Patients with human papillomavirus (HPV)-positive squamous cell carcinoma of the oropharynx (SCCOP) enjoy better treatment outcomes than patients suffering from HPV-negative head and neck cancer. To maintain the integrity and utility of future clinical trials, HPV-positive SCCOP must be studied as a distinct entity. The discovery of HPV-positive disease has (1) convoluted comparison of current phase II trial data to historical controls, (2) made formal stratification for HPV infection status an imperative for future phase III trial design, and (3) drawn focus toward opportunities for personalization of treatment intensity. This review discusses these research issues.
- •
Patients with human papillomavirus (HPV)-positive squamous cell carcinoma of the oropharynx (SCCOP) have favorable clinical outcomes relative to patients with HPV-negative SCCOP.
- •
Patient selection with validated biomarkers is critical for personalized clinical trial design.
- •
HPV-associated SCCOP represents an opportunity for careful treatment deintensification to reduce toxicity. New radiotherapy, less invasive surgery, and systemic treatment options serve to make this a feasible goal.
- •
New approaches are required to identify and treat HPV-negative SCCOP at the highest risk for treatment failure.
Introduction
Squamous cell carcinoma of the head and neck (SCCHN) is a heterogeneous disease arising from various subsites within the head and neck region. As cigarette use has declined over the past 3 decades in the United States, there has been a gratifying reduction in the incidence of oral cavity, larynx, and hypopharynx cancers. In stark contrast, the incidence of SCCOP has been steadily increasing, and the literature demonstrates that the common etiologic factor for this new population of patients is pharyngeal infection with oncogenic strains of HPV. The epidemiologic and pathophysiologic differences between HPV-positive SCCOP and HPV-negative SCCOP, as well as methods of HPV detection, are discussed in the articles by Fakhry, Westra and Pai elsewhere in this issue. This article focuses on the impact of HPV infection in clinical trials and patient care.
The ultimate goal of therapeutic clinical trials is to develop novel, safe, and efficacious therapies for patients that can improve clinical outcomes. To bring experimental agents from laboratories to clinics, they have to undergo multiple testing steps through preclinical studies and phase I, II, and III clinical trials. Once safety and efficacy are established in phase I and II trials, randomized phase III clinical trials are conducted for a comparison with the current standard of care. To avoid bias in the treatment arms and allow comparisons from trials to trials, clearly defining the study population through strict eligibility criteria is important. Traditionally the eligibility criteria included general clinical parameters, such as age, tumor histology, disease stage, performance status, and hematological and nonhematological end-organ function measures. It has been known, however, that these clinical parameters include heterogeneous patient populations with various outcomes given the same treatment. With these limitations in patient selection, various biomarkers were developed in an attempt to further molecularly define the disease with information gained from laboratory studies in addition to existing clinical parameters. One of the most powerful prognostic biomarkers found to date is HPV status in SCCHN.
Retrospective identification of improved treatment response and control of HPV-positive tumors in completed phase II trials is summarized in Table 1 . One of the first studies showing this prognostic difference was a secondary analysis of the Eastern Cooperative Oncology Group (ECOG) phase II clinical trial E2399. This study was originally designed to determine the role of induction chemotherapy followed by concurrent chemoradiotherapy for organ preservation in resectable stage III or IV squamous cell carcinoma of the larynx or oropharynx. Patients with HPV-positive tumors had significantly higher response rate and longer progression-free survival (PFS) and overall survival (OS). In a multivariate analysis adjusting for age, tumor stage, ECOG performance status, and tumor HPV status, HPV status was an independent prognostic variable.
| Study | Subsite | Treatment Regimen | N (n) | HPV Status | Progression-Free Survival | Overall Survival | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Pos (%) | Neg (%) | HPV+ | HPV− | Log Rank ( P Value) | HPV+ | HPV− | Log Rank ( P Value) | ||||
| E2399 Fakhry et al | L or OP | Sequential therapy: paclitaxel/carboplatin radiation 70 Gy + paclitaxel | 105 (96) | 38 (40) | 58 (60) | 2-y (86%) | 2-y (53%) | 0.02 | 2-y (95%) | 2-y (62%) | 0.005 |
| UMCC 9921 Worden et al | OC, OP, HP, or L | Sequential therapy: cisplatin/5-fluorouracil or carboplatin/5-fluorouracil Responders: radiation + cisplatin or Nonresponders: surgery + radiation | 66 (42) | 27 (64) | 15 (36) | Favorable prognosis for HPV+ ( P = .004) | Favorable prognosis for HPV+ ( P = .008) | ||||
| Kies et al | OC, OP, HP, L, or NP | Sequential therapy: carboplatin/paclitaxel/cetuximab Early stage: radiation 66–72 Gy or Late stage: 66–72 Gy + cisplatin or carboplatin | 47 (26) | 12 (46) | 14 (54) | Favorable prognosis for HPV+ ( P = .012) | Favorable prognosis for HPV+ ( P = .046) | ||||
With the preliminary but powerful evidence that the HPV status may have a significant impact in patient care and outcome, secondary post hoc analyses of tumor samples from several phase III clinical trials have been conducted ( Table 2 ). The first study to confirm these findings was from the phase III Radiation Therapy Oncology Group (RTOG) 0129 trial. This trial was originally designed to compare standard-fractionation radiotherapy and cisplatin with accelerated-fractionation radiotherapy and cisplatin. Of the 721 patients enrolled, 323 (45%) patients had the oropharyngeal primary subsite with available tissue samples for HPV testing. Among these patients, 206 of 323 (64%) were HPV positive according to in situ hybridization and 214 of 323 (66%) were positive for p16 protein overexpression by immunohistochemistry as a surrogate marker of biologically active HPV infection. Consistent with the earlier studies, the demographics of patients with HPV-positive tumors were distinct from the patients with HPV-negative tumors. These patients were of younger age, had less-extensive tobacco exposure histories, and enjoyed better performance status. Furthermore, the patients with HPV-positive tumors had smaller primary tumors and more-extensive lymph node metastasis relative to patients with HPV-negative tumors. Although the clinical outcomes between the 2 comparison arms were not significantly different, secondary analysis confirmed significantly superior survival in patients with HPV-positive tumors versus HPV-negative disease.
| Study | Subsite | Treatment Regimen | N (n) | HPV or p16 Status | Progression-Free Survival | Overall Survival | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Pos (%) | Neg (%) | HPV+/p16+ | HPV−/p16− | Log Rank ( P Value) | HPV+/p16+ | HPV−/p16− | Log-Rank ( P Value) | ||||
| RTOG 0129 Ang et al | OC, OP, HP, or L | Standard-fractionated radiation + cisplatin vs hyperfractionated radiation + cisplatin (no survival differences seen between the 2 treatment arms) | 743 (323) | HPV: 206 (64) p16: 215 (68) | HPV: 117 (36) p16: 101 (32) | 3-y HPV (74%) 3-y p16 (74%) | 3-y HPV (43%) 3-y p16 (38%) | HPV <0.001 p16 <0.001 | 3-y HPV (82%) 3-y p16 (84%) | 3-y HPV (57%) 3-y p16 (51%) | HPV <0.001 p16 <0.001 |
| DAHANCA 5 Lassen et al | OP, HP, NP, or L | Radiation | 195 (156) | p16: 35 (22) | p16: 121 (78) | 5-y a p16 (70%) | 5-y a p16 (40%) | NA | 5-y p16 (62%) | 5-y p16 (26%) | 0.0003 |
| Radiation + nimorazole | 219 (175) | p16: 49 (28) | p16: 126 (72) | ||||||||
| DAHANCA 6&7 Lassen et al | OC, OP, HP, NP, or L | 5 Fractions/wk radiation | 726 (385) | p16: 84 (22) | p16: 301 (78) | 5-y a p16 (78%) | 5-y a p16 (64%) | p16 0.001 | 5-y p16 (62%) | 5-y p16 (47%) | <0.0001 |
| 6 Fractions/wk radiation | 750 (409) | p16: 95 (23) | p16: 314 (77) | ||||||||
| TROG 02.02 Rischin et al | OC, OP, HP, or L | Radiation + cisplatin vs radiation + cisplatin + tirapazamine (no survival differences seen between the 2 treatment arms) | 861 (185) | p16: 106 (57) | p16: 79 (43) | 2-y b p16 (87%) | 2-y b p16 (72%) | p16 0.003 | 2-y p16 (91%) | 2-y p16 (74%) | 0.004 |
| TAX 324 Posner et al | OC, OP, HP, or L | Induction chemotherapy regimen Docetaxel/cisplatin/5-fluorouracil vs cisplatin/5-fluorouracil (no survival differences seen between the 2 treatment arms in the subset of HPV-tested patients) | 501 (111) | HPV: 56 (50) | HPV: 55 (50) | 5-y HPV (78%) | 5-y HPV (28%) | <0.0001 | 5-y HPV (82%) | 5-y HPV (35%) | <0.0001 |
Moving forward: head and neck clinical trial design in the HPV era
Patients with HPV-positive tumors, therefore, are a clinically distinct subgroup within SCCHN. Study of such patients in future clinical trials requires consideration of several key issues:
- 1.
How can current phase II trial data be reliably compared with historical controls?
- 2.
How can HPV infection status best be incorporated into the statistical design of new phase III trials?
- 3.
How is personalizing treatment for patients with favorable HPV-positive cancer and high-risk HPV-negative disease best addressed?
Issue 1: Control Data from Older Clinical Trials are Beset by New Confusion
This first issue stems from the increasing incidence of HPV infection and HPV-associated cancers in the context of decreased tobacco use. The incidence of HPV-related cancer increased from 1973 to 2004 whereas the overall incidence of HPV-unrelated cancer declined. Many clinical trials were conducted between 1980 and 2000, and these trials continue to provide standard-of-care control treatments and reference data for calculation of statistical power and sample size for current trials. Tumor tissue samples from most of these trials are not available for HPV testing and it is difficult to accurately infer the HPV status based on the clinical characteristics, such as younger age, smoking history, and better performance status. It is difficult to directly apply results from these series to outcomes data generated from clinical trials performed in the HPV era.
Issue 2: New Clinical Trial Designs Must Directly Address HPV Infection Status
It is mandatory that patient enrollment in competing treatment arms of phase III trials be balanced for important prognostic characteristics to permit valid comparisons of treatment outcomes. Study arms containing unequal proportions of HPV-positive patients and HPV-negative patients with differing baseline relative risks for treatment failure would introduce severe bias and obscure final interpretation. For instance, if a phase III clinical trial is performed to study an experimental drug, this drug may falsely seem superior to current standard of care if the study cohort treated with the experimental drug contained more favorable-risk HPV-positive patients. In addition, difference in drug efficacy may be biased by either known or uncharacterized biologic differences between HPV-positive tumors and HPV-negative tumors. To categorically avoid these biases, growing consensus among investigators exists to support the opening of phase II/III trials that strictly study either HPV-positive disease or HPV-negative disease using HPV status as an integral biomarker.
Issue 3: Can Treatment Intensity Be Personalized According to HPV Infection Status to Improve Therapeutic Ratio?
Currently treatments of SCCHN have a severe impact on quality of life. Although aggressive radiotherapy and chemoradiotherapy regimens have improved survival outcomes, such treatment yields severe post-treatment morbidity. Older radiotherapy techniques often resulted in restrictive fibrosis of the alimentary tract, which impaired swallowing and airway protection. Current intensity-modulated radiotherapy (IMRT) techniques permit protection of normal tissues adjacent to tumors but still result in significant dysphagia and oral debilitation. Emerging data demonstrate that reducing the dose to the uninvolved larynx and pharyngeal axis improves post-treatment rehabilitation. Two reports based on prospectively gathered data demonstrate that modest reductions in radiation dose to oral cavity, superior pharyngeal constrictor muscles, and/or larynx are potentially associated with preserved long-term swallowing function.
Thus, HPV-associated SCCOP represents a unique opportunity for carefully selected radiation dose shielding of novel dysphagia-specific anatomy to improve on-treatment therapeutic index and long-term quality of life. Additional focus has also been placed on reducing the overall intensity of combined modality treatment. It has been proposed that rational radiation dose reduction to responsive primary and nodal disease, coupled with avoidance of concurrent chemoradiotherapy, in patients triaged by induction chemotherapy to single-modality radiation in cases of complete response to induction is a promising approach. The rationale for use of induction chemotherapy is further strengthened by data confirming that the incidence of distant metastatic failure of HPV-positive disease is no lower than for high-risk HPV-negative disease. Thus, this strategy has the potential to:
- 1.
Reduce morbidity associated with concurrent chemoradiotherapy
- 2.
Improve systemic disease control through intensification of chemotherapy doses during the induction phase of treatment
- 3.
Maintain gains in locoregional disease control in a patient population at low risk for such a failure pattern.
Biomarkers: EGFR
HPV infection status has an impact on the potential utility of other candidate biomarkers of SCCHN outcomes. Tumor HPV status alone has limited relevance to non-oropharyngeal SCCHN, cannot discriminate prognosis with ideal specificity, and cannot predict outcomes for specific treatments. Fortunately, a growing list of complementary candidate biomarkers promises to improve predictive capability. A key example is epidermal growth factor receptor (EGFR), a cell-surface tyrosine kinase receptor critical to epithelial development and maintenance. Although some studies suggest poor prognosis after surgery or cytotoxic therapy in tumors with EGFR gene dosage, this result has not been consistently reproduced and has never been correlated with EGFR protein expression. A potentially more satisfying strategy would be to combine EGFR measures with mechanistically related markers of parallel signaling pathways or HPV infection status in clinical head and neck tumor specimens. Early studies suggested that HPV infection is inversely correlated with EGFR protein expression and that EGFR expression status may retain prognostic relevance regardless of HPV infection status. A more recent series has subsequently confirmed increased EGFR gene copy number status (as detected by fluorescence in situ hybridization) to remain largely confined to HPV-unassociated (eg, p16-negative) cancers. Nonetheless, this study also showed that p16 expression supersedes EGFR-specific markers on multivariate analysis. Definitive prospective corroboration remains necessary but taken together these findings suggest a potential need to regularly combine at least HPV-specific and EGFR-specific biomarkers to guide future clinical strategies.
Imaging-based biomarkers: FDG-PET/CT
HPV infection status also has particular relevance to the ultimate utility of imaging-based biomarkers of SCCHN. Considerable interest has focused on fludeoxyglucose F 18 (FDG)–positron emission tomography (PET)/CT monitoring of SCCHN response to radiotherapy. Several groups have found that FDG-PET post-treatment restaging provides high negative predictive power ; accordingly, there is growing acceptance of withholding consolidative neck dissection after radiotherapy in the absence of residual FDG-avid adenopathy, although some investigators argue that expert clinical interpretation of serial CT imaging could achieve similar results. A potentially more effective and efficient approach would be to emphasize identification of specific clinical situations where FDG-PET/CT diagnostic yield may be optimized. FDG-PET/CT utility has been studied in the context of other important clinical parameters, in particular HPV infection status, through a bayesian, risk-based approach classically used by clinicians choosing between alternative diagnostic tests in specific patients. FDG-PET/CT seems to provide little value over CT alone in radiation response assessment for unselected patients with locally advanced SCCHN. Nonetheless, FDG-PET/CT can significantly improve assessment of treatment response in high-risk patients, such as those with HPV-unassociated disease. These results provide critical impetus to incorporate risk-based stratification strategies informed of HPV infection status into FDG-PET/CT response assessment of locally advanced SCCHN. Such an individualized, context-specific approach will be relevant to any current or future imaging-based biomarker.
How Can These Issues Be Reconciled in Future Trials?
Currently ongoing trials that use HPV-positive status as an integral biomarker for eligibility criteria are summarized in Table 3 . The first national cooperative group trial to test this concept is ECOG-E1308. This phase II study tests the role of induction chemotherapy followed by de-escalation of radiation dose and substitution of platinum chemotherapy with cetuximab. After the completion of induction chemotherapy, patients with complete response receive dose-reduced radiation with cetuximab. Patients with less than complete response receive standard curative doses of radiation with concurrent cetuximab. The RTOG has also opened a phase III clinical trial (RTOG 1016) directly comparing concurrent radiation/cisplatin with concurrent radiation/cetuximab. This trial will answer whether de-escalation of treatment away from conventional cytotoxic chemotherapy will result in decreased toxicity without compromising survival outcomes. There are several smaller institutional trials being conducted throughout North America and Europe designed to answer similar questions with varying approaches.
