Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Jul 15.
Published in final edited form as: Cancer Causes Control. 2014 Feb 2;25(4):437–450. doi: 10.1007/s10552-014-0346-x

Single nucleotide polymorphisms in nucleotide excision repair genes, cancer treatment, and head and neck cancer survival

Annah B Wyss 1, Mark C Weissler 2, Christy L Avery 3, Amy H Herring 4, Jeannette T Bensen 5, Jill S Barnholtz-Sloan 6, William K Funkhouser 7, Andrew F Olshan 8,✉
PMCID: PMC4096829  NIHMSID: NIHMS596148  PMID: 24487794

Abstract

Purpose

Head and neck cancers (HNC) are commonly treated with radiation and platinum-based chemotherapy, which produce bulky DNA adducts to eradicate cancerous cells. Because nucleotide excision repair (NER) enzymes remove adducts, variants in NER genes may be associated with survival among HNC cases both independently and jointly with treatment.

Methods

Cox proportional hazards models were used to estimate race-stratified (White, African American) hazard ratios (HRs) and 95 % confidence intervals for overall (OS) and disease-specific (DS) survival based on treatment (combinations of surgery, radiation, and chemotherapy) and 84 single nucleotide polymorphisms (SNPs) in 15 NER genes among 1,227 HNC cases from the Carolina Head and Neck Cancer Epidemiology Study.

Results

None of the NER variants evaluated were associated with survival at a Bonferroni-corrected alpha of 0.0006. However, rs3136038 [OS HR = 0.79 (0.65, 0.97), DS HR = 0.69 (0.51, 0.93)] and rs3136130 [OS HR = 0.78 (0.64, 0.96), DS HR = 0.68 (0.50, 0.92)] of ERCC4 and rs50871 [OS HR = 0.80 (0.64, 1.00), DS HR = 0.67 (0.48, 0.92)] of ERCC2 among Whites, and rs2607755 [OS HR = 0.62 (0.45, 0.86), DS HR = 0.51 (0.30, 0.86)] of XPC among African Americans were suggestively associated with survival at an uncorrected alpha of 0.05. Three SNP-treatment joint effects showed possible departures from additivity among Whites.

Conclusions

Our study, a large and extensive evaluation of SNPs in NER genes and HNC survival, identified mostly null associations, though a few variants were suggestively associated with survival and potentially interacted additively with treatment.

Keywords: Head and neck cancer DNA repair, Nucleotide excision repair, Chemotherapy, Radiation, Survival

Background

An estimated 53,640 incident head and neck cancer (HNC) cases and 11,520 associated deaths occured in the United States during 2013 [1]. Comprising tumors of the oral cavity, pharynx, and larynx, HNC is a relatively fatal disease [2, 3]. Among individuals with oral and pharyngeal cancers in the United States, five-year survival rates are 61.7 and 63.2 % for White men and women, respectively, and 37.2 and 51.2 % for African American men and women, respectively [2, 3]. HNC was historically treated with surgery and/or radiation [4, 5]. However, following a series of clinical trials in the 1990s, advanced tumors (stages 3 and 4) are increasingly treated with concurrent or induction radiation and chemotherapy [4, 5]. Other tumor characteristics (e.g., location and size) and the patients’ demographics (e.g., age) can also influence treatment decisions and outcomes [6].

Emerging literature suggests that genetic factors may also impact treatment response and survival among cancer patients [7, 8]. In order to initiate cell death (apoptosis) of cancerous cells, radiation and platinum-based chemotherapy are known to cause bulky DNA adducts, among other types of DNA damage [7, 9]. Since nucleotide excision repair (NER) is the pathway primarily responsible for removing DNA adducts, functional NER processes may lessen the efficacy of cancer treatment [7]. This hypothesis has led some researchers to describe DNA repair, including NER, as a ‘‘double-edged sword’’ or ‘‘Janus, the two-faced Roman god,’’ since functional genes are thought to protect against cancer incidence, but may mitigate the effectiveness of cancer treatments thus decreasing survival [7].

Although it is hypothesized that the effects of NER variants on survival may be dependent on treatment, previous epidemiologic studies on the effects of single nucleotide polymorphisms (SNPs) in NER genes and treatment on HNC mortality have been inconsistent [9–17]. For example, some studies conducted among patients receiving radiation reported null associations for rs13181 in excision repair cross-complementing 2 (ERCC2) and survival [9, 12]. Other studies showed evidence for significant differences in survival across genotypes of rs13181 [10, 14, 15], including a study which found the referent genotype (AA) was associated with worse survival among individuals treated with radiation and better survival among those not receiving radiation [17]. However, previous studies have been based on small sample sizes, predominantly European-descent populations, and a limited number of variants in NER genes [9–17]. The present study extends the literature by estimating main and joint effects of treatment (combinations of surgery, radiation, and chemotherapy) and 84 SNPs across 15 NER genes on survival in a large, racially diverse group of HNC cases.

Methods

Study population

The Carolina Head and Neck Cancer Epidemiology (CHANCE) Study is a population-based case–control study of 1,389 cases and 1,396 controls from North Carolina (NC) [18–22]. For the present analysis, we compared survival among cases by treatment and genotype. All cases were 20–80 years of age and were identified from the NC Central Cancer Registry between 1 January 2002 and 28 February 2006 using rapid case ascertainment [18–22]. Cases with tumors in the oral cavity, oropharynx, hypo-pharynx, larynx, and HNC not otherwise specified (NOS) were included, while tumors of the salivary glands, naso-pharynx, nasal cavity, and nasal sinuses were excluded [18–22]. Self-reported demographic and behavioral information and biologic samples (~90 % blood, ~10 % buccal cells) were collected during a nurse-administered interview [18–22]. We excluded cases who self-reported race other than White or African American (n = 26, 1.9 %) because of sparse data, as well as lip cancers (n = 21, 1.3 %) because of etiologic differences. Cases who did not provide a biologic sample were also excluded; this comprised 52 cases (3.7 %) who were deceased at time of interview (i.e. proxy interviews) and eight cases (0.6 %) who provided in-person interviews but no biologic sample. Finally, cases whose samples were insufficient for genotyping or whose samples did not otherwise meet quality control criteria (n = 55, 4.0 %) were excluded. Our analysis included 1,227 HNC cases (922 White cases and 305 African American cases).

SNP selection and genotyping

Illumina GoldenGate assay with Sentrix Array Matrix and 96-well standard microtiter plates was used for genotyping [20–23]. As described previously [22], 71 tag SNPs in eight NER genes were selected based on a case–control study of HNC at The University of Texas MD Anderson Cancer Center (r2 ≥ 0.80, minor allele frequency (MAF) ≥ 0.05, 1–2 kb flanking region, CEU population) and 58 SNPs in 12 NER genes were selected based on other cancer studies and/or potential function (Online Resource 1). Of the 129 NER SNPs, variants with poor signal intensity or genotype clustering (14 SNPs) or a MAF less than 0.05 (30 SNPs among Whites and 36 SNPs among African Americans) were excluded (Online Resource 1) [21-22]. The majority of excluded SNPs were candidate SNPs selected based on previous literature or function (Online Resource 1). Genotype frequencies for the remaining SNPs were consistent with Hardy–Weinberg equilibrium (HWE) among CHANCE controls at a Bonferroni-corrected 0.0006 alpha level, and scatter plots showed reasonable clustering; therefore, no SNPs were excluded for HWE violations [22]. Our analysis included 84 SNPs in 14 NER genes among Whites and 79 SNPs in 15 NER genes among African Americans.

Treatment

First-course treatment information was abstracted from patients’ medical records. Information included whether the patient received surgery, radiation, and chemotherapy, including types of chemotherapy drugs: carboplatin, paraplatin, cisplatin, 5-FU, taxol, taxotere, docetaxel, paclit-axel, ifosfamide, and other [21]. Information on treatment start and end dates and whether radiation and chemotherapy were administered concurrently was not available for a large proportion of individuals. Therefore, combinations of treatment were generated from dichotomous variables for surgery, radiation, and chemotherapy regardless of timing. Information on tumor histology and stage was also abstracted from medical records [21]. Tumor grade was not uniformly available for all cases and therefore not considered.

HNC survival

CHANCE data were linked to the National Death Index (NDI) based on name, social security number, date of birth, sex, race, and state of residence to identify deaths through 2009, including date of death, location of death, and cause of death [21, 24]. Death records with HNC listed as an underlying cause of death were considered disease-specific deaths. For overall survival models, follow-up started at date of diagnosis for all cases and ended at date of death for individuals who died, or censoring on 31 December 2009 for individuals who were still alive [21]. For disease-specific survival models, follow-up started at date of diagnosis for all cases and ended at date of death for individuals who died of HNC, or censoring at date of death for individuals who died from causes other than HNC or 31 December 2009 for individuals who were still alive [21].

Statistical analysis

Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95 % confidence intervals (CIs) for the independent and joint effects of treatment and SNPs on HNC survival among Whites and African Americans separately. To evaluate the proportionality of hazards, we examined adjusted log-negative log plots by treatment and by genotype separately. In addition, we assessed the significance of including an interaction term for time and treatment or genotype in models. If log-negative log plots indicated a violation of the proportional hazards assumption and interaction terms with time were significant (p <>0.05), accelerated failure time (AFT) models were fit to explore robustness of results. This was the case for four SNPs in Whites (rs3731068, rs744154, rs3136085, and rs3136172) and three SNPs in African Americans (rs4150360, rs2020955, and rs13181). However, because p values for the AFT models were similar to those obtained from Cox models (i.e., the same set of significant SNP-HNC survival associations resulted from both approaches), results from the Cox models without an interaction term between SNPs and time are presented for simplicity. Absolute differences in HNC survival by treatment or genotype were also assessed via Kaplan–Meier plots, with cumulative survival calculated as the proportion of cases alive at each time point and log rank tests used to assess differences in survival.

Treatment

Treatment was modeled as a categorical variable with six groups: surgery only; radiation only; surgery and radiation; radiation and chemotherapy; surgery, radiation, and chemotherapy; and other (no treatment, chemotherapy only, or surgery and chemotherapy without radiation). Surgery only was used as the referent category because few individuals received no treatment (n = 9, 0.7 %). Even fewer individuals received chemotherapy only or chemotherapy with surgery without radiation (n = 4, 0.3 %), so these individuals were combined with individuals receiving no treatment into a single category labeled ‘‘other treatment.’’ In a separate model, we also considered receiving platinum-based chemotherapy drugs (carboplatin, paraplatin, or cisplatin, n = 464) versus not receiving platinum-based chemotherapy drugs (i.e., not receiving chemotherapy, n = 754, or only receiving non-platinum-based chemotherapy drugs, including 5 FU, taxol, taxotere, docetaxel, paclitaxel, or ifosfamide, n = 9). All treatment models were stratified by race and adjusted for sex, age (categorical), tumor stage (stages I, II, III, IV), tumor site (oral cavity, oropharynx, hypopharynx, larynx, NOS), education (high school or less, some college, and college or more), duration of cigarette smoking (years), and lifetime consumption of alcohol (categorical milliliters of ethanol).

SNPs

In agreement with previous CHANCE publications [22], SNPs were defined using a dominant genetic model and the referent allele for both Whites and African Americans was assigned to be the major allele based on controls from the overall study population. Race-stratified models included a single SNP at a time, with p values corrected using the Bonferroni method (0.05/84 = 0.0006 among Whites and 0.05/79 = 0.0006 among African Americans). The false discovery rate (FDR) approach to correcting for multiple comparisons was also considered as a supplementary analysis [25]. Further, SNP-survival associations with p values below 0.05 but not significant at a Bonferroni- or FDR-corrected alpha level were considered as ‘‘suggestive’’ associations. SNP models were adjusted for sex and age (including their interaction), as well as ancestry (proportion African ancestry). As described in previous studies of cancer among Whites and African Americans in NC, 145 ancestral informative markers (AIMS) were used to estimate the proportion of African and European ancestry of each participant based on Fisher’s information criterion (FIC) [20–22, 26–28]. Models did not include education, treatment, smoking or drinking because these variables were not considered confounders (i.e. these variables are not believed to effect germline variation), and models did not include tumor stage or site because these variables were considered causal intermediates.

Joint effects

Joint effects models included nine indicator variables for the interaction between treatment and a single SNP at a time: (1) surgery only, variant genotype; (2) radiation only, referent genotype; (3) radiation only, variant genotype; (4) surgery and radiation, referent genotype; (5) surgery and radiation, variant genotype; (6) radiation and chemotherapy, reference genotype; (7) radiation and chemotherapy, variant genotype; (8) surgery, radiation and chemotherapy, referent genotype; and (9) surgery, radiation and chemotherapy, variant genotype. Individuals receiving other treatment were excluded from this model due to small cell counts. Only joint effect estimates among Whites are presented because small cell counts among African Americans prohibited reliable estimation. Since both genetic and treatment exposures were assessed, models were adjusted for sex, age, tumor stage, tumor site, education, cigarette smoking, alcohol drinking, and ancestry. We also considered joint effects models for SNPs and platinum-based chemotherapy which included three disjoint indicator variables (no platinum-based chemotherapy, variant genotype; platinum-based chemotherapy, referent genotype; and platinum-based chemotherapy, variant genotype) and were adjusted for the covariates previously mentioned plus surgery (yes/no) and radiation (yes/no). Interactions between SNPs and treatments were assessed on the additive scale using the relative excess risk for interaction (RERI) with 95 % CIs calculated using the Hosmer and Lemeshow method [29]. Statistical analyses were performed using SAS 9.3 (Cary, NC) [30].

Results

Demographics

Of the 1,227 HNC cases in CHANCE, 545 (44.4 %) linked with the NDI through 2009 (Table 1). The remaining 682 (55.6 %) were assumed to be alive as of 31 December 2009. The median and mean follow-up times were 919.7 and 764.0 days, respectively, among individuals who died and 2,137.4 and 2,086.0 days, respectively, among those whowere alive. Among the 545 cases who died, just under half (n = 227, 41.7 %) had HNC listed as an underlying cause of death. Among these disease-specific deaths, the median and mean follow-up times were 729.7 and 594.0 days, respectively.

Table 1.

Demographic characteristics of head and neck cancer cases, Carolina Head and Neck Cancer Epidemiology (CHANCE) study, 2002–2009

Characteristic Alive
Overall deaths
Disease- specific deaths
n % n % n %
Total 682 55.6 545 44.4 227
Sex
 Male 519 76.1 419 76.9 169 74.4
 Female 163 23.9 126 23.1 58 25.6
Race/ethnicity
 White 539 79.0 383 70.3 169 74.4
 African American 143 21.0 162 29.7 58 25.6
Age at diagnosis
 20–49 152 22.3 87 16.0 42 18.5
 50–54 110 16.1 79 14.5 30 13.2
 55–59 125 18.3 82 15.0 35 15.4
 60–64 118 17.3 87 16.0 37 16.3
 65–69 78 11.4 90 16.5 33 14.5
 70–74 64 9.4 71 13.0 27 11.9
 75–80 35 5.1 49 9.0 23 10.1
Education
 High school or less 362 53.1 392 71.9 157 69.2
 Some college 195 28.6 99 18.2 48 21.1
 College or more 125 18.3 54 9.9 22 9.7
Tumor site
 Oral cavity 81 11.9 91 16.7 38 16.7
 Oropharynx 205 30.1 128 23.5 54 23.8
 Hypophaynx 16 2.3 39 7.2 15 6.6
 NOS 134 19.6 90 16.5 42 18.5
 Larynx 246 36.1 197 36.1 78 34.4
Stage
 I 195 28.6 84 15.4 17 7.5
 II 119 17.4 101 18.5 38 16.7
 III 118 17.3 93 17.1 40 17.6
 IV 250 36.7 267 49.0 132 58.1
Surgery
 No 269 39.4 268 49.2 119 52.4
 Yes 413 60.6 277 50.8 108 47.6
Radiation
 No 177 26.0 105 19.3 37 16.3
 Yes 505 74.0 440 80.7 190 83.7
Chemotherapy
 No 428 62.8 326 59.8 130 57.3
 Yes 254 37.2 219 40.2 97 42.7
Mean follow-up time (days) 2,137.4 919.7 729.7
Median follow-up time (days) 2,086.0 764.0 594.0
Open in a new tab

Modest variation by sex was observed when comparing cases who were living and dead after study follow-up (Table 1). However, a higher proportion of cases who died were diagnosed between 65 and 80 years of age (38.5% vs. 25.9%),were African American (29.7 vs. 21.0 %) or had a high school education or less (71.9 vs. 53.1 %) compared to living cases. With respect to tumor site, similar proportions had a diagnosis of laryngeal cancer (36.1 %). In contrast, 23.5 % of cases who died had oropharyngeal cancer compared to 30.1 % of cases who were living. As expected [31, 32], the distribution of tumor stage also varied by survival status, with living cases tending to have lower tumor stage.

Treatment

Among Whites, individuals who received only radiation tended to have worse overall (HR = 1.59, 95 % CI = 1.08, 2.34) and disease-specific survival (HR = 2.47, 95 % CI = 1.34, 4.56) compared to individuals who were treated with surgery alone (Table 2). Individuals receiving ‘‘other’’ treatment (i.e. no treatment, chemotherapy only or chemotherapy and surgery) also appeared to have poorer overall and disease-specific survival, though estimates were imprecise since they were based on few individuals (data not shown). In a separate model, receiving platinum-based chemotherapy appeared to be associated with better overall (HR = 0.71, 95 % CI = 0.52, 0.95 among Whites and HR = 0.77, 95 % CI = 0.48, 1.20 among African Americans) and disease-specific survival (HR = 0.63, 95 % CI = 0.41, 0.97 among Whites and HR = 0.45, 95 % CI = 0.19, 1.02 among African Americans) (Table 2).

Table 2.

Hazard ratios for cancer treatment and head and neck cancer survival in the Carolina Head and Neck Cancer Epidemiology (CHANCE) study

Cancer treatment Whites
African Americans
Overall deaths HNC deaths Alive Overall survival HR (95 % CI)a,b Disease-specific survival HR (95 % CI)a,b Overall deaths HNC deaths Alive Overall survival HR (95 % CI)a,b Disease-specific survivala,b HR (95 % CI)
Surgery only 68 21 144 1.00 (Referent) 1.00 (Referent) 26 9 31 1.00 (Referent) 1.00 (Referent)
Radiation only 77 36 83 1.59 (1.08, 2.34) 2.47 (1.34, 4.56) 24 7 31 1.00 (0.52, 1.92) 0.90 (0.26, 3.05)
Surgery and radiation 85 35 106 1.19 (0.81, 1.73) 1.28 (0.69, 2.36) 38 16 32 1.10 (0.63, 1.94) 1.13 (0.44, 2.86)
Radiation and chemotherapy 102 50 118 1.19 (0.78, 1.81) 1.48 (0.78, 2.81) 55 19 36 1.00 (0.56, 1.80) 0.52 (0.19, 1.39)
Surgery, radiation, chemotherapy 43 22 87 0.98 (0.61, 1.59) 1.26 (0.61, 2.58) 16 5 12 0.77 (0.36, 1.64) 0.36 (0.09, 1.39)
Platinum-based chemotherapy
Did not receive platinum-based chemotherapy 239 98 336 1.00 (Referent) 1.00 (Referent) 93 34 95 1.00 (Referent) 1.00 (Referent)
Did receive platinum-based chemotherapy 144 71 203 0.71 (0.52, 0.95) 0.63 (0.41, 0.97) 69 24 48 0.77 (0.48, 1.22) 0.45 (0.19, 1.02)
Open in a new tab

HR hazards ratio, CI confidence interval

a

HRs adjusted for age and sex (including pairwise interaction), tumor stage, tumor site, education, cigarette smoking, and alcohol drinking

b

Platinum-based chemotherapy HRs adjusted for age and sex (including pairwise interaction), tumor stage, tumor site, education, cigarette smoking, alcohol drinking, and surgery and/or radiation. Fifty-seven White cases and 20 African American cases missing information on alcohol drinking and therefore dropped from models

SNPs

Among Whites, four SNPs were modestly associated with only overall survival and three SNPs were modestly associated with both overall and disease-specific survival at an uncorrected 0.05 alpha level (Table 3). However, after correcting the alpha level using the Bonferroni and FDR methods, no SNPs were statistically significantly associated with either outcome. Among the SNPs associated with both outcomes at an uncorrected 0.05 alpha level, two tag SNPs were in linkage disequilibrium (LD) (r2 = 0.92, CEU population) on ERCC4, also known as xeroderma pigmentosum F (XPF) [33]. Specifically, rs3136038 (TT + TC vs. CC) and rs3136130 (TT + GT vs. GG) were suggestively associated with a similarly reduced hazards of overall (HR = 0.79, uncorrected 95 % CI = 0.65, 0.97 and HR = 0.78, uncorrected 95 % CI 0.64, 0.96, respectively) and disease-specific death (HR = 0.69, uncorrected 95 % CI = 0.51, 0.93, and HR = 0.68, uncorrected 95 % CI = 0.50, 0.92, respectively). In addition, rs50871 (TT + TC vs. CC), a tag SNP on ERCC2 also known as XPD, was suggestively associated with decreased hazards of overall (HR = 0.80, uncorrected 95 % CI = 0.64, 1.00) and disease-specific death (HR = 0.67, uncorrected 95 % CI = 0.48, 0.92). Stratifying by tumor stage, associations for rs3136038 or rs3136130 and survival were strongest among stage four cases and rs50871 among stage three cases (Online Resource 2). Figure 1 shows Kap-lan–Meier plots for these SNPs.

Table 3.

Hazard ratios for single nucleotide polymorphisms (SNPs) in nucleotide excision repair (NER) genes and survival among head and neck cancer cases in the Carolina Head and Neck Cancer Epidemiology (CHANCE) study, Whites

Gene SNP Coded allele
Overall deaths/deaths from HNC/ alive
Overall survival
Disease-specific survival
Referent (A) Variant (B) AA AB + BB HR (95 % CI)a p value HR (95 % CI)a p value
ERCC3 (XPB) rs4150496 G A 168 67 233 215 102 305 0.94 (0.77, 1.15) 0.55 1.10 (0.81, 1.51) 0.53
rs1011019 C T 191 84 271 192 85 268 1.02 (0.83, 1.24) 0.88 1.03 (0.76, 1.39) 0.87
rs4150434 G A 230 110 318 153 59 221 0.99 (0.80, 1.21) 0.91 0.80 (0.58, 1.11) 0.18
rs4150416 T G 165 69 245 217 99 292 1.10 (0.90, 1.35) 0.34 1.20 (0.88, 1.64) 0.24
rs4150407 A G 131 47 187 252 122 352 0.99 (0.80, 1.23) 0.94 1.32 (0.94, 1.85) 0.11
rs4150403 G A 303 139 433 80 30 106 1.02 (0.79, 1.31) 0.88 0.81 (0.55, 1.21) 0.31
rs4150402 G A 191 84 271 192 85 268 1.02 (0.83, 1.24) 0.88 1.03 (0.76, 1.39) 0.87
XPC rs2228001 A C 135 58 202 248 111 336 1.13 (0.92, 1.40) 0.24 1.16 (0.84, 1.60) 0.36
rs3731143 T C 333 147 485 50 22 54 1.18 (0.87, 1.59) 0.29 1.16 (0.73, 1.82) 0.53
rs2228000 C T 213 94 311 168 74 228 1.04 (0.85, 1.28) 0.68 1.04 (0.76, 1.41) 0.82
rs3731124 A C 215 94 306 168 75 233 1.02 (0.83, 1.26) 0.83 1.08 (0.79, 1.47) 0.63
rs13099160 A G 335 146 479 48 23 60 1.13 (0.83, 1.53) 0.44 1.17 (0.75, 1.82) 0.50
rs3731093 T C 321 141 455 59 26 79 1.03 (0.78, 1.36) 0.83 0.98 (0.64, 1.49) 0.91
rs3731089 G A 321 141 457 62 28 82 1.05 (0.80, 1.38) 0.74 1.02 (0.68, 1.53) 0.94
rs2733537 A G 167 73 249 216 96 290 1.07 (0.87, 1.31) 0.53 1.06 (0.78, 1.44) 0.71
rs3731068 C A 257 111 367 126 58 172 1.09 (0.88, 1.35) 0.45 1.16 (0.84, 1.60) 0.36
rs2607755 T C 100 46 142 283 123 397 0.99 (0.79, 1.25) 0.95 0.96 (0.68, 1.35) 0.80
rs1902658 G A 99 46 136 284 123 402 0.97 (0.77, 1.22) 0.78 0.92 (0.65, 1.29) 0.61
ERCC8 rs3117 T C 126 51 211 257 118 328 1.22 (0.98, 1.51) 0.07 1.37 (0.99, 1.91) 0.06
CDK7 rs2972388 A G 110 50 156 273 119 383 0.99 (0.79, 1.24) 0.94 0.95 (0.68, 1.33) 0.76
XPA rs3176757 C T 256 114 353 127 55 186 0.96 (0.77, 1.19) 0.71 0.93 (0.67, 1.29) 0.68
rs3176753 T C 381 169 537 0.47 (0.06, 3.70) 0.47
rs2808667 C T 345 153 469 36 16 70 0.70 (0.50, 1.00) 0.05 0.74 (0.44, 1.25) 0.26
rs2805835 G C 301 133 426 82 36 113 1.00 (0.78, 1.28) 0.97 0.94 (0.65, 1.36) 0.74
rs3176689 A T 267 118 355 116 51 184 0.90 (0.72, 1.12) 0.34 0.91 (0.65, 1.26) 0.57
rs3176683 T C 342 152 476 41 17 63 0.85 (0.61, 1.18) 0.33 0.80 (0.48, 1.33) 0.40
rs3176658 C T 294 137 405 89 32 134 0.93 (0.73, 1.18) 0.55 0.72 (0.49, 1.07) 0.10
rs1800975 G A 184 87 236 185 77 280 0.88 (0.71, 1.08) 0.22 0.78 (0.57, 1.07) 0.12
RAD23B rs1805330 C T 306 136 458 77 33 81 1.34 (1.04, 1.72) 0.03 1.28 (0.87, 1.88) 0.21
rs1805329 C T 257 113 333 126 56 206 0.81 (0.65, 1.00) 0.05 0.83 (0.60, 1.14) 0.25
ERCC6 rs2228529 A G 252 108 345 127 59 186 0.94 (0.76, 1.17) 0.58 1.03 (0.75, 1.42) 0.86
rs2228527 A G 251 107 347 132 62 192 0.96 (0.78, 1.19) 0.70 1.06 (0.78, 1.46) 0.70
rs4253132 T C 294 132 429 89 37 110 1.08 (0.85, 1.37) 0.55 1.01 (0.69, 1.46) 0.98
rs2228528 G A 265 115 372 117 53 167 1.05 (0.84, 1.31) 0.67 1.09 (0.78, 1.51) 0.61
DDB2 (XPE) rs2029298 A G 170 68 255 213 101 284 1.05 (0.86, 1.29) 0.63 1.31 (0.96, 1.79) 0.09
rs4647709 C T 317 137 449 66 32 90 1.06 (0.81, 1.39) 0.68 1.21 (0.82, 1.79) 0.34
rs2291120 T C 283 126 402 100 43 137 1.11 (0.88, 1.39) 0.39 1.06 (0.75, 1.50) 0.73
rs1685404 G C 172 71 246 211 98 293 1.08 (0.88, 1.32) 0.48 1.20 (0.88, 1.64) 0.24
rs2957873 A G 257 109 386 126 60 153 1.11 (0.90, 1.38) 0.33 1.29 (0.94, 1.78) 0.11
rs326224 G A 279 119 404 104 50 135 1.05 (0.84, 1.32) 0.65 1.24 (0.89, 1.74) 0.20
rs2306353 G A 281 121 415 102 48 124 1.10 (0.87, 1.38) 0.43 1.25 (0.89, 1.75) 0.20
rs326222 C T 190 81 294 193 88 245 1.09 (0.89, 1.34) 0.39 1.21 (0.89, 1.64) 0.22
rs901746 A G 190 81 295 193 88 244 1.10 (0.90, 1.34) 0.37 1.22 (0.89 1.65) 0.21
ERCC5 (XPG) rs2296147 T C 130 45 150 250 123 387 0.78 (0.62, 0.97) 0.02 1.11 (0.79, 1.58) 0.54
rs4771436 T G 227 99 336 156 70 203 1.09 (0.89, 1.35) 0.39 1.13 (0.83, 1.55) 0.43
rs1047768 C T 122 62 197 261 107 342 1.20 (0.97, 1.50) 0.10 0.96 (0.70, 1.32) 0.82
rs3818356 C T 227 99 336 155 69 202 1.09 (0.89, 1.34) 0.41 1.12 (0.82, 1.53) 0.46
rs4150351 A C 254 114 341 129 55 198 0.84 (0.68, 1.04) 0.12 0.81 (0.59, 1.13) 0.22
rs4150355 C T 177 66 225 206 103 314 0.86 (0.70, 1.05) 0.14 1.15 (0.84, 1.57) 0.37
rs4150360 T C 106 55 169 277 114 370 1.18 (0.94, 1.48) 0.15 0.93 (0.67, 1.29) 0.67
rs4150383 G A 255 112 375 128 57 164 1.11 (0.90, 1.38) 0.34 1.13 (0.82, 1.56) 0.47
rs4150386 A C 304 133 420 79 36 119 1.03 (0.80, 1.33) 0.81 1.05 (0.72, 1.53) 0.80
rs17655 C G 223 112 332 160 57 207 1.15 (0.94, 1.41) 0.18 0.81 (0.59, 1.11) 0.19
rs873601 A G 190 93 274 193 76 265 1.05 (0.86, 1.28) 0.65 0.84 (0.62, 1.14) 0.25
rs4150393 A G 296 132 406 87 37 133 0.89 (0.70, 1.13) 0.34 0.88 (0.61, 1.26) 0.48
rs876430 C T 191 94 274 192 75 265 1.04 (0.85, 1.27) 0.72 0.82 (0.60, 1.11) 0.19
rs1051677 T C 303 129 432 80 40 106 1.06 (0.83, 1.36) 0.63 1.28 (0.89, 1.82) 0.18
rs1051685 A G 302 141 434 81 28 104 1.04 (0.81, 1.34) 0.74 0.79 (0.52, 1.19) 0.26
ERCC4 (XPF) rs3136038 C T 184 86 218 199 83 321 0.79 (0.65, 0.97) 0.03 0.69 (0.51, 0.93) 0.02
rs1799798 G A 320 143 437 63 26 102 0.90 (0.69, 1.19) 0.47 0.83 (0.55, 1.27) 0.40
rs744154 C G 208 96 272 175 73 267 0.88 (0.72, 1.08) 0.21 0.77 (0.57, 1.05) 0.10
rs3136085 G C 205 96 270 178 73 269 0.88 (0.72, 1.08) 0.23 0.75 (0.55, 1.02) 0.06
rs3136130 G T 184 86 216 199 83 323 0.78 (0.64, 0.96) 0.02 0.68 (0.50, 0.92) 0.01
rs1800067 G A 322 144 456 61 25 83 0.99 (0.75, 1.31) 0.96 0.89 (0.58, 1.37) 0.61
rs3136172 A G 195 91 263 188 78 276 0.92 (0.75, 1.13) 0.43 0.79 (0.58, 1.08) 0.14
RAD23A rs2974752 A G 135 58 198 235 106 326 1.04 (0.84, 1.29) 0.72 1.10 (0.80, 1.52) 0.57
ERCC2 (XPD) rs13181 T G 154 73 227 224 94 310 1.07 (0.87, 1.31) 0.53 0.94 (0.69, 1.28) 0.69
rs238418 C A 156 74 226 227 95 313 1.05 (0.86, 1.29) 0.63 0.92 (0.68, 1.25) 0.59
rs1799787 C T 196 87 276 187 82 263 1.02 (0.83, 1.25) 0.85 1.00 (0.74, 1.35) 0.99
rs3916874 G C 209 96 268 174 73 271 0.85 (0.69, 1.04) 0.12 0.75 (0.55, 1.02) 0.07
rs238416 G A 152 68 217 231 101 321 1.00 (0.81, 1.23) 1.00 0.99 (0.73, 1.35) 0.96
rs50872 C T 226 91 305 157 78 232 0.95 (0.77, 1.16) 0.59 1.16 (0.86, 1.58) 0.34
rs50871 T G 110 55 132 273 114 407 0.80 (0.64, 1.00) 0.05 0.67 (0.48, 0.92) 0.01
rs238407 A T 121 54 142 262 115 396 0.81 (0.65, 1.01) 0.06 0.83 (0.60, 1.15) 0.27
rs3810366 C G 79 31 99 304 138 439 0.89 (0.69, 1.15) 0.36 1.09 (0.74, 1.62) 0.66
ERCC1 rs735482 A C 278 131 410 105 38 129 1.13 (0.90, 1.42) 0.28 0.86 (0.60, 1.23) 0.40
rs2336219 G A 278 131 410 105 38 129 1.13 (0.90, 1.42) 0.28 0.86 (0.60, 1.23) 0.40
rs3212964 G A 280 132 412 103 37 127 1.13 (0.90, 1.41) 0.31 0.84 (0.58, 1.21) 0.34
rs3212955 A G 209 94 319 174 75 220 1.15 (0.94, 1.41) 0.18 1.10 (0.81, 1.50) 0.54
rs3212948 C G 154 73 228 229 96 311 1.06 (0.86, 1.30) 0.58 0.93 (0.68, 1.26) 0.64
rs3212930 T C 241 108 335 142 61 204 0.95 (0.77, 1.18) 0.66 0.93 (0.68, 1.27) 0.64
LIG1 rs156641 G A 151 74 219 232 95 320 1.02 (0.83, 1.26) 0.84 0.86 (0.63, 1.17) 0.33
rs20580 C A 97 46 140 286 123 399 1.03 (0.82, 1.30) 0.78 0.94 (0.66, 1.32) 0.70
rs20579 C T 294 127 397 89 42 142 0.90 (0.71, 1.15) 0.41 1.01 (0.71, 1.44) 0.94
Open in a new tab

Confidence intervals presented not corrected for multiple comparisons. Significant associations using a dominant genetic model (p <>0.05) highlighted in bold. None significant at a Bonferroni-corrected level (p <0.0006)

HR hazards ratio, CI confidence interval

a

HR for dominant genetic model (AB + BB vs. AA). HRs adjusted for age and sex (including pairwise interaction) and ancestry (proportion African ancestry)

Fig. 1.

Fig. 1

Open in a new tab

Kaplan–Meier plots for overall (OS) and disease-specific (DS) survival by select genotype among White HNC cases in the Carolina Head and Neck Cancer Epidemiology (CHANCE) study. Solid line represents individuals with referent genotype, while dashed line represents individuals with variant genotype

Among African Americans, two SNPs were associated with overall survival and four SNPs were associated with disease-specific survival at an uncorrected 0.05 alpha level, but none were significantly associated with survival at Bonferroni-corrected or FDR-corrected levels (Table 4). Only one tag SNP was associated with both overall and disease-specific survival at an uncorrected 0.05 alpha level. Specifically, rs2607755 (CC + CT vs. TT) on XPC was suggestively associated with reduced hazards of overall (HR = 0.62, uncorrected 95 % CI = 0.45, 0.86) and disease-specific death (HR = 0.51, uncorrected 95 % CI = 0.30, 0.86). This association was strongest among cases with stage 4 tumors (Online Resource 2). Figure 2 shows Kaplan–Meier plots for this SNP.

Table 4.

Hazard ratios for single nucleotide polymorphisms (SNPs) in nucleotide excision repair (NER) genes and survival among head and neck cancer cases in the Carolina Head and Neck Cancer Epidemiology (CHANCE) study, African Americans

Gene SNP Coded allele
Overall deaths/deaths from HNC/alive
Overall survival
Disease-specific survival
Referent (A) Variant (B) AA AB + BB HR (95 % CI)a p value HR (95 % CI)a p value
ERCC3 (XPB) rs4150496 G A 95 30 83 66 28 59 0.98 (0.71, 1.36) 0.93 1.40 (0.82, 2.38) 0.22
rs4150459 G A 101 35 89 61 23 54 1.08 (0.78, 1.50) 0.65 1.14 (0.66, 1.98) 0.63
rs1011019 C T 94 34 89 68 24 54 1.06 (0.77, 1.47) 0.72 1.08 (0.62, 1.86) 0.78
rs4150434 G A 126 46 106 36 12 37 0.84 (0.57, 1.24) 0.38 0.79 (0.41, 1.55) 0.50
rs4150416 T G 39 13 46 122 44 97 1.32 (0.91, 1.90) 0.14 1.47 (0.78, 2.80) 0.23
rs4150407 A G 43 13 41 119 45 102 1.13 (0.79, 1.62) 0.49 1.35 (0.72, 2.53) 0.35
rs4150402 G A 94 34 89 68 24 54 1.06 (0.77, 1.47) 0.72 1.08 (0.62, 1.86) 0.78
XPC rs2228001 A C 91 32 89 71 26 54 1.18 (0.85, 1.63) 0.32 1.36 (0.79, 2.32) 0.26
rs2228000 C T 137 50 114 25 8 29 0.77 (0.50, 1.21) 0.26 0.65 (0.30, 1.42) 0.28
rs3731124 A C 136 49 116 26 9 27 0.91 (0.60, 1.40) 0.67 0.75 (0.37, 1.53) 0.43
rs3731093 T C 140 54 123 19 3 19 0.86 (0.51, 1.45) 0.57 0.44 (0.14, 1.43) 0.17
rs3731089 G A 140 54 123 22 4 20 0.91 (0.56, 1.48) 0.70 0.52 (0.19, 1.47) 0.22
rs2733537 A G 115 46 97 47 12 46 0.86 (0.60, 1.23) 0.40 0.59 (0.30, 1.13) 0.11
rs2607755 T C 70 27 41 92 31 102 0.62 (0.45, 0.86) 0.004 0.51 (0.30, 0.86) 0.01
rs1902658 G A 30 10 23 132 48 120 0.94 (0.62, 1.43) 0.78 0.87 (0.43, 1.76) 0.71
ERCC8 rs3117 T C 72 21 54 90 37 89 0.81 (0.59, 1.12) 0.20 1.15 (0.66, 1.99) 0.62
CDK7 rs2972388 A G 77 30 83 85 28 60 1.36 (0.98, 1.87) 0.06 1.04 (0.61, 1.77) 0.88
CCNH rs2266691 A G 140 52 117 22 6 26 0.83 (0.53, 1.32) 0.44 0.54 (0.23, 1.28) 0.16
rs2266692 G T 133 46 104 29 12 39 0.68 (0.45, 1.03) 0.07 0.80 (0.41, 1.53) 0.49
XPA rs3176757 C T 126 42 110 36 16 33 1.00 (0.68, 1.46) 0.98 1.42 (0.78, 2.59) 0.25
rs3176753 T C 123 40 102 39 18 41 0.89 (0.62, 1.29) 0.54 1.17 (0.66, 2.08) 0.58
rs3176748 A G 132 53 118 30 5 25 1.03 (0.68, 1.56) 0.88 0.40 (0.16, 1.02) 0.05
rs3176658 C T 141 51 119 21 7 24 0.89 (0.55, 1.44) 0.64 0.88 (0.38, 2.00) 0.76
rs1800975 G A 100 32 87 55 54 51 1.05 (0.74, 1.47) 0.80 1.44 (0.82, 2.55) 0.21
RAD23B rs1805330 C T 93 32 90 69 26 53 1.05 (0.76, 1.47) 0.76 1.15 (0.67, 1.99) 0.61
ERCC6 rs2228529 A G 123 48 111 39 10 30 1.17 (0.81, 1.69) 0.41 0.74 (0.37, 1.49) 0.40
rs2228527 A G 115 44 103 47 14 40 1.04 (0.74, 1.48) 0.82 0.81 (0.44, 1.50) 0.51
rs4253132 T C 97 35 94 65 23 49 1.13 (0.82, 1.56) 0.46 1.17 (0.69, 1.99) 0.57
rs2228528 G A 107 40 105 54 18 38 1.46 (1.04, 2.04) 0.03 1.22 (0.69, 2.14) 0.50
DDB2 (XPE) rs2029298 A G 43 16 47 119 42 96 1.23 (0.86, 1.76) 0.26 1.27 (0.71, 2.30) 0.42
rs1685404 G C 86 31 78 76 27 65 1.07 (0.77, 1.47) 0.70 0.97 (0.56, 1.66) 0.90
rs2957873 A G 49 20 41 112 38 102 0.95 (0.67, 1.35) 0.78 0.79 (0.46, 1.38) 0.41
rs326224 G A 43 14 37 119 44 106 1.12 (0.78, 1.60) 0.54 1.21 (0.66, 2.22) 0.55
rs2306353 G A 60 21 46 102 37 97 0.88 (0.64, 1.22) 0.45 0.93 (0.54, 1.59) 0.78
rs326222 C T 26 12 28 136 46 115 1.17 (0.76, 1.80) 0.47 0.88 (0.46, 1.68) 0.70
rs901746 A G 35 15 30 127 43 113 0.96 (0.66, 1.41) 0.85 0.79 (0.43, 1.44) 0.44
ERCC5 (XPG) rs2296147 T C 101 32 92 60 25 51 1.07 (0.77, 1.48) 0.70 1.42 (0.84, 2.42) 0.19
rs2296148 C T 122 45 106 39 13 37 0.96 (0.66, 1.39) 0.81 0.93 (0.49, 1.75) 0.82
rs4771436 T G 110 35 94 52 23 49 0.90 (0.64, 1.26) 0.54 1.24 (0.73, 2.12) 0.43
rs1047768 C T 66 21 50 96 37 93 0.87 (0.63, 1.20) 0.39 1.06 (0.61, 1.84) 0.83
rs2020915 G A 103 42 102 59 16 41 1.22 (0.88, 1.70) 0.24 0.82 (0.45, 1.49) 0.52
rs3818356 C T 110 35 96 52 23 47 0.93 (0.66, 1.31) 0.68 1.27 (0.75, 2.18) 0.37
rs4150355 C T 119 39 99 43 19 44 0.93 (0.64, 1.34) 0.68 1.16 (0.65, 2.05) 0.62
rs4150360 T C 11 5 8 151 53 135 0.79 (0.42, 1.50) 0.48 0.65 (0.25, 1.70) 0.38
rs4150383 G A 131 42 111 31 16 32 0.88 (0.59, 1.31) 0.53 1.34 (0.75, 2.39) 0.33
rs17655 C G 44 19 46 118 39 97 1.09 (0.76, 1.56) 0.65 0.84 (0.47, 1.50) 0.56
rs873601 A G 14 8 16 148 50 127 1.16 (0.66, 2.02) 0.61 0.76 (0.35, 1.62) 0.47
rs876430 C T 15 8 17 147 50 126 1.17 (0.68, 2.00) 0.58 0.84 (0.39, 1.79) 0.65
rs1051677 T C 116 43 116 46 15 27 1.31 (0.92, 1.86) 0.14 1.15 (0.63, 2.11) 0.64
rs1051685 A G 77 30 61 85 28 82 0.85 (0.62, 1.15) 0.29 0.72 (0.43, 1.20) 0.20
ERCC4 (XPF) rs3136038 C T 43 17 49 119 41 94 1.26 (0.88, 1.81) 0.20 1.13 (0.64, 2.01) 0.67
rs744154 C G 114 40 107 48 18 36 1.16 (0.82, 1.64) 0.42 1.27 (0.72, 2.26) 0.41
rs3136085 G C 92 34 81 70 24 62 1.06 (0.76, 1.46) 0.75 1.01 (0.59, 1.75) 0.96
rs3136091 C G 131 45 124 31 13 19 1.27 (0.84, 1.90) 0.25 1.55 (0.82, 2.94) 0.18
rs3136130 G T 39 16 36 123 42 107 1.04 (0.72, 1.51) 0.84 0.90 (0.50, 1.63) 0.74
rs3136172 A G 112 39 104 50 19 39 1.12 (0.80, 1.59) 0.51 1.24 (0.70, 2.19) 0.46
rs2020955 T C 101 37 92 61 21 51 1.11 (0.80, 1.55) 0.53 1.05 (0.60, 1.83) 0.86
RAD23A rs2974752 A G 39 12 38 116 42 100 1.11 (0.77, 1.62) 0.57 1.29 (0.67, 2.48) 0.45
rs11558955 A G 135 47 121 27 11 22 0.99 (0.65, 1.51) 0.97 1.14 (0.58, 2.22) 0.71
ERCC2 (XPD) rs13181 T G 86 31 87 76 27 55 1.24 (0.90, 1.70) 0.19 1.18 (0.70, 2.00) 0.54
rs238418 C A 5 4 3 157 54 139 0.87 (0.34, 2.27) 0.78 0.27 (0.09, 0.82) 0.02
rs1799787 C T 120 40 115 42 18 28 1.36 (0.95, 1.95) 0.09 1.65 (0.93, 2.91) 0.09
rs3916874 G C 142 53 126 20 5 17 0.94 (0.58, 1.54) 0.82 0.59 (0.23, 1.52) 0.27
rs238416 G A 129 44 114 32 14 29 1.04 (0.70, 1.55) 0.83 1.33 (0.72, 2.45) 0.36
rs50872 C T 116 42 107 46 16 36 1.08 (0.76, 1.54) 0.66 0.98 (0.54, 1.76) 0.94
rs50871 T G 120 36 109 42 22 34 1.07 (0.74, 1.54) 0.73 1.93 (1.11, 3.35) 0.02
rs238407 A T 114 39 112 48 19 31 1.22 (0.86, 1.72) 0.27 1.36 (0.77, 2.38) 0.29
rs3810366 C G 108 39 104 54 19 39 1.13 (0.81, 1.59) 0.47 1.04 (0.59, 1.83) 0.89
ERCC1 rs735482 A C 83 31 73 79 27 70 0.94 (0.69, 1.29) 0.72 0.86 (0.51, 1.46) 0.58
rs2336219 G A 84 31 73 78 27 70 0.93 (0.67, 1.27) 0.63 0.87 (0.51, 1.47) 0.59
rs3212964 G A 111 43 96 49 15 47 0.98 (0.69, 1.38) 0.89 0.76 (0.41, 1.39) 0.37
rs3212955 A G 82 31 77 80 27 66 0.92 (0.67, 1.27) 0.62 0.92 (0.54, 1.55) 0.75
rs3212948 C G 5 3 4 157 55 139 0.64 (0.24, 1.73) 0.38 0.36 (0.09, 1.37) 0.13
rs3212935 A G 81 32 62 81 26 81 0.84 (0.61, 1.16) 0.28 0.65 (0.38, 1.10) 0.11
rs3212930 T C 131 44 119 31 14 24 1.09 (0.73, 1.62) 0.68 1.43 (0.77, 2.64) 0.26
LIG1 rs156641 G A 129 43 105 33 15 38 0.76 (0.51, 1.14) 0.19 1.13 (0.61, 2.09) 0.70
rs20580 C A 33 8 29 129 50 113 1.00 (0.68, 1.48) 0.99 1.62 (0.76, 3.45) 0.21
rs20579 C T 83 26 67 79 32 76 0.91 (0.66, 1.26) 0.58 1.32 (0.77, 2.25) 0.31
rs439132 A G 88 27 84 74 31 59 1.10 (0.79, 1.52) 0.57 1.48 (0.86, 2.55) 0.15
Open in a new tab

Confidence intervals presented not corrected for multiple comparisons. Significant associations using a dominant genetic model (p <>0.05) highlighted in bold. None significant at a Bonferroni-corrected level (p <>0.0006)

HR hazards ratio, CI confidence interval

a

HR for dominant genetic model (AB + BB vs. AA). HRs adjusted for age and sex (including pairwise interaction) and ancestry (proportion African ancestry)

Fig. 2.

Fig. 2

Open in a new tab

Kaplan–Meier plots for overall (OS) and disease-specific (DS) survival by select genotypes among African American HNC cases in the Carolina Head and Neck Cancer Epidemiology (CHANCE) study. Solid line represents individuals with referent genotype, while dashed line represents individuals with variant genotype

Joint effects

At an uncorrected 0.05 alpha level, four SNPs appeared to interact super-additively with radiation only, six SNPs appeared to interact super-additively with radiation and chemotherapy, and one SNP appeared to interact sub-additively with surgery, radiation, and chemotherapy, with respect to overall survival among Whites (Online Resource 3). Of these suggestive interactions, one SNP-radiation and two SNP-radiation, chemotherapy interactions were significant at a Bonferroni-corrected 0.0006 alpha level. Specifically, rs2972388 of cyclin-dependent kinase 7 (CDK7) interacted super-additively with radiation only (RERI = 1.07, uncorrected 95 % CI = 0.55, 1.60) and with radiation and chemotherapy (RERI = 0.72, uncorrected 95 % CI = 0.33, 1.10). In addition, rs2974752 of RAD23 homolog A (RAD23A) interacted super-additively with radiation and chemotherapy (RERI = 0.80, uncorrected 95 % CI = 0.36, 1.24). However, when disease-specific survival was considered, no SNP-treatment interactions were significant at a Bonferroni or FDR level among Whites (data not shown). Among African Americans, no SNP-treatment interactions appeared to be significant at a Bonferroni-corrected alpha level with respect to overall survival, though a super-additive interaction between rs1902658 of XPC and radiation, chemotherapy was significant when FDR was considered (RERI = 0.75, uncorrected 95 % CI = 0.29, 1.21). However, interaction estimates among African Americans were considered unreliable due to relatively low cell counts and are therefore not presented. With platinum-based chemotherapy, 10 SNPs suggested additive interactions at an uncorrected alpha level with respect to overall survival among Whites, but none were significant after correction for multiple comparisons (Online Resource 4).

Discussion

We detected mostly null associations between 84 SNPs in 15 NER genes and survival among White and African American HNC cases. Identifying null associations is important for following-up early positive associations, avoiding publication bias, and informing future meta-analyses [34]. To account for multiple comparisons, we principally used the Bonferroni approach, which though widely used in genetic epidemiology assumes independence of tests [25, 35, 36]. Given the correlated nature of SNPs, including some SNPs in our study, using the Bonferroni correction may be overly conservative potentially resulting in false negatives [25, 35, 36]. Therefore, we also considered the FDR approach as well as highlighted SNP-survival associations with p values below an uncorrected 0.05 alpha level as suggestive associations warranting further investigation.

Among Whites, we found that rs3136038 (near the 5’ end) and rs3136130 (intron 5) of ERCC4 were suggestively associated with improved overall and disease-specific survival [37, 38]. These SNPs are in LD with each other as well as several other untyped SNPs near or in introns or the 3′UTR of ERCC4 (r2 >0.80, CEU population) [33]. rs50871 of ERCC2 intron 11 was also suggestively associated with improved survival among Whites [37, 38]. Among African Americans, rs2607755, which is located in intron 2 of XPC and is in LD with other intronic SNPs and the missense SNP rs2227998 (Arg687Ser, r2 = 0.86, YRI population), was suggestively associated with improved overall and disease-specific survival [33, 37, 38]. The ERCC4 enzyme helps create an incision at the 5′ end of DNA adducts, while ERCC2 operates as a component of the transcription factor II H (TFIIH) subunit to denature the double helix in preparation for incisions [39, 40]. The XPC enzyme acts early in the NER pathway to recognize and bind with DNA adducts [39, 40]. Assuming that it is minor alleles which mitigate functional NER effects, thereby facilitating cancer treatment effects, one may expect variant genotypes of intronic SNPs, especially SNPs in regulatory regions or in LD with SNPs in coding regions, to be associated with improved survival, as was suggested by our findings for rs3136038, rs3136130, and rs50871 among Whites and rs2227998 among African Americans.

Although no previous HNC studies examined rs3136038 or rs3136130 on ERCC4, two studies assessed nine other ERCC4 SNPs (rs1799799, rs1799801, rs3136105, rs3136146, rs3136152, rs3136155, rs3136166, rs3136189, rs3136202), many of which were in LD with the SNPs in our study (r2 >0.80, CEU population) [10, 16, 33]. While five of these SNPs were not associated with progression-free survival among HNC cases, four SNPs appeared to be associated with worse progression-free survival contrary to our study [10, 16]. Among esophageal cancer cases, a study by Lee et al. [41] did assess rs3136038 reporting better overall survival associated with the genotype TT, though HRs were not statistically significant, similar to our study. Further, ERCC4 protein expression has been found to be elevated in HNC cell lines and displayed cisplatin resistance [42]. With respect to ERCC2, no previous studies have considered the effects of rs50871 on HNC survival. Rather, rs13181 and rs1799793 (which are not in LD with rs50871, CEU population) are the most commonly studied SNPs in ERCC2, with some studies reporting near null associations between these SNPs and survival among HNC cases [9, 10, 12, 14, 15, 33]. In our study, rs13181 was not associated with survival. Finally, no previous studies have considered associations between rs2607755 of XPC and survival, nor have any studies considered association between any NER variants and survival among African American HNC cases. Only one previous study has investigated a single variant in XPC, rs2228001 (which is not in LD with rs2607755, YRI population), noting no association with overall survival [9, 33]. Likewise, we did not find an association between rs2228001 and survival.

Since radiation and platinum-based chemotherapy are known to cause DNA damage repaired by NER genes [7, 9], we also considered associations between SNPs and survival among HNC cases in the context of treatment. Accounting for multiple comparisons using the Bonferroni method, we found that interactions between rs2972388, a synonymous SNP in CDK7, and radiation only, as well as radiation and chemotherapy, were more than additive with respect to overall survival among Whites [37, 38]. In addition, rs2974752, located near RAD23A and in LD with other SNPs near or in introns of this gene (r2 >0.80, CEU population), interacted super-additively with radiation and chemotherapy [33, 37, 38]. However, these SNP-treatment interactions were not significant at a Bonferroni level when disease-specific survival was considered and should therefore be interpreted with some caution. Further, genotype frequencies for rs2972388 and rs2974752 were consistent with HWE at a Bonferroni-corrected 0.0006 alpha level, but not at a 0.05 alpha level. 0.05 alpha level. No previous studies have considered CDK7 or RAD23A SNPs in relation to treatment and HNC survival. Only one previous study has compared NER SNP-survival associations across strata of treatment regimens [17]. Specifically, Zhong et al. [17] analyzed the effect of rs13181 in ERCC2 on survival among 275 HNC cases receiving radiotherapy and 210 cases not receiving radiotherapy. Among cases with stage 3 and 4 tumors, the referent genotype (AA) was associated with poorer overall survival among those treated with radiation, but better survival among those who did not receive radiation [17]. Among cases with stage 1 and 2 tumors who did not receive radiation, rs13181 was not associated with survival [17].

With a population-based study of 1,227 HNC cases, the present analysis included more than double the number of HNC cases of the next largest study [10]. Study populations of previous publications were mostly hospital-based and ranged from 47 to 531 HNC cases [10, 13]. Further, the present study population included 922 White cases and 305 African Americans cases which allowed for estimation of race-specific HRs. Linkage disequilibrium is known to vary by ancestral populations and distinct differences in survival by race occur in the United States [2, 3, 33]. Yet, prior to this study, no studies had considered NER SNP-survival associations among African American HNC cases. Another contribution of our study was the broad evaluation of NER variants which included a large number of SNPs that have not been previously evaluated. Previous studies have collectively examined approximately 18 SNPs in six NER genes and survival among HNC cases [9–17]. Our study included 84 SNPs across 15 NER genes.

Although our study included the largest study population and broadest array of NER SNPs to date to our knowledge, a few limitations should be noted. We were unable to include proxy interviews (52 cases, 3.7 %) in our analysis since these occurred for individuals who died prior to interview and therefore did not provide a biologic sample. If SNPs were related to aggressive tumors, then estimates for SNP-survival associations may be slightly attenuated to the null [21]. When SNP-survival associations were stratified by stage, associations were strongest among cases with stage 3 and 4 tumors. Further, follow-up was started at date of diagnosis, rather than date of interview (i.e., date of blood draw), since diagnosis is a more clinically informative time point. To assess the potential immortal person-time bias this may have introduced [43], we conducted sensitivity analyses with follow-up starting at date of interview. No material differences in results were noted across models, though rs50871 was suggestively associated with overall survival among Whites with an unadjusted p value of 0.05 in the primary analysis, but with a p value of 0.06 in the sensitivity analysis, and the interaction between rs1902658 and radiation, chemotherapy with respect to overall survival among African Americans was significant at an FDR level in the primary analysis, but at a Bonferroni level in the sensitivity analysis.

Other potential limitations include the following. First, because both tagging and candidate SNPs were included and selected based on European-descent populations, variation captured across some genes was limited, especially among African Americans. For this reason, haplotypes were not explored. Second, treatment was considered solely as the first-course combinations of dichotomous variables for surgery, radiation, and chemotherapy abstracted from medical records. Although treatment is fairly standardized based on tumor stage and site as well as other patient demographics [4, 5], information on duration of treatment (e.g., start and end dates) and timing of treatments combinations (e.g., concurrent chemotherapy) may have been informative. Third, treatment-SNP joint effect estimates were imprecise among Whites and unreliable among African Americans due to small cell counts. Fourth, models were adjusted for cigarette and alcohol information that was ascertained at baseline based on behaviors prior to diagnosis since information on behavioral risk factors following diagnosis was not uniformly available. Further, we did not have information on human papillomavirus infection, a known predictor of survival among cases with oropharyngeal tumors [44]. Finally, we did not have access to information on recurrent tumors and therefore did not consider disease-free or relapse-free survival.

In summary, most NER variants did not appear to be associated with survival among HNC cases. However, three SNPs in Whites (rs3136038 and rs3136130 of ERCC4 and rs50871 of ERCC2) and one SNP among African Americans (rs2607755 of XPC) were suggestively associated with better overall and disease-specific survival. Therefore, it is recommended that future genetic epidemiology studies of HNC survival include these SNPs for replication. In addition, two SNPs appeared to possibly interact additively with radiation with or without chemotherapy among Whites. While our study is the largest to date, it is only the second to consider NER SNP-treatment joint effects on HNC survival. Therefore, additional studies with even larger sample sizes are needed to evaluate gene–environment interactions more precisely. Further studies focusing on African American and other diverse populations are recommended.

Supplementary Material

Supplementary

Acknowledgments

This work was supported by the US National Institutes of Health (NIH), National Cancer Institute (NCI) [R01-CA90731-01; 2T32CA009330-21-26], and National Institute of Environmental Health Sciences (NIEHS) [T32ES007018; P30ES010126]. Dr. Avery was supported in part by grant R00-HL-098458 from the National Heart, Lung, and Blood Institute (NHBLI). The authors thank Dr. Robert Millikan for his substantial contributions to the conceptual development and analyses of this paper. We also thank Dr. Anne Hakenewerth for assistance in research development and Ms. Kathy Wisniewski for programming support.

Footnotes

Conflict of interest

The authors declare that they have no conflict of interest.

Electronic supplementary material The online version of this article (doi:10.1007/s10552-014-0346-x) contains supplementary material, which is available to authorized users.

Contributor Information

Annah B. Wyss, Email: alayman@email.unc.edu, Department of Epidemiology, Gillings School of Global Pubic Health, University of North Carolina at Chapel Hill, 2101B McGavran-Greenberg Hall, CB 7435, Chapel Hill, NC 27599, USA

Mark C. Weissler, Department of Otolaryngology/Head and Neck Surgery, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Christy L. Avery, Department of Epidemiology, Gillings School of Global Pubic Health, University of North Carolina at Chapel Hill, 2101B McGavran-Greenberg Hall, CB 7435, Chapel Hill, NC 27599, USA

Amy H. Herring, Department of Biostatistics, Gillings School of Global Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, Carolina Population Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Jeannette T. Bensen, Department of Epidemiology, Gillings School of Global Pubic Health, University of North Carolina at Chapel Hill, 2101B McGavran-Greenberg Hall, CB 7435, Chapel Hill, NC 27599, USA, Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Jill S. Barnholtz-Sloan, Case Comprehensive Cancer Center, Case Western Reserve University, Cleveland, OH, USA

William K. Funkhouser, Department of Pathology and Laboratory Medicine, School of Medicine, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

Andrew F. Olshan, Email: andy_olshan@unc.edu, Department of Epidemiology, Gillings School of Global Pubic Health, University of North Carolina at Chapel Hill, 2101B McGavran-Greenberg Hall, CB 7435, Chapel Hill, NC 27599, USA, Carolina Population Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA

References

  • 1.American Cancer Society. [Accessed on 9 Feb 2013];Cancer facts and figures 2013. 2013 http://www.cancer.org/acs/groups/content/@epidemiologysurveilance/documents/document/acspc-031941.pdf.
  • 2.National Cancer Institute. [Accessed on 9 Dec 2013];Relative survival by survival time by race/ethnicity oral cavity and pharynz, all age, male, 1992–2009. 2013 http://seer.cancer.gov/faststats/selections.php?#Output.
  • 3.National Cancer Institute. [Accessed on 9 Dec 2013];Relative survival by survival time by race/ethnicity oral cavity and pharynz, all age, female, 1992–2009. 2013 http://seer.cancer.gov/faststats/selections.php?#Output.
  • 4.Vokes EE, Weichselbaum RR, Lippman SM, Hong WK. Head and neck cancer. N Engl J Med. 1993;328(3):184–194. doi: 10.1056/NEJM199301213280306. [DOI] [PubMed] [Google Scholar]
  • 5.Forastiere A, Koch W, Trotti A, Sidransky D. Head and neck cancer. N Engl J Med. 2001;345(26):1890–1900. doi: 10.1056/NEJMra001375. [DOI] [PubMed] [Google Scholar]
  • 6.National Cancer Institute. [Accessed on 10 Dec 2013];General information about lip and oral cavity cancer. 2013 http://www.cancer.gov/cancertopics/pdq/treatment/lip-and-oral-cavity/Patient.
  • 7.Gazdar AF. DNA repair and survival in lung cancer—the two faces of Janus. N Engl J Med. 2007;356(8):771–773. doi: 10.1056/NEJMp068308. [DOI] [PubMed] [Google Scholar]
  • 8.Braakhuis BJ, Brakenhoff RH, Leemans CR. Treatment choice for locally advanced head and neck cancers on the basis of risk factors: biological risk factors. Ann Oncol. 2012;23(Suppl 10):173–177. doi: 10.1093/annonc/mds299. [DOI] [PubMed] [Google Scholar]
  • 9.Carles J, Monzo M, Amat M, Jansa S, Artells R, Navarro A, et al. Single-nucleotide polymorphisms in base excision repair, nucleotide excision repair, and double strand break genes as markers for response to radiotherapy in patients with Stage I to II head-and-neck cancer. Int J Radiat Oncol Biol Phys. 2006;66(4):1022–1030. doi: 10.1016/j.ijrobp.2006.06.029. [DOI] [PubMed] [Google Scholar]
  • 10.Azad AK, Bairati I, Samson E, Cheng D, Mirshams M, Qiu X, et al. Validation of genetic sequence variants as prognostic factors in early-stage head and neck squamous cell cancer survival. Clin Cancer Res. 2012;18(1):196–206. doi: 10.1158/1078-0432.CCR-11-1759. [DOI] [PubMed] [Google Scholar]
  • 11.De Castro G, Jr, Pasini FS, Siqueira SA, Ferraz AR, Villar RC, Snitcovsky IM, et al. ERCC1 protein, mRNA expression and T19007C polymorphism as prognostic markers in head and neck squamous cell carcinoma patients treated with surgery and adjuvant cisplatin-based chemoradiation. Oncol Rep. 2011;25(3):693–699. doi: 10.3892/or.2011.1133. [DOI] [PubMed] [Google Scholar]
  • 12.Gal TJ, Huang WY, Chen C, Hayes RB, Schwartz SM. DNA repair gene polymorphisms and risk of second primary neoplasms and mortality in oral cancer patients. Laryngoscope. 2005;115(12):2221–2231. doi: 10.1097/01.mlg.0000183736.96004.f7. [DOI] [PubMed] [Google Scholar]
  • 13.Grau JJ, Caballero M, Campayo M, Jansa S, Vargas M, Alos L, et al. Gene single nucleotide polymorphism accumulation improves survival in advanced head and neck cancer patients treated with weekly paclitaxel. Laryngoscope. 2009;119(8):1484–1490. doi: 10.1002/lary.20254. [DOI] [PubMed] [Google Scholar]
  • 14.Mahimkar MB, Samant TA, Kannan S, Tulsulkar J, Pai PS, Anantharaman D. Polymorphisms in GSTM1 and XPD genes predict clinical outcome in advanced oral cancer patients treated with postoperative radiotherapy. Mol Carcinog. 2012;51(Suppl 1):E94–E103. doi: 10.1002/mc.21868. [DOI] [PubMed] [Google Scholar]
  • 15.Quintela-Fandino M, Hitt R, Medina PP, Gamarra S, Manso L, Cortes-Funes H, et al. DNA-repair gene polymorphisms predict favorable clinical outcome among patients with advanced squamous cell carcinoma of the head and neck treated with cis-platin-based induction chemotherapy. J Clin Oncol. 2006;24(26):4333–4339. doi: 10.1200/JCO.2006.05.8768. [DOI] [PubMed] [Google Scholar]
  • 16.Vaezi A, Wang X, Buch S, Gooding W, Wang L, Seethala RR, et al. XPF expression correlates with clinical outcome in squamous cell carcinoma of the head and neck. Clin Cancer Res. 2011;17(16):5513–5522. doi: 10.1158/1078-0432.CCR-11-0086. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Zhong S, Nukui T, Buch SC, Diergaarde B, Weissfeld LA, Grandis JR, et al. Effects of ERCC2 Lys751Gln (A35931C) and CCND1 (G870A) polymorphism on outcome of advanced-stage squamous cell carcinoma of the head and neck are treatment dependent. Cancer Epidemiol Biomark Prev. 2011;20(11):2429–2437. doi: 10.1158/1055-9965.EPI-11-0520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Divaris K, Olshan AF, Smith J, Bell ME, Weissler MC, Funkhouser WK, et al. Oral health and risk for head and neck squamous cell carcinoma: the Carolina Head and Neck Cancer Study. Cancer Causes Control. 2010;21(4):567–575. doi: 10.1007/s10552-009-9486-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Stingone JA, Funkhouser WK, Weissler MC, Bell ME, Olshan AF. Racial differences in the relationship between tobacco, alcohol, and squamous cell carcinoma of the head and neck. Cancer Causes Control. 2013;24(4):649–664. doi: 10.1007/s10552-012-9999-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Hakenewerth AM, Millikan RC, Rusyn I, Herring AH, North KE, Barnholtz-Sloan JS, et al. Joint effects of alcohol consumption and polymorphisms in alcohol and oxidative stress metabolism genes on risk of head and neck cancer. Cancer Epidemiol Biomark Prev. 2011;20(11):2438–2449. doi: 10.1158/1055-9965.EPI-11-0649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Hakenewerth AM, Millikan RC, Rusyn I, Herring AH, Weissler MC, Funkhouser WK, et al. Effects of polymorphisms in alcohol metabolism and oxidative stress genes on survival from head and neck cancer. Cancer Epidemiol. 2013;37(4):479–491. doi: 10.1016/j.canep.2013.03.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Wyss AB, Herring AH, Avery C, Weissler MC, Bensen JT, Barnholtz-Sloan JS, et al. Single Nucleotide Polymorphisms in Nucleotide Excision Repair Genes, Cigarette Smoking, and the Risk of Head and Neck Cancer. Cancer Epidemiol Biomark Prev. 2013;22(8):1428–1445. doi: 10.1158/1055-9965.EPI-13-0185. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Illumina. [Accessed on 25 Oct 2011];GoldenGate assay workflow. 2006 http://www.illumina.com/documents/products/workflows/workflow_goldengate_assay.pdf.
  • 24.Centers for Disease Control and Prevention. [Accessed on 10 Dec 2013];National death index. 2013 http://www.cdc.gov/nchs/ndi.htm.
  • 25.Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B (Methodol) 1995;1995:289–300. [Google Scholar]
  • 26.Pfaff CL, Barnholtz-Sloan J, Wagner JK, Long JC. Information on ancestry from genetic markers. Genet Epidemiol. 2004;26(4):305–315. doi: 10.1002/gepi.10319. [DOI] [PubMed] [Google Scholar]
  • 27.Barnholtz-Sloan JS, McEvoy B, Shriver MD, Rebbeck TR. Ancestry estimation and correction for population stratification in molecular epidemiologic association studies. Cancer Epidemiol Biomark Prev. 2008;17(3):471–477. doi: 10.1158/1055-9965.EPI-07-0491. [DOI] [PubMed] [Google Scholar]
  • 28.Barnholtz-Sloan JS, Shetty PB, Guan X, Nyante SJ, Luo J, Brennan DJ, et al. FGFR2 and other loci identified in genome-wide association studies are associated with breast cancer in African-American and younger women. Carcinogenesis. 2010;31(8):1417–1423. doi: 10.1093/carcin/bgq128. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Hosmer DW, Lemeshow S. Confidence interval estimation of interaction. Epidemiology. 1992;3(5):452–456. doi: 10.1097/00001648-199209000-00012. [DOI] [PubMed] [Google Scholar]
  • 30.SAS Institute Inc. SAS 9.3. SAS Institute Inc; Cary, NC: 2011. [Google Scholar]
  • 31.National Cancer Institute. [Accessed on 10 Dec 2013];SEER stat fact sheets: oral cavity and pharynx cancer. 2013 http://seer.cancer.gov/statfacts/html/oralcav.html.
  • 32.National Cancer Institute. [Accessed on 10 Dec 2013];SEER stat fact sheets: larynx cancer. 2013 http://seer.cancer.gov/statfacts/html/laryn.html.
  • 33.Johnson AD, Handsaker RE, Pulit S, Nizzari MM, O’Donnell CJ, de Bakker PIW. SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap. Bioinformatics. 2008;24(24):2938–2939. doi: 10.1093/bioinformatics/btn564. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ioannidis JP, Ntzani EE, Trikalinos TA, Contopoulos-Ioannidis DG. Replication validity of genetic association studies. Nat Genet. 2001;29(3):306–309. doi: 10.1038/ng749. [DOI] [PubMed] [Google Scholar]
  • 35.Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I. Controlling the false discovery rate in behavior genetics research. Behav Brain Res. 2001;125(1):279–284. doi: 10.1016/s0166-4328(01)00297-2. [DOI] [PubMed] [Google Scholar]
  • 36.Rice TK, Schork NJ, Rao DC. Methods for handling multiple testing. Adv Genet. 2008;60:293–308. doi: 10.1016/S0065-2660(07)00412-9. [DOI] [PubMed] [Google Scholar]
  • 37.National Center for Biotechnology Information, National Library of Medicine. [Accessed on Feb 2013];Database of single nucleotide polymorphisms (dbSNP) 2013 http://www.ncbi.nlm.nih.gov/SNP/
  • 38.Meyer LR, Zweig AS, Hinrichs AS, Karolchik D, Kuhn RM, Wong M, et al. The UCSC genome browser database: extensions and updates 2013. Nucleic Acids Res. 2013;41(D1):D64–D69. doi: 10.1093/nar/gks1048. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Friedberg EC. How nucleotide excision repair protects against cancer. Nat Rev Cancer. 2001;1(1):22–33. doi: 10.1038/35094000. [DOI] [PubMed] [Google Scholar]
  • 40.Friedberg EC, Walker GC, Siede W, Wood RD, Schultz RA. DNA repair and mutagenesis. ASM Press; Washington, DC: 2006. [Google Scholar]
  • 41.Lee JM, Yang PW, Yang SY, Chuang TH, Tung EC, Chen JS, et al. Genetic variants in DNA repair predicts the survival of patients with esophageal cancer. Ann Surg. 2011;253(5):918–927. doi: 10.1097/SLA.0b013e318216f374. [DOI] [PubMed] [Google Scholar]
  • 42.Koberle B, Ditz C, Kausch I, Wollenberg B, Ferris RL, Albers AE. Metastases of squamous cell carcinoma of the head and neck show increased levels of nucleotide excision repair protein XPF in vivo that correlate with increased chemoresistance ex vivo. Int J Oncol. 2010;36(5):1277–1284. doi: 10.3892/ijo_00000612. [DOI] [PubMed] [Google Scholar]
  • 43.Lash TL, Cole SR. Immortal person-time in studies of cancer outcomes. J Clin Oncol. 2009;27(23):e55–e56. doi: 10.1200/JCO.2009.24.1877. [DOI] [PubMed] [Google Scholar]
  • 44.Ragin CC, Taioli E. Survival of squamous cell carcinoma of the head and neck in relation to human papillomavirus infection: review and meta-analysis. Int J Cancer. 2007;121(8):1813–1820. doi: 10.1002/ijc.22851. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary
NIHMS596148-supplement-Supplementary.docx (101KB, docx)

RESOURCES