Abstract 911O
Background
p16INK4a (p16) immunostaining is the most widely implemented technique in clinical settings for determining HPV causation and HPV-related prognosis biomarker of oropharyngeal cancer (OPC). A subset of p16+ OPC are HPV-; and their prognosis is still unclear. The aim of this study is to clearly define the proportion, determinants and prognosis of OPC patients who are p16+/HPV-.
Methods
We established an international consortium comprising 13 cohorts of OPC patients with data on p16, HPV, demographics, tobacco/alcohol use and clinical data. A centralized individual patient data reanalysis was performed. Multivariate models were used to evaluate factors associated with HPV status as defined by different HPV-assessment methods. Proportional-hazards models were used to compare the risk of death (OS) among HPV-related and un-related OPC.
Results
In total 7702 patients from 9 different countries were included. The percentage of positive cases was 49.7%, 47.9% and 44.3% for p16+, HPV+ and p16+/HPV+, respectively. Among p16+ cases, 10.9% were HPV-. This proportion differed significantly by cohorts and geographic areas (p-value<0.001) and was lowest in the highest prevalence areas (Table). Compared to p16-/HPV- tumors, p16+/HPV+was the biomarker with strongest prognostic value (aHR 0.28, 95%CI 0.25-0.31), followed by p16+/HPV- (aHR=0.65, 95%CI 0.56-0.76), and p16-/HPV+ (HR=0.72, 95%CI 0.60-0.86). Disease-free/OPC-specific survival analyses will be presented at the congress. Table: 911O
p16+/HPV- by region and cohort from the EPIC-OPC study
| Region/Cohort | Total N | P16+ N | P16+/HPV- | |
| N | (%)* | |||
| North America | 186 | 135 | 2 | 1.5 |
| Canada-Toronto | 186 | 135 | 2 | 1.5 |
| Northern Europe | 6380 | 3320 | 351 | 10.6 |
| Denmark-Copenhagen | 2169 | 1324 | 123 | 9.3 |
| UK-Birmingham | 816 | 499 | 58 | 11.6 |
| UK-Liverpool | 252 | 152 | 12 | 7.9 |
| UK-Belfast | 232 | 95 | 11 | 11.6 |
| The Netherlands-Amsterdam/Rotterdam | 1203 | 388 | 48 | 12.4 |
| Germany-Giessen | 704 | 235 | 40 | 17.0 |
| Germany-Cologne | 205 | 111 | 17 | 15.3 |
| Germany-Kiel | 126 | 58 | 11 | 19.0 |
| Sweden-Stockholm | 539 | 375 | 24 | 6.4 |
| Switzerland-Zurich | 134 | 83 | 7 | 8.4 |
| Southern Europe | 1136 | 370 | 63 | 17.0 |
| France-Paris | 275 | 275 | 275 | 12.7 |
| Spain-Barcelona | 861 | 95 | 28 | 29.5 |
| All | 7702 | 3825 | 416 | 10.9 |
*p16+/HPV- percent among p16 tested.
Conclusions
p16+/HPV+ tumors showed the highest OS magnitude of association compared with other biomarkers combinations. Using p16 immunostaining alone, 11% OPC patients would be incorrectly classified as HPV-related OPC according to TNM-8 staging, and with risk of misclassification for de-escalation in clinical trials, particularly in regions with lower attributable fractions of HPV.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
Catalan Institute of Oncology.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.