
AM
,
Zhang
L
,
Walker
C
,
Gray
S
,
Thakker
N
, et al
.
Loss of CSMD1 expression is associated with high tumour grade and poor survival in invasive ductal breast carcinoma
.
Breast Cancer Res Treat
2010
;
121
:
555
63
.
31.
Xiu
B
,
Chi
Y
,
Liu
L
,
Chi
W
,
Zhang
Q
,
Chen
J
, et al
.
LINC02273 drives breast cancer metastasis by epigenetically increasing AGR2 transcription
.
Mol Cancer
2019
;
18
:
187
.
32.
Wisnieski
F
,
Geraldis
JC
,
Santos
LC
,
Leal
MF
,
Calcagno
DQ
,
Gigek
CO
, et al
.
Differential regulation of LRRC37A2 in gastric cancer by DNA methylation
.
Epigenetics
2022
;
17
:
110
6
.
33.
Zhang
Y
,
Chen
D
,
Yang
M
,
Qian
X
,
Long
C
,
Zheng
Z
.
Comprehensive analysis of competing endogenous RNA network focusing on long noncoding RNA involved in cirrhotic hepatocellular carcinoma
.
Anal Cell Pathol
2021
;
2021
:
5510111
.
34.
Gravlee
CC
.
How race becomes biology: embodiment of social inequality
.
Am J Phys Anthropol
2009
;
139
:
47
57
.
35.
Mukherjee
A
,
Idigo
AJ
,
Ye
Y
,
Wiener
HW
,
Paluri
R
,
Nabell
LM
, et al
.
Geographical and racial disparities in head and neck cancer diagnosis in South-Eastern United States: using real-world electronic medical records data
.
Health Equity
2020
;
4
:
43
51
.
36.
Stockley
TL
,
Oza
AM
,
Berman
HK
,
Leighl
NB
,
Knox
JJ
,
Shepherd
FA
, et al
.
Molecular profiling of advanced solid tumors and patient outcomes with genotype-matched clinical trials: the Princess Margaret IMPACT/COMPACT trial
.
Genome Med
2016
;
8
:
109
.
37.
Taniguchi
H
,
Sen
T
,
Rudin
CM
.
Targeted therapies and biomarkers in small cell lung cancer
.
Front Oncol
2020
;
10
:
741
.
38.
Wong
DT
,
Todd
R
,
Tsuji
T
,
Donoff
RB
.
Molecular biology of human oral cancer
.
Crit Rev Oral Biol Med
1996
;
7
:
319
28
.
39.
Henriksson
M
,
Luscher
B
.
Proteins of the Myc network: essential regulators of cell growth and differentiation
.
Adv Cancer Res
1996
;
68
:
109
82
.
40.
Knoepfler
PS
.
Myc goes global: new tricks for an old oncogene
.
Cancer Res
2007
;
67
:
5061
3
.
41.
Field
JK
,
Spandidos
DA
,
Stell
PM
,
Vaughan
ED
,
Evan
GI
,
Moore
JP
.
Elevated expression of the c-myc oncoprotein correlates with poor prognosis in head and neck squamous cell carcinoma
.
Oncogene
1989
;
4
:
1463
8
.
42.
Dang
CV
.
MYC on the path to cancer
.
Cell
2012
;
149
:
22
35
.
43.
Arighi
E
,
Borrello
MG
,
Sariola
H
.
RET tyrosine kinase signaling in development and cancer
.
Cytokine Growth Factor Rev
2005
;
16
:
441
67
.
44.
Lin
C
,
Lu
W
,
Ren
Z
,
Tang
Y
,
Zhang
C
,
Yang
R
, et al
.
Elevated RET expression enhances EGFR activation and mediates EGFR inhibitor resistance in head and neck squamous cell carcinoma
.
Cancer Lett
2016
;
377
:
1
10
.
45.
Poeta
ML
,
Manola
J
,
Goldwasser
MA
,
Forastiere
A
,
Benoit
N
,
Califano
JA
, et al
.
TP53 mutations and survival in squamous-cell carcinoma of the head and neck
.
N Engl J Med
2007
;
357
:
2552
61
.
46.
Albert
M
,
Helin
K
.
Histone methyltransferases in cancer
.
Semin Cell Dev Biol
2010
;
21
:
209
20
.
47.
Wang
J
,
Xiu
J
,
Baca
Y
,
Battaglin
F
,
Arai
H
,
Kawanishi
N
, et al
.
Large-scale analysis of KMT2 mutations defines a distinctive molecular subset with treatment implication in gastric cancer
.
Oncogene
2021
;
40
:
4894
905
.
48.
Velmurugan
BK
,
Yeh
KT
,
Hsieh
MJ
,
Yeh
CM
,
Lin
CC
,
Kao
CY
, et al
.
UNC13C Suppress tumor progression via inhibiting EMT pathway and improves survival in oral squamous cell carcinoma
.
Front Oncol
2019
;
9
:
728
.
49.
Yanaihara
N
,
Nishioka
M
,
Kohno
T
,
Otsuka
A
,
Okamoto
A
,
Ochiai
K
, et al
.
Reduced expression of MYO18B, a candidate tumor-suppressor gene on chromosome arm 22q, in ovarian cancer
.
Int J Cancer
2004
;
112
:
150
4
.
50.
Chen
WS
,
Bindra
RS
,
Mo
A
,
Hayman
T
,
Husain
Z
,
Contessa
JN
, et al
.
CDKN2A copy number loss is an independent prognostic factor in HPV-negative head and neck squamous cell carcinoma
.
Front Oncol
2018
;
8
:
95
.