The incidence of tonsillar cancer and the proportion of human papillomavirus (HPV) positive tonsillar cancer cases have increased in the last decades. In parallel, treatment for tonsillar cancer has been intensified e.g., by accelerated radiotherapy, and chemotherapy, resulting in more side effects. Patients with HPV-positive tonsillar cancer have better prognosis than those with HPV-negative tumors, and the former group could hypothetically benefit from reduced, less-toxic treatment without compromising survival. Here, we therefore evaluated possible differences in overall and disease-specific survival after different oncological treatments in 153 patients with HPV DNA- and P16-positive tonsillar cancer who were diagnosed and treated with intent to cure between 2000 and 2007, in Stockholm, Sweden. Of these patients, 86 were treated with conventional radiotherapy, 40 were treated with accelerated radiotherapy and 27 were treated with chemoradiotherapy. There were no significant differences in overall or disease-free survival between the groups. However, there was a trend, implying a beneficial effect of the intensified treatment, with chemoradiotherapy being better than radiotherapy despite that more patients had stage IV disease in the former group; and accelerated radiotherapy being better than conventional radiotherapy. This needs to be followed further in larger more homogenous groups of patients. In conclusion, patients with HPV-positive tonsillar cancer treated with conventional- or accelerated radiotherapy or chemoradiotherapy disclosed similar survival rates. The trend for better survival and less metastasis after intensified treatment underlines the need for large prospective studies comparing less intense to more intense treatment (chemoradiotherapy).
The incidence of tonsillar squamous cell carcinoma (SCC) has increased in the last decades, in Sweden, USA as well as in other countries1–4 and human papilloma virus (HPV) has been recognized as a risk factor. The prevalence of HPV in tonsillar cancer has increased dramatically in Stockholm, Sweden in the last decade and by now, over 90% of all tonsillar cancers is HPV-positive.5 It is furthermore well established that HPV-positive and HPV-negative oropharyngeal SCC (OSCC), including tonsillar SCC, may represent separate entities with alternative mechanisms for carcinogenesis and different clinical features. The clinically most important difference being the significantly improved survival of the patients with HPV-positive SCC as compared to those with HPV-negative cancer.6–11 This favorable outcome for patients with HPV-positive cancer seems to be independent of treatment modality i.e., regardless of oncological12, 13 or surgical treatment.6, 14
Treatment for tonsillar cancer patients has, at many centers including Stockholm, been intensified from conventional once-a-day radiotherapy only, toward combined treatment strategies including hyperfractionated, accelerated radiotherapy, interstitial brachy-boost radiotherapy and induction as well as concomitant chemotherapy. Also the use of targeted therapy is increasing. This intensified treatment implies more side effects,15 which is why some authors have proposed to investigate if it is possible to reduce the oncological treatment for patients with HPV-positive tonsillar cancer without compromising survival.7, 16 So far, response and survival after different treatments for patients with only HPV-positive tonsillar cancer exclusively have not been investigated systematically.
The purpose of this study was therefore to investigate overall and disease-free survival for patients with HPV-positive tonsillar squamous cell carcinoma in relation to the different oncological treatment they received (conventional radiotherapy, accelerated radiotherapy and chemoradiotherapy). In addition, we evaluated whether the development of distant metastasis differed between the three treatment regimes. Earlier studies have suggested that p16INK4a analysis needs to be added to HPV-DNA analysis to optimize its prognostic value which is why p16 analysis was also carried out7, 10, 17 and only patients with samples that were both HPV-DNA and p16 positive were considered HPV-positive and included in the survival analysis.
Patients, Material and Methods
ICD-10 codes C09.0-C09.9 were used to identify all 290 patients diagnosed with tonsillar squamous cell carcinoma in the County of Stockholm, Sweden, between 2000 and 2007 using the Swedish Cancer Registry and the local cancer registry. Of these patients, 264 were treated with the intention to cure, and from this latter group 211 pretreatment biopsies were available and tested for HPV by PCR and for p16 by immunohistochemistry. Notably, some patients in this study (those also overlapping with ICD-7 code 145.0) have previously been included in our previous studies.2, 5 The study was conducted according to ethical permissions 2005/431-31/4, 2005/1330-32 and 2009/1278-31/4 from the Ethical Committee at Karolinska Institutet, Stockholm, Sweden.
Treatment modalities and evaluation of response to treatment
Treatment was categorized as (i) conventional radiotherapy (2.0 Gy/day, for 6.5–7 weeks, total dose: 68 Gy) or (ii) accelerated radiotherapy (1.1 + 2.0 Gy/day for 4.5 weeks, total dose: 68 Gy) or (iii) chemotherapy consisting of Cisplatin 75 mg/kvm and 5-fluorouracil 1,000 mg/kvm with or without docetaxel [(Taxotere®) 75 mg/kvm], followed by either conventional or accelerated radiotherapy. Treatment response was evaluated by clinical examination or by histopathology of biopsies taken under general anesthesia after completed treatment. Patients were followed up by clinical examination every 3 months during the first 3 years and then every 6 months for a total of 5 years. All patients were evaluated for a minimum of 2 years and the follow-up period for the patients ranged from 1 to 119 months. For the survival analysis, patients were categorized into two age groups, either below or above the age of 59 (mean and median age for the whole group). For overall survival, patients categorized in Stages 1 or 2 at the time of diagnosis were pooled in one group and compared to patients categorized in Stages 3 and 4. For disease-free and metastasis-free survival, patients categorized in Stages 1–3 at the time of diagnosis were compared with patients categorized in Stage 4.
Detection of HPV DNA
HPV DNA was detected as previously described.5 Briefly, DNA was extracted from 30-μm paraffin-embedded tonsillar tumor biopsy slices. A PCR setting, using general primer pairs (GP5+/6+ and CPI/IIG) and HPV-16 type-specific primers, was used to analyze presence of HPV DNA. Type-specific primers and sequencing were used for HPV typing. A housekeeping gene S14 PCR was used to verify amplifiable DNA.
Immunohistochemical analysis of P16INK4a
Immunohistochemistry was performed with the p16INK4a primary monoclonal mouse anti-human p16INK4a antibody (clone E6H4) (DakoCytomation A/S, Carpinteria, CA).18 The staining was performed in an automated deparaffinization and staining machine (BenchMark® XT, Venetana Medical system, Tuscon, AZ). Serial 5-μm sections of formalin-fixed and paraffin-embedded tissues were used with a final haematoxylin and eosin-stained section to confirm the histopathological diagnosis and the representativeness of the cut-off levels. The sections were processed in BenchMark® XT for deparaffinization (98°C, for 30 min) and then rehydrated through graded alcohol. Epitope retrieval was performed by heating and then the sections were treated with peroxidase blocking reagent. The slides were incubated for 32 min (42°C) with 100 μl of the monoclonal antibody p16INK4a, followed by incubation with visualization reagent. The reaction was developed using chromogen solution (DAB) and counterstained with haematoxylin. As a negative antibody control monoclonal mouse igA2 was used. Immunohistochemical staining was evaluated by light microscopy and was graded using a 4-tier scale where 0–1 (absent or weak staining) was regarded as negative and 2–3 (strong or very strong staining) was regarded as positive. A positive control section from a cervical carcinoma in situ was included with all samples.
Overall survival was defined as the time from date of diagnosis to the date of death or to the date of censorship (the last day of follow-up). Disease-free survival was defined as time from the date of diagnosis to the date of the last known occasion that the patient was disease-free or the date of disease recurrence (local, regional or distant recurrence). Death without documented recurrence was censored at the date of death. The different treatments evaluated were: conventional radiotherapy versus accelerated radiotherapy and chemoradiotherapy versus radiotherapy (conventional and accelerated pooled together). Kaplan–Meier curves were used to present survival data for patients and the log-rank test was used in univariate analysis. In multivariable analyses, a Cox proportional hazards model was used to adjust for covariates. The proportional hazards assumption was evaluated with Schoenfeld residuals. A p value of ≤0.05 was considered statistically significant. Two-sided p-values were reported. All analyses were performed in a statistical computer program (Stata). The associations of treatment with sex or age were calculated using Chi-square test while the association between treatment and stage, histopathological differentiation or TNM-status was calculated using Fishers exact test.
Study group, treatment groups and survival analysis
Between 2000 and 2007, 264 patients diagnosed with tonsillar SCC were treated with the intention to cure in the Stockholm area and from these patients 211 tumor pretreatment biopsies were available for HPV and p16 analysis. In total 153 patients had HPV-DNA positive and p16-positive tumors and were included in the comparison between the different treatment modalities (Table 1). No patient received surgery (except for biopsy or diagnostic tonsillectomy) before oncological treatment.
Table 1. Characteristics of patients and their tonsillar cancer samples obtained during 2000–2007
The characteristics of the patients and their tumors according to treatment are listed in Table 1. The group of patients (n = 86) receiving conventional radiotherapy was larger than the groups receiving accelerated radiotherapy (n = 40), or receiving chemoradiotherapy (n = 27). Patients in the three different treatment groups were similar, except for stage, where patients treated with chemoradiotherapy were significantly higher staged than patients in the other two groups (p = 0.002, Fishers exact test). For the analysis between chemoradiotherapy and radiotherapy groups, all 153 patients were included. For the analysis between accelerated radiotherapy and conventional radiotherapy groups, 126 patients were included.
The mean follow-up time for all the study groups was 54 months and the median follow-up time was 49 months (range 1–119 months). There were 27 (31%) deaths among patients in the conventional radiotherapy group, 8 (20%) deaths among patients who received accelerated radiotherapy and 4 (15%) deaths in the chemoradiotherapy group. Nine patients treated with conventional radiotherapy, one patient treated with accelerated radiotherapy and two patients treated with chemoradiotherapy died free of disease.
Conventional versus accelerated radiotherapy with regard to differences in survival
There was no significant difference in overall or disease-free survival between the conventional or accelerated radiotherapy groups based on Kaplan–Meier estimates with regard to overall survival (log rank test p = 0.207, Fig. 1) or disease-specific survival (log rank test p = 0.569, Fig. 1).
In addition, a univariate analysis was performed to evaluate factors potentially associated with overall and disease-free survival; age, sex, tumor stage and treatment were analyzed. Age above median and mean age was associated with worse overall (p = 0.001), but not with disease-free survival (Table 2). For disease-free survival, patients categorized in Stage 4 had a significantly worse prognosis compared to patients categorized in Stages 1–3 (p = 0.030), (Table 2). No other variable showed any significant correlation to prognosis.
Table 2. Uni- and multivariate analysis comparing the conventional radiotherapy (RT) and accelerated radiotherapy (accRT) groups
To assess the independent predictive value of all factors for overall- and disease-free survival, a multivariable analysis using the Cox proportional hazards model was performed. The following variables were included in the model: age, sex, treatment and tumor stage and the results are shown in Table 2. Age above median and mean age was found to be an independent and significant negative prognostic factor for overall survival, but not for disease-free survival (p = 0.001). For disease-free survival patients categorized in Stage 4 had a significantly worse prognosis compared with patients categorized in Stages 1–3 (p = 0.025). Other included variables were not significant.
Radiotherapy versus chemoradiotherapy with regard to differences in survival
There was no significant difference in overall or disease-free survival between the radiotherapy versus the chemoradiotherapy groups as demonstrated with regard to overall survival (log rank test p = 1.0, Fig. 2), or disease-free survival (log rank test p = 0.357, Fig. 2).
A univariate analysis was performed to evaluate factors potentially associated with overall and disease-free survival; age, sex, tumor stage and treatment were analyzed. Age above median and mean age was associated with worse overall (p = 0.002), but not disease-free survival (Table 3). For disease-free survival, patients categorized in Stage 4 had a significantly worse prognosis compared to patients categorized in Stages 1–3 (p = 0.044). Other included variables were not significant.
Table 3. Uni- and multivariate analysis comparing the radiotherapy (RT) and chemoradiotherapy (CRT) groups
To assess the independent predictive value of all factors for overall- and disease-free survival, a multivariable analysis using the Cox proportional hazards model was also performed. The same variables were included in the model: age, sex, treatment and tumor stage; and the results are shown in Table 3. Age above median and mean age was found to be an independent and significant negative prognostic factor for overall survival (p = 0.001). For disease-free survival patients categorized in Stage 4 had a significantly worse prognosis compared to patients categorized in Stages 1–3 (p = 0.025). Other included variables were not significant.
The conventional radiotherapy group was also compared separately with the chemotherapy group regarding all the above parameters, but there were no significant differences from when comparing both radiotherapy groups to the chemotherapy group (data not shown).
Development of distant metastasis between treatment groups
The numbers of patients that developed distant metastases during the course of the disease, i.e., M0 at diagnosis, in the three treatment groups was also analyzed. In total, 8/86 (9.3%) in the group treated with conventional RT; 5/40 (12.5%) in the group treated with accelerated RT, and 1/27 (3.7%) in the chemoradiotherapy group, developed distant metastases during the investigated period. To evaluate a possible inhibitory effect on the development of distant metastasis by the addition of chemotherapy, we compared the radiotherapy groups and the chemoradiotherapy group and found no significant differences in metastasis-free survival between the groups (log-rank test p = 0.450, Fig. 3). A uni- and a multivariate analysis were then also performed, but no significant differences were found (Table 3).
In this study, overall and disease-specific survival, as well as survival free of distant metastasis was compared in patients with HPV-positive tonsillar cancer, treated with intention to cure by (i) conventional radiotherapy, or (ii) accelerated radiotherapy or (iii) chemoradiotherapy. There was no significant difference in overall or disease-specific survival between patients treated with conventional radiotherapy versus accelerated radiotherapy, or between patients treated with chemoradiotherapy versus the joint group of patients receiving conventional or accelerated radiotherapy.
When comparing patients treated with conventional radiotherapy versus accelerated radiotherapy, the lack of improved results from accelerated radiotherapy is in accordance with the ARTSCAN-trial results, even if that study was not stratified for HPV-status.19
However, a trend favoring chemoradiotherapy was observed in that fewer patients (3.7%), treated with chemoradiotherapy developed distant metastases as compared to those receiving conventional or accelerated radiotherapy (9.3 and 12.5% respectively), but over time this result was not significant. In addition, there was a nonsignificant trend for both better overall- and disease specific survival for patients with HPV-positive tonsillar cancer treated with the more intense treatment. This is visualized for accelerated radiotherapy versus conventionally fractionated radiotherapy in Figure 1, and for chemoradiotherapy versus radiotherapy in Figure 2. This trend needs a longer follow-up time to confirm its relevance, but it suggests that so far there is not enough evidence to reduce treatment intensity for patients with HPV-positive tonsillar cancer and additional studies are warranted.
Others have also analyzed treatment response for OSCC stratified for different treatments and do not find any difference between overall disease-specific survival after radiotherapy compared with chemoradiotherapy.7, 20 However, in the study by Pedruzzi et al.,20 the presence of HPV was not accounted for, which would be of interest. Moreover, Pedruzzi et al.20 report an overall 5-year survival of only 17.6%, which is strikingly lower than for any treatment group in our study. We have no explanation for this difference in outcome, but it could be due to the inclusion of a higher proportion of smoking patients and possibly more patients with HPV-negative tumors in their material. Ang et al.7 report similar survival rates for patients with HPV-positive oropharyngeal cancer (including tonsillar cancer and other subsites) as we do, but all their patients received chemotherapy (cisplatin) in addition to either conventional or accelerated radiotherapy.
One limitation of this study is that the treatment groups are not completely homogenous. The patients treated with chemoradiotherapy, consisted of patients which were treated both with conventional and accelerated radiotherapy. However, in this study there were no major differences in outcome for patients with HPV positive tonsillar SCC with regard to the different radiation treatments.
Another limitation is that, in spite of the fact that there are a relatively large number of patients included in this study, the sample size required for a nonequivalent study between two arms to show a 4–5% difference in survival would be 700 patients. This could imply that there are survival benefits between the compared groups but the sample size of this study is too small to reveal significant differences.
There was a significant stage difference between the groups treated with radiotherapy and the group of patients treated with chemoradiotherapy. Patients in the latter group were more often staged in Stages 3 or 4, which may have affected choice of treatment and it is possible that the favorable effect of this treatment is clouded by the fact that these patients were higher staged.
This study is retrospective which limits the number of variables that can be included in the statistical analyses. Reliable data on patient performance status, smoking habits, loco-regional control and comorbidity were not available. This highlights the need for prospective, controlled studies with larger numbers of patients.
Moreover, previous reports have led to the conclusion that it is necessary to stratify for HPV when performing clinical studies, especially with regard to response to treatment. It is also possible that patients with HPV-negative tonsillar cancer may benefit the most from intensified treatment, but studies on these patients is increasingly difficult, since their incidence is decreasing and they accounted for less than 10% of all tonsillar cancer cases in 2006–2007, in Stockholm, Sweden.5
In conclusion, in this study we find no major differences between overall- or disease-free survival in patients with HPV-positive tonsillar cancer treated with conventional- or accelerated radiotherapy or chemoradiotherapy. However, there are trends for the more intensive treatments to be of benefit, both in terms of survival and in developing distant metastasis and thus we cannot today identify a subgroup of patients for whom we safely can suggest a less intense treatment is acceptable without additional prospective studies.
The authors thank David Nordfeldt and Henrik Hellborg for valuable data analysis and statistical analysis.