Analysis of the effect of p16 and tobacco pack-years (p-y) on overall (OS) and progression-free survival (PFS) for patients with oropharynx cancer (OPC) in Radiation Therapy Oncology Group (RTOG) protocol 9003

Reporter: Samuel Swisher-McClure, MD
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 7, 2010

Authors: M. L. Gillison, The Ohio State University, Columbus, OH


  • There is a growing body of evidence that Human Papillomavirus (HPV) infection plays an important role in the pathogenesis of a subset of head and neck squamous cell carcinomas (SCC).
    • HPV related tumors are characterized by the presence of HPV DNA within the tumor and are predominantly observed in younger patients with cancers of the oropharynx (Gillison et al. JNCI, 2000).
  • Several retrospective analyses have been published suggesting that patients with HPV positive tumors have better prognosis than HPV negative tumors (Weinberger et al JCO, 2006).
  • A prospective clinical trial published by Fakhry et al. (JNCI, 2008) enrolled 96 patients with locally advanced SCC of the oropharynx and larynx treated with induction chemotherapy followed by concurrent chemoradiation. The authors reported a marked difference in 2 year overall survival favoring the HPV positive patients [95% vs. 62%, p = 0.005; Hazard ratio (HR) for death of 0.36, 95% CI 0.15-0.85].
  • Ang et al. (NEJM, 2010) recently published an analysis of 720 patients enrolled in the RTOG 0129 trial, demonstrating that patient HPV status and tobacco pack years are independent predictors of overall survival (OS) and progression free survival (PFS) in patients treated with concurrent chemoradiation.
  • In the current analysis, the authors evaluated whether similar differences were observed when OPC patients were treated with radiotherapy alone in RTOG 9003.


  • The RTOG 9003 was a multi-institutional phase III randomized trial which compared 4 different radiation fractionation schemes (standard fractionation, hyperfractionation, accelerated fractionation with split course, or accelerated fractionation with concomitant boost) in the treatment of patients with locally advanced head and neck cancer. The results of this trial have been previously published by Fu et al. (IJROBP, 2000).
  • The current study was an unplanned retrospective analysis of patients enrolled in the RTOG 9003 trial, to evaluate HPV status and tobacco use as prognostic factors in HNC patients treated with radiation alone.
  • The analysis included only patients with OPC since prior work indicates that this is the predominant subsite for HPV related HNC.
  • Tumor HPV status was determined by a surrogate, p16 immunihistochemistry (?70% expression positivity).
  • The IHC was performed on archived tumor specimens that remained available from the RTOG 9003.
  • Patient tobacco use data was self-reported in surveys that were obtained at the time of the trial.
  • The primary endpoint of the analysis was 5 year OS by p16 status.
  • Secondary Endpoints included PFS and patterns of failure.
  • Five-year OS and PFS were estimated (with 95% confidence intervals [CI]) by the Kaplan- Meier method and compared by log-rank test.
  • Hazard ratios (HR) for OS (death) and PFS (progression or death) were estimated by Cox-proportional hazards models.


  • 646 of the 1,068 patients enrolled in the RTOG 9003 trial had OPC and were potentially eligible for consideration in this analysis.
  • 29% (190/646) of the OPC patients had available tissue that could be evaluated for p16, and 39% of these were found to be p16 positive.
  • OS/PFS outcomes were similar for patients with/without p16 data.
  • Patients with p16 positive tumors were more likely to have better Zubrod Performance status (PS), absent anemia, and T1 stage.
  • There were no differences in assigned treatment between p16 positive and negative patients.
  • P 16 positive patients were more likely to be non-smokers (21 % vs. 5%, p = 0.0007).
  • Overall median pack years of tobacco use was high (38 p-y), but was lower in the p16 positive group (29 vs. 45.9 p-y, p=0.016).
  • With a median f/u of 9.3 years, p16-pos OPC patients had better 5 year OS (49.0 vs. 19.6%, p<0.0001; HR 0.43, 95%CI 0.31-0.61) and 5 year PFS (43.6 vs.19.0%, p<0.0001; HR 0.45, 95% CI 0.32-0.63) when compared to p16 negative patients.
  • At 5-years, local-regional failure was lower for p16 positive OPC (30.5 vs. 54.9%, p<0.001).
  • However, there were no significant differences between p16 positive and p 16 negative patients in rates of distant metastases (11.1 vs. 13.0%, p=0.71) or second primary tumors (SPT, 13.8 vs. 11.4%, p=0.40).
  • When pack years of tobacco use were analyzed as a continuous variable, the hazard of death (OS) or progression and death (PFS) significantly increased by 1% per pack year of tobacco smoking (for both OS and PFS, HR 1.01, 95% CI 1.00-1.01, p<0.002), after adjustment for Zubrod PS, T stage, N stage and p16.

Authors' Conclusions

  • P 16 status and tobacco pack-years are strong independent predictors of OS and PFS in patients with OPC treated with radiotherapy alone.
    • Patients with p 16 positive tumors have approximately half the risk of death compared to p 16 negative patients.
    • Risk of death increases by approximately 1% per pack year of tobacco use.
  • The relative hazard of death or progression by p 16 status and pack years of tobacco use appears to be independent of treatment with radiotherapy or chemoradiation.
  • The absolute rates for OS, PFS, and SPT observed in this patient group may be compromised in part by higher cumulative pack years of tobacco use in patients enrolled a decade before those in the RTOG 0129 trial.

Clinical Implications

  • There is a growing body of evidence in the literature that HPV status is a significant prognostic factor for patients with head and neck cancer.
  • The current study was an unplanned retrospective subset analysis of patients enrolled in the RTOG 9003 trial, to evaluate HPV status and tobacco use as prognostic factors in HNC patients treated with radiation alone.
  • The authors used appropriate statistical methods to control for potential confounders between the two comparison groups. However, the results of the study are still limited by the retrospective nature of the study and that evaluable tissue was only available for 29% of the patients in the original RTOG 9003 trial.
  • The results of the study further support the findings previously observed in patients receiving combined chemoradiation in the RTOG 0129. HPV status and pack-years of tobacco use were again found to be strong independent predictors of OS, and PFS.
  • The prognostic value of HPV status and tobacco use history appears to be independent of therapy with either chemoradiation or radiotherapy alone.
  • These factors should therefore be considered when designing future clinical trials for head and neck cancer patients. Further efforts to provide education regarding the relationship of HPV to cancer development may also be important from a public and social health standpoint.

Frequently Asked Questions