Intensity-modulated radiotherapy outcomes for oropharyngeal squamous cell carcinoma patients stratified by p16 status




Patients with oropharyngeal squamous cell carcinoma (OPSCC) treated with intensity-modulated radiotherapy (IMRT) were stratified by p16 status, neck dissection, and chemotherapy to correlate these factors with outcomes.


A total of 112 patients with OPSCC treated with IMRT from 2002 to 2008 were retrospectively analyzed. All patients received RT to 66-70 Gray. Forty-five of the tumors were p16 positive (p16+), 27 were p16 negative (p16−), and 41 had unknown p16 status. Sixty-two patients had postradiation neck dissections. Nine patients with p16− tumors and 28 patients with p16+ tumors received chemotherapy. The distribution of T, N, and stage grouping among the p16+ and p16− patients was not significantly different, and 87.5% patients had stage III/IV disease.


The median follow-up was 26.3 months. For patients with p16+ tumors, p16− tumors, and the overall cohort, the actuarial 3-year locoregional progression-free survival rate was 97.8%,73.5%, and 90.5% respectively (P = .006) and the disease-free survival rate was 88.2%, 61.4%, and 81.7%, respectively (P = .004). Patients with p16+ tumors had an 89.5% and 87.5% pathologic complete response (CR) on neck dissection with and without chemotherapy, respectively. In contrast, patients with p16− tumors had a 66.7% and 25.0% pathologic CR on neck dissection with and without chemotherapy, respectively.


In this series, p16 status was found to be a significant predictive biomarker and patients with p16+ tumors had much better outcomes than patients with p16− tumors. Further investigation is warranted to determine whether less intense therapy is appropriate for selected patients with p16+ OPSCC, whereas more aggressive strategies are needed to improve outcomes in patients with p16− disease. Cancer 2010. © 2010 American Cancer Society.

Greater than 35,000 cases of oral cavity and pharynx cancer were estimated to occur in the United States in 2008.1 Oropharyngeal squamous cell carcinoma (OPSCC) incidence has increased in the United States in nontraditional populations (ie, nonethanol and nontobacco users), secondary to human papillomavirus (HPV) -associated malignancies.2, 3 HPV-associated tumors represent a distinct epidemiologic, clinical, and molecular entity.4 Sexual transmission of HPV is a causative factor in virtually all squamous cell carcinomas of the cervix, and more recently HPV infection in the upper aerodigestive tract has been associated with head and neck squamous cell carcinomas (HNSCC), most commonly in the oropharynx.5 Data from multiple studies have indicated that HPV-associated tumors account for 20% to 75% of OPSCC.6-8 HPV-16 is the most common viral subtype associated with malignant transformation, ranging between 90% and 95% of all cases.4 HPV DNA can be directly assayed in biopsies by polymerase chain reaction (PCR)9 or in situ hybridization.10 Greater than 86% of HPV-associated tumors overexpress p16INK4A (p16), a cyclin-dependent kinase inhibitor, and only 3% of HPV-nonassociated tumors overexpress p16.11 p16 status can be identified by immunohistochemical techniques and has been reported to be an independent prognostic factor.11 In 1 study, p16 immunohistochemistry followed by PCR detection of HPV in p16-positive (p16+) cases has yielded a sensitivity and specificity of 100% with a false-positive rate estimated to be 2%.12

HPV-associated tumors have been reported to be more sensitive to chemotherapy and RT than non–HPV-associated tumors and p16 status is a significant prognostic indicator for disease-free survival (DFS) with reported 5-year DFS of 84% and 46% for patients with p16+ and p16-negative (p16−) tumors, respectively.11 The prediction for DFS based on p16 expression was superior to all clinicopathologic parameters normally used for treatment decisions and assessment of prognosis.11 In a meta-analysis of 37 studies reporting on HPV status and outcome for HNSCC, patients with HPV+ OPSCC had a 28% reduced risk of death compared with patients with HPV− OPSCC.13 An Affimatrix Human U133A GeneChip analysis of HPV+ and HPV− HNSCC indicated that these tumors had different gene expression profiles14; however, it remains unknown why p16+ tumors are more radiation responsive. It may be due to higher proliferation rates of these poorly differentiated tumors because radiation preferentially kills mitotically active cells.15 Despite the extensive published prognostic importance of HPV status, and the finding that the National Comprehensive Care Network clinical practice guidelines recommend determining either HPV or p16 status for OPSCC patients,16 we recently reported a survey demonstrating that approximately 40% of responding radiation oncologists obtained either of these pathologic assays.17

Radiotherapy (RT) techniques for OPSCC have undergone technical advancements over the past decade with the development of intensity-modulated RT (IMRT) and image guidance RT. IMRT can deliver a conformal dose distribution to target volumes while sparing adjacent nontarget tissues, causing less biologic injury to normal adjacent structures while maintaining similar or improved locoregional tumor control rates.18-20 Additional studies have reported reduced late toxicities after IMRT compared with non-IMRT techniques, resulting in improved quality in life.21, 22 Survey results from responding US radiation oncologists revealed that IMRT utilization for HNSCC was 32.1% in 2002,23 80.5% in 2004,24 and 100% in 2009.17 Despite the widespread adoption of IMRT to treat OPSCC patients, to the best of our knowledge there are only single-institution retrospective publications of the outcomes of OPSCC patients treated with IMRT.25-31 Some studies included patients treated postoperatively, and all ranged between 50 and 74 patients, making our series (to the best of our knowledge) the largest reported series of OPSCC patients definitively treated with IMRT to date and the only study reporting stratification of p16 tumor status for outcome analysis.

Patients with advanced stage OPSCC are commonly treated with platinum-based chemoradiation with 70 Gray (Gy) in 35 fractions to macroscopic disease and 50 Gy in 25 fractions to the uninvolved bilateral cervical lymphatics, followed by neck dissection(s) if they have locally advanced lymph node disease.16 Despite advances in technology and utilization of IMRT, treatment still causes significant acute toxicity, long-term morbidity, reduced functional status, and a poor quality of life for many patients. We analyzed outcomes of OPSCC patients treated with IMRT and stratified patients based on p16 status to determine its influence on outcomes. Our goal was to elucidate patterns of failure for OPSCC patients and to determine whether p16 status could be used to select patients with p16− disease for more aggressive protocols and patients with p16+ disease for less aggressive protocols.


Patient Data

Data were retrospectively collected on an institutional Human Investigations Committee approved protocol. Patient care was prospectively discussed by members of the Head and Neck Cancer Tumor Board. Between January 1, 2002 and May 31, 2008, 118 patients were treated with definitive IMRT for histologically confirmed OPSCC. Patients with a history of prior head and neck irradiation or cancer, neck dissection before irradiation, or distant metastases at diagnosis were excluded from the analysis, leaving 112 patients. Patients were treated with multiple chemotherapy regimens during this interval, including some treated on an institutional phase 1 pilot study of induction and concurrent carboplatin and capecitabine with IMRT.32 Generally, most patients with T3 to T4 primary tumors or N2 to N3 lymph node disease who were fit enough received chemotherapy. Prospectively planned neck dissections were performed 6 weeks after RT.

RT Planning

An aquaplast mask was used for immobilization, and computed tomography (CT) simulation was performed. The gross (macroscopic) tumor volume (GTV) was comprised of the primary tumor and pathologically involved lymph nodes based on pretreatment imaging. The clinical target volume (CTV) included the GTV with margin for microscopic spread and the draining cervical lymph node levels at risk for microscopic spread.33-35 Retropharyngeal and ipsilateral level II to IV lymph nodes were treated in all patients. Bilateral level II to IV lymph nodes were treated (except for a few patients with early tonsillar primary tumors who only received ipsilateral lymph node irradiation). Level V and level IB lymph nodes were treated in patients with pathologic ipsilateral adenopathy. The high level II lymph node volumes were treated to the jugular foramen on the side ipsilateral to the primary tumor and in any neck that contained known pathologic adenopathy. If the contralateral neck was without pathologic adenopathy, the jugulodigastric lymph node was the most superior lymph node treated.33 A 3-mm expansion of the CTV accounted for setup uncertainties to create the planning treatment volume (PTV). This expansion was based on published data and a retrospective institutional analysis of mean HNSCC patient setup inaccuracies determined by pretreatment CT image guidance.36 Organs at risk contoured included the brainstem, spinal cord, mandible, oral cavity, parotid glands, and larynx.

Patients were treated at a dose of 1.8 to 2.0 Gy per day, 5 days per week. The mean dose of at least 1 parotid gland was kept below 26 Gy.37 Patients received a dose to the primary tumor of 66 to 70 Gy and 50 to 70 Gy to the pathologic lymph nodes depending on the extent of lymph node disease and whether a postradiation neck dissection was prospectively planned. Clinically uninvolved cervical lymph nodes (N0 lymph nodes) were treated to 50 Gy. A 3-dimensional (3D) conformal or IMRT technique was used to boost the macroscopic disease with 14 to 20 Gy after 50 to 56 Gy was delivered with IMRT. IMRT plans were optimized so that 95% of the PTV received the prescribed dose.

During the early phase of this series, IMRT was performed with Corvus planning software (Nomos Corporation, Sewickly, Pa) and delivered using a “step and shoot” technique on a Varian 2300 LINAC (Varian Medical Systems, Palo Alto, Calif) using 7 to 9 coplanar fields. After December 2004, patients were treated on a Hi-Art Helical Tomotherapy Unit (TomoTherapy Inc., Madison, Wis). We previously published the dosimetric differences between these techniques.38 RT-induced mucositis, dermatitis, and xerostomia toxicity grading was based on the Radiation Therapy Oncology Group/ European Organization for Research and Treatment of Cancer (RTOG/EORTC) toxicity grading system.

Tissue Microarray and Immunohistochemistry

Archived pretreatment primary tumor biopsies were available for 72 patients in an institutional tumor bank. An adequate tissue specimen was not available for all 112 patients because the majority of tissue samples were archived specimens. A tissue microarray was created, and p16 status was determined by immunohistochemistry as specified by Reimers.11 Tumors were considered positive for p16 when strong nuclear and cytoplasmic staining was present in >60% of cells.11 To overcome the variability in interpretation of immunohistochemical staining patterns, preparations of HeLa cells that we have verified contain HPV genomes were used as positive controls.39 Others have also used these as controls for aberrant p16 gene product expression.40, 41

Statistical Analysis

Survival was calculated from date of diagnosis to date of death or last follow-up. Social security death index was searched for patients with unknown vital status. Winks SDS statistical software (version 6 [Texasoft, Cedar Hill, Tex]) was used for calculating actuarial survival curves. The Student t test, chi-square test, and Fisher exact test were used to assess differences in patient characteristics and outcomes using statistical functions in Microsoft Excel 2003 (Microsoft Corporation, Redmond, Wash). The log-rank statistic was used to identify differences between survival rates. A P value threshold of .05 was used for statistical significance.



Demographic characteristics of patients are shown in Table 1. Ninety-eight patients (87.5%) had stage III/IV disease at diagnosis (according to the sixth edition of the American Joint Committee on Cancer staging manual). Tonsillar and base of tongue tumors constituted 91.9% of primary tumors.

Table 1. Demographics of Patients Stratified by p16-Statusa
 Allp16 Status KnownP
  • NOS indicates not otherwise specified.

  • a

    Percentages represent distributions for each characteristic within each subgroup.

  • b

    Statistical comparisons between p16-positive and p16-negative patient groups were determined by the Student t test.

  • c

    Statistical comparisons between p16-positive and p16-negative patient groups were determined by the Fisher exact test.

  • d

    Statistical comparisons between p16-positive and p16-negative patient groups were determined by chi-square analysis.

No.112 (100.0%)72(64.3%) 
  p16 negativep16 positive 
p16 status 27 (24.1%)45 (40.2%) 
Mean age, y57.361.957.6 
Gender   .1520c
 Male91 (81.3%)20 (74.1%)39 (86.7%) 
 Female21 (18.8%)7 (25.9%)6 (13.3%) 
Primary site   .0201d
 Tonsil51 (45.5%)15 (55.6%)20 (44.4%) 
 Base of tongue52 (46.4%)8 (29.6%)24 (53.3%) 
 Soft palate/tonsil pillar8 (7.1%)4 (14.8%)0 (0.0%) 
 Lateral pharyngeal wall1 (0.9%)0 (0.0%)1 (2.2%) 
Primary tumor classification   .6862d
 T132 (28.6%)9 (33.3%)17 (37.8%) 
 T247 (42.0%)12 (44.4%)20 (44.4%) 
 T317 (15.2%)2 (7.4%)5 (11.1%) 
 T416 (14.3%)4 (14.8%)3 (6.7%) 
Lymph node classification   .0800d
 N019 (17.0%)6 (22.2%)4 (8.9%) 
 N118 (16.1%)6 (22.2%)7 (15.6%) 
 N2a12 (10.7%)1 (3.7%)6 (13.3%) 
 N2b42 (37.5%)13 (48.1%)16 (35.6%) 
 N2c18 (16.1%)1 (3.7%)10 (22.2%) 
 N33 (2.7%)0 (0.0%)2 (4.4%) 
Stage group   .2691d
 I4 (3.6%)0 (0.0%)0 (0.0%) 
 II10 (8.9%)5 (18.5%)2 (4.4%) 
 III20 (17.9%)6 (22.2%)10 (22.2%) 
 IVA73 (65.2%)15 (55.6%)31 (68.9%) 
 IVB5 (4.5%)1 (3.7%)2 (4.4%) 
Histology   .2108d
 Well differentiated10 (8.9%)2 (7.4%)6 (13.3%) 
 Moderately differentiated37 (33.0%)11 (40.7%)12 (26.7%) 
 Poorly differentiated53 (47.3%)9 (33.3%)25 (55.6%) 
 NOS12 (10.7%)5 (18.5%)2 (4.4%) 

Patient characteristics for those with p16+ and p16− disease were similar, with no statistically significant differences noted with regard to age, gender, T stage, or stage group. The p16− patients had more tonsil and soft palate cancers than the p16+ patients (P = .02). There were no soft palate tumors in the p16+ group. Differences between the N stages of the 2 groups was not statistically significant (P = .08), but patients with p16− disease demonstrated a trend toward a lower lymph node stage at presentation, with 44.4% of patients staged N0 or N1 versus 24.5% for patients with p16+ disease. The p16+ tumors exhibited more poorly differentiated histology (55.6%, in contrast to 33.3% for p16− tumors); however, this did not reach statistical significance (P = .21).

Chemotherapy and Neck Dissection

Sixty-eight patients (60.7%) received chemotherapy. Approximately one-third with p16− disease and 62.2% of patients with p16+ disease received chemotherapy with the agents shown in Table 2. In both groups, patients treated with chemotherapy trended toward higher stage grouping (p16−, P = .09; p16+, P = .03). For patients with p16+ disease, 96.4%who were treated with chemotherapy were free of disease at the time of last follow-up, whereas 82.4% of patients who did not receive chemotherapy were disease-free (median follow-up, 29.2 months). For patients with p16− tumors, only 66.6% who were treated with chemotherapy and 38.8% of patients who did not receive chemotherapy were free of disease at the time of last follow-up (median follow-up, 22.2 months).

Table 2. Chemotherapy
  1. PEG indicates percutaneous endoscopic gastrostomy; PICC, peripherally inserted central catheter; 5-FU, 5-fluorouracil.

PEG tube placed during treatment46 (41.0%)
PICC/central line28 (25.0%)
Chemotherapy68 (60.7%)
Cisplatin51 (45.5%)
Carboplatin15 (13.3%)
Intravenous 5-FU25 (22.3%)
Capecitabine26 (23.2%)
Cetuximab3 (2.6%)
Docetaxel/paclitaxel17 (15.1%)
Induction chemotherapy59 (52.6%)
Average no. of induction chemotherapy cycles2
Average no. of concurrent chemotherapy cycles3
Concurrent chemotherapy61 (54.4%)

Sixty-two (55.3%) patients underwent a postradiation neck dissection; either bilateral or unilateral. The majority of these were selective or modified radical neck dissection(s) as shown in Table 3.

Table 3. Neck Dissection Characteristicsa
Planned Postradiation Neck Dissections in Patients with N+ DiseaseNo.
  • SND indicates selective neck dissection; MRND, modified radical neck dissection; RND, radical neck dissection.

  • a

    Some patients underwent 2 separate types of neck dissections and both are accounted for in the table.

Total no. patients with neck dissections62

RT Toxicity

Grade 3 mucositis was the peak grade in 81.0% of patients, and 70.9% of patients experienced peak grade 2 dermatitis. Peak xerostomia was grade 1 in 75.2% of patients at the end of treatment, 81.3% at 6 months after RT, and 71.8% 1 year after RT.

Clinical, Radiologic, and Pathologic Response After RT

Patients were evaluated both clinically and radiologically after completing RT (on average, 29.5 and 31.6 days, respectively) (Table 4). Clinical complete response (CR) of the primary tumor was noted in 93.4% of patients, while radiologic primary CR was observed in 71.0%. On clinical evaluation of the involved lymph nodes, 70.7% of patients had achieved a clinical CR, whereas 45.5% had achieved a radiologic CR (Table 4).

Table 4. Comparison of Clinical, Radiologic, and Pathologic Response
ResponseClinical ResponseRadiologic ResponsePathologic Response
  1. NA indicates not available; CR, complete response; PR, partial response.

Primary CR100 (93.4%)71 (71%) 
Primary PR7 (6.5%)25 (25%) 
Primary disease progression or no response04 (4%) 
Lymph nodesN=99N=90N=62
Lymph node CR70 (70.7%)41 (45.5%)48 (77.4%)
Lymph node PR29 (29.2%)38 (42.2%)14 (22.6%)
Lymph node disease progression or no response2 (2.0%)11 (12.2%)NA
Average time from radiation to clinical examination/imaging29.5 d31.6 dNA

Sixty-two patients with pretreatment cervical lymph node disease had postradiation neck dissections; 48 of these patients (77.4%) achieved a pathologic CR (no viable cancer in the specimen). Of the patients treated with neck dissection and who received chemotherapy, 84.4% had a CR versus 58.8% of patients treated without chemotherapy (P = .05). A total of 66.7% of patients who had p16− disease and received chemotherapy had a pathologic CR compared with 25.0% of patients with p16− disease treated without chemotherapy; however, the numbers were too small to reach statistical significance (P = .16) (Table 5). There was no statistical difference in the pathologic response rate for patients with p16+ disease treated with or without chemotherapy (P = .80). Overall, 88.9% of patients with p16+ disease had a pathologic CR on neck dissection versus 42.9% of those with p16− disease (P = .003).

Table 5. Three-Year Survival and PFS Stratified by p16 Status
OutcomeOverall N=112p16+ N=45p16- N=27Pa
  • PFS indicates progression-free survival; p16+, p16 positive; p16-, p16 negative; ChemoRT, chemoradiotherapy; RT, radiotherapy; DFS, disease-free survival; CSS, cause-specific survival; LN, lymph nodes; CR, complete response.

  • a

    P values shown are comparison between the p16-positive cohort versus the p16-negativer cohort.

Local PFS    
 RT alone85.5%94.1%70.5%.241
Regional PFS    
 RT alone88.5%90.5%79.4%.402
Locoregional PFS    
 RT alone85.2%94.1%70.4%.245
 RT alone74.0%75.3%58.9%.30
Overall survival    
 RT alone60.1%65.7%44.3%.174
 RT alone77.0%77.8%63.3%.420
Distant metastasis-free survival
 RT alone82.8%89.5%68.1%.140
LN pathologic CR    
 RT alone58.8%87.5%25.0%.020
N0 LN control    
 RT alone95.5%100%88.9%.257

Of the 62 patients who underwent neck dissection(s), 1 patient with unknown p16 status had disease in lymph node levels which was clinically determined to be free of disease before treatment. Thus, 98.4% of patients who underwent a neck dissection had pathologic confirmation of control of the N0 lymph node volumes. Of the patients who had pathologic lymph nodes on presentation (n = 31) and who did not undergo neck dissection, 96.8% had no observed lymph node recurrence on clinical follow-up. Of all patients without a neck dissection (n = 50), only 2 have developed a disease recurrence in the N0 cervical levels, and thus control of the observed N0 neck was determined to be 97.3% for all patients, 100% for patients with p16+ disease, 92.6% for patients with p16− disease, and 88.9% for patients with p16− disease who did not receive chemotherapy.


The median follow-up was 26.3 months for the entire cohort, 29.2 months for patients with p16+ tumors, and 22.2 months for patients with p16− tumors. Calculated actuarial local, regional, locoregional progression-free, disease-free, cause-specific, distant metastasis-free, and overall survival at 3 years are shown in Table 5 and Figure 1 stratified by p16 status and whether patients were treated with chemotherapy. Within the follow-up period, 11.1% of patients with p16− tumors who received chemotherapy progressed with metastatic disease compared with 27.8% of p16− patients who did not receive chemotherapy (P = .32).

Figure 1.

(Top Left) Locoregional control is shown for the overall cohort, p16-positive (p16+) patients, and p16-negative (p16−) patients (P = .006). (Top Right) Overall survival is shown for the overall cohort, p16+ patients, and p16− patients (P <.001). (Middle Left) Cause-specific survival is shown for the overall cohort, p16+ patients, and p16− patients (P = .013). (Middle Right) Disease-free survival is shown for the overall cohort, p16+ patients, and p16− patients (P = .004). (Bottom Left) Metastasis-free survival is shown for the overall cohort, p16+ patients, and p16− patients (P = .026). The legend is shown in the panel depicting the metastasis-free survival curve (bottom left).

Outcomes Based on p16 Status and Neck Dissection

The 27 patients with p16+ disease who underwent neck dissection had a median follow-up of 35.2 months. This group had statistically higher N stage on presentation than the 18 p16+ patients who did not undergo neck dissection (P = .02). The 14 patients whose tumors were p16− and underwent a neck dissection had a median follow-up of 33.2 months. This group had statistically greater N stage than the 13 patients with p16− tumors who did not undergo neck dissection (P = .04). Of patients with known p16 status undergoing neck dissection, only 57.1% of the patients with p16− disease were free of disease at the time of last follow-up compared with 92.6% of the patients with p16+ disease (P = .012). In contrast, of those without neck dissection and with p16+tumors, 16 of 18 (88.8%) were without evidence of disease, compared with 9 of 13 (69.2%) with p16− disease (P = .18). The median follow-up was 21.2 months for patients with p16+ disease and 19.2 months for patients with p16− disease without a neck dissection.


Large single-institution retrospective reports remain the main source of outcome information for OPSCC patients. Hodge et al reported the outcomes of 195 OPSCC patients treated from 1995 to 2005, with outcomes stratified by radiation technique and treatment era, including patients treated with IMRT (n = 52), 3D conformal techniques after the IMRT era (n = 38), and 3D conformal techniques before the IMRT era (n = 105).31 With a median follow-up of 30.4 months, they reported a 3-year locoregional control rate of 96.1%, 78.1%, and 67.7%, respectively, and cautioned that retrospective analyses are hampered by treatment selection bias and clinical trends in utilization of chemotherapy and neck dissection.31 In addition, the improved results for patients treated later could reflect a higher percentage of patients with HPV-associated tumors secondary to what Sturgis has questioned may be an emerging epidemic of HPV-associated HNSCC.7 Our clinical outcomes are similar to reports of other investigators as summarized in Table 6.25-31

Table 6. Published Outcomes of Oropharynx Carcinomas Using IMRT
StudyMedian Follow-Up, MonthsTotal No. of Patients (%) of Patients With Stage III or IV Oropharyngeal CarcinomaConcurrent Chemoradiation, No. (%)3-Year Local PFS Rate3-Year Locoregional PFS Rate3-Year Distant Metastasis-Free Survival3-Year DFS
  • IMRT indicates with intensity-modulated radiotherapy; PFS, progression-free survival; DFS, disease-free survival; NS, not specified; RT, radiotherapy.

  • a

    Indicates all patients had small (<4 cm) primary tumors of the oropharynx).

Current study27112 patients (88%)61 patients (54%)91%90%88%81%
De Arruda 2006251850 patients (92%)46 patients (86%)2-y: 98%2-y: 88%2-y: 84%NS
Huang 2008263371 patients (100%)71 patients (100%)94%90%NS81%
Chao 2004273374 patients (93%)17 patients (22%)NS4-y: 87% (78% in definitive vs 95% postoperative RT)4-y: 90% (84% in definitive vs 94% postoperative RT)4-y: 81%
Garden 200728a4551 patients (84%)4 patients (8%)2-y: 96%2-y: 94%NSNS
Yao 20062927.366 patients (93%)46 patients (70%)91.9%98.8%80.4%64.4%
Sanguineti 20083032.650 patients (88%)0 patients (0%)93.8%85.1%NSNS
Hodge 20073123.852 patients (86%)28 patients (54%)NS96.1%NSNS

At the time of presentation, patients with p16− and p16+ tumors were statistically similar with respect to clinical T stage and N stage, although those with p16+ tumors trended toward higher N stage. Regardless, patients with p16+ tumors still had better outcomes. All patients were treated blinded to p16 tumor status. We were unable to obtain p16 status on all tissue samples, but the similarity of the group with unknown p16 status to the overall cohort suggests that the absence of these patients from the studies of p16+ versus p16− patients should not impact the findings of the study. At the time of last follow–up, 91.1% of patients with p16+ tumors were without evidence of disease whereas only 62.9% of patients with p16− tumors remained free of disease. Our data confirm that p16 status is a significant prognostic indicator for locoregional control and survival after IMRT-based radiation, which is consistent with other studies.11, 42, 43

All patients undergoing neck dissection had statistically higher N stage on presentation than the patients not undergoing neck dissection, regardless of p16 status. In the patients with p16− disease, 42.9% of neck dissection specimens contained viable tumor, whereas only 11.1% of patients with p16+ disease contained viable tumor. Further investigation is warranted for selecting patients for postradiation neck dissection based on p16 status as a prognostic factor combined with other factors, such as lymph node fluorodeoxyglucose uptake in post-treatment positron emission tomography–CT scans.44

Approximately one-third of patients with p16− disease and 62.2% of patients with p16+ disease were treated with chemotherapy. Chemotherapy was used more in patients with p16+ tumors because they trended toward higher lymph node stage. The p16+ patients who received chemotherapy were directly compared with p16− patients who received chemotherapy in Table 5 to ensure that higher chemotherapy use in the p16+ patients was not a confounding factor to their improved outcomes. Regardless, the p16+ cohort still had statistically significant improved outcomes. Our 3-year locoregional control rate for p16+ patients treated with chemoradiation was 100% versus 77.8% for p16− patients treated with chemoradiation (P = .011). Furthermore, a meta-analysis of 93 randomized trials and 17,346 patients demonstrated only an absolute benefit of 4.5% at 5 years for chemotherapy.45

For patients with p16+ tumors, 96.4% of patients who were treated with chemotherapy were free of disease at the time of last follow-up, whereas 82.4% of patients who did not receive chemotherapy were free of disease (median follow-up, 29.2 months). For patients with p16− tumors, only 66.6% treated with chemotherapy and 38.8% who did not receive chemotherapy were disease free at the most recent follow-up (median follow-up, 22.2 months). Because patients with p16− disease who did not receive chemotherapy had significantly poorer outcomes, clinical protocols using more aggressive chemotherapy, neck dissections, and radiographic surveillance are warranted in these patients.

In our cohort, we had excellent control of the electively irradiated N0 lymph node volumes. On observed and pathologic evaluation of the N0 lymph node volumes, overall there was 97.3% control in the total cohort and 100% control in the p16+ cohort. This is consistent with other published series, including a meta-analysis reporting that 45 to 50 Gy results in >95% N0 lymph node control.46 Despite using IMRT, there is still room to improve on the radiation-induced toxicity profile. Approximately 80% of our cohort experienced grade 3 mucositis during treatment and peak grade 1 xerostomia at 6 months after RT. Currently, we have excellent control of the N0 neck, especially in patients with p16+ disease treated with or without chemotherapy, and we are developing a clinical trial aimed at reducing toxicity in this patient population through adaptive RT techniques to reduce the PTV expansion of the N0 lymph node volumes in patients treated with image-guided IMRT and concurrent cisplatin.


In the current series, patients with p16+ tumors had better outcomes than patients with p16− tumors and p16 status was found to be a significant prognostic indicator of locoregional control and survival. Clinical trials using p16 status for selection criteria are warranted to improve outcomes. Specifically, further investigation is warranted to determine if strategies to reduce toxicity are appropriate for selected OPSCC patients with p16+ disease, whereas more aggressive strategies are needed to improve outcomes in patients with p16− disease. When feasible, patients with OPSCC should have HPV DNA analysis or p16 immunohistochemistry performed on their biopsy specimens because of the prognostic information it provides.


The authors made no disclosures.