In the present study, we analyzed the association of tumor p16 expression with prognosis in a retrospectively collected cohort of patients with HNSCC treated with RT+EGFR inhibitors or RT+CT. A strong correlation between HPV16 status and p16 immunostaining was found, as well as a significant benefit in OS and DFS for p16-positive OPSCC patients. Our most interesting result is the significant benefit in OS and DFS for p16-positive patients when treated with RT+EGFR inhibitors compared with conventional RT+CT (p = 0.01), whereas this benefit was not observed in p16-negative patients (p = 0.9). From 2000 to 2011, more than 304 patients were treated in our center with RT plus platinum-based CT. To ensure this study provided results from a representative sample of our patients, we assessed the OS rates of the entire group. The median OS of these 304 patients was 37 months (range of 21–53 months), with 2- and 5-year OS rates of 56% and 37%, respectively. These results are similar to those observed for the group of 108 patients analyzed in this study, who had a median OS of 32 months (range of 18–46 months) and 2- and 5-year OS rates of 58% and 35%, respectively.
We found HPV16 DNA in 11% of the patients and in 15% of the OPSCCs. Both of these prevalences are lower compared with data from a recent meta-analysis (22-34% in HNSCC and 30-41% in OPSCC) [8, 10, 25], possibly due to the epidemiologic profile of our population, which had a high proportion of former/current and heavy tobacco users.
Several studies have confirmed a better outcome for patients with HPV-positive HNSCC compared with HPV-negative patients. The majority of these studies have been performed in OPSCC, so this benefit is well established in this subset of tumors [13, 14, 17–21]. A recent meta-analysis of 34 studies published by Dayyani et al. has shown a benefit in OS for HPV-positive patients compared to HPV-negative patients (HR = 0.4; 95% CI 0.2 to 0.6; p = 0.0001), including all locations of HNSCC . According with these data, several studies have shown that among patients with HNSCC and OPSCC managed with different treatment modalities, those with p16-positive tumors have a better prognosis in terms of response, recurrence and survival than patients with p16-negative tumors [14, 22–24]. Recently, the results of a phase III clinical trial (TROG 02.02) with concurrent RT+CT with or without tirapazamine has shown promising results in the subgroup of patients with OPSCC using p16 immunohistochemistry as a prognostic factor. Patients with p16-positive tumors had better 2-year OS and failure-free survival (FFS) rates compared with patients with p16-negative tumors (OS 91% vs. 74%; p = 0.04; FFS 87% vs. 72%; p = 0.003). Cox regression analysis of OS, including the prognostic factors of hemoglobin, T category, N category, and ECOG performance status, demonstrated that p16 status was the only significant factor in multivariable analysis (HR = 0.4; 95% CI 0.2 to 0.9; p = 0.03) .
The HPV-related beneficial outcome in this neoplasm indicates that these tumors respond well to conventional treatment approaches. The biologic basis for this observation is under investigation, but it could be due to some extent to an increased sensitivity to RT+CT of functional, non-mutated p53  and the absence of field cancerization related to tobacco/alcohol exposure. Moreover, an increased sensitivity to apoptosis has been observed in E6/E7-positive human keratinocytes when exposed to cisplatin-based CT [4, 27]. To the best of our knowledge, only one study has reported differences in survival in patients treated with EGFR inhibitors according to p16 or HPV16 status  That phase II trial, which applied induction CT based on paclitaxel, carboplatin and cetuximab followed by either RT, concomitant RT+CT, or surgery, for locally advanced (LA)-HNSCC, shows a robust benefit in survival for HPV-positive patients over HPV-negative patients, with improved progression-free survival (p = 0.012) and OS (p = 0.046). The current study supports this better outcome for HPV-positive patients regardless the treatment received.
Our study found a better outcome in terms of DFS and OS in p16-positive patients treated with RT+EGFR inhibitors compared with those treated with conventional RT+CT. The reason for these findings remains unclear. An inverse relationship between HPV status and EGFR expression has been demonstrated in several recent head and neck cancer studies [15, 29–31]. Most of them suggest that a combination of HPV and EGFR may more accurately predict the outcome of patients than either alone. One recent study examined the prognostic significance of EGFR in relation to HPV in a large cohort of patients with OPSCC. That study reported that HPV was a predictor of loco-regional recurrence, event-free survival, and OS after adjustment for clinicopathological variables, including EGFR. This study showed that HPV-negative/EGFR-positive patients had an adjusted 13-fold increased risk of having a loco-regional failure and more than a 4-fold increased risk of dying compared with HPV-positive/EGFR-negative patients . Although several preclinical studies have shown that HPV16 E6 and E7 expression sensitizes human keratinocytes to apoptosis caused by cisplatin, etoposide and mitomycin C by mechanisms that are not fully understood , little information is available regarding EGFR inhibitors and HPV-related cell tumors.
The five-year survival data of the study of Bonner et al. on RT plus cetuximab for locoregionally advanced head and neck cancer show some interesting associations . Cetuximab seemed to provide the most benefit for patients with oropharyngeal tumors, T1–3 tumors, advanced nodal stage, and high Karnofsky performance status, characteristics associated with HPV-related tumors. These subgroups represent small numbers of patients and might represent spurious findings, so further work should be performed to test the consistency of these results.
Our study has some limitations, such as the retrospective analysis, the small sample size, the limited number of p16-positive patients (18) and the fact that p16 expression was not an independent prognostic factor in the multivariable analysis. The retrospective setting and the possibility of selection bias may be considered pitfalls, and although patient characteristics between treatment groups were not completely well balanced (more women and less alcohol consumption in the group treated with RT+EFGR inhibitors), we should point that the most important clinicopathological variables (tumor size, N category and ECOG performance status) were well distributed between both arms of treatment. The issue of sample size is another limitation because the number of p16-positive patients in each arm of treatment was small (8 RT+EGFR and 10 RT+CT), but despite the limited sample size, the survival benefit in p16-positive patients treated with RT+EGFR inhibitors compared with RT+CT reached statistical significance.