Clinical and prognostic differences in oropharyngeal squamous cell carcinoma in USA and Denmark, two HPV high-prevalence areas

Background: Uncertainty persists regarding clinical and treatment variations crucial to consider when comparing high human papillomavirus (HPV)-prevalence oropharyngeal squamous cell carcinoma (OPSCC) cohorts for accurate patient stratification and replicability of clinical trials across different geographical areas. Methods: OPSCC patients were included from The University of Texas MD Anderson Cancer Center (UTMDACC), USA and from The University Hospital of Copenhagen, Denmark from 2015–2020, (n = 2484). Outcomes were 3-year overall survival (OS) and recurrence-free interval (RFI). Subgroup analyses were made for low-risk OPSCC patients (T1–2N0M0) and high-risk patients (UICC8 III-IV). Results: There were significantly more HPV-positive (88.2 % vs. 63.1 %), males (89.4 % vs. 74.1 %), never-smokers (52.1 % vs. 23.7 %), lower UICC8-stage (I/II: 79.3 % vs. 68 %), and fewer patients treated with radiotherapy (RT) alone (14.8 % vs. 30.3 %) in the UTMDACC cohort. No difference in the adjusted OS was observed (hazard ratio [HR] 1.21, p = 0.23), but a significantly increased RFI HR was observed for the Copenhagen cohort (HR: 1.74, p = 0.003). Subgroup analyses of low- and high-risk patients revealed significant clinical and treatment differences. No difference in prognosis was observed for low-risk patients, but the prognosis for high-risk patients in the Copenhagen cohort was worse (OS HR 2.20, p = 0.004, RFI HR 2.80, p = 0.002). Conclusions: We identified significant differences in clinical characteristics, treatment modalities, and prognosis between a Northern European and Northern American OPSCC population. These differences are important to consider when comparing outcomes and for patient stratification in clinical trials, as reproducibility might be challenging.


Introduction
Human papillomavirus (HPV) is a key factor in the rising incidence of oropharyngeal squamous cell carcinoma (OPSCC), but with geographical variation [1][2][3][4].How HPVpositivity is defined varies, but the expression of p16 is widely accepted as a surrogate marker for HPV-positivity.However, double p16/HPV-positivity more accurately classify biologically active HPV infection, better predicts patient outcomes, and standardized HPV testing has been recommended [4,5].
Given the remarkable treatment response [6] and the toxicities associated with treatment [7], attempts have been made to identify low-risk HPV-positive (HPV+) patients suitable for treatment de-intensification [8,9].Some well described factors associated with low-risk OPSCC include limited smoking history, p16 + status, and low tumor burden [6].
Both the U.S. and Denmark are HPV high-prevalence areas [10,11].However, the healthcare systems in Denmark and the U.S. exhibit significant differences.Denmark operates under a public, universal healthcare system and all cancer diagnoses, treatments, and follow-ups are conducted with the standardized cancer care packages available within the public health care system.The healthcare system in the U.S. is larger and more complex, with patients being treated at private or public care centers based on a variety of factors, including access to health insurance.
The consideration of potential clinical and demographic variations is crucial when interpreting the prognosis across different patient populations, for accurate patient stratification and management and to ensure reproducibility and replicability of clinical trials in diverse geographical areas.
The aim of this study was to identify potential differences in clinical characteristics, treatment modalities given, and prognosis among OPSCC patients in two HPV highprevalence areas with distinct populations and health care systems: The Copenhagen Oropharyngeal Cancer Database (COHOC) at the University Hospital of Copenhagen, Denmark and the Stiefel OPSCC Database at The University of Texas MD Anderson Cancer Center, USA (UTMDACC).Second, we wished to identify if such differences impact the prognosis for a subgroup of low-risk (T1T2N0M0) and high-risk (III-IV) patients potentially eligible for de-escalated or intensified therapy, respectively.

Study design, setting and population
This retrospective cohort study was approved by The Regional Scientific Ethical Committee (H-20072877) and the Danish Data Protection Agency and followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline [12].The Stiefel OPSCC Database is approved by The Institutional Review Board of the University of Texas MD Anderson Cancer Center (UTMDACC) (protocol PA 14-0947).
The study cohort consisted of all patients with a new OPSCC diagnosed from 2015-2020 treated with curative intent from two independent OPSCC cohorts: The COHOC Database is a population-based, retrospective cohort, comprising patients diagnosed with OPSCC from 2000-2020 in Eastern Denmark and was previously well described [13][14][15].The Stiefel OPSCC Database is a prospective, longitudinal cohort initiated at UTMDACC.Since 2015, 1671 OPSCC patients have been enrolled with detailed characterization of exposures, HPVstatus, diagnostic/staging, treatment, disease control, and longitudinal collection of validated clinician-and patient-reported survivorship outcomes from diagnosis through 5-years of follow-up.

Variables
The primary outcomes were 3-year overall survival (OS) and 3-year recurrence-free interval (RFI).OS was defined as time from diagnosis to death of any cause and patients alive were censored at the last date of follow-up or 3 years after diagnosis.RFI was defined as the time from primary diagnosis to the diagnosis of recurrence (locoregional and/or distant) and patients without recurrence were censored at last day of follow-up, date of death or 3-years after diagnosis.HPV-positivity was defined as being positive for both HPV DNA and p16 in the Copenhagen cohort, while HPV-positivity was defined as being positive to HPV or p16 in the UTMDACC cohort.Details on HPV detection, p16 immunohistochemistry, and measurements are provided in Supplement 1.

Treatment
Therapeutic decisions were made during a multidisciplinary treatment planning conference based on factors such as disease stage, anatomical subsite, comorbidity, functional considerations, and the preferences of both patients and clinicians.In Denmark, treatment decisions are in accordance with the Danish Head and Neck Cancer Group (DAHANCA) national guidelines [16,17], while treatment decisions generally follow the National Comprehensive Cancer Network (NCCN) guidelines at UTMDACC [18].Details on treatment guidelines are provided in Supplement 1.

Statistics
Statistical analysis was performed in R statistics version 4.1.3.Continuous variables were reported as median values with an interquartile range (IQR) and categorical variables as frequencies.To test for significance, Pearson's chi-square test was used for the binomial categorical covariables, Fisher's exact test was used for small sample sizes, and a t-test was used for the quantitative covariables distributed in two groups.We considered a p-value < 0.05 to be statistically significant.Age and follow-up time were included as continuous variables and the other variables were considered categorical variables.
OS and RFI were evaluated by Kaplan-Meier curves and with the log-rank method and by uni-and multivariable Cox regression analyses (packages Survminer and Survival).For the total OPSCC cohort, the OS (372 events) and RFI (267 events) multivariable analyses were adjusted for treatment center, age (cut-off >60 years), gender, HPV-status, smoking status, tumor location, T-, N-and M-class and treatment modality.For model control, proportionality was tested with Score residuals.Details on subgroup analyses are provided in Supplement 1.

HPV status
Overall, 63.1 % of the OPSCC patients were HPV+ in the Copenhagen cohort, while 88.2 % were HPV+ in the UTMDACC cohort.In the UTMDACC cohort, 68.9 % did not have both HPV and p16 status available, while this was the case in 3.1 % of the patients in the Copenhagen cohort.In the Copenhagen cohort, 0.4 % and 10.8 % in the UTMDACC were tested for HPV DNA only, while 1.9 % in the Copenhagen cohort and 46.7 % in the UTMDACC cohort were tested for p16 only.See Table 1.

Treatment modalities
Radiation-based treatment was predominant in both cohorts, but significantly more patients received radiotherapy (RT) alone in the Copenhagen cohort (30.3 % vs. 14.8 %).In the Copenhagen cohort, 51.0 % received concurrent chemotherapy (CRT) vs. 49.8 % in the UTMDACC cohort.Neoadjuvant chemotherapy + RT/CRT was given to 16.7 % of patients in the UTMDACC cohort.Significantly more patients received surgery alone in the Copenhagen cohort (16.0 % vs. 8.7 %), while significantly more received surgery and post-operative adjuvant therapy in the UTMDACC cohort (9.8 % vs. 2.7 %). See Table 1.
The multivariable OS analysis revealed no significant difference between the two cohorts (hazard ratio [HR] 1.21, p = 0.23).A significantly increased HR was observed for HPV−, current smoking, T3-and T4-class, N2 and N3-class, M1-class, and RT.See Table 2.
The multivariable recurrence analysis showed a significantly increased HR in the Danish cohort (HR: 1.74, p = 0.003).Also, HPV−, T2-T4-class, N1-N3-class, M1-class, RT and neoadjuvant chemotherapy + RT/CRT had a significantly increased HR.See Table 3.The analysis was also performed for patients with either progression or second primary tumors, which did not significantly impact the conclusions; data not shown.

Subgroup analysis of RT single modality patients
See Results (Supplement 1).

Discussion
This study, which included 2484 patients with OPSCC from Eastern Denmark and UTMDACC treated between 2015-2020, revealed significant differences in OS and RFI driven by clinical and treatment differences.
Although RT was predominant in both cohorts, significantly more patients received RT alone in the Copenhagen cohort, which is noteworthy, since the Copenhagen cohort presented at higher UICC-stage, T-and N-class.In line with this, RT remains the predominant treatment modality in most European centers [19].
The UTMDACC cohort included more clinical characteristics associated with a better prognosis: more HPV+ OPSCC patients, lower T-and N-burden and less smoking [20] but also more men which are associated with a worse prognosis [21].When adjusting for these differences, an independent increased OS HR was not observed, but the risk of recurrence was significantly higher in the Copenhagen cohort (RFI HR 1.74, p = 0.003).Discrepancies in follow-up regimens and divergent definitions of recurrence may also attribute to the observed disparities.Although speculative, it could also indicate a different biology of HPV+ OPSCCs.In line with our findings, a recent study including a Danish and a Toronto cohort observed significant clinical and treatment differences, which were reflected in differences in locoregional failure and OS [22].
In low-risk patients, significant clinical and treatment differences were observed with 41.7 % in the Copenhagen cohort receiving RT alone versus 19.6 % in the UTMDACC cohort, yet the prognosis remained equally good.Several studies have investigated de-escalated treatment to minimize treatment-related toxicities for a subgroup of patients with HPV+ OPSCC [8,9,[23][24][25].It is important to keep in mind that substantial clinical differences exist when selecting patients and validating de-escalating trials in various geographical areas.This study underlines the importance of including multiple factors like TNM-classification, HPV− and smoking status, but also indicates that factors like age and tumor location could be considered for stratification.
Conversely, a sub-analysis of high-risk patients revealed significant clinical and treatment differences influencing the prognosis.The UTMDACC cohort had significantly better OS and RFI.OS was also significantly associated with current smoking, stage, and treatment modality given.RT alone was given to a substantial part of the Copenhagen patients (37.1 % vs 7.2 %), despite that CRT is recommended for this group [16,17].The difference in use of concurrent chemotherapy is likely a result of factors impacting treatment selection not included in this study, e.g., comorbidities, as the Copenhagen cohort is older and include more smokers.But a difference in practice patterns cannot be excluded.Neoadjuvant chemotherapy is not used in Denmark but was given to 16.7 % of the patients in the UTMDACC cohort (corresponding to 39.1 % of UTMDACC high-risk patients), which is in line with data from The National Cancer Database [26].The role of neoadjuvant therapy in locally advanced OPSCC cancer remains controversial [27], however, it may be associated with a decreased risk of distant metastasis [28].A recent study suggests that neoadjuvant therapy is associated with an improved OS and lower risk of distant metastasis in patients with OPSCC [29].
A notable difference in treatment between MDA and Denmark is the use of hypoxic modification with nimorazole.The DAHANCA 5 randomized trial, investigated the effect of nimorazole, showed a significantly better loco-regional control rate and lower cancer-related death in patients receiving nimorazole compared to placebo [30].Nimorazole has since been standard of care for all Danish OPSCC patients.The benefit of nimorazole may primarily be present in HPV− OPSCC, as reanalysis showed no significant benefit in HPV/p16-positive tumors [31].
Disparities in the healthcare system might introduce a selection bias contributing to the differences observed in this study, with more HPV-OPSCC patients and more high-risk patients in the Copenhagen cohort.Studies have shown that insurance status is predictive of clinical characteristics such as tumor stage and comorbidities at diagnosis as well as oncological outcome [32][33][34].A recent Copenhagen study comprising a part of the Copenhagen cohort suggests that low socioeconomic status negatively impact the OS, largely due to differences in clinical characteristics at diagnosis, including smoking status, comorbidities, and clinical stage [35].Future studies including socioeconomic status, comorbidities, and performance status would provide further insight into this matter.Disparities in the healthcare system might also contribute to the differences observed in treatment modalities, availability, and regimen of follow-up care.

Limitations
A recent study, where the Copenhagen cohort comprised the majority of the study cohort, has proven double HPV/p16-positivity to be superior in prognostication and has led to the recommendation of including both p16 and HPV-status in areas with high discordance, which is also supported by the ESMO guidelines [4,5].In this study, HPV-positivity was defined in the UTMDACC as p16 + or HPV+, which might overestimate the HPV-positive prevalence, although only 1.9 % had discordant p16/HPV status [4].The included cohorts might not be representative of the Northern American and European cohorts overall, and the results may not apply to other regions.Socioeconomic status, performance score, and comorbidities were not included which potentially impacts treatment decisions and have shown to influence both OS and RFI and which we hypothesize differs greatly in the two cohorts.

Conclusion
Significant demographic, clinical, and treatment differences influencing the prognosis were identified between two high HPV-prevalence OPSCC cohorts representing two different health care systems: a Copenhagen cohort from a public, universal health care system and a UTMDACC cohort from an insurance-based health care system.Despite significant clinical and treatment differences, the prognosis for low-risk OPSCC patients was equally good, while the prognosis was significantly better for patients with high-risk OPSCC in the UTMDACC cohort.
Our study suggests that significant clinical differences exist between a Northern European and Northern American OPSCC population, which is important to consider when comparing outcomes and for patient stratification in clinical trials, as reproducibility might be challenging.However, to fully understand possible intercontinental variations, more data are needed.Eur J Cancer.Author manuscript; available in PMC 2024 August 29.

Table 1
Characteristics of 2484 patients with OPSCC in Eastern Denmark and UTM-DACC, USA from 2015-2020.

Table 2
Multi-and univariable analysis for overall survival for OPSCC in Eastern Denmark and UTMDACC from 2015-2020.
c Adjusted for center, age group, gender, HPV status, smoking status, T-site location, stage UICC8 and treatment regimen.Eur J Cancer.Author manuscript; available in PMC 2024August 29.

Table 3
Multi-and univariable analysis for recurrence-free interval for OPSCC in Eastern Denmark and UTMDACC from 2015-2020.
d Significant value.eAdjusted for center, age group, gender, HPV status, smoking status, T-site location, stage UICC8 and treatment regimen.