Incidence of head and neck squamous cell carcinoma among subjects at high-risk for lung cancer: results from the Pittsburgh Lung Screening Study

Abstract

Background

Earlier detection and diagnosis of head and neck squamous cell carcinoma (HNSCC) should lead to improved outcomes. However, to date, no effective screening strategy has been identified. We evaluated whether it would be useful to screen subjects targeted for lung cancer screening also for HNSCC.

Methods

Medical records, death certificates, and cancer registry and questionnaire data were used to determine the number of observed incident HNSCC cases in the Pittsburgh Lung Screening Study (PLuSS), a cohort of ≥ 50 year old current and ex-smokers with ≥ 12.5 pack-year smoking history. Expected number of cases was estimated using stratum-specific incidence rates obtained from 2000-2011 SEER data. Standardized incidence ratio (SIR) was calculated to examine the difference between observed and expected number of cases.

Results

Twenty-three (0.64%) of the 3,587 at-risk participants in PLuSS developed HNSCC over in total 32,201 person-years of follow-up. This was significantly higher than expected based on incidence rates obtained from SEER (13.70 cases expected; SIR=1.68, 95% confidence interval: 1.06-2.52). Excess burden of HNSCC in PLuSS was 28.9 cases per 100,000 person-years. Observed incident cases were significantly more often male, started smoking at a younger age, smoked more per day and had more pack-years of smoking than the rest of the PLuSS at-risk participants.

Conclusion

Our results provide a rationale for offering head and neck cancer screening along with CT screening for lung cancer. Randomized controlled trials that assess the effectiveness of adding examination of the head and neck area to lung cancer screening programs are warranted.

Introduction

Head and neck cancer is the world’s sixth most common type of cancer. More than 600,000 new patients are diagnosed and about 350,000 individuals die from this disease worldwide each year.¹ In the U.S. alone, there are expected to be approximately 55,000 incident cases and 12,000 deaths due to head and neck cancer in 2014.² Most cancers of the head and neck are squamous cell carcinomas (HNSCC) and affect the oral cavity, pharynx or larynx. Tobacco use and alcohol consumption are the major risk factors for HNSCC. Approximately 75% of all HNSCCs are associated with these lifestyle factors, which have a greater than multiplicative effect on risk of HNSCC in combination.^3,4 Oral infection with human papillomavirus (HPV), especially HPV type 16, is an important independent risk factor for, in particular, the development of oropharyngeal squamous cell carcinomas.⁵

Early detection of head and neck cancer is associated with reduced morbidity and improved survival. For patients diagnosed with localized disease, estimated 5-year survival rates are 83% for oral cavity and pharyngeal cancer and 75% for laryngeal cancer.⁶ Unfortunately, the majority of new head and neck cancer cases are currently diagnosed with regional or distant disease when 5-year survival rates are, respectively, 61% and 37% for oral cavity and pharyngeal cancer and 43% and 35% for laryngeal cancer.⁶ Moreover, these rates have remained relatively constant the past few decades.⁶ Thus, there is a strong need to develop strategies that will result in the identification of head and neck cancer at an earlier stage when it can be curatively treated.

Regular screening for head and neck cancer may result in earlier detection and diagnosis and, thus, in lower disease-related morbidity and mortality. The head and neck area is generally easily accessible to visual and tactile inspection and the primary screening test is a systematic clinical examination of the area. However, given that only a relatively small number of individuals will develop head and neck cancer, screening for this cancer in the general population has not proven effective.^7-10 Stratification of head and neck cancer risk based on known risk factors and targeting only individuals at high-risk for screening may be more feasible and successful.^9,10

Recent results from the U.S. National Lung Screening Trial (NLST) showed that screening of subjects with low-dose computed tomography (CT) reduces lung cancer mortality.¹¹ Driven primarily by these results, the U.S. Preventive Services Task Force (USPSTF) as well as the American Cancer Society and several other organizations are now recommending annual screening for lung cancer with low-dose CT in individuals aged 55 to 74 years who have a smoking history of at least 30 pack-years and currently smoke or have quit within the past 15 years.^12,13 Given that cigarette smoking is also one of the primary risk factors for head and neck cancer, we hypothesized that it may be efficacious to screen for head and neck cancer in individuals at risk for smoking-related lung cancer as well. In this study, as a first step to determine if this would be useful, we used data from the Pittsburgh Lung Screening Study (PLuSS)^14,15 and the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute to evaluate whether incident HNSCC cases are more common among those targeted for lung cancer screening than in the general U.S. population.

Methods

Study Population

PLuSS is a community-based research cohort of current and ex-smokers who are being followed for lung cancer that has been described in detail previously.^14,15 Briefly, between January 2002 and April 2005, PLuSS used mass mailings and advertisements to enroll 3,754 volunteers, mainly from southwestern Pennsylvania, with the following characteristics: 1) age 50-79 years old; 2) no personal history of lung cancer; 3) current or ex-smoker of at least 10 cigarettes per day for at least 25 years and, if ex-smoker, stopped smoking not more than 10 years before enrollment; 4) no chest computed tomography within 12 months; and, 5) body weight less than 400 pounds. Study eligibility was assessed by telephone interview; of 5,033 eligible persons, 1,279 (25.4%) ultimately did not agree to an enrollment visit. Of the 3,754 volunteers enrolled, 92.9% were white and more than 95% had a cigarette smoking history at enrollment of at least 20 pack-years. Follow-up of PLuSS participants is ongoing. All participants provided written informed consent and the University of Pittsburgh Institutional Review Board approved the study.

Data Collection

Participants completed a self-administered, risk factor questionnaire at enrollment. This questionnaire covered demographic data, cigarette smoking history, medical history (including personal history of cancer), and current medical problems. In addition, each year, participants are asked to complete an Annual Study Update (ASU) questionnaire. This questionnaire asks participants among other things if, during the previous 12 months: 1) they have received a diagnosis of a cancer from a health care provider; 2) what type or types of cancer had been diagnosed; and, 3) to record the cancers and dates of diagnosis on the questionnaire. Information on head and neck cancer occurrence among PLuSS participants was collected from these questionnaires and by reviewing death certificates and searching the University of Pittsburgh Medical Center Network Cancer Registry (http://cancerregistrynetwork.upmc.com/). Subsequently, medical records, pathology reports and death certificates were used to confirm a diagnosis of HNSCC (subsites included: oral cavity, pharynx and larynx) during time enrolled in PLuSS. Head and neck cancers that were not confirmed HNSCC (N=7; six were other types of cancer located in the head and neck area and for one we were unable to obtain records that could confirm that it was squamous cell carcinoma) were not counted as an observed incident HNSCC case. For the purposes of this study, the cutoff date for follow-up data collection was March 19, 2014.

Statistical Analysis

Differences in characteristics between observed incident HNSCC cases and the rest of the at-risk PLuSS population were assessed using t-tests for continuous variables and Fisher’s exact tests for categorical variables. Person-years of follow-up was defined as the time elapsed between date of enrollment and date of last completed ASU (for those without HNSCC) or date of diagnosis (for those diagnosed with incident HNSCC). To evaluate whether the observed number of incident HNSCC cases in the PLuSS cohort differs from the expected number in the general US population, we estimated the standardized incidence ratio (SIR; the ratio of observed to expected incident HNSCC cases). The corresponding 95% confidence interval (CI) and P value were determined using Byar’s approximation.¹⁶ The number of HNSCC cases expected was calculated using stratum-specific incidence rates obtained from the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute (www.SEER.cancer.gov). Specifically, we used the following SEER*Stat Database to estimate sex-race-age-specific HNSCC incidence rates: Incidence - SEER 18 Regs Research Data + Hurricane Katrina Impacted Louisiana Cases, Nov 2013 Sub (2000-2011) <Katrina/Rita Population Adjustment>. This database covers the largest geographic coverage available, approximately 27.8% of the US population (based on 2010 census).¹⁷ As in PLuSS, both sexes and all races (categories used: white, non-white) were included while age was limited to 50 years or older (categories used: 50-54, 55-59, 60-64, 65-69, 70-74, 75-79, 80-84, and 85≥). Anatomical sites were limited to oral cavity and pharynx (includes: lip, tongue, salivary gland, floor of mouth, gum and other mouth, nasopharynx, tonsil, oropharynx, hypopharynx, and other oral cavity and pharynx), and larynx; histology was restricted to all malignant squamous cell carcinoma types (includes: papillary squamous cell carcinoma, papillary squamous cell carcinoma inverted, squamous cell carcinoma NOS, squamous cell carcinoma keratinizing NOS, squamous cell carcinoma large cell nonkeratinizing, squamous cell carcinoma small cell nonkeratinizing, squamous cell carcinoma spindle cell, squamous cell carcinoma adenoid, squamous cell carcinoma micro-invasive, squamous cell carcinoma & Grade III, and squamous cell carcinoma with horn formation, and squamous cell carcinoma clear cell type). We subsequently calculated the expected number of incident HNSCC cases by multiplying each SEER sex-race-age-specific incidence rate by the total number of person-years of follow-up accumulated in PLuSS in that particular sex-race-age category (32 categories in total: [2 (sex) × 2 (race) × 8 (age)]) and then adding up the results. P values <0.05 were considered statistically significant. Analyses were conducted with use of the SAS® statistical software package (SAS version 9.3, SAS Institute Inc., Cary, NC).

Results

Seventeen (0.45%) of the in total 3,754 PLuSS participants indicated a past history of head and neck cancer on their risk factor questionnaire completed at enrollment and were excluded from the current study. Of the remaining 3,737 participants at risk for incident HNSCC, 150 (4.0%) were lost to follow up and never completed an ASU questionnaire leaving a total of 3,587 participants at risk. Twenty-three (0.64%) of the at-risk participants developed HNSCC over in total 32,201 person-years of follow-up (median follow-up of 9.5 years per participant; interquartile range: 8.6-10.2 years) for a HNSCC incidence rate calculated at 71.4 cases per 100,000 person-years.

Using the 2000-2011 sex-race-age-specific HNSCC incidence rates obtained from SEER, the estimated expected number of incident HNSCC cases was 13.70. The observed number of incident HNSCC cases was significantly higher than the expected number, SIR=23/13.70=1.68 (95% CI: 1.06-2.52; P=0.027). Excess burden of HNSCC in the at-risk PLuSS population is 28.9 cases per 100,000 person-years.

Characteristics of the at-risk population by status are presented in Table 1. Mean age at enrollment and smoking status did not differ significantly between the incident HNSCC cases and the other at-risk PLuSS participants. Compared to the at-risk PLuSS participants not diagnosed with incident HNSCC, the 23 incident HNSCC cases were significantly more often male, started smoking at a younger age, smoked more cigarettes per day and had more pack-years of smoking. Regarding pack-years, the interquartile range was 42.5-90.8 among the incident cases and 33.0-61.3 among the other at-risk participants (not in Table). The majority of cases were diagnosed with cancer of the oral cavity or pharynx (56.5%). Median time in PLuSS before being diagnosed with HNSCC was 5.1 years (interquartile range 1.3-7.0 years; not in Table).

Table 1.

Characteristics of the PLuSS at-risk population, by status

	Incident HNSCC cases	Other at-risk PLuSS participants	Pa
	Ntotal=23	Ntotal=3564
Age at enrollment (in years, mean ± sd)	58.0 ± 5.6	59.2 ± 6.8	0.38
Sex, N (%)
male	19 (82.6)	1814 (50.9)
female	4 (17.4)	1750 (49.1)	0.003
Race, N (%)
white	23 (100)	3349 (94.0)
non-white	0 (0)	215 (6.0)	0.40
Smoking statusb, N (%)
current	16 (69.6)	2139 (60.0)
former	7 (30.4)	1425 (40.0)	0.40
Age start smoking (in years, mean ± sd)	15.7 ± 3.3	17.5 ± 3.8	0.03
Pack-years of smokingb (mean ± sd)	71.1 ± 34.4	51.5 ± 25.7	0.01
Pack-years of smokingb, N (%)
<30	0 (0)	664 (18.6)
30-44	6 (26.1)	935 (26.2)
45-49	5 (21.7)	967 (27.1)
≥60	12 (52.2)	998 (28.0)	0.02
Cigarettes per day, N (%)
1-19	3 (13.0)	1138 (31.9)
20-29	8 (34.8)	1521 (42.7)
30-39	6 (26.1)	576 (16.2)
≥40	6 (25.1)	329 (9.2)	0.01
Age at diagnosis [in years, median (range)]	63.0 (52.0-75.2)	-
HNSCC subsite, N (%)
oral cavity	4 (17.4)	-
oropharynx	6 (26.1)	-
hypopharynx	3 (13.0)	-
larynx	10 (43.5)	-

^aFisher’s exact test for categorical variables; t-test for continuous variables.

^bAt enrollment.

Discussion

Regular screening of asymptomatic subjects for head and neck cancer may result in earlier detection and lower morbidity and mortality. However, because head and neck cancer is relatively rare, screening of the general population is impractical. In line with this, free annual screening programs have generally not been successful.^7,8 In this study, we evaluated if it would be effective to screen subjects targeted for lung cancer screening also for head and neck cancer by examining whether incident HNSCC cases are more common among those at risk for smoking-related lung cancer than in the general U.S population. In total, we identified and confirmed 23 incident HNSCC cases in the PLuSS at-risk population, a cohort of current and ex-smokers with ≥12.5 pack years of smoking history at enrollment, during 32,201 person-years of follow-up. This was statistically significantly more than expected based on sex-race-age specific incidence rates in the general U.S. population (N_expected=13.70), and the excess burden of HNSCC in PLuSS was 28.9 cases per 100,000 person-years.

Consistent with the established risk factors for HNSCC,^3,4 incident HNSCC cases in PLuSS were significantly more often male, started smoking at a younger age, smoked more cigarettes per day and had also smoked more pack-years than the other PLuSS participants. Information on alcohol consumption and HPV infection was not available for PLuSS participants and we were therefore unable to assess differences in these risk factors. Based on pack-years smoked and smoking status only, all observed cases would have been targeted for lung cancer screening according to the current recommendations.^12,13 However, three incident cases were diagnosed with HNSCC when they were younger than 55 years of age.

Overall, our results suggest that it may indeed be useful to examine individuals who come in for lung cancer screening also for head and neck cancer. Although tobacco use plays a role in the majority of HNSCC,^3,4 it should be noted that individuals who are non-smokers but at high-risk for HNSCC due to heavy alcohol consumption or persistent HPV infection will be missed when head and neck cancer screening is offered to those at high-risk for lung cancer only. About 4% of HNSCCs is due to alcohol alone and approximately 25% is attributed to HPV infections.^4,18

A strength of the current study is that only subjects with confirmed incident HNSCC were counted as an observed case. However, because we relied on self-report and searches of death certificates and the cancer registry database for the initial identification of participants with HNSCC, we may have missed some cases. Thus, the incidence rate and excess burden of HNSCC in PLuSS may actually be higher than calculated. Data from SEER was used to estimate the expected number of cases. Although the population covered by SEER is comparable to the general U.S. population regarding measures of education and poverty, the SEER population does tend to contain a higher proportion of foreign-born individuals.¹⁹ In addition, PLuSS participants are primarily from southwestern Pennsylvania and volunteered to participate in a study.^14,15 They may not be representative of the at-risk cigarette smoking population in general.

Based on the accessibility of the head and neck area for physical examination, the possibility to identify premalignant lesions and early-stage tumors, and the easily identifiable risk factors, HNSCC seems particularly suited for screening programs that target high-risk subjects. However, in its most recent recommendation statement on oral cancer the USPSTF concluded that the evidence for screening is still insufficient.²⁰ Harms related to overdiagnosis and overtreatment, including unnecessary patient anxiety, co-morbidities and differing rates of detection at different head and neck cancer subsites may all present additional barriers to screening. In order to move forward, it will therefore be important to determine the accuracy of current screening methods and whether head and neck cancer screening actually reduces mortality and/or morbidity among those at high-risk and is cost-effective.

To conclude, our results suggest that it may be beneficial to screen subjects at high-risk for smoking-related lung cancer also for HNSCC. Randomized controlled trials that assess the benefits and harms of adding examination of the head and neck area to lung cancer screening programs and aid determination of the most appropriate risk cohort and screening methods are warranted.

Abbreviations used

ASU

annual study update

CI

confidence interval

CT

computed tomography

HNSCC

head and neck squamous cell carcinoma

HPV

human papilloma virus

NLST

National Lung Screening Trial

PLuSS

Pittsburgh Lung Screening Study

sd

standard deviation

SEER

Surveillance, Epidemiology, and End Results

SIR

standardized incidence ratio

USPSTF

U.S. Preventive Services Task Force