Keywords head and neck cancer - oropharyngeal squamous cell carcinoma - HPV - human papillomavirus
viruses
Introduction
The Human Papillomavirus (HPV) was first identified in 1949, and molecular evidence
for HPV as the etiologic agent of head and neck squamous cell carcinoma (HNSCC) appeared
in 1983.[1 ] Since that time, the incidence of HPV-related (HPV + ) oropharyngeal squamous cell
carcinoma (OPSCC) has increased dramatically, from roughly 16% of all OPSCCs in 1984
to 1988 to 72% by 2000 to 2004.[2 ] It has now become clear that HPV+ OPSCC has distinct characteristics from non-HPV
related (HPV-) OPSCC.
Delay in the diagnosis of HNSCC is associated with poorer outcomes,[3 ] and is often stratified into patient delay (symptomatic onset to first health care
professional appointment) and professional delay (first consultation to pathologic
diagnosis). A subset of patients with HNSCC has a prolonged delay from onset of their
symptoms to diagnosis; such delays can be attributed to a variety of social and disease
factors. A large survey of HNSCC patients demonstrated that 9% of patients with squamous
cell carcinoma (SCCA) outside of the oral cavity presented for an initial health care
appointment 6 months or more after their initial symptoms.[4 ] A 2015 study from Canada cited several factors contributing to delays in diagnosis
in HNSCC, including inappropriate treatment for infectious or reflux etiology, inappropriate
reassurances by health care provider, lack of primary care physician knowledge of
signs and symptoms of HNSCC, patient lack of knowledge of risk factors and symptoms
of HNSCC, and patients taking alternative medicines.[5 ] A recent study cited that most patients with HPV- OPSCC visited at least two providers
before diagnosis. Evaluation by three or more providers or being prescribed analgesia
prior to diagnosis was associated with significant delays in diagnosis of nearly a
year.[6 ] Additionally, patients with HPV+ OPSCC who were diagnosed at >12 months from symptom
onset were more likely to have T4 disease and higher overall American Joint Committee
on Cancer (AJCC) clinical stage at presentation than patients diagnosed < 12 months
from symptom onset.[7 ]
The aim of the present study was to further characterize HPV+ OPSCC by assessing for
the presence of diagnostic delay compared with HPV- OPSCC and evaluating the role
of various management factors that potentially impact time to diagnosis.
Materials and Methods
The present study has undergone formal review and approval by our institutional review
board (IRB), with the approval number 13457. We reviewed 554 cases of OPSCC treated
from January 2002 through December 2014. p16 immunohistochemistry (IHC) was used as
an indicator of HPV status for all patients from biopsy or surgical tissue when available.
Since 2008, p16 staining has been routinely performed in HNSCC biopsies or surgical
specimens; a tissue micro array (TMA) was used to obtain p16 status for a portion
of patients treated prior to this date. A total of 213 cases were excluded due to
lack of p16 IHC. A total of 57 cases were excluded due to inadequate documentation
or history of prior HNSCC. Ultimately, 284 cases were included. All patients underwent
treatment planning by the Head and Neck Tumor Board. Medical records were reviewed
to obtain patient demographics, smoking and alcohol history, presenting symptoms,
medical comorbidities, and to assess for delays in diagnosis. Univariate chi-squared,
t -test, and Fisher exact tests were utilized to compare these characteristics for p16+
versus p16- OPSCC. All tumors were staged according to AJCC 7th Edition staging guidelines.
Alcohol use was determined by consumption of > 5 alcoholic beverages per week.[6 ] To assess factors independently associated with diagnostic delay in this population,
a multivariable logistic regression analysis was performed (IBM SPSS Statistics for
Windows, IBM Corp., Armonk, NY, USA) using characteristics that were statistically
more likely to be present in patients who experienced diagnostic delay versus those
who did not.
Diagnostic delay was assessed in two ways. First, overall time to diagnosis ([Fig. 1A ]) was measured in months from the time the patient began experiencing symptoms until
a pathologic diagnosis was made. Overall time to diagnosis was used as a continuous
variable with higher values representing greater diagnostic delay. Second, management
factors impacting diagnosis ([Fig. 1B ]) were evaluated as a categorical variable, with any of the following representing
diagnostic delay: 1) multiple nondiagnostic fine needle aspirate (FNA) biopsies; 2)
two or more courses of antibiotic therapy; 3) surgery with incorrect preoperative
diagnosis; 4) evaluation by an otolaryngologist without further workup; or 5) surgery
without definitive postoperative diagnosis.
Fig. 1 Flowchart for the diagnosis of patients with OPSCC. The time between patients beginning to experience symptoms and the arrival at a pathologic
diagnosis was termed overall time to diagnosis. The time between presentation to a
physician and the arrival at a pathologic diagnosis highlights the management factors
impacting diagnosis.
Results
Patient and Tumor Characteristics
[Table 1 ] shows patient characteristics for p16+ and p16- cohorts. There was no significant
difference in the mean age at diagnosis or gender distribution between p16+ (58.6
years, old 22.5% female) and p16- OPSCC (58.2 years old, 16.4% female). A total of
93.3% of patients with p16+ tumors were white, compared with only 66.3% of the p16-
group (p < 0.001). In patients with p16+ tumors, there were significantly lower rates of tobacco
use (65.1 versus 92.1% of p16-; p < 0.001) and alcohol use defined as > 5 drinks weekly (30.3 versus 61.8% of p16-;
p < 0.001).
Table 1
Patient characteristics
All
p-value
p16 Status
p-value
p16-
p16
Patients
284
(100.0%)
89
(31.3%)
195
(68.7%)
Age
Mean Age (years old)
58.3
n
= 284
0.21a
58.6
58.2
0.69a
Mean Age with Dx Delay (years old)
57.0
n
= 112
57.6
56.8
0.70a
Gender
Females
52
(18.3%)
−
20
(22.5%)
32
(16.4%)
0.22b
Males
232
(81.7%)
69
(77.5%)
163
(83.6%)
Tobacco
Tobacco user
209
(73.6%)
−
82
(92.1%)
127
(65.1%)
< 0.001b
Tobacco nonuser
75
(26.4%)
7
(7.9%)
68
(34.9%)
Alcohol
Use
114
(40.1%)
−
55
(61.8%)
59
(30.3%)
< 0.001b
Nonuse
170
(59.9%)
34
(38.2%)
136
(69.7%)
(a) Student t -test.
(b) Chi-squared test.
As shown in [Table 2 ], 99.5% of p16+ tumors originated in the base of the tongue (33.3%) and tonsils (66.2%);
13 of 14 OPSCC cases originating outside of these locations were p16-. p16+ tumors
were more likely to present with smaller (T1-T2) primaries (72.3 versus 59.6% of p16-;
p = 0.03) and with more advanced nodal stage (90.8% N+ versus 74.2% of p16-).
Table 2
Tumor characteristics
All
p16 Status
p-value
p16-
p16
284
(100.0%)
89
(31.3%)
195
(68.7%)
Primary Site
Base of tongue
88
(31.0%)
23
(25.8%)
65
(33.3%)
< 0.001a
Tonsil
182
(64.1%)
53
(59.6%)
129
(66.2%)
Otherb
14
(4.9%)
13
(14.6%)
1
(0.5%)
Primary Stage
T1/2
194
(68.3%)
53
(59.6%)
141
(72.3%)
0.03a
T3/4
90
(31.7%)
36
(40.4%)
54
(27.7%)
Nodal Stage
N0
41
(14.4%)
23
(25.8%)
18
(9.2%)
< 0.001a
N1/2a
62
(21.8%)
12
(13.5%)
50
(25.6%)
N2b/2c/3
181
(63.7%)
54
(60.7%)
127
(65.1%)
Stage Group
I/II
23
(8.1%)
13
(14.6%)
10
(5.1%)
0.007a
III/IV
261
(91.9%)
76
(85.4%)
185
(94.9%)
(a) Chi-squared test.
(b) Soft palate, lateral oropharyngeal wall, posterior oropharyngeal wall.
Presenting Symptoms
[Table 3 ] summarizes the presenting symptoms for patients with p16- and p16+ OPSCC. Compared
with patients with p16- tumors, patients with p16+ OPSCC presented more often with
neck mass (85.1 versus 57.3%; p < 0.001) and less often with throat pain/odynophagia (24.6 versus 51.7%; p < 0.001) or weight loss (19.0 versus 41.6%; p < 0.001). Although they did not achieve statistical significance, otalgia (33.3 versus
44.9%; p = 0.06) and shortness of breath (4.1 versus 10.1%; p = 0.06) were also noted less frequently in patients with p16+ tumors.
Table 3
Presenting symptoms and diagnostic process
p16-
p16
p-value
n = 89
n = 195
Neck Mass
50
(56.2%)
166
(85.1%)
< 0.001a
Throat Pain
45
(50.6%)
48
(24.6%)
< 0.001a
Dysphagia
33
(37.1%)
64
(32.8%)
0.48a
Otalgia
40
(44.9%)
65
(33.3%)
0.06a
Weight loss
37
(41.6%)
37
(19.0%)
< 0.001a
Dyspnea
9
(10.1%)
8
(4.1%)
0.06b
Fatigue
8
(9.0%)
12
(6.2%)
0.45b
Voice Change
15
(16.9%)
28
(14.4%)
0.59a
Dysguesia
1
(1.1%)
4
(2.1%)
1.0b
Overall Time to Diagnosis
Mean symptom onset until pathology diagnosis in months (SD)
4.75
(5.03)c
4.65
(5.96)d
0.90e
Management Factors Impacting Diagnosis
None
64
(71.9%)
108
(55.4%)
0.008a
Any
25
(28.1%)
87
(44.6%)
Occult Primary
6
(6.7%)
37
(19.0%)
0.007b
Cystic Lymph Node(s)
1
(1.5%)f
17
(9.6%)g
0.02b
Necrotic Lymph Node(s)
40
(60.6%)f
22
(12.4%)g
0.05a
Nondiagnostic FNA (any)
3
(3.4%)
23
(11.8%)
0.02b
Nondiagnostic FNA (number)
0
86
(96.6%)
172
(88.2%)
0.74a
1
2
(2.2%)
18
(9.2%)
≥2
1
(1.1%)
5
(2.6%)
Abbreviations: FNA, fine needle aspirate; SD, standard deviation.
(a) Chi-squared test.
(b) Fisher exact test.
(c) Reported for 71 patients.
(d) Reported for 132 patients.
(e) Student t -test.
(f) Of 66 N+ patients.
(g) Of 177 N+ patients.
Diagnostic Findings
A summary of the diagnostic process is presented in [Table 3 ]. Patients with p16+ OPSCC experienced diagnostic delay more frequently due to management
factors prior to presentation to the Head and Neck Oncologic Surgery service (44.6
versus 28.1% of p16-; p = 0.008) ([Fig. 1B ]). p16+ primary tumors were more likely to be undetectable by physical examination
of the head and neck including flexible laryngoscopy (19.0 versus 6.7% of p16-; p = 0.007). Patients with p16+ OPSCC more often had cystic lymph node metastasis (8.8
versus 1.1% of p16-; p = 0.02) while patients with p16- tumors had a higher rate of necrotic nodes (44.9
versus 33.3% of p16 + ; p = 0.05). Patients with p16+ OPSCC underwent more frequently nondiagnostic FNA biopsy
of a cervical nodal mass (11.8 versus 3.4% of p16-; p = 0.03). Five patients in the p16+ group underwent ≥ 2 nondiagnostic FNA biopsies.
[Table 4 ] compares patients who did versus who did not experience diagnostic delay due to
management factors. There was no significant difference in age, gender, or race. Patients
who experienced diagnostic delay were 1) more likely to have p16+ tumors (77.7% delayed
versus 62.8% not delayed; p = 0.006); 2) less likely to consume alcohol (33.0% delayed versus 44.8% not delayed;
p = 0.047); 3) more likely to exhibit weight loss (33.0% delayed versus 21.5% not delayed;
p < 0.001); 4) more likely to have an occult primary tumor (25.0% of delayed versus
8.7% of not delayed; p < 0.001); and 5) more likely to undergo a nondiagnostic FNA (17.0% delayed versus
4.1% not delayed; p = 0.001). Using the significant associations shown in [Table 3 ], a multivariate analysis was performed to detect independently associated variables
and the results are shown in [Table 5 ]. Alcohol consumption was not interpedently associated with diagnostic delay. Weight
loss, occult primary tumor, any nondiagnostic FNA, and p16 positivity were independently
associated with delay in diagnosis ([Table 5 ]).
Table 4
Analysis of factors potentially associated with diagnostic delay
No diagnostic delay
Diagnostic Delay
p-value
n = 173
n = 112
Neck Mass
130
(75.6%)
86
(76.8%)
0.817
Throat Pain
57
(33.1%)
36
(32.1%)
0.86
Dysphagia
55
(32.0%)
42
(37.5%)
0.34a
Otalgia
64
(37.2%)
41
(36.6%)
0.92a
Weight loss
37
(21.5%)
37
(33.0%)
< 0.001a
Dyspnea
11
(6.4%)
6
(5.4%)
0.72b
Fatigue
11
(6.4%)
9
(8.0%)
0.61b
Voice Change
26
(15.1%)
28
(15.2%)
0.99a
Dysguesia
2
(1.2%)
3
(2.7%)
0.38b
P16 status
P16-
64
(37.2%)
25
(22.3%)
0.006a
P16+
108
(62.8%)
87
(77.7%)
Occult Primary
15
(8.7%)
28
(25.0%)
< 0.001b
Cystic Lymph Node(s)
7
(4.1%)
17
(9.8%)
0.074b
Necrotic Lymph Node(s)
67
(39.0%)
38
(33.9%)
0.39a
Nondiagnostic FNA (any)
7
(4.1%)
19
(17.0%)
0.001b
Tobacco Use
132
(77.2%)
77
(68.8%)
0.12
Alcohol Use
77
(44.8%)
37
(33.0%)
0.047
Abbreivation: FNA, fine needle aspirate.
(a) Chi-squared test.
(b) Fisher exact test.
(c) Student t -test.
Table 5
Multivariate analysis of factors associated with diagnostic delaya
OR
95%CI
p-value
p16+
2.08
1.13–3.94
0.022
Weight Loss
2.80
1.55–5.16
0.001
Occult Primary Tumor
2.88
1.40–6.07
0.005
Nondiagnostic FNA
3.60
1.41–10.08
0.010
Alcohol consumption
0.82
0.47–1.44
0.495
Abbreviations: CI, confidence interval; FNA, fine needle aspirate; OR, odds ratio.
Discussion
Our findings corroborate previous studies demonstrating that HPV+ OPSCC occurs more
often in patients who are white, nonsmokers, and nonalcohol users. In addition, p16+
OPSCC presented more often with a neck mass and less often with odynophagia and weight
loss compared with p16- OPSCC. We found no significant difference in rates of dysphagia.
Our results also concur with previous studies demonstrating the increased incidence
of cystic lymph node metastasis in HPV+ OPSCC. Weight loss, occult primary tumor,
nondiagnostic FNA, and p16 positivity were independently associated with diagnostic
delay.
Our study examines a large cohort of OPSCC patients and helps characterizing the differences
in presenting symptoms and evaluation in HPV+ OPSCC. In addition, we identified several
clinical events that delay diagnosis in the HPV+ OPSCC population, including nondiagnostic
FNA, repeated trials of empiric therapy, and, occasionally, open surgical procedures
prior to defining the diagnosis. Studies of delays in diagnosis encounter many limitations.
Calculating time from presenting symptoms to initial consultation is difficult and
subject to recall bias. We elected to use both temporal and clinical markers to evaluate
diagnostic delays in our patients. While deviating from a purely time-based measure
may be less objective, this allows a focus on specific clinical events that contribute
to delays in time to diagnosis. Despite this approach, the study is retrospective
and limited by the quality and completeness of prior documentation, and is potentially
impacted by multiple uncontrolled variables, including clinician charting, introduction
of new treating and/or referring clinicians during the study period etc. Regardless
of the limitations, the study demonstrates that patients with HPV+ OPSCC are at risk
for delay in diagnosis and treatment. It also suggests that targeted education of
appropriate care providers around the unique presentation of HPV+ OPSCC as compared
with the “classic” HPV- OPSCC could improve time to diagnosis and treatment.
Dhooge et al. reviewed 127 cases of HNSCC from all subsites but with roughly 50% laryngeal
SCCa and found that 15.7% of the patients had a time interval from symptom to diagnosis > 12
months. Subsite location can affect time to diagnosis; several studies have found
that laryngeal SCCa is associated with a longer delay in diagnosis compared with oropharyngeal
cancer.[9 ]
[10 ]
Crucial in the interpretation of these results in the context of other studies is
the distinction between statistical and clinical significance. Kowalski et al. studied
the time interval needed for clinical upstaging in 69 cases who experienced delays
in treatment compared with 138 controls who were treated shortly after diagnosis in
Brazil. The delay in treatment of the cases was due to initial treatment refusal (53.7%)
or a long waiting period for hospital admission (40%). The shortest median time for
clinical upstaging was for resectable stage IV to unresectable stage IV, which took
an average of 3.6 months. The median time for clinical upstaging from stage I to stage
II was 6.3 months; and from III to IV, it was 8.5 months.[11 ] This data provides a gross measure for the degree of delay in diagnosis which is
clearly clinically significant.
Goy et al performed a systematic review of studies evaluating diagnostic delay in
head and neck cancer. Given the heterogeneity of cancer site and measurement of delay
among the 27 studies reviewed, the group did not aggregate the data as a meta-analysis.
There were five studies evaluating pharyngeal cancer. Three studies demonstrated no
relationship between diagnostic delay and more advanced stage at diagnosis, two studies
found a positive relationship between delay in diagnosis and more advanced stage at
diagnosis. Although very few studies (when including all cancer sites) demonstrated
that delay in diagnosis was associated with more advanced stage at diagnosis, delay
in diagnosis did predict survival in many of the studies examined. This suggests that
the TMN staging system is perhaps not sensitive enough to detect the actual clinical
effects of delays in diagnosis.[3 ]
Seoane et al. performed a systematic review and meta-analysis of the impact of diagnostic
delay on survival in HNSCC, which included 10 studies. The pooled data showed equivocal
results, with a trend between patient/professional delay and worsened mortality which
did not reach statistical significance. For all subsites, total delay was not associated
with worsened mortality (pooled relative risk [RR]: 1.04; confidence interval [CI]:
1.01–1.07), but pharyngeal site had the highest association between diagnostic delay
and mortality (pooled RR 1.68; CI: 1.22–2.31).[12 ] Despite these equivocal results, it is commonly agreed in the field that more rapid
diagnosis of malignancy will reduce treatment burden and morbidity and improve survival.
More recently, Schutte et al. performed a systematic review on the impact of time
to diagnosis and treatment in head and neck cancer, which included 51 studies. They
found delay in HNC diagnosis and treatment is associated with higher stage and worse
survival. They found that an increased delay was associated with decreased overall
survival in 8 of 14 studies investigating the effect of time delays on survival. Similar
outcomes were found for disease-specific survival.[13 ]
Huang et al. evaluated factors that contribute to delay in diagnosis in HPV- and HPV+
OPSCC. In alignment with our findings, their review of 304 patients with OPSCC showed
HPV+ OPSCC presented more often with asymptomatic neck mass and underwent more nondiagnostic
FNAs. Nearly 10% of patients with HPV+ OPSCC in their cohort required > 12 months
from first presentation to diagnosis. They attributed the delay in ⅓ of this subset
to be clinician-related and ⅔ to be patient related.[14 ] Similarly, Davis et al. reported 6 patients with delayed diagnosis of HPV+ OPSCC,
where nondiagnostic or negative FNAs (n = 2) and negative direct laryngoscopy and biopsy (n = 1) were taken as sufficient evidence of a benign cyst and further biopsies were
deferred. They reported a median time from development of a unilateral neck mass to
presentation for alternate opinion and suspected HPV-OPSCC as 42 months, ranging from
3 months to 7 years.[15 ]
HPV+ OPSCC is known to have distinct tumorigenesis from HPV- OPSCC.[16 ] HPV+ OPSCC is characterized by a younger patient population and association with
nonsmoking status,[17 ]
[18 ] a propensity to cause malignancy in oropharyngeal sites compared with other sites
of the head and neck,[19 ]
[20 ] smaller primary tumor size,[21 ] cystic lymph node metastasis,[22 ] improved response to radiotherapy,[23 ]
[24 ] improved survival compared with HPV- OPSCC,[17 ]
[25 ] and decreased risk of second primary malignancy.[26 ]
[27 ]
Delays in diagnosis in HPV-related OPSCC is likely multifactorial. First, the presenting
demographics, tumor characteristics, and symptoms of HPV+ OPSCC are distinct from
“classic” HPV- disease. HPV+ OPSCC presents more often with a neck mass and less often
with odynophagia, otalgia, dyspnea, or weight loss. Many patients with HPV+ tumors
are nonsmokers. When otherwise healthy patients present with a neck mass and without
throat pain, they are more likely to be treated for an infectious process, than immediately
biopsied or referred to a specialist. If the neck mass persists after antibiotic therapy,
imaging may show a cystic node, which can be mistaken for a branchial cleft cyst or
other developmental anomaly, despite the fact that these generally present earlier
in life. The overall incidence of malignancy in a cystic neck mass is only between
10 and 22%,[28 ]
[29 ] but increases to as high as 80% in patients > 40 years old.[30 ] Therefore, according to their 2017 clinical practice guidelines on the evaluation
of a neck mass in adults, the American Academy of Otolaryngology – Head and Neck Surgery
recommends work-up of a cystic neck mass until a definitive diagnosis is obtained.[31 ]
As shown by multivariable analysis, both the presence of nondiagnostic FNA and an
occult primary tumor were independently associated with diagnostic delay. p16+ OPSCC
was more often associated with one or more nondiagnostic FNA lymph nodes biopsies,
which could result in further delay in referral or give false reassurance. HPV+ OPSCC
primary tumors were also more frequently undetectable on physical exam and flexible
laryngoscopy at the time of evaluation by an Otolaryngologist, potentially resulting
in further delay despite specialist evaluation. Franco et al. performed a retrospective
review of 100 sequential patients with HNSCC and found the most common reason for
diagnostic delay was a prolonged wait to evaluation in being seen in the otolaryngology
clinic after referral placement (28.6%) followed by diagnostic error by the referring
physician (22%) and delayed referral of a symptomatic patient to the otolaryngology
clinic (16.2%); these findings further highlight the importance of raising awareness
and targeted education of appropriate care providers.[32 ]
Conclusions
Patients with HPV+ OPSCC experienced a significantly higher rate of diagnostic delay
that is likely associated with inappropriate treatment (such as antibiotics) and procedures
(repeat FNA, excisional biopsy) and delays in time to treatment. Additional research
and targeted education of appropriate care providers may improve time to diagnosis
in this population.
