Emergence of HPV16-Positive Oropharyngeal Cancer in Black Patients Over Time: University of Maryland 1992–2007 Free

Squamous cell carcinoma of the head and neck (HNSCC) spans diverse anatomic sites, including the oral cavity, oropharynx, larynx, hypopharynx, and nasopharynx. While traditional risk factors for HNSCC include smoking and alcohol, it has become apparent over the last 2 to 3 decades that the human papilloma virus (HPV) has an etiologic role in the development of squamous cell carcinoma of the oropharynx (OPSCC), with predilection for tumors involving the base of tongue or tonsils (1, 2). HPV-positive OPSCC has a distinct clinical phenotype occurring predominantly in nonsmokers and nondrinkers of male sex and younger age than HPV-negative OPSCC (1–3). Although these tumors are more often poorly differentiated with basaloid features and patients typically present in stage III and IV with more advanced nodal stage (4, 5), HPV-positive oropharyngeal cancers have been shown to be more responsive to therapy with improved outcome (6–9).

Recent analyses of HPV-positive OPSCC have shown significant disparities by race between black and white patients (10–13). Our group analyzed 237 specimens available from patients treated on the TAX 324 trial and showed that whites had a significantly higher proportion of HPV16-positive tumors than black patients (34% vs. 4%, P < 0.001; ref. 11). More recently, Chernock and colleagues examined the biopsy specimens of 174 patients with stage III/IV OPSCC and found that a significantly higher percentage of white patients tested positive for high-risk HPV (63.5% vs. 11.5%, P ≤ 0.001) or p16 (83.1% vs. 34.6%, P < 0.001) than black patients (10).

While disparities in the percentage of HPV-positive OPSCC have been reported, fewer data exist on trends over time in HPV-positive OPSCC by race. Using the Surveillance, Epidemiology, and End Results (SEER) Residual Tissue Repositories Program, Chaturvedi and colleagues observed an increase in the prevalence of HPV16-positive OPSCC over time for both whites and blacks from 1984 to 2004, with this trend reaching significance only for whites (P ≤ 0.001; ref. 3). Other studies have looked at changes in incidence of HPV-positive OPSCC over time by race; however, HPV-positive HNSCC was inferred on the basis of potentially associated anatomic locations, including the base of tongue, tonsils, or other sites in the oropharynx (14, 15).

To further investigate changes over time in the percentage of HPV16-positive OPSCC between black and white patients, we analyzed tissue samples from patients with oropharyngeal cancer treated here at the University of Maryland Greenebaum Cancer Center (UMGCC; Baltimore, MD) over a 15-year time period.

After Institutional Review Board (IRB) approval, we performed a retrospective chart review of data from patients treated at UMGCC for HNSCC from 1992 to 2007. We identified 467 cases of OPSCC. Diagnosis was confirmed by biopsy. Data collected included race, gender, age at diagnosis, smoking status, alcohol use, disease occurrence (first primary, second primary, or recurrence), tumor stage, nodal stage, overall staging, treatment received, and year of diagnosis. Baseline characteristics were obtained through chart review and gender, race, and smoking, and drinking status were based on patient reporting. Tumor–node–metastasis (TNM) staging (16) was used to define tumor, nodal, and overall staging. Any missing values were recorded as unknown. Of the 467 cases, formalin-fixed, paraffin-embedded (FFPE) tumor tissue was available for HPV16 testing from 200 patients.

Tissue sections were reviewed by a pathologist to guide core punches containing primarily tumor tissue. DNA extraction was performed using the QIAamp DNA FFPE Tissue Kit (Qiagen). DNA was quantified using the Quant-iT dsDNA Assay Kit, High Sensitivity (Invitrogen) and stored at −80°C in aliquots.

PCR was performed for the E6 (forward ATGTTTCAGGACCCACAGGA, reverse CAGCTGGGTTTCTCTACGTGTT) and E7 (forward ATGCATGGAGATACACCTAC, reverse CATTAACAGGTCTTCCAAAG) genes of HPV16. Using 2 ng DNA, 40 cycles of standard 3-step PCR (annealing temperature: 55°C) were performed. A negative control (no DNA) was included in every PCR run. Only cases positive for both genes were scored HPV16-positive. Cases that were discordant between the genes were excluded. In the case of ambiguity interpreting either gene, both genes were amplified again from a freshly diluted DNA aliquot using a different primer set (overlapping but not nested with the original primer set). Cases that remained ambiguous were excluded. HPV16 status could be scored for 194 of 200 cases.

While other HPV types, including HPV18, 31, 33, and 35, can also be found in OPSCC, HPV16 accounts for the majority of cases of HPV-associated OPSCC, accounting for upward of 87% of cases worldwide (17). To ensure we were not missing a significant amount of other HPV types, we performed an exploratory analysis of HPV16-negative tumor samples from 27 black patients. HPV genotyping on these DNA samples was performed using the LINEAR ARRAY HPV Genotyping Test (Roche Molecular Diagnostics), which tests for 37 high- and low-risk HPV types.

Baseline characteristics were compared between the 200 patients with tissue available for HPV16 testing and the remainder of the OPSCC patient population (n = 267) using univariable and multivariable logistic regression models. For the 200 patients with known HPV16 status, the Fisher exact and the Fisher–Freeman–Halton with Monte Carlo simulation tests were used to compare demographic and clinical characteristics between HPV16-positive and HPV16-negative patients, HPV16-positive whites versus HPV16-positive black patients, HPV16-positive white patients versus HPV16-negative white patients, and HPV16-positive black patients versus HPV16-negative black patients. Continuous variables were compared using the t test. Patients with unknown status or value of a characteristic were excluded from tables and the corresponding and relevant analysis. All statistical tests were 2-sided and done at 0.05 Type I error.

Overall survival (OS) functions were estimated and compared using the Kaplan–Meier approach and the log-rank tests. The univariable and multivariable Cox regression models were applied to identify possible predictors as well as the magnitude of their effect on survival parameters (HR, hazard of death). OS was defined from the date of diagnosis until the date of death from any cause or censored at the date last known alive. Median OS times with the corresponding 95% confidence intervals (CI) were estimated. The homogeneity test was applied for multiple (4) stratified 2 × 2 tables to assess whether ORs)for acquiring HPV16-positive status across 4 distinct time periods were constant. Because of the estimated common OR across these 4 time periods, the second common OR test was used to estimate and compare OR for distinct ethnic groups of patients. Statistical analyses were conducted using S-plus (TIBCO, v. 8.2) and StatXact (Cytel st., v.8.0).

A total of 467 patients with OPSCC were identified between 1992 and 2007, and 200 of these patients had tissue available for testing for HPV16. The median age of the tested population was 56 years, 81% of patients were male, and racial distribution was 62.5% white, 36.5% black, and 1% other races. The majority of cases (93.5%) represented a first primary diagnosis, with most patients presenting with locally advanced disease with 62% presenting as stage IVA. The largest proportion of our patients (47%) was treated with chemoradiation alone. Compared with our entire population of OPSCC, the tested patients were younger and more likely to consume alcohol or both alcohol and tobacco. (Supplementary Table S1).

Of patients with tissue available for testing, 67 (33%) were positive for HPV16, 127 (64%) were negative, and 6 (3%) could not be analyzed for technical reasons. Comparison of HPV16-positive versus -negative patients is shown in Table 1. Patients with HPV16-associated OPSCC were significantly more likely to be white and male and were significantly less likely to be smokers, drinkers, or both smokers and drinkers. They also had a significantly lower tumor (T) stage and higher nodal (N) stage than their HPV16-negative counterparts. There were no differences in baseline characteristics between white and black HPV16-positive patients (Supplementary Table S2). Compared with HPV16-negative white patients, HPV16-negative black patients were significantly younger and more likely to be male, drinkers, and combined drinkers and smokers (Table 2).

Differences in percentage of HPV16-positive oropharyngeal cancers by race were evaluated over time, from 1992 until 2007. For each 4-year time period examined, whites had a significantly higher percentage of HPV16-positive oropharyngeal cancers than blacks (P < 0.0001). When our entire cohort of tested patients was analyzed, there was a significant increase in the percentage of HPV16-positive oropharyngeal cancers over time from 15.6% in 1992 to 1995 to 43.3% in 2004 to 2007 (P = 0.01; Fig. 1). However, when trends over time for each individual race were examined, while both exhibited an increase in the percentage of HPV16-positive patients over time, neither of these trends reached significance (Fig. 1). Importantly, we observed the emergence of HPV16-positive OPSCC in blacks treated at UMGCC between 1992 and 2007, increasing from no cases in 1992 to 1995 to 17.7% of OPSCC in 2004 to 2007.

Survival was analyzed in patients tested for HPV16 with a first primary diagnosis (n = 187). Blacks had a significantly worse survival than all whites (HR, 1.9; 95% CI, 1.4–2.7; P = 0.0002), and HPV16-positive patients had a significantly better OS than HPV16-negative patients (median OS, 8.1 vs. 1.5 years; P < 0.0001). Survival analysis by race and HPV16 status is shown in Fig. 2. Both white HPV16-positive and black HPV16-positive patients had a significantly better OS than their respective white and black HPV16-negative cohorts (Fig. 2C and D, respectively). While there were no differences in OS between HPV16-positive black and HPV16-positive white patients (median OS, 8.1 vs. 8.1 years; Fig. 2A), HPV16-negative whites had a significantly better OS than HPV16-negative black patients (median OS, 2.3 vs. 0.9 years; P = 0.02; Fig. 2B).

A multivariable Cox regression model was used to assess whether differences in characteristics between patients tested for HPV16 (Table 3) and between black and white HPV16-negative patients (Table 4) accounted for the survival differences between these groups. Race, age, gender, overall stage, and smoking and drinking were included in the model. Among the entire tested population, all factors, except for stage, were significantly associated with survival, with race and HPV status having the strongest association. Among HPV-negative patients, race was highly significant when controlling for these other factors (P = 0.003; Table 4).

As described in Materials and Methods, given the low number of HPV16-positive black patients, to ensure we were not missing a significant amount of other HPV types, we analyzed HPV16-negative tumor samples from 27 black patients using LINEAR ARRAY HPV Genotyping Test (Roche Molecular Diagnostics), to evaluate for 37 high- and low-risk HPV strains. With appropriate positive controls, we detected only one other HPV strain (HPV58) in a single patient.

Our study uniquely evaluated changes over time in percentage of HPV16-positive oropharyngeal cancers between white and black patients based on pathologic confirmation of HPV status. In our entire population of patients, there was a significant increase in HPV16-positive OPSCC over time. While the increase in percentage of HPV16-positive OPSCC was not significant when white and black patients were analyzed separately, importantly, during this time period, we observed the emergence of HPV16-positive OPSCC in black patients treated at UMGCC, increasing from no cases in 1992 to 1995 to 17.7% of black OPSCC cases in 2004 to 2007.

To our knowledge, only one other study has examined prevalence of HPV-positive OPSCC over time by race with HPV status confirmed by analysis of tumor samples. In this study, Chaturvedi and colleagues examined specimens from the SEER Repositories Tissue Program and observed a significant increase in the prevalence of HPV-positive OPSCC in whites from 20% to more than 60% from 1984 to 2004. With a total of 50 HPV-positive OPSCC black patients found during the years examined, they observed the emergence and subsequent increase in prevalence in blacks, with an increase from no cases in 1984 to 1989 to 20% in 1995 to 1999. However, this trend did not reach statistical significance (3). Although our data were from a single institution in one geographic region, we similarly observed an increase in HPV16-positive oropharyngeal cancers in whites and the emergence (albeit ∼6 years later than by SEER data) and subsequent increase of HPV16-positive OPSCC in black patients between 1992 and 2007. Our ability to detect a significant difference for each individual race is limited by a smaller sample size, including only 10 HPV16-positive black patients found during the years studied. Our results also showed that for all time periods between 1992 and 2007, white patients had a significantly higher percentage of HPV16-positive tumors. Confirming our initial work as well as subsequent studies (10–13, 18, 19), we found that HPV16-positive oropharyngeal cancer is significantly more common in whites than in blacks.

Risk factors for HPV-positive OPSCC include increased number of vaginal, oral, and oral–anal sexual partners, leading to increased exposure and subsequent infection of oropharyngeal tissue (2, 18, 20). Increasing trends of HPV-related disease over time have been attributed to changes in sexual practices over the previous decades, including increasing number of sexual partners, decreased age at sexual debut, as well an increase in the number of people performing oral sex (21, 22). Differences in the prevalence of HPV-positive OPSCC between blacks and whites may be partially attributed to differences in sexual practices with whites engaging in more oral–genital sex than blacks (23, 24). Data from interviews conducted during a national survey of adolescent males in the United states ages 15 to 19 years showed that 42% of white males had performed oral sex on a female compared with only 21% of black males. Black males had a significant change in percentage who received oral sex from a female increasing 25% to 57% from 1988 to 1995, whereas there was no significant change over the same time period in white males (48% to 50%) in this study (23). Similarly, data from the 2009–2010 National Health and Nutrition Examination Survey (NHANES) that included more than 4,000 patients found that across all age cohorts, white men and women were significantly more likely to have ever performed oral sex than other races. White men for example were significantly more likely to have performed oral sex at sexual debut (43.5% vs. 17.6%; P < 0.001) and have more than 5 lifetime oral sex partners (38.8% vs. 20.7%; P < 0.0001). Although blacks reported less oral sexual behaviors, in both blacks and whites, younger age cohorts (20–29 and 30–44 years) were more likely to have ever performed oral sex than older age cohorts (45–59 and 60–69 years; ref. 25). Increased oral sex behaviors in whites may partially explain the consistent disparity in prevalence of HPV-positive OPSCC by race. While more data are needed on changes in sexual behaviors over time by race, the delayed emergence of HPV-positive OPSCC compared with whites and our observation of an increasing prevalence of HPV16-positive OPSCC in blacks in more recent years may reflect changes in sexual practices over time in blacks.

Despite the differences in oral sexual behaviors described above, no disparity in prevalence of oral HPV infection was found in the NHANES cohort (25, 26). In a cross-sectional study as part of NHANES, which involved evaluating oral rinse samples in healthy adults for HPV DNA of oral exfoliated cells, a trend toward a significantly higher prevalence of any oral HPV infection was seen in black versus white participants (10.5% vs. 6.5%, respectively; P = 0.06), with only a small disparity in oral HPV16 infection (2.4% in whites vs. 1.9% in blacks), and race was not independently associated with oral HPV infection after controlling for sexual behavior (25, 26). This suggests that differences in sexual practice accounting for higher exposure in white patients do not appear to lead to significant differences in the prevalence of oral HPV infection between blacks and whites. Surprisingly, despite similar rates of HPV infection, HPV-positive OPSCC is at least 3-fold more common in whites than in blacks (3, 10, 11, 13). This suggests that other not-yet-defined factors, which may influence the persistence of oral HPV infection and the subsequent development of cancer following HPV infection, may also contribute to the racial disparity in HPV-associated OPSCC. In fact, a study of 140 patients showed that whites were significantly more likely to have HPV “active” HNSCC (defined as HPV DNA positive and p16 positive) than blacks (27), as growing data suggest that it is in tumors in which both HPV and p16 are positive, that HPV is the driver of oncogenesis, leading to an improved prognosis (28, 29). Further research is needed to determine the reasons for a consistently higher proportion of HPV-associated OPSCC among white patients compared to black.

Similar to the experience in the literature, our HPV16-positive patients had a significantly improved OS compared with HPV16-negative patients (6–9). Our study also compared HPV16-positive and -negative patients by race and found that both black and white HPV16-positive patients had a significantly improved OS compared with their respective HPV16-negative racial counterparts. There was no difference in OS between HPV16-positive black and HPV16-positive white patients. Interestingly, HPV16-negative blacks had a significantly worse OS than HPV16-negative white patients. The reasons for this disparity are not defined.

Worsham and colleagues compared survival by HPV status and race in 121 patients with OPSCC (13). While they similarly found that HPV-positive patients had a significantly improved OS compared with HPV negative in their entire cohort and among blacks, in contrast, they found no difference in OS between HPV-positive and -negative white patients. While whites had a higher percentage of HPV-positive tumors, reasons for the lack of impact of HPV status on survival among whites in this study are unclear. Similar to our findings, there was no difference between black and white HPV-positive patients, and black HPV-negative patients had a significantly shorter survival than white HPV-negative patients (13). However, other studies have shown no difference in OS between HPV-negative white and black patients (11, 27). This includes our previous analysis of the TAX 324 trial which showed no difference in survival between white HPV-negative patients and all black patients, where all but one of the black patients were HPV-negative patients (11).

While there were significant differences in baseline characteristics between HPV16-negative black and white patients in our population (younger age, male predominance, and higher proportion of smokers and drinkers among black patients), race was still significantly associated with worse survival among HPV-negative patients after controlling for these factors as well as TNM stage. In a parallel analysis, we looked at survival by race in more than 500 black and white patients with stage III and IV tumors of the larynx, hypopharynx, and oral cavity treated during the same time period at our institution and saw no difference in survival by race for those anatomic sites (data not shown). This suggests that in our population, currently undefined biologic differences may be contributing to the worse outcomes seen in patients with HPV16-negative black oropharyngeal cancer. Further analysis is needed to better characterize reasons for disparities in OS between HPV-negative patients with oropharyngeal cancer by race.

Black race was associated with a significantly worse OS among our entire cohort with tissue available for HPV testing. Given that HPV16-negative white patients had improved OS compared with blacks, to evaluate the effects of race and HPV status on OS in our patient population, we carried out a multivariate analysis. While both race and HPV were independently associated with survival, given the magnitude of the improved prognosis in HPV-associated OPSCC seen in our population, and the lack of difference in survival between white and black HPV-positive patients, it is more likely that the higher incidence of HPV-positive OPSCC among white patients, rather than race alone, accounts for their significantly improved OS compared with black patients in our entire cohort.

Our article has a number of potential limitations. While our population was diverse, our absolute number of HPV-positive black patients was low, albeit similar to other comparisons in the literature (10, 13). In addition, changes in proportions of HPV-positive OPSCC for both whites and blacks are affected by changes in the incidence of OPSCC. Thus, a limitation of our retrospective study is that we cannot determine incidence rates for our population. Another potential limitation is that as testing was confined to HPV16, cases involving other HPV types would have been missed. However, HPV16 accounts for the majority of HPV types worldwide (17). In addition, because of a low number of HPV16-positive black patients, we carried out our linear array analysis on specimens from 27 black patients that were negative for HPV16 by PCR and found only 1 specimen positive for another HPV type.

Our retrospective analysis of patients with oropharyngeal cancer diagnosed from 1992 to 2007 shows a significantly higher percentage of HPV16-positive tumors among whites compared to blacks for the entire population as well as over each 4-year time period examined. Trends for an increase in HPV16-positive OPSCC over time were seen not only in white patients but also in blacks, and we observed the emergence of HPV16-positive OPSCC in black patients treated at UMGCC during the time period examined. HPV16-positive OPSCC was associated with a significantly improved OS for both races, and among HPV16-negative patients, black race was associated with a worse outcome even after controlling for selected baseline characteristics. Our analysis highlights the need for further research to better explain the poor prognosis seen in HPV16-negative blacks and the consistent disparity in the prevalence of HPV-positive oropharyngeal cancer between black and white patients.

No potential conflicts of interest were disclosed.

Conception and design: D.P. Zandberg, O.G. Goloubeva, K.J. Cullen

Development of methodology: D.P. Zandberg, O.G. Goloubeva, L.M. Schumaker, K.J. Cullen

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): D.P. Zandberg, S. Liu, L.M. Schumaker, K.J. Cullen

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): D.P. Zandberg, S. Liu, O.G. Goloubeva, L.M. Schumaker, K.J. Cullen

Writing, review, and/or revision of the manuscript: D.P. Zandberg, O.G. Goloubeva, L.M. Schumaker, K.J. Cullen

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D.P. Zandberg, S. Liu, L.M. Schumaker, K.J. Cullen

Study supervision: D.P. Zandberg, K.J. Cullen

This work was supported by the Orakawa Foundation. K.J. Cullen received NIH grant funding (1 P30 CA 134274-01). D.P. Zandberg received NIH grant funding (K12 CA126849).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.