Abstract
BACKGROUND
HPV16E6 antibodies may be an early marker of HPV-driven oropharyngeal cancer (HPV-OPC) diagnosis and recurrence.
METHODS
We identified 161 incident OPCs diagnosed at the University of Pittsburgh (2003–2013) with pre-treatment serum; 112 had pre-existing clinical HPV testing with p16 immunohistochemistry (p16-IHC) and HPV in situ hybridization (HPV-ISH) (87 dual positives [HPV-OPC], 25 dual negatives [HPV-negative]); 62 had at least 1 post-treatment serum sample. 86 out of 161 tumors were available for additional HPV16DNA/RNA testing (45 dual positive [HPV16-OPC], 19 dual negative [HPV16-negative). HPV16E6 antibody testing was conducted using multiplex serology. We evaluated: a) the sensitivity/specificity of HPV16E6 serology to distinguish HPV-OPC and HPV16-OPC from HPV-negative OPC, b) HPV16E6 antibody decay post-treatment using linear models accommodating correlations in variance estimates and c) pre- and post-treatment HPV16E6 levels and risk of recurrence using Cox proportional hazards models.
RESULTS
Of 87 HPV-OPCs, 78 were HPV16E6 seropositive; sensitivity: 89.7% (95%CI: 81.3%–95.2%) and 24 of 25 HPV-negative OPCs were HPV16E6 seronegative; specificity: 96.0% (95%CI: 79.6%–99.9%). 42 of 45 HPV16-OPCs were HPV16E6 seropositive; sensitivity: 93.3% (95%CI: 81.7%–98.6%) and 18 of 19 HPV16-negative OPCs were HPV16 E6 seronegative; specificity: 94.7% (95%CI: 74.0%–99.9%). Post-treatment HPV16E6 antibody levels did not decrease significantly from baseline (P=0.475, median follow-up: 307 days) and were not associated with risk of recurrence. However, pre-treatment HPV16E6 seropositivity was associated with an 86% reduced risk of local/regional recurrence; hazard ratio: 0.14 (95%CI: 0.03–0.68), P=0.015.
CONCLUSIONS
HPV16E6 antibodies may have potential clinical utility for the diagnosis and/or prognosis of HPV-OPC.
Keywords: Human papillomavirus, HPV, serology, HPV16 E6 antibodies, antibody sensitivity, recurrence
INTRODUCTION
The incidence of human papillomavirus-driven oropharyngeal cancer (HPV-OPC) has risen by 225% since the 1980s within the US and now accounts for 75% of all OPCs[1].
While treatment for HPV-OPC is often curative, severe treatment-related morbidity includes permanent difficulty swallowing/eating[2]. Recently, there has been a focus toward reducing this morbidity through developing de-escalated treatment protocols as well as developing screening methods for early detection. However, implementation of de-escalated treatment protocols is hindered by the inability to identify patients at highest risk for recurrence, for whom de-escalation would be ill-advised and development of screening methods have been hindered by the inability to identify those at highest risk for OPC.
HPV16E6 antibody positivity has recently been identified as a promising biomarker for both OPC diagnosis as well as prognosis. A prospective European-based study found that 34.8% of patients with OPC were seropositive for HPV16E6 compared to only 0.6% of cancer-free controls. Importantly, HPV16E6 antibodies were present more than 10 years prior to diagnosis[3]. We recently confirmed these findings in a US-based cohort[4]. Several studies also suggested that both pre- and post-treatment HPV16E6 levels may be associated with risk of HPV-OPC recurrence[5–8].
In many of these serum-based studies,[3, 5–7, 9, 10] information regarding the HPV status of the tumors was not available or HPV status was assessed using non-gold standard methods. Thus, the sensitivity of HPV16 E6 seropositivity to detect the subset of tumors caused by HPV infection is unknown. Before considering HPV16E6 antibodies as a potential tool for early detection and/or prognosis of HPV-OPC, larger studies using highly specific HPV16E6 serologic assays and gold standard methods for assigning HPV tumor status are needed.
Using a large sample of OPC tumors (N=161) from a clinic-based US population, the objectives of this study were to: a) determine both the sensitivity and specificity of HPV16E6 seropositivity to detect HPV-OPC and b) determine the association between pre- and post-treatment HPV16E6 levels and risk of HPV-OPC recurrence using gold standard methods for HPV serology and assignment of HPV tumor status. We used the same highly specific multiplex serology assay used in the initial discovery and validation of HPV16E6 seropositivity as an early marker of OPC [3, 4]. Additionally, to minimize misclassification, we used both the clinical and research gold standard methods of assigning HPV tumor status, combined p16 immunohistochemistry (IHC) and HPV in situ hybridization (ISH) and combined HPV DNA and RNA analysis, respectively.
MATERIALS AND METHODS
Study Population
Incident, previously untreated, cases of OPC were identified using an IRB-approved protocol (UPCI 99-069) that collects baseline serum at the University of Pittsburgh Cancer Institute; all patients provided written informed consent.
Since the establishment of the biobank in 2000, a total of 1,462 patients with OPC (ICD10 codes: C01.0, C02.4, C05.1, C05.2, C05.8, C09.0, C09.1, C09.8, C09.9, C10.0, C10.2, C10.3, C10.8, C10.9, C14.0) were treated; 939 (64%) were enrolled as part of the tissue banking study. Of the 860 incident cases, 720 (84%) had serum samples collected prior to treatment, of which, 112 (17%) had concurrent p16 IHC and HPV ISH test results available as part of clinical management.
HPV antibody testing was conducted on pre-treatment serum samples from all incident cases of HPV-OPC (p16 and HPV ISH dual positives, N=87) and HPV-negative OPC (p16 and HPV ISH dual negatives, N=25). Testing of serial samples was conducted for the subset of patients (N=35 of 87) who had 1–2 post-treatment samples available. Additionally, we chose to test serial samples from all HPV indeterminate OPCs (discordant or missing p16 IHC and HPV ISH results) for which post-treatment samples were available under the assumption that the majority would be HPV16E6 seropositive pre-treatment; thus increasing our power to assess changes in HPV16E6 levels post-treatment and risk of recurrence. However, serum samples from 22 HPV-indeterminate patients were erroneously tested under the belief they had post-treatment serial samples (Supplemental Figure 1).
Eighty-six tumors were recovered for HPV16-type specific DNA and RNA testing. Those with concordant HPV16 DNA/RNA positive results were defined as HPV16-driven (HPV16-OPC); those with concordant negative results were defined as HPV16-negative and those with discordant results were considered indeterminate.
Laboratory Methods
Serologic Testing
Multiplex serologic testing was performed at the German Cancer Research Center (Heidelberg, Germany) by staff blinded to the HPV status of the patients[11–14]. Antigens were affinity-purified, bacterially expressed fusion proteins with N-terminal Glutathione S-transferase. Samples were analyzed for antibodies against HPV16 L1, E1, E2, E4, E6, and E7 and the E6 protein from HPV6, HPV11, HPV18, HPV31, HPV33, HPV45 and HPV58. Antibody levels were quantified at 1:100 serum dilution as median fluorescence intensity (MFI) and dichotomized as positive or negative based on previously defined cutpoints[3, 15].
Clinical HPV testing
As part of clinical management, p16 overexpression and HPV DNA ISH were available for a subset of patients. p16 IHC was conducted on deparaffinized tissue sections using the monoclonal antibody clone G175–405 against p16INK4 (dilution 1:200; BD Pharmingen, San Diego, CA); immunoreactivity in ≥70% of cells was considered positive[16, 17]. ISH was performed with a probe set specific for 10 HPV types (Dako Cytomation, Carpinteria, CA).
HPV type-specific testing
Formalin-fixed paraffin-embedded (FFPE) tissue samples were consecutively sectioned for DNA and RNA analysis; the first and the last sections were stained by hematoxylin and eosin to verify the presence of tumor[18]. DNA and RNA were extracted from FFPE sections as previously described[18]; water samples were included to monitor possible cross-contamination. HPV DNA analysis was carried out using Multiplex Papillomavirus Genotyping[19, 20] homogenously amplifying and specifically detecting 51 HPV types, three subtypes and cellular beta-globin as internal DNA quality control. Samples that were positive for HPV and/or positive for cellular beta-globin DNA were considered DNA valid. HPV RNA analysis, i.e. detection of viral transcripts, was performed by HPV type-specific RT-PCR and hybridization[18] assays developed for 20 distinct HPV types. These assays amplify HPV E6*I and ubiquitin C (ubC) cDNA as a cellular mRNA quality control. Specimens that were HPV E6*I and/or ubC mRNA-positive were considered RNA valid.
Statistical Analyses
Patient characteristics were evaluated overall and by HPV tumor status. The proportion of HPV-OPC cases seropositive for HPV16 antibodies (sensitivity) and the proportion of HPV-negative OPCs seronegative for HPV16 antibodies (specificity) was calculated for all HPV16 proteins analyzed; confidence intervals (CIs) were estimated using an exact binomial method. Overall sensitivity and specificity of HPV16E6 antibodies for HPV16-OPC and overall OPC (regardless of HPV tumor status) were calculated using the same methods. The association between pre-treatment HPV16E6 seropositivity and a) risk of overall recurrence (local/regional and distant metastasis combined); b) risk of local/regional and metastasis separately; and c) overall survival was evaluated by estimating hazard ratios (HRs) and 95%CIs from Cox proportional hazards models; years since cancer diagnosis was used as the time variable. For risk of overall recurrence, follow up ended at time (years) of first recurrence (local/regional or distant) or time of censoring; censoring events were death or last follow-up. For risk of local/regional recurrence, follow-up time ended at time of first local/regional recurrence (the event of interest) or time of diagnosis of metastasis, death or last follow-up (all censoring events); to estimate risk of metastasis, follow-up time ended at diagnosis of first metastasis, local/regional recurrence, death or last follow-up. To assess risk of death from any cause, the only censoring event was at time of last follow-up. We chose Cox regression for estimating risk of recurrence given that Prentice et al showed that the cause-specific hazard can be estimated by treating all other competing events as censoring events [21]. This approach does not require any independence assumption of the main outcome and competing events to obtain valid and unbiased relative risk estimates. The proportional hazard assumption was confirmed for each model. HPV16E6 antibody decay was evaluated in linear models accommodating correlations over time in the variance estimate [22, 23]. We also modelled HPV16E6 antibody levels coded in tertiles as a time dependent variable in a Cox regression model to incorporate post-treatment HPV16E6 antibody levels to assess risk of recurrence among all patients HPV16E6 seropositive pre-treatment. All analyses were performed by STATA IC version 14 and SAS version 9.4.
RESULTS
Patient characteristics
A total of 161 patients were included in this analysis; 112 (69.6%) with prior clinical testing for HPV (25 [22.3%] HPV-negative OPC, 87 [77.7%] HPV-OPC) and 49 (30.4%) with indeterminate tumor status (Table 1). Overall, the median age at diagnosis was 58 (interquartile range [IQR]:51–64); the majority was male (80.8%), Caucasian (96.9%) and advanced stage (77.0% stage III-IV; AJCC 7th edition). Compared to patients with HPV-negative OPC, patients with HPV-OPC tended to be younger, male, non-smokers, non-drinkers and advanced stage (AJCC 7th edition). Although, according to the AJCC 8th edition guidelines, the majority of HPV-OPC patients would be considered stage I (78.2%). Patients with indeterminate HPV tumor status most closely resembled those with HPV-OPC.
Table 1.
Patient characteristics overall and by HPV tumor status
| HPV-Tumor Status | ||||
|---|---|---|---|---|
|
|
||||
| Overall | HPV-negative1 | HPV-OPC2 | Indeterminate3 | |
|
|
|
|||
| Characteristics | N=161 | N=25 | N=87 | N=49 |
| N (%) | N (%) | N (%) | N (%) | |
| Age at Diagnosis (years) | ||||
| Median (IQR) | 58 (51–64) | 61 (57–68) | 54 (50–62) | 59 (53–63) |
| Year of Cancer Diagnosis | ||||
| 2003–2006 | 22 (13.7) | 1 (4.0) | 5 (5.7) | 16 (32.7) |
| 2007–2010 | 56 (34.8) | 9 (36.0) | 31 (35.6) | 16 (32.7) |
| 2011–2013 | 83 (51.6) | 15 (60.0) | 51 (58.6) | 17 (34.7) |
| Gender | ||||
| Male | 130 (80.8) | 18 (72.0) | 74 (85.1) | 38 (77.6) |
| Female | 31 (19.3) | 7 (28.0) | 13 (14.9) | 11 (22.5) |
| Race | ||||
| Caucasian | 156 (96.9) | 24 (96.0) | 83 (95.4) | 49 (100) |
| Asian | 4 (2.5) | 0 (0.0) | 3 (3.5) | 0 (0.0) |
| African American | 1 (0.6) | 1 (4.0) | 1 (1.2) | 0 (0.0) |
| Smoking History | ||||
| No | 45 (28.1) | 2 (8.0) | 24 (27.9) | 19 (38.8) |
| Yes | 115 (71.9) | 23 (92.0) | 62 (72.1) | 30 (61.2) |
| Pack years4 | ||||
| Median (IQR) | 40 (19.5–52.5) | 45 (20–60) | 30 (20–45) | 41 (16–50) |
| Alcohol History | ||||
| No | 40 (25.2) | 2 (8.3) | 23 (26.7) | 15 (30.6) |
| Yes | 119 (74.8) | 22 (91.7) | 63 (73.3) | 34 (69.4) |
| Stage (AJCC 7th Edition) | ||||
| I | 17 (10.6) | 5 (20.0) | 7 (8.1) | 5 (10.2) |
| II | 20 (12.4) | 8 (32.0) | 7 (8.1) | 5 (10.2) |
| III | 28 (17.4) | 1 (4.0) | 18 (20.7) | 9 (18.4) |
| IV | 96 (59.6) | 11 (44.0) | 55 (63.2) | 30 (61.2) |
| Stage (AJCC 8th Edition)5 | ||||
| I | -- | -- | 68 (78.2) | -- |
| II | -- | -- | 8 (9.2) | -- |
| III | -- | -- | 10 (11.5) | -- |
| Treatment | ||||
| Radiation + Chemo6 | 67 (41.6) | 7 (28.0) | 35 (40.2) | 25 (51.0) |
| Surgery + Radiation + Chemo6 | 51 (31.7) | 6 (24.0) | 31 (35.6) | 14 (28.6) |
| Surgery | 27 (16.8) | 11 (44.0) | 9 (10.3) | 7 (14.3) |
| Surgery + Radiation | 10 (6.2) | 1 (4.0) | 7 (8.1) | 2 (4.1) |
| Surgery + Chemo | 5 (3.1) | 0 (0.0) | 4 (4.6) | 1 (2.0) |
| Radiation | 1 (0.6) | 0 (0.0) | 1 (1.2) | 0 (0.0) |
| Serial Samples | ||||
| No | 99 (61.5) | 25 (100) | 52 (59.8) | 22 (44.9) |
| Yes | 62 (38.5) | -- | 35 (40.2) | 27 (55.1) |
| 2 serial samples | 38 (23.6) | -- | 22 (25.3) | 16 (32.7) |
| 3 serial samples | 24 (14.9) | -- | 13 (14.9) | 11 (22.5) |
HPV ISH and p16 double negative
HPV ISH and p16 double positive
Includes 22 patients missing both HPV ISH and p16 test results, 9 patients with p16 results and no HPV ISH results, 11 patients with HPV ISH results and no p16 results and 9 patients with positive p16 results and negative HPV ISH results
24% missing overall (28 out of 116)
1 patient could not be staged
Includes patients who received other treatments (Radiation + Chemo: 4 HPV-driven, 3 HPV-indeterminant; Surgery + Radiation + Chemo: 1 HPV-indeterminate)
Sensitivity and specificity of HPV16 serology for HPV-OPC
Of the 6 HPV16 proteins assessed, HPV16E6 antibodies were the most sensitive; 78 out of 87 HPV-OPC patients were HPV16E6 seropositive; sensitivity 89.7% (95%CI:81.3%–95.2%, Table 2). 24 out of 25 HPV-negative OPC patients were HPV16E6 seronegative, specificity 96.0% (95%CI:79.6%–99.9%). Sensitivity and specificity of the other HPV16 antibodies ranged from 35.6% to 83.9% and 80.0% to 96.0%, respectively. Sensitivity of HPV serology for HPV-OPC increased when HPV16 seropositivity was expanded to include seropositivity for ≥3 early proteins; sensitivity 93.1% (95%CI:85.6%–97.4%). Of the 9 HPV-OPCs without detectable HPV16E6 antibodies; 8 were seropositive against at least 1 other HPV16 E protein, 4 were seropositive against ≥1 E6 proteins from oncogenic non-HPV16 types, and 1 patient was seronegative for all HPV proteins tested. The 1 HPV16E6 seropositive patient with an HPV-negative tumor was seroreactive against all HPV16 proteins evaluated (Supplemental Table 1). Sensitivity for HPV16E6 seropositivity for OPC (regardless of HPV status) was 67.7% (95%CI:59.9%–74.8%, Supplemental Table 2).
Table 2.
Sensitivity and specificity of HPV16 serology for HPV-OPC
| Tumor Status (N=112) | ||||
|---|---|---|---|---|
|
|
||||
| HPV Serology | HPV-OPC1 (N=87) |
HPV-Negative2 (N=25) |
Sensitivity (95% CI) | Specificity (95%CI) |
| N (%) | N (%) | |||
| HPV16E6 | ||||
| Seropositive | 78 (89.7) | 1 (4.0)3 | 89.7% (81.3%–95.2%) | 96.0% (79.6%–99.9%) |
| Seronegative | 9 (10.3)4 | 24 (96.0) | ||
| HPV16 E7 | ||||
| Seropositive | 62 (71.3) | 5 (20.0) | 71.3% (60.6%–80.5%) | 80.0% (59.3%–93.2%) |
| Seronegative | 25 (28.7) | 20 (80.0) | ||
| HPV16 L1 | ||||
| Seropositive | 61 (70.1) | 1 (4.0) | 70.1% (59.4%–79.5%) | 96.0% (79.6%–99.9%) |
| Seronegative | 26 (29.9) | 24 (96.0) | ||
| HPV16 E1 | ||||
| Seropositive | 56 (64.4) | 1 (4.0) | 64.4% (53.4%–73.9%) | 96.0% (79.6%–99.9%) |
| Seronegative | 31 (35.6) | 24 (96.0) | ||
| HPV16 E2 | ||||
| Seropositive | 73 (83.9) | 2 (8.0) | 83.9% (74.4%–90.9%) | 92.0% (74.0%–99.0%) |
| Seronegative | 14 (16.1) | 23 (92.0) | ||
| HPV16 E4 | ||||
| Seropositive | 31 (35.6) | 2 (8.0) | 35.6% (25.6%–46.6%) | 92.0% (74.0%–99.0%) |
| Seronegative | 56 (64.4) | 23 (92.0) | ||
| HPV16 E6 and/or 3 Early Proteins | ||||
| Seropositive | 81 (93.1) | 1 (4.0) | 93.1% (85.6%–97.4%) | 96.0% (79.6%–99.9%) |
| Seronegative | 6 (6.9) | 24 (96.0) | ||
HPV ISH and p16 double positive
HPV ISH and p16 double negative
Patient seropositive for all 6 HPV16 proteins (see Supplemental Table 1)
8 out of 9 patients seroreactive against ≥1 other HPV16 proteins, 4 out of 9 seroreactive against E6 proteins from non-HPV16 types (Supplemental Table 1)
Sensitivity and specificity of HPV16 serology for HPV16-OPC
HPV16 type-specific DNA/RNA testing was performed on all recovered tumors (total: 86 out of 161; 49 out of 87 HPV-OPC and 16 out of 25 HPV-negative). Of the 49 retrieved HPV-OPCs, 35 had concordant positive HPV16 DNA/RNA results (HPV16-OPC), 3 had concordant negative HPV16 DNA/RNA results (HPV16-negative) and 11 had invalid and/or discordant test results (HPV16-indeterminate). Of the 16 retrieved HPV-negative tumors, 13 were HPV16-negative, 1 was HPV16-OPC and 2 were HPV16-indeterminate (Supplemental Figure 2). Thirty-four out of 36 HPV16-OPC patients were HPV16E6 seropositive, sensitivity 94.4% (95%CI:81.3%–99.3%). Fifteen out of 16 HPV16-negative OPC patients were HPV16E6 seronegative; specificity 93.8% (95%CI:69.8%–99.8%, Supplemental Figure 2). When considering all retrieved tumors regardless of prior p16 IHC and HPV ISH testing (N=64) sensitivity and specificity estimates were 93.3% (95%CI:81.7%–98.6%) and 94.7% (95%CI:74.0%–99.9%), respectively, (Table 3). Sensitivity of HPV16E6 antibodies by varying definitions of HPV tumor status are presented in Supplemental Table 3.
Table 3.
Sensitivity and specificity of HPV16 serology defining HPV16 tumor status by concordant HPV16 DNA and HPV16 RNA test results. Sensitivity and specificity were calculated among the: a) retrieved OPC tumors with prior clinical HPV testing and concordant HPV16 DNA/RNA results (N=52) and b) all retrieved OPC tumors regardless of prior clinical HPV testing with concordant HPV16 DNA/RNA results (N=64).
| HPV16 Tumor Status | ||||
|---|---|---|---|---|
|
|
||||
| HPV Serology | HPV16-OPC1 | HPV16-Negative2 | Sensitivity (95% CI) | Specificity (95% CI) |
| Retrieved OPC tumors with prior clinical HPV testing | ||||
| HPV16E6 | N (%) | N (%) | ||
| Seropositive | 34 (94.4) | 1 (6.2) | 94.4% (81.3%–99.3%) | 93.8% (69.8%–99.8%) |
| Seronegative | 2 (5.6) | 15 (93.8) | ||
| Retrieved OPC tumors regardless of prior clinical HPV testing | ||||
| HPV16E6 | N (%) | N (%) | ||
| Seropositive | 42 (93.3) | 1 (5.3) | 93.3% (81.7%–98.6%) | 94.7% (74.0%–99.9%) |
| Seronegative | 3 (6.7) | 18 (94.7) | ||
HPV16 DNA and HPV16 RNA dual positive
HPV16 DNA and HPV16 RNA dual negative
Of the 10 patients with discordant HPV16E6 serology and clinical HPV tumor status, 6 tumors were retrieved; all originated from HPV-OPC patients who were HPV16E6 seronegative. One tumor was HPV16-driven, two were HPV16 DNA positive and 3 were driven by other HPV types, Supplemental Table 1.
Sensitivity and specificity of HPV E6 serology for non-HPV16-driven tumors was also assessed. Of 3 HPV18-OPCs, 1 was HPV18E6 seropositive (sensitivity: 33%) and 78 out of 79 HPV18-negatives were HPV18E6 seronegative (specificity: 99%). Of 2 HPV33-OPCs, all were HPV33E6 seropositive (sensitivity: 100%) and of 80 HPV33-negative OPCs, 50 were HPV33E6 seronegative (specificity: 63%).
HPV16 Antibody Levels and Risk of Recurrence
Eleven (12.6%) out of 87 HPV-OPC patients recurred (6 local/regional recurrences; 5 distant metastases; median follow-up 3.6 years [IQR:2.1–5.2]). Pre-treatment HPV16E6 seropositivity was associated with an 86% reduced risk of local/regional recurrence; hazard ratio (HR):0.14 (95%CI:0.03–0.68), P=0.015. No association between pre-treatment HPV16E6 seropositivity and distant metastasis or overall survival was observed (Table 4). A similar HR estimate was observed for local/regional recurrence when restricting to patients with HPV16-OPC, although due to the smaller sample size, it was not significant (P=0.065).
Table 4.
Association between pre-treatment HPV16E6 seropositivity and prognosis among the 87 patients with HPV-OPC tumors (unadjusted Hazard Ratios).
| Outcome | Hazard Ratio (95% CI)* | P-Value |
|---|---|---|
| Local, regional, distant recurrence | 0.36 (0.09–1.35) | 0.129 |
| Local/regional recurrence | 0.14 (0.03–0.68) | 0.015 |
| Metastasis | Not Estimable | -- |
| Overall Survival | 0.93 (0.21–4.16) | 0.929 |
Unadjusted hazard ratios
Change in HPV16E6 antibodies levels was evaluated among 53 patients with serial samples who were HPV16E6 seropositive at diagnosis. HPV16E6 MFI levels did not significantly decrease over time (Figure 1); unadjusted estimate −161.2 MFI/year (95%CI:−724.6–402.1), P=0.575. Seven of the 53 patients recurred (5 local/regional recurrences and 2 distant metastases). Change in HPV16E6 levels post-treatment were not associated with recurrence of any type, all P-values>0.05 (Supplemental Figure 3).
Figure 1.
Change in HPV16E6 MFI levels post-treatment among the 53 OPC patients who were HPV16E6 seropositive at diagnosis and whom had a serial post-treatment serum sample. The red regression line denotes the average change in antibody levels over time.
DISCUSSION
We used a highly specific multiplex serology assay and both the clinical and research gold standard methods for assigning HPV tumor status to assess the sensitivity of HPV16E6 serology for HPV-OPC and HPV16-OPC. Using the clinical gold standard method of concordant p16 IHC and HPV ISH testing, sensitivity and specificity of HPV16E6 antibodies for HPV-OPC were high; 89.7% and 96.0%, respectively. Sensitivity was further increased when considering tumors driven specifically by HPV16, as defined by the research gold standard method of concordant HPV16 DNA/RNA testing; sensitivity: 93.3%. Although we did not observe a significant decrease in HPV16E6 antibody levels post-treatment nor an association between change in antibody levels post-treatment and risk of recurrence, HPV-OPC patients who failed to mount an HPV16E6 antibody response at diagnosis were significantly more likely to experience a local/regional recurrence. These results suggest that HPV16E6 antibodies may be a potential biomarker for both HPV-OPC diagnosis and prognosis.
This is the largest study to date to assess the sensitivity of HPV16E6 antibodies for HPV-OPC. While, several small studies (4 to 66 HPV-OPC tumors) reported the sensitivity of HPV16E6 serology for HPV-OPC, many relied on sub-optimal methods for assigning HPV tumor status and/or unvalidated HPV16E6 antibody assays. Only 5 studies[9, 10, 24–26] measured HPV16E6 antibody levels using the same highly specific multiplex serology assay as the prospective studies by Kreimer et al that reported the initial discovery and validation of HPV16E6 antibodies as an early marker for OPC[3, 4]. In the largest of these studies, Holzinger et al reported that 63 out of 66 HPV16-OPCs evaluated (defined using the research gold standard method of dual HPV16DNA/RNA positivity) were HPV16E6 seropositive (sensitivity=96%)[9]. Taken together, these results as well as our own suggest that the vast majority of patients with HPV16-OPC mount an HPV16E6 antibody response by the time of diagnosis.
Five small studies assessed changes in HPV16E6 antibody levels post-treatment among HPV-OPC patients[5–7, 10, 27]. All 5 reported HPV16E6 antibody levels decreased post-treatment; 2 reported an association between stable or increasing HPV16E6 antibody levels and recurrence[5, 7]. In contrast, we did not observe a decrease in HPV16E6 MFI levels post-treatment nor an association between change in HPV16E6 levels post-treatment and recurrence. This may be due to differences in serological assays and/or the small number of recurrences among our HPV16E6 seropositive patients.
We did however observe an association between pre-treatment HPV16E6 seronegativity and increased risk of recurrence. Four prior studies evaluated pre-treatment HPV16E6 levels and risk of recurrence; all reported different conclusions[7, 8, 10, 27]. A study of 43 HPV-OPC patients found that a log10-unit increase in pre-treatment HPV16E6 antibody level was associated with an increased risk of recurrence (HR:5.42, 95%CI:1.1–25.7)[7]. In contrast, a study of 96 HPV-OPC patients found that pre-treatment HPV16E6 seropositivity was associated with a 70%-decreased risk of recurrence or death (HR:0.3, 95%CI:0.1–0.9)[8]. Finally, two found no association between HPV16E6 antibody levels and recurrence[10, 27]. These studies are difficult to compare given the differences in serologic assays and methods used for assigning HPV tumor status. A limitation of the previous studies is that they all relied on non-gold standard methods for assigning HPV tumor status, which may have resulted in misclassification. Furthermore, 3 of the 4 studies used ELISA-based assays to measure HPV16E6 antibodies[7, 8, 27], which may have lower sensitivity and specificity than the multiplex serologic assay used in this current analysis. Dahlstrom et al previously reported a sensitivity and specificity of 79% and 56%, respectively, of HPV16E6 seropositivity for HPV-OPC as measured by ELISA[8].
A strength of our study was our large sample size and the use of the HPV serologic assay that has shown the highest sensitivity and specificity to date, as well as accepted gold standard methods for assigning HPV tumors status. By using gold-standard methods, we reduced the potential for misclassification and more precisely assessed the performance of the HPV16E6 biomarker. However, we had limited power to evaluate change in post-treatment HPV16E6 antibodies and risk of recurrence both due to the low recurrence rate and the limited number of serial samples. Additionally, due to the above-mentioned reasons, we could not conduct multivariable Cox regressions controlling for important baseline characteristics such as smoking, stage and treatment.
Taken together these findings may inform efforts for early detection, treatment and post-treatment surveillance of HPV-OPC. Our finding that HPV16E6 antibodies are present in the majority of patients with HPV-OPC at the time of diagnosis, coupled with the prospective findings that HPV16E6 antibodies are induced more than 10 years prior to diagnosis suggest that HPV16E6 antibodies may have potential clinical utility for early diagnosis of HPV-OPC. These results also suggest that HPV16E6 antibodies may have potential utility for pre-treatment risk stratification, allowing patients at low-risk for recurrence to be safely de-escalated and for high-risk patients to receive more intensive post-treatment surveillance. While the multiplex serology assay is currently for research use only, an ELISA-based could be developed that would allow for HPV16E6 antibody detection within standard hospital based laboratories. However, it is important to note that additional larger studies using gold standard methods for determination of HPV serology and HPV tumor status are needed to confirm these findings before moving to prospective trials.
Supplementary Material
Acknowledgments
Funding Information
This study was supported by the Intramural Research Program of the US National Institutes of Health (National Cancer Institute, Division of Cancer Epidemiology and Genetics), and the US National Institute of Health grants R01 CA206517, P50 CA097190, T32 CA060397 and the UPMC Tumor Microenvironment/Virus Oncology Center.
Footnotes
Conflicts of Interest
The authors have no conflicts of interest to declare.
Author Contributions
Conceptualization: All authors
Data curation: All authors
Formal analysis: All authors
Funding acquisition: KLK, ARK, RLF
Investigation: All authors
Project Administration: KLK, RLF, TW, MP
Resources: RLF, MP, TW
Supervision: ARK, RLF, MP, TW
Validation: RLF, MP, TW
Visualization: KLK
Writing (original draft): KLK
Writing (review & editing): All authors
References
- 1.Chaturvedi AK, Engels EA, Pfeiffer RM, et al. Human papillomavirus and rising oropharyngeal cancer incidence in the United States. J Clin Oncol. 2011;29(32):4294–301. doi: 10.1200/JCO.2011.36.4596. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Nguyen NP, Sallah S, Karlsson U, et al. Combined chemotherapy and radiation therapy for head and neck malignancies: quality of life issues. Cancer. 2002;94(4):1131–41. doi: 10.1002/cncr.10257. [DOI] [PubMed] [Google Scholar]
- 3.Kreimer AR, Johansson M, Waterboer T, et al. Evaluation of human papillomavirus antibodies and risk of subsequent head and neck cancer. J Clin Oncol. 2013;31(21):2708–15. doi: 10.1200/JCO.2012.47.2738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kreimer AR, Johansson M, Yanik EL, et al. Kinetics of the Human Papillomavirus Type 16 E6 Antibody Response Prior to Oropharyngeal Cancer. J Natl Cancer Inst. 2017;109(8) doi: 10.1093/jnci/djx005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Koslabova E, Hamsikova E, Salakova M, et al. Markers of HPV infection and survival in patients with head and neck tumors. Int J Cancer. 2013;133(8):1832–9. doi: 10.1002/ijc.28194. [DOI] [PubMed] [Google Scholar]
- 6.Rubenstein LM, Smith EM, Pawlita M, et al. Human papillomavirus serologic follow-up response and relationship to survival in head and neck cancer: a case-comparison study. Infect Agent Cancer. 2011;6:9. doi: 10.1186/1750-9378-6-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Fakhry C, Qualliotine JR, Zhang Z, et al. Serum Antibodies to HPV16 Early Proteins Warrant Investigation as Potential Biomarkers for Risk Stratification and Recurrence of HPV-Associated Oropharyngeal Cancer. Cancer Prev Res (Phila) 2016;9(2):135–41. doi: 10.1158/1940-6207.CAPR-15-0299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dahlstrom KR, Anderson KS, Cheng JN, et al. HPV Serum Antibodies as Predictors of Survival and Disease Progression in Patients with HPV-Positive Squamous Cell Carcinoma of the Oropharynx. Clin Cancer Res. 2015;21(12):2861–9. doi: 10.1158/1078-0432.CCR-14-3323. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Holzinger D, Wichmann G, Baboci L, et al. Sensitivity and specificity of antibodies against HPV16 E6 and other early proteins for the detection of HPV16-driven oropharyngeal squamous cell carcinoma. Int J Cancer. 2017 doi: 10.1002/ijc.30697. [DOI] [PubMed] [Google Scholar]
- 10.Zhang Y, Waterboer T, Haddad RI, et al. Human papillomavirus (HPV) 16 antibodies at diagnosis of HPV-related oropharyngeal cancer and antibody trajectories after treatment. Oral Oncol. 2017;67:77–82. doi: 10.1016/j.oraloncology.2017.02.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Waterboer T, Sehr P, Pawlita M. Suppression of non-specific binding in serological Luminex assays. J Immunol Methods. 2006;309(1–2):200–4. doi: 10.1016/j.jim.2005.11.008. [DOI] [PubMed] [Google Scholar]
- 12.Waterboer T, Sehr P, Michael KM, et al. Multiplex human papillomavirus serology based on in situ-purified glutathione s-transferase fusion proteins. Clin Chem. 2005;51(10):1845–53. doi: 10.1373/clinchem.2005.052381. [DOI] [PubMed] [Google Scholar]
- 13.Ribeiro KB, Levi JE, Pawlita M, et al. Low human papillomavirus prevalence in head and neck cancer: results from two large case-control studies in high-incidence regions. Int J Epidemiol. 2011;40(2):489–502. doi: 10.1093/ije/dyq249. [DOI] [PubMed] [Google Scholar]
- 14.Sitas F, Egger S, Urban MI, et al. InterSCOPE study: Associations between esophageal squamous cell carcinoma and human papillomavirus serological markers. J Natl Cancer Inst. 2012;104(2):147–58. doi: 10.1093/jnci/djr499. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Clifford GM, Shin HR, Oh JK, et al. Serologic response to oncogenic human papillomavirus types in male and female university students in Busan, South Korea. Cancer Epidemiol Biomarkers Prev. 2007;16(9):1874–9. doi: 10.1158/1055-9965.EPI-07-0349. [DOI] [PubMed] [Google Scholar]
- 16.Singhi AD, Westra WH. Comparison of human papillomavirus in situ hybridization and p16 immunohistochemistry in the detection of human papillomavirus-associated head and neck cancer based on a prospective clinical experience. Cancer. 2010;116(9):2166–73. doi: 10.1002/cncr.25033. [DOI] [PubMed] [Google Scholar]
- 17.Argiris A, Heron DE, Smith RP, et al. Induction docetaxel, cisplatin, and cetuximab followed by concurrent radiotherapy, cisplatin, and cetuximab and maintenance cetuximab in patients with locally advanced head and neck cancer. J Clin Oncol. 2010;28(36):5294–300. doi: 10.1200/JCO.2010.30.6423. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Halec G, Schmitt M, Dondog B, et al. Biological activity of probable/possible high-risk human papillomavirus types in cervical cancer. Int. J. Cancer. 2013;132(1):63–71. doi: 10.1002/ijc.27605. [DOI] [PubMed] [Google Scholar]
- 19.Schmitt M, Bravo IG, Snijders PJ, et al. Bead-based multiplex genotyping of human papillomaviruses. J Clin Microbiol. 2006;44(2):504–12. doi: 10.1128/JCM.44.2.504-512.2006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Schmitt M, Dondog B, Waterboer T, et al. Homogeneous amplification of genital human alpha papillomaviruses by PCR using novel broad-spectrum GP5+ and GP6+ primers. J Clin Microbiol. 2008;46(3):1050–9. doi: 10.1128/JCM.02227-07. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Prentice RL, Kalbfleisch JD, Peterson AV, Jr, et al. The analysis of failure times in the presence of competing risks. Biometrics. 1978;34(4):541–54. [PubMed] [Google Scholar]
- 22.Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics. 1986;42(1):121–30. [PubMed] [Google Scholar]
- 23.Zeger SL, Liang KY, Albert PS. Models for longitudinal data: a generalized estimating equation approach. Biometrics. 1988;44(4):1049–60. [PubMed] [Google Scholar]
- 24.Anantharaman D, Gheit T, Waterboer T, et al. Human papillomavirus infections and upper aero-digestive tract cancers: the ARCAGE study. J Natl Cancer Inst. 2013;105(8):536–45. doi: 10.1093/jnci/djt053. [DOI] [PubMed] [Google Scholar]
- 25.Lopez RV, Levi JE, Eluf-Neto J, et al. Human papillomavirus (HPV) 16 and the prognosis of head and neck cancer in a geographical region with a low prevalence of HPV infection. Cancer Causes Control. 2014;25(4):461–71. doi: 10.1007/s10552-014-0348-8. [DOI] [PubMed] [Google Scholar]
- 26.Smith EM, Rubenstein LM, Haugen TH, et al. Complex etiology underlies risk and survival in head and neck cancer human papillomavirus, tobacco, and alcohol: a case for multifactor disease. J Oncol. 2012;2012:571862. doi: 10.1155/2012/571862. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Spector ME, Sacco AG, Bellile EL, et al. E6 and E7 antibody levels are potential biomarkers of recurrence in patients with advanced stage human papillomavirus positive oropharyngeal squamous cell carcinoma. Clin Cancer Res. 2016 doi: 10.1158/1078-0432.CCR-16-1617. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.