Degree project, 30 ECTS 2020/05/22
The prognostic utility of p16 overexpression and
Human Papillomavirus DNA presence in base of
tongue cancer patients: A retrospective cohort
study in Region Örebro County.
Version 2
Author:
Max Waenerlund, Bachelor of Medical Science School of Medical Sciences, Örebro University
Supervisors:
Anna Oldaeus-Almerén, MD Department of Otolaryngology, Örebro University Hospital
Gisela helenius, PhD Faculty of Medicine and Health, Örebro University. Department of Laboratory Medicine, Örebro University Hospital.
Word count
Abstract: 245 Manuscript: 3477
Abstract
Background: The overall good prognosis for Human Papillomavirus (HPV) driven base of the
tongue cancer has prompted an increasing interest in whether this group could benefit from a less aggressive treatment regime. Different studies have drawn different conclusions as to which laboratory test should be used to identify these patients, using the surrogate marker p16, analyzing for HPV presence or both.
Aim: The main purpose of this study was to investigate the presence of HPV-DNA and p16 in
base of tongue cancer patients and their respective prognostic value, both used individually as well as combined.
Material and methods: This was a retrospective cohort study consisting of 40 patients
diagnosed with base of tongue cancer. The follow-up period was 5 years. Survival analysis was performed both depending on the combined results from the p16 immunohistochemistry analysis and the HPV DNA PCR, as well as separately.
Results: Five-year survival rates were 73.9% for p16(+) and 17,6% for p16(-) subjects
(p<0.001), 60.7% for HPV-DNA(+) and 25.0% for HPV-DNA(-) subjects (p=0.025). Five-year survival rates when combining p16 and HPV-DNA were 73.9% for p16(+)/HPV DNA(+), 25.0% for p16(-)/HPV DNA(-) and 0.0% for p16(-)/HPV DNA(+) (p<0.001).
Conclusion: Our results add to previous research that p16 is a strong predictor of prognosis for
base of tongue cancer patients, and could have the clinical implication of serving as a reliable tool for clinicians when determining prognosis and identifying patients who could benefit from treatment de-escalation in the future.
Abbreviations:
AJCC: American Joint Committee on Cancer HR: Hazard Ratio
HPV: Human Papillomavirus
HPV-AF: Human Papillomavirus Attributable Fraction IHC: Immunohistochemistry
ISH: In Situ Hybridization KM: Kaplan-Meier
OPSCC: Oropharyngeal Squamous Cell Cancer PCR: Polymerase Chain Reaction
Introduction
Human papilloma virus (HPV) is a known causative agent of cancer in the oropharynx (tonsils and base of the tongue). There are over a dozen different oncogenic HPV types, the most common being HPV 16 and 18 who account for 30-40% of all cancers in the oropharynx [1, 2]. HPV is a small, naked capsule, DNA virus. It is primarily a sexually transmissible infection. It reaches its target cells, keratinocytes or basal cells, through small cracks in the skin or mucosa. When an oncogenic HPV enters the cell, it incorporates its DNA into the DNA of the host cell. The viral genes of importance regarding oncogenesis are E6 and E7 which then are transcribed to their respective proteins. The E6 and E7 proteins then target the important growth suppressor proteins p53 and the p105 retinoblastoma gene product and render them inactive. This results in an uncontrolled growth of the cell and allows for the virus to spread further. HPV also has a way of hiding in the cells, making it undetectable for the immune system, and can stay latent for longer periods of time [3].
Oropharyngeal squamous cell cancer (OPSCC) has increased in recent decades, and it has been described as an epidemic caused by an increased incidence of HPV [4–7]. The global incidence of OPSCC was 96,000 in 2012 and the HPV-attributable fraction (HPV-AF) was 29,000 (30%). The HPV-AF is higher in more developed countries such as Europe and North America, around 40% [8]. It has also been speculated that the number of driven OPSCCs will surpass HPV-driven cervical cancers by 2020 [9]. HPV negative OPSCC is associated with smoking and/or excessive alcohol consumption and afflicts older patients. HPV positive OPSCC patients are younger (54 years median age) and do not necessarily have any of the other risk factors, compared to those who are HPV negative[10]. HPV positive cancer patients seem to respond better to treatment, and as such has a better prognosis than those who are HPV negative [11– 14]. The 5-year survival rate for HPV positive OPSCC is 80% while it is 25-50% for HPV negative OPSCC patients [15]. Alcohol and tobacco use seem to worsen the prognosis for HPV positive OPSCC patients [16, 17]. Both higher education and living in rural areas seem to increase the risk for acquiring HPV-driven cancer [18, 19].
It is now acknowledged that HPV positive OPSCC is a biologically different form of cancer than HPV negative OPSCC. These two forms have, as mentioned above, different prognosis and could also entail different approaches to treatment. Research has suggested that HPV negative tumors seem to be less sensitive to chemotherapy and radiotherapy, and could as such benefit from surgery as the primary tool, while HPV positive tumors seem to benefit more from
radiotherapy and possibly concurrent chemotherapy [20, 21]. Furthermore, in order to mitigate the risk of late side effects with treatment/radiotherapy, for instance osteoradionecrosis and trismus, there has been an increasing focus on whether HPV positive cancer patients could benefit from a less aggressive treatment regime. As such, some treatment de-escalation trials have been introduced [22]. However, the choice of treatment in standard care today does not differ depending on HPV status, consisting mainly of radiotherapy and sometimes surgery (primarily neck dissection) and chemotherapy [23].
In 2016, the AJCC/UICC published their 8th edition of the TNM-classification of oropharyngeal cancer. This new system takes HPV into account by using p16 overexpression as a surrogate for HPV presence. The overall implication of this is that if a tumor has p16 overexpression, it will be considered to be in a lower stage than if it did not have p16 overexpression. P16 is an antiproliferative protein that slows down the process of placing the cell in the S phase in the cell cycle. P16 is overexpressed in most HPV-induced oropharyngeal cancers, as a result of the HPV-DNA integration into the host cell and the resulting production of its oncogenic proteins mentioned above [20, 24].
P16 overexpression is analyzed by immunohistochemistry and is regarded as a measurement of transcriptionally active HPV [21]. Other ways of determining the presence of HPV is analysis of HPV-RNA and DNA by use of PCR and in situ hybridization (ISH) [20]. The rationale behind using p16 as a surrogate instead of analysis of HPV-DNA or RNA is its affordability and accessibility, and the technical simplicity of the analysis. Previous studies have shown conflicting evidence regarding the accuracy of p16 as a standalone test for HPV presence. Some have indicated that p16 is not only an accurate marker, but is also of superior significance when it comes to predicting the prognosis, while other studies have found that a subset of patients who are p16 positive but HPV negative have a worse prognosis, and also highlighting the problems with subjective interpretations of immunohistochemistry [21, 25–30].
In sum, some patients with base of tongue cancer have a better prognosis than others. It is important to be able to identify these patients seeing that they could benefit from a less aggressive treatment regime in order to reduce the risk of long-term side effects from treatment. While some studies have suggested that p16 is the most important predictor of prognosis, other have suggested that patients who are p16 positive but HPV-DNA negative have a worse prognosis. These contrasting evidences highlight the need of further research into how one
should interpret the results from HPV-DNA analysis and p16 immunohistochemistry when determining prognosis and tailoring treatment for individual patients in the future.
Aim
The main purpose of this study was to investigate the presence of HPV-DNA and p16 in base of tongue cancer patients and their respective prognostic value, both used individually as well as combined. An underlying aim of this study was to explore whether gender and socioeconomic factors affect the risk of acquiring HPV-driven base of tongue cancer.
Material and Methods
Study design and population
This was a retrospective cohort study. The inclusion criterion was confirmed invasive squamous cell carcinoma of the base of tongue, ICD C01.9, in Region Örebro County (period 1988-2014, N=42). Study subjects were excluded if their medical records could not be found or if analysis for p16 or HPV DNA could not be made. The follow-up period was 5 years. The patients were identified by a register of all head and neck cancer patients diagnosed and treated at Örebro University Hospital, a register which was established in 1988 and has been continually updated ever since. All patients in the cohort had their cases brought up at a multidisciplinary round prior to treatment.
The medical records of each study subject were examined between January and February 2020. Data extracted from the medical records were: time of multidisciplinary round; time of death; smoking (never smoker, active smoker, previous smoker as defined by at least 1 year smoke free); alcohol overconsumption (never, active and previous overconsumption); age upon diagnosis; highest acquired education (primary school, high school, university); rural vs urban living situation; living with someone or living alone.
Age were grouped in two categories, above and below 60 years of age. This stratification was used since the median age for HPV-driven oropharyngeal cancer is below 60 years and above 60 years for non-HPV-driven oropharyngeal cancer. This stratification has been used in previous similar studies [17, 19].
HPV DNA and p16 analysis
Additionally, previously collected tissue samples from diagnosis were gathered from the Örebro tissue bank. All samples were formalin-fixed, paraffin embedded and analyzed for HPV-DNA presence by use of PCR, and p16 overexpression by use of immunohistochemistry.
HPV-DNA extraction and PCR
The QIAamp DNA FFPE Tissue Kit from Qiagen was used in order to purify the formalin-fixed and paraffin-embedded samples and acquire pure DNA. Then, PCR was performed using the anyplex II HPV28 kit from Seegene which can identify 19 high risk and 9 low risk HPV types.
P16 Immunohistochemistry
Each sample was colored for p16 overexpression using the CINtec® Histology
immunohistochemistry test from Roche. A 70% coloring of active tumor tissue was used as the criterion for positivity. Each sample was compared to a negative control in order to verify the validity of the test. Analyses were performed by a pathologist, MW and an otolaryngologist. A pathologist was consulted in difficult and borderline cases.
Statistics
First, the survival was analyzed depending on p16 and HPV DNA separately. Second, the survival was analyzed depending on the combined results from the HPV-DNA PCR results and p16 immunohistochemistry, yielding four groups: P16 positive(+)/HPV-DNA(+); P16(+)/HPV-DNA negative(-); P16(-)/HPV-DNA(-); P16(-)/HPV-DNA(+). Third, further univariate analysis was performed for smoking, alcohol overconsumption, gender and age.
Survival curves and survival rates were calculated using the Kaplan-meier (KM) method and the log rank test was used to determine statistically significant differences. To examine effect size and adjust for confounders, cox proportional hazards regression was used. Checking of the assumption of proportional hazard was made by creating time-by covariate interactions for each variable that would be adjusted for, in which case a p-value below 0.05 was the criterion for non-proportionality and thus the assumption deemed violated.
Active and previous alcohol overconsumption were merged post-hoc when adjusting for this factor.
Gender, living alone or with someone, urban vs rural living situation and education level were included in a binary logistic regression with HPV-DNA as the dependent variable in order to analyze whether any of those factors were overrepresented among the HPV-DNA(+) study subjects.
Missing data
Available-case analysis was implemented for univariate analysis and complete-case analysis for multivariate analysis. Missing data largely consisted of covariates that could not be found in the medical records such as risk factors and socioeconomic factors, largely due to old medical records and poor documentation for these factors. The missing data is as such treated as missing at random, and should not introduce any major bias in that regard.
Ethical considerations
This study was mainly made with the purpose of quality assurance at the department of otolaryngology, Örebro University Hospital. Clinical information was retrieved within the scope of an ongoing study at the department which has received authorization through an institutional review board-approved retrospective research protocol (ref 2015/548/2). Biobank approval was obtained.
Results
Forty-two patients were diagnosed and treated at Örebro University Hospital during the study period. Two were excluded because their medical records could not be found. As such, forty study subjects were included in the study. There was no loss to follow-up. There was missing data for some study subjects regarding smoking, alcohol overconsumption, education level and living status. Baseline characteristics as divided by the combined results from p16 and HPV-DNA are listed in table 1.
Of all study subjects, 57.5% were p16(+) and 70.0% were HPV-DNA(+). There were 23 study subjects in the p16(+)/HPV(+) group, 0 study subjects in the p16(+)/HPV(-) group, 12 study subjects in the p16(-)/HPV(-) group and 5 study subjects in the p16(-)/HPV(+) group.
Survival analysis
The overall 5-year survival rate was 50.0%. The 5-year survival rate was 73.9% for p16(+) study subjects and 17.6% for p16(-) study subjects, the difference was statistically significant (p<0.001 (log rank)). The 5-year survival rate was 60.7% for HPV-DNA(+) study subjects and
25.0% for HPV-DNA(-) study subjects, the difference was statistically significant (p=0.025 (log rank)).
Table 1 - Patients in each group according to combined Human papilloma virus (HPV) DNA results and p16 immunohistochemistry results. No study subject was p16(+)/ HPV DNA(-). Missing data in parenthesis.
Tot P16(+)/HPV(+) P16(-)/HPV(-) P16(-)/HPV(+) N 40 23 12 5 Age (median) 63.4 (SD 11.3) 58.6 68.7 72.6 Age category <60 years 16 (40.0%) 14 1 1 >60 years 24 (60.0%) 9 11 4 Man 30 (75.0%) 20 8 2 Woman 10 (25.0%) 3 4 3 Smoking (2) Never smoker 9 (23.7%) 8 0 1 Smoker 15 (39.5%) 7 6 2 Previous smoker 14 (36.8%) 8 4 2 Alcohol (4) No overconsumption 32 (88.9%) 19 9 4 Active Overconsumption 3 (8.3%) 1 1 1 Previous overconsumption 1 (2.8%) 1 0 0
Rural vs urban living (1)
Rural 24 (61.5%) 12 8 4
Urban 15 (38.5%) 11 3 1
Living alone or with someone (4)
Living alone 16 (44.4%) 9 5 2
Living with someone 20 (55.6%) 14 4 2
Education (12)
Primary school 8 (28.6%) 2 4 2
High school 14 (50.0%) 13 1 0
University 6 (21.4%) 5 1 0
No of events (death) 20 6 9 5
Alcohol overconsumption had a statistically significant correlation with poorer survival according to the log rank test (p=0.014). However, only three study subjects were labeled as active over consumers and only one as a previous consumer. Active smokers had worse survival than never-smokers, and previous smokers had worse survival compared to active smokers, however this was not statistically significant (p=0.052). Further univariate analysis of survival
depending on age and gender was made. Female gender and study subjects over 60 years of age had slightly poorer survival, but these differences were not statistically significant (fig. 1).
Figure 1 – Univariate analysis of overall 5-year survival depending on Human papilloma virus (HPV) DNA PCR results, p16 immunohistochemistry, age group, tobacco use and alcohol overconsumption. P-value from Log rank test included for each curve. N=number of study subjects. E=number of events (death from any cause).
HPV-DNA + (N=29, E=5) HPV-DNA – (N=12, E=5) N = 41 P = 0.063 HPV-DNA + (N=28, E=5) HPV-DNA – (N=12, E=5) N = 40 P = 0.073 P16 + (N=23, E=2) P16 – (N=17, E=8) N = 40 P = 0.004 <60 (N=16, E=3) >60 (N=24, E=7) N = 40 P = 0.392 No overconsumption (N=32, E=7) Active overconsumption (N=3, E=2) Previous overconsumption (N=1, E=0) N = 36 P = 0.171 HPV-DNA + (N=28, E=11) HPV-DNA – (N=12, E=9) N = 40 P = 0.025 P16 + (N=23, E=6) P16 – (N=17, E=14) N = 40 P < 0.001 <60 (N=16, E=6) >60 (N=24, E=14) N = 40 P = 0.247
Never smoker (N=9, E=1) N=38
Active smoker (N=15, E=8) P=0.052
Previous smoker (N=14, E=9)
Never (N=32, E=13) Active (N=3, E=3) Previous (N=1, E=1) N = 36 P = 0.014 Man (N=30, E=13) Woman (N=10, E=7) N = 40 P = 0.160
Neither p16, HPV-DNA, or any of the confounders showed a statistically significant interaction with time, allowing for the proceeding with a cox proportional hazards regression (table 2). When fitted in a univariate cox proportional hazards regression, there was more than a five-fold increase in risk of mortality for p16(-) study subjects as compared to p16(+) study subjects, which was statistically significant (HR 5.543; 95% CI, 2.095-14.665, p=0.001). HPV-DNA negativity was associated with almost a three-fold increase in risk of mortality as compared to HPV-DNA positivity, which was statistically significant (HR 2.618; 95% CI, 1.080-6.345, ,p=0.033) (table 2).
Table 2 - Cox proportional hazards regression with hazard ratios (HR), 95% confidence intervals (CI) and p-values, for p16 overexpression (+) as compared to no overexpression and Human papillomavirus (HPV) DNA positive (+) as compared to HPV DNA negative. Time by covariate interactions for all covariates in italics.
HR (Exp(B)) 95,0% CI for Exp(B) p-value
Lower Upper P16(-)1 5.543 2.095 14.665 0.001 HPV-DNA(-)1 2.618 1.080 6.345 0.033 P16(-)2 4.614 1.308 16.279 0.017 HPV-DNA(-)2 2.198 0.635 7.609 0.214 P16*time 1.001 0.937 1.070 0.974 HPV-DNA*time 1.013 0.951 1.078 0.697 Age*time 0.961 0.888 1.040 0.329 Gender*time 0.979 0.918 1.044 0.520 Alcohol overconsumption*time 1.091 0.956 1.244 0.196 Active smoking*time3 1.153 0.867 1.534 0.327 Previous smoking*time3 1.099 0.830 1.456 0.510 1 = univariate analysis. (N=40)
2 = adjusted for age, gender, smoking and alcohol overconsumption. (N=36) 3 = as compared to never-smokers.
When adjusted for age, gender, alcohol overconsumption and smoking, p16 negativity was still associated with poorer survival but the effect was decreased (HR 4.614; 95% CI, 1.308-16.279, p=0.017). When adjusted for the same confounders, HPV-DNA(-) study subjects also had slightly poorer survival as compared to HPV-DNA(+) study subjects, but the effect was decreased and not statistically significant (HR 2.198; 95% CI; 0.635-7.609, p=0.214).
KM curves showed that the group with the combined results p16(+)/HPV DNA(+) had the best survival (73.9% 5-year survival rate) followed by p16(-)/HPV DNA(-) (25.0% 5-year survival rate) and p16(-)/HPV DNA(+) (0.0% 5-year survival rate) respectively (fig. 2). The difference
in survival depending on these groups was statistically significant with a p-value from the log rank test less than 0.001. However, there was a marked difference in mean age between these groups as can be seen in table 1.
Figure 2 - 5-year overall survival depending on the combined results from p16 and human papilloma virus (HPV) DNA analysis. Blue line = p16+/HPV-DNA+. Red line = p16-/HPV-DNA-. Green line = p16-/HPV-DNA+. Result from log rank test was p<0.001.
The groups did not exhibit a statistically significant interaction with time, however p16(-)/HPV-DNA(+) had a near statistically significant correlation with time (P=0.096). When adjusted for confounders and compared to p16(+)/HPV-DNA(+), p16(-)/HPV-DNA(+) was still associated with statistically significant worse prognosis (HR 3.812; 95% CI 1.339-29.115, p=0.020), while p16(-)/HPV-DNA(-) was just above the significance margin (HR 6.243; 95% CI 0.958-15.161, p=0.057) (table 3). P16+/HPV-DNA + (N=23, E=6) P16-/HPV-DNA - (N=12, E=9) P16-/HPV-DNA+ (N=5, E=5) N = 40 P <0.001
Table 3 - Cox proportional hazards regression with hazard ratios (HR), 95% confidence intervals (CI) and p-values for study subjects with the combination of no p16 overexpression (-) and human papillomavirus (HPV) DNA negativity (-), as well as p16(-) and HPV DNA positivity
HR (Exp(B)) 95,0% CI for Exp(B) p-value
Lower Upper P16(-)/HPV-DNA(-)1,2 4.578 1.615 12.977 0.004 P16(-)/HPV-DNA(+)1,2 10.449 2.883 37.871 <0.001 P16(-)/HPV-DNA(-)2,3 3.812 0.958 15.161 0.057 P16(-)/HPV-DNA(+)2,3 6.243 1.339 29.115 0.020 P16(-)/HPV-DNA(-)*time2 1.003 0.938 1.071 0.940 P16(-)/HPV-DNA(+)*time2 1.217 0.966 1.532 0.096 1 = univariate analysis (N=40) 2 = compared to p16(+)/HPV-DNA(+)
3 = adjusted for age, gender, smoking and alcohol overconsumption (N=36)
The odds of acquiring HPV DNA positive base of tongue cancer was 21 times greater when having high school education as the highest education level compared to primary school. (OR 21.305; 95% CI 1.110-408.914; p=0.042), (table 5). There was no statistically significant increased risk for study subjects living in a rural area, living alone, university or depending on gender.
Table 4 – Multivariate binary logistic regression model on the correlation between risk factors and Human Papilloma Virus (HPV)-driven base of tongue cancer. N=28.
N Sig. OR (Exp(B)) 95% C.I.for EXP(B)
Lower Upper
Rural area (reference) 18
Urban area 10 .442 3.249 .161 65.505
Living alone (reference) 11
Living with someone 17 .154 8.110 .455 144.479
Elementary school (reference) 9 High school 14 .042 21.305 1.110 408.914 University 5 .999 1304804255.453 .000 . Woman (reference) 5 Man 23 .744 .563 .018 17.824
Discussion and conclusions
The main finding of this study was that both p16 positive and HPV-DNA positive cases had a better prognosis than their negative counterparts. However, p16 had a larger effect size and a higher degree of statistical significance as compared to HPV-DNA. When corrected for confounders p16 still had a statistically significant correlation with survival, while HPV-DNA did not. This suggests that p16 corresponds better with prognosis than HPV presence.
Furthermore, when p16 and HPV-DNA results were combined, HPV-DNA(+)/p16(+) was associated with the best survival, followed by HPV-DNA(-)/p16(-) and HPV-DNA(+)/p16(-) respectively, this difference was statistically significant. By looking at the survival curves, one could see that p16 positivity encompassed most cases with a good prognosis and p16 negativity encompassed most cases with a bad prognosis. HPV-DNA positivity however encompassed cases with both good and bad prognosis. The fact that p16(-)/HPV-DNA(+) subjects had the worst prognosis is surprising and interesting, but one should have in mind that this was the smallest group (N=5) with the highest mean age. When adjusting for confounders, there was still a statistically significant correlation with mortality for this group (i.e., p16(-)/HPV-DNA(+), but not for those who were double negative. It should be noted that the relatively large ammount of variables in relation to the sample size here weakens the statistical power. Making the adjustment model less valid.
Our results go in line with previous research suggesting that p16 is of superior significance when determining prognosis compared to HPV presence [21, 25, 28]. The interest in looking at survival when combining results from p16 and HPV DNA came from two recent studies where poor survival had been suggested for patients who are p16(+)/HPV-DNA(-), something that would question the use of p16 as a standalone test for prognosis in oropharyngeal cancer [29, 30]. Seeing that no study subject in our material had the combination p16(+)/HPV-DNA(-), we could not add any new knowledge concerning the suggested worse prognosis for those patients, other than HPV-DNA did not provide additional prognostic information than p16 alone in our material. One should note that those studies only performed HPV-DNA analysis after a positive p16 test, which means that they could not identify patients who were HPV-DNA(+) but p16(-), which is the group that had the worst prognosis in our material. Here, our study could offer some new knowledge. Therefore, by combining our results with previous findings, it seems that single positivity for either p16 or HPV is associated with a worse prognosis. This is an area calling for further research.
Less researched is the effect of smoking and alcohol consumption on survival in HPV(+) patients. Although some studies have suggested a worse survival for those patients [16, 17], we could not see a statistically significant effect of smoking on survival after diagnosis. However, this analysis was not stratified depending on HPV presence. This could be a manifestation of the small study population but could also mean that other mechanisms are at play that was not included in this study.
We found that there was an increased risk of HPV-associated base of tongue cancer among those who had high school as the highest acquired education as compared to elementary school. This goes in line with previous studies on the subject[18, 19]. While our study had three levels of education (primary school, high school and university) other studies have chosen two levels of education (high school or not). This makes comparisons somewhat difficult. Furthermore, the sample size for this analysis was low due to substantial missing data on education level (12 missing). This was due to old medical records with poor documentation on socioeconomic factors. This resulted in large odds ratios and wide confidence intervals. This makes it inappropriate to draw any conclusions from the odds ratios presented here. We can however conclude that there were signs of risk differences, which provides arguments for further research on bigger datasets.
The main limitation of this study was the small study population. This limits the external validity of the study in general, but in particular the results from the hazard ratios presented here, which should therefore be interpreted with caution. A manifestation of this was the generally wide confidence intervals. There were also missing data on alcohol (4 missing) and smoking habits (2 missing), which means that the multivariate analysis had a smaller sample size than the univariate analysis. This means that the univariate and multivariate analyses are not entirely consequential in relation to each other.
Additionally, age was adjusted for in a binary fashion, above and below 60. This grouping has been used in previous, similar studies; it is possible that another stratification could make the adjustment for age more accurate. Indeed, the group of study subjects who were p16(-) and HPV(+) had a high median age, but seeing as 4 of 5 were older than 60 years old, it is questionable that adjustment using the binary age variable provided an accurate adjustment for age in this particular group. Therefore, it is possible that age could stand for some, or all, of the effect on survival in the group that was p16(-)/HPV-DNA(+).
Finally, this study consisted of a cohort in a Swedish region with a relatively large portion living in rural areas. It is possible that the results would have been different in a more urban setting or in a non-occidental country.
Our results highlight the need for further research on the survival depending on the combination of p16 and HPV presence. Such research should first, include more patients. Due to the low incidence of base of tongue cancer, data collecting within a reasonable time period could be problematic. Therefore, a multicenter study would help achieve a sufficiently large patient population, while also providing a more demographically heterogenous population. One would then both get enhanced statistical power but also making adjustment for more confounders possible. Second, it would also be interesting to adjust for age using a different stratification. Third, due to the retrospective nature of the study our material includes patients from a large time span. This could mean that some differences in survival could be explained by changes in treatment protocols, changes in patient behavior or awareness of disease that could make people search medical advice in an earlier or later stage of the disease. Future studies could benefit from adjusting for date of diagnosis.
In conclusion, our results add to previous research that p16 is a strong predictor of prognosis in base of tongue cancer patients and could have the clinical implication of being a reliable contributor when identifying patients who could benefit from treatment de-escalation in the future. We encourage further research regarding survival depending on the combined results from p16 and HPV-DNA analyses on bigger datasets.
The authors of this study are in the midst of gathering the same type of data on tonsillar cancer patients from the same center and study period, with an aim to perform the same analyses. This study can as such be viewed as a pilot for a coming, bigger study.
Acknowledgments
I owe my deepest gratitude to my supervisors Anna Oldaeus-Almerén and Gisela Helenius. They have been essential in the creation of this thesis and have provided excellent support and patient guidance throughout the whole process.
I also express my sincere gratitude to our bioinformatician Anders Wirén for crucial assistance in the statistical analysis.
References
[1] Muñoz N, Bosch FX, de Sanjosé S, et al. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med 2003; 348: 518–527. [2] Muñoz N, Castellsagué X, de González AB, et al. Chapter 1: HPV in the etiology of
human cancer. Vaccine 2006; 24: S1–S10.
[3] Murray PR, Rosenthal KS, Pfaller MA. Medical Microbiology. Philadelphia: Elsevier, 2016.
[4] Marur S, D’Souza G, Westra WH, et al. HPV-associated head and neck cancer: a virus-related cancer epidemic. Lancet Oncol 2010; 11: 781–789.
[5] Deschler DG, Richmon JD, Khariwala SS, et al. The ‘new’ head and neck cancer patient-young, nonsmoker, nondrinker, and HPV positive: evaluation. Otolaryngol--Head Neck Surg Off J Am Acad Otolaryngol-Head Neck Surg 2014; 151: 375–380.
[6] Chaturvedi AK, Engels EA, Pfeiffer RM, et al. Human papillomavirus and rising oropharyngeal cancer incidence in the United States. J Clin Oncol Off J Am Soc Clin Oncol 2011; 29: 4294–4301.
[7] Näsman A, Attner P, Hammarstedt L, et al. Incidence of human papillomavirus (HPV) positive tonsillar carcinoma in Stockholm, Sweden: an epidemic of viral-induced carcinoma? Int J Cancer 2009; 125: 362–366.
[8] de Martel C, Plummer M, Vignat J, et al. Worldwide burden of cancer attributable to HPV by site, country and HPV type. Int J Cancer 2017; 141: 664–670.
[9] Reuschenbach M, Tinhofer I, Wittekindt C, et al. A systematic review of the HPV‐ attributable fraction of oropharyngeal squamous cell carcinomas in Germany. Cancer Med 2019; 8: 1908–1918.
[10] You EL, Henry M, Zeitouni AG. Human papillomavirus–associated oropharyngeal cancer: review of current evidence and management. Curr Oncol 2019; 26: 119–123. [11] Kimple RJ, Smith MA, Blitzer GC, et al. Enhanced radiation sensitivity in HPV-positive
head and neck cancer. Cancer Res 2013; 73: 4791–4800.
[12] Fakhry C, Zhang Q, Nguyen-Tan PF, et al. Human papillomavirus and overall survival after progression of oropharyngeal squamous cell carcinoma. J Clin Oncol Off J Am Soc Clin Oncol 2014; 32: 3365–3373.
[13] Ang KK, Harris J, Wheeler R, et al. Human Papillomavirus and Survival of Patients with Oropharyngeal Cancer. N Engl J Med 2010; 363: 24–35.
[14] Fakhry C, Westra WH, Li S, et al. Improved survival of patients with human papillomavirus-positive head and neck squamous cell carcinoma in a prospective clinical trial. J Natl Cancer Inst 2008; 100: 261–269.
[15] Näsman A, Du J, Dalianis T. A global epidemic increase of an HPV-induced tonsil and tongue base cancer – potential benefit from a pan-gender use of HPV vaccine. J Intern Med 2020; 287: 134–152.
[16] Alabi O, O’Neill JP. ‘Good cancer gone bad’: a narrative review of HPV oropharyngeal cancer and potential poor outcomes. Eur Arch Rhino-Laryngol Off J Eur Fed Oto-Rhino-Laryngol Soc EUFOS Affil Ger Soc Oto-Oto-Rhino-Laryngol - Head Neck Surg. Epub ahead of print 24 April 2020. DOI: 10.1007/s00405-020-05991-z.
[17] Hafkamp HC, Manni JJ, Haesevoets A, et al. Marked differences in survival rate between smokers and nonsmokers with HPV 16-associated tonsillar carcinomas. Int J Cancer 2008; 122: 2656–2664.
[18] Zahnd WE, Rodriguez C, Jenkins WD. Rural-Urban Differences in Human Papillomavirus-associated Cancer Trends and Rates. J Rural Health Off J Am Rural Health Assoc Natl Rural Health Care Assoc 2019; 35: 208–215.
[19] Benard VB, Johnson CJ, Thompson TD, et al. Examining the association between socioeconomic status and potential human papillomavirus-associated cancers. Cancer
2008; 113: 2910–2918.
[20] Pannone G, Santoro A, Papagerakis S, et al. The role of human papillomavirus in the pathogenesis of head & neck squamous cell carcinoma: an overview. Infect Agent Cancer 2011; 6: 4.
[21] Lewis JJr. p16 Immunohistochemistry as a standalone test for risk stratification in oropharyngeal squamous cell carcinoma. - PubMed - NCBI. Head Neck Pathol 2012; 75– 82.
[22] O’Sullivan B, Huang SH, Siu LL, et al. Deintensification Candidate Subgroups in Human Papillomavirus–Related Oropharyngeal Cancer According to Minimal Risk of Distant Metastasis. J Clin Oncol 2013; 31: 543–550.
[23] Huvud- och halscancer. Nationellt vårdprogram.,
https://www.cancercentrum.se/globalassets/cancerdiagnoser/huvud-och-hals/vardprogram/nationellt-vardprogram-huvud-halscancer.pdf (2019).
[24] Sano D, Oridate N. The molecular mechanism of human papillomavirus-induced carcinogenesis in head and neck squamous cell carcinoma. Int J Clin Oncol 2016; 21: 819–826.
[25] Ukpo OC, Flanagan JJ, Ma X-J, et al. High-risk human papillomavirus E6/E7 mRNA detection by a novel in situ hybridization assay strongly correlates with p16 expression and patient outcomes in oropharyngeal squamous cell carcinoma. Am J Surg Pathol 2011; 35: 1343–1350.
[26] Bishop JA, Ma X-J, Wang H, et al. Detection of transcriptionally active high-risk HPV in patients with head and neck squamous cell carcinoma as visualized by a novel E6/E7 mRNA in situ hybridization method. Am J Surg Pathol 2012; 36: 1874–1882.
[27] Lewis JS, Thorstad WL, Chernock RD, et al. p16 positive oropharyngeal squamous cell carcinoma:an entity with a favorable prognosis regardless of tumor HPV status. Am J Surg Pathol 2010; 34: 1088–1096.
[28] El-Naggar AK, Westra WH. p16 expression as a surrogate marker for HPV-related oropharyngeal carcinoma: a guide for interpretative relevance and consistency. Head Neck 2012; 34: 459–461.
[29] Nauta IH, Rietbergen MM, van Bokhoven A a. JD, et al. Evaluation of the eighth TNM classification on p16-positive oropharyngeal squamous cell carcinomas in the Netherlands and the importance of additional HPV DNA testing. Ann Oncol Off J Eur Soc Med Oncol 2018; 29: 1273–1279.
[30] Sathasivam HP, Santambrogio A, Andoniadou CL, et al. Prognostic utility of HPV specific testing in addition to p16 immunohistochemistry in oropharyngeal squamous cell carcinoma. Ann Oncol 2018; 29: 2144–2145.
Appendix I – Populärvetenskaplig sammanfattning
Tungbascancer har traditionellt sett drabbat äldre individer som har en lång historia av rökning- och alkoholkonsumtion. På senare decennier har det skett en ökning av tungbascancer hos yngre individer, och då ej kopplat till rökning och alkohol. Denna ökning har berott på en ökning av humant papillomvirus, som annars är känt för att orsaka cancer i livmoderhalstappen. P16 är ett kroppseget protein vars produktion ökar i tungbascancertumörer som är orsakade av humant papillomvirus. Virusassocierad tungbascancer är generellt sett förenat med god prognos och p16 används som en markör för att identifiera patienter med virusassocierad tungbascancer. Ett annat alternativ för att fastställa att tumören är driven av humant papillomvirus är att leta efter virus-DNA i tumören. Förutom att ge en idé om prognosen, tänker man att man i framtiden kan ge en mindre intensiv behandling till patienter med bättre prognos för att minska risken för biverkningar av behandlingen. Tidigare studier är inte helt överens om p16 ska användas ensamt som en prognostisk markör, eller om man även ska testa för virus-DNA för att identifiera patienter som kan dra nytta av en mindre intensiv behandling. I denna studie jämförde vi överlevnad beroende på p16 och virus-DNA. Våra resultat visade en statistiskt säkerställd koppling till överlevnad för p16, men inte för virus-DNA när vi justerade för andra riskfaktorer. Detta tyder på att p16 är en bättre prognostisk markör än virus-DNA och kan således fungera som ett bra verktyg för att identifiera patienter med bättre prognos.
Appendix II – Etiska överväganden
Delar av den data som hämtats för att kunna genomföra denna studie har skett genom journalgranskning. Inga av studiedeltagarna har blivit tillfrågade om medverkande. Detta inskränker givetvis på den personliga integriteten. All data har kodats och databasen är pseudonymiserad. Det har funnits en identifieringsnyckel som förvarats på annan plats än databasen och som dessutom varit krypterad. På så vis är risken för att man kan identifiera individer utifrån databasen mycket liten.
Individerna som ingår i studien kommer inte kunna dra nytta av resultaten från denna studie, eftersom resultatet främst har implikationer för prognos och i framtiden eventuellt val av behandling. Däremot kan individer som drabbas av den studerade sjukdomen i framtiden dra nytta av resultaten i och med att de bidrar till att ha så bra verktyg som möjligt för att kunna göra tillförlitliga prognoser, vilket i slutändan kan leda till så bra individanpassad behandling i framtiden.
Ett annat etiskt dilemma är att detta arbete har gjorts inom ramen för kvalitetsgranskande arbete på klinik. Detta innebär att resultaten från denna specifika studie inte kan kallas forskning och därmed inte publiceras. Detta innebär en sorts semantisk problematik där det å ena sidan egentligen kan tolkas som forskning, men ej får kallas vid det namnet. Det faktum att denna studie, i sin nuvarande form, ej kan publiceras innebär att kunskapen som den bidrar med ej blir tillgänglig för alla och därför ej kan gagna den här patientgruppen. Detta går emot artikel 27 i FNs deklaration om de mänskliga rättigheterna.
Appendix III – Cover letter
Date: 2020-05-21
Title: The prognostic utility of p16 overexpression and HPV-DNA presence in base of the tongue cancer: A retrospective cohort study in Region Örebro County
Corresponding author: Max Waenerlund
Dear Editor-in-Chief, Stephen A. Cannistra, M.D.,
Attached here is a manuscript with the title: ”The prognostic utility of p16 overexpression and HPV-DNA presence in base of the tongue cancer: A retrospective cohort study in Region Örebro County”. We hope you will consider this manuscript for publication in the journal of clinical oncology. We wanted to build on recent studies, more specifically ”Nauta IH, Rietbergen MM, van Bokhoven A a. JD, et al. Evaluation of the eighth TNM classification on p16-positive oropharyngeal squamous cell carcinomas in the Netherlands and the importance of additional HPV DNA testing. Ann Oncol Off J Eur Soc Med Oncol 2018; 29: 1273–1279”, that had found that oropharyngeal cancer patients who where positive for p16 but negative for HPV had a significantly worse prognosis than those who were positive for both, something that questioned the use of p16 as a standalone test for prognosis in oropharyngeal cancer patients. We did a retrospective cohort study on 40 study subjects with base of tongue cancer.
Our key findings were:
• P16 corresponded better with prognosis than HPV-DNA.
• Those who were negative for p16 but positive for HPV-DNA had a statistically significant worse prognosis than those who were positive or negative for both.
Our results add to the debate on how one should interpret p16 overexpression and HPV presence when determining prognosis and tailoring treatments in the future for this group of patients.
All authors have approved the submission of this manuscript. The results have not been previously published and are not being considered for publication elsewhere.
Sincerely,
Max Waenerlund maxwaen@gmail.com
