Risk of second primary cancer among patients with head and neck cancers: A pooled analysis of 13 cancer registries

Authors


Abstract

The objective of the study was to assess the risk of second primary cancers (SPCs) following a primary head and neck cancer (oral cavity, pharynx and larynx) and the risk of head and neck cancer as a SPC. The present investigation is a multicenter study from 13 population-based cancer registries. The study population involved 99,257 patients with a first primary head and neck cancer and contributed 489,855 person-years of follow-up. To assess the excess risk of SPCs following head and neck cancers, we calculated standardized incidence ratios (SIRs) by dividing the observed numbers of SPCs by the expected number of cancers calculated from accumulated person-years and the age-, sex- and calendar period-specific first primary cancer incidence rates in each of the cancer registries. During the observation period, there were 10,826 cases of SPCs after head and neck cancer. For all cancer sites combined, the SIR of SPCs was 1.86 (95% CI = 1.83–1.90) and the 20-year cumulative risk was 36%. Lung cancer contributed to the highest proportion of the SPCs with a 20-year cumulative risk of 13%. Excess second head and neck cancer risk was observed 10 years after diagnosis with lymphohaematopoietic cancers. The most common SPC following a first primary head and neck cancer was lung cancer. However, the highest excess of SPCs was in the head and neck region. These patterns were consistent with the notion that the pattern of cancer in survivors of head and neck cancer is dominated by the effect of tobacco smoking and alcohol drinking. © 2008 Wiley-Liss, Inc.

A significant improvement in locoregional control of head and neck squamous cell carcinoma has been seen over the last decades, due to the introduction of new surgical techniques, improved radiotherapy techniques and the use of chemotherapy. However, 5-year relative survival from head and neck cancer has not increased greatly over the period of 1985–1994.1, 2 One of the reasons for the lack of improvement in overall survival is the frequent development of second primary cancers (SPCs) that occur more often in head and neck cancer patients than in patients with cancers of other sites3 and lead to poor prognosis.4–6

Head and neck cancers (oral cavity, pharynx and larynx) comprise around 5% of all cancer cases worldwide and result in 6% of cancer deaths in men and 3% in women.7 SPCs after head and neck cancers are mostly found in the aerodigestive tract, including the lung and esophagus, which leads to a significant decrease in survival.5 In contrast to the overall 5-year survival rate for head and neck cancer patients of approximately 50%, the 5-year survival rate in head and neck cancer patients who developed SPC was around 20% after the SPC was diagnosed.4, 5 The 5-year survival rates after SPC diagnosis were above 30% if the SPC was also a head and neck cancer and decreased to 8% if the SPC was outside the head and neck area.4

Multiple cancers within a person provide a unique clue to the understanding of cancer etiology and mechanisms.8, 9 Multiple cancers could be due to common risk factors or other predispositions or treatment for the first cancer. If the risk increases over time with a special pattern, the association may be more likely due to the treatment for the first cancer. For example, with alkylating agent treatment, leukemia risk begins to increase and peaks at 5 to 10 years, whereas therapy-associated solid cancer risk increases typically after 10 years.10 Otherwise, if both associations treating the index cancer as the primary cancer and as the SPC with a specific cancer are observed, then common risk factors may be suggested for the association.

The present investigation is a multicenter study including cancer data from 13 population-based cancer registries in Europe, Canada, Australia and Singapore. The data were used to assess the incidence of SPCs among 99,257 patients who were diagnosed with a first head and neck cancer between 1943 and 2000. The long follow-up and large sample size enables us to look at the risks for rare SPCs. In addition, we assessed the risk of second head and neck cancers after a first cancer of various types to determine whether associations are likely to be caused by common risk factors or treatment.

Material and methods

An international multicenter study was initiated to incorporate cancer registries that have been in operation for at least 25 years to conduct a systematic analysis of SPCs.11, 12 The registries include New South Wales in Australia, British Columbia, Manitoba and Saskatchewan in Canada, Denmark, Finland, Iceland, Norway, Scotland, Singapore, Slovenia, Sweden and Zaragoza in Spain. These registries had cancer data covering different time periods within the period of 1943–2000. A high degree of completeness of ascertainment by the registry is suggested by an appropriate level of microscopically verified cases, a low proportion of cases registered from death certificates only, and consistent inclusion in the cancer incidence in 5 continents monograph series.13

Data were provided from each cancer registry on all first primary cancers, including age and sex of the subject, diagnosis and date of the first primary, follow-up for mortality and diagnosis and date of the second primary, if any. In this study, registries used different cancer codes, which were systematically converted into International Classification of Disease, Ninth Revision (ICD-9). Here, we have analyzed the occurrence of SPCs in patients who survived from head and neck cancers (ICD-9 codes: tongue: 141; mouth: 143–145; oropharynx: 146; hypopharynx: 148; pharynx unspecified: 149 and larynx: 161). Skin of lip was coded as 173, and thus not included in the present study. Seventy-nine percent of invasive lip cancer arose on the vermilion border (ICD-9: 140.0 and 140.1) based on SEER data 1992–1999.14 Because of the large proportion of external lip cancer and its different etiology, we excluded lip cancer (ICD-9: 140) from the analysis. Cancers of salivary glands (ICD-9: 142) were not included in the analysis as they have a different etiology from head and neck cancer. Nasopharyngeal cancer (ICD-9: 147) was the subject of a separate study15 thus we excluded them from the primary cancers. Coding of multiple primaries in the cancer registries has followed a common set of rules proposed by the International Association of Cancer Registries and the International Agency for Research on Cancer.16 According to the rules, a primary cancer is one that originates in a primary site or tissue and is thus neither an extension, nor a recurrence, nor a metastasis. SPCs from the same site as the first primaries (same ICD-9) and nonmelanoma skin cancer (ICD-9: 173) were excluded from the analysis because of the inconsistency in reporting across the registries. Other tobacco-related cancers but lung and esophageal cancer included sites of the nasopharynx (ICD-9: 147), nose and nasal cavity (160), stomach (151), liver (155), pancreas (157), cervix (180), bladder (188), kidney (189) and leukemia (204–208).17 Other alcohol related cancers included colorectal (153, 154), liver (155) and female breast (174) cancers.18

To assess the potential excess occurrence of SPCs after having a first primary head and neck cancer, the number of SPCs observed was compared to the expected number of cancers to estimate the standardized incidence ratios (SIRs). SIRs adjusted for age, sex, year and registry were calculated using indirect standardization methods.11, 12 The expected number was calculated from accumulated person-years and the age-, sex- and calendar period-specific (depending on the registries) first primary cancer incidence rates in each of the cancer registries. All cases of first primary head and neck cancers were followed up for SPC from the date of head and neck cancer diagnosis (1943–2000) to the date of SPC (1943–2000), date of death or end of follow-up (1992–2000) depending on the coverage of each registry when the study closed. The same methodology was adopted for the reverse analysis of head and neck cancer as a SPC following a first primary tumor in the other sites. The cumulative risk of developing a SPC after the first head and neck cancer within a specific period was estimated by the Kaplan-Meier method. The differences (heterogeneity) across groups of age at diagnosis or follow-up time were estimated by Poisson Regression.

Results

The study population of 99,257 patients with a first primary head and neck cancer contributed 489,855 person-years of follow-up (mean follow-up time: 4.9 years). The population was composed of 15,985 (16.1%) tongue, 22,378 (22.5%) mouth, 10,871 (11.0%) oropharyngeal, 8,288 (8.3%) hypopharyngeal, 1,599 (1.6%) unspecified pharyngeal and 40,190 (40.5%) laryngeal cancer patients (Table I). Of the first primary head and neck cancer patients, the majority were male and 56,350 (56.8%) were younger than 65 years of age at diagnosis. The majority of cases were from Europe (72.6%), whereas smaller proportions were from Canada (11.3%), Australia (13.1%) and Singapore (3.0%).

Table I. Characteristics of First Head and Neck Cancer Cases from 13 Cancer Registries
 Tongue (141)Mouth (143–145)PharynxLarynx (161)Total
N%N%N%N%N%
Total15,98522,37820,70440,19099,257
Sex
 Female5,83936.58,34537.35,27825.54,80712.024,26924.5
 Male10,14663.514,03362.715,42674.535,38388.074,98875.5
Age at diagnosis (year)
 <564,59828.85,65025.25,74227.79,42923.525,41925.6
 56–654,30426.96,15627.56,68332.313,78834.330,93131.2
 66–743,80223.85,41124.25,07424.511,07327.625,36025.5
 75+3,28120.55,16123.13,20515.55,90014.717,54717.7
Calendar period at diagnosis
 ∼19753,47321.74,51120.24,13620.09,89524.622,01522.2
 1975–19833,92524.65,63425.25,33225.810,78826.825,67925.9
 1984–19904,00125.05,76525.85,33525.89,89624.624,99725.2
 1991∼4,58628.76,46828.95,90128.59,61123.926,56626.8
Follow-up period
 <12 months5,57034.86,79330.49,08943.98,14920.329,60129.8
 1–4 year5,97137.48,51538.17,93338.314,73636.737,15537.4
 5–9 year2,24014.03,68316.52,0579.98,05420.016,03416.2
 10+ year2,20413.83,38715.11,6257.89,25123.016,46716.6
Registries
 Australia NSW (1972–1997)2,45415.43,00413.42,95714.34,59111.413,00613.1
 Canada British Columbia (1970–1998)1,3658.51,7938.01,6037.72,6526.67,4137.5
 Canada Manitoba (1970–1998)4202.65582.54622.28472.12,2872.3
 Canada Saskatchewan (1967–1998)2611.63421.52711.36111.51,4851.5
 Denmark (1943–1997)2,10513.24,02718.03,12015.17,62319.016,87517.0
 Finland (1953–1998)2,02812.72,0479.12,37611.55,83314.512,28412.4
 Iceland (1955–2000)800.51180.5470.21770.44220.4
 Norway (1953–1999)2,02812.7292613.12,20310.63,6719.110,82810.9
 Singapore Chinese (1968–1992)4923.15962.75962.91,3023.22,9863.0
 Slovenia (1961–1998)9285.81,1895.32,25310.92,3295.86,6996.7
 Spain Saragoza (1978–1998)3692.33171.43401.61,8484.62,8742.9
 Sweden (1961–1998)1,90211.92,85912.82,45711.94,08310.211,30111.4
 UK, Scotland (1975–1996)1,5539.72,60211.62,0199.84,62311.510,79710.9

During the follow-up period, 10,826 SPCs occurred in the patients (Table I). Singapore patients had a lower proportion of SPC development whereas patients in the other study populations had fairly similar proportions. The proportions of patients who developed a SPC following a first primary head and neck cancer were high after first primary laryngeal cancer (14%).

The 20-year cumulative risk of SPC after head and neck cancer was 36%, ranging from 33% in mouth and tongue cancers to 46% in hypopharyngeal cancer. Figures 1a–1c shows the cumulative risk of various second primaries after first primary tongue and mouth, pharyngeal and laryngeal cancers. The 5-, 10- and 20-year cumulative risks of second UADT (upper aerodigestive tract, including the oral cavity, pharynx, larynx and esophagus) and lung cancers were 5.4%, 10.5% and 16.5% for first primary tongue and mouth cancer patients, 7.3%, 14.2% and 21.5% for first primary pharyngeal cancer patients and 4.6%, 9.9% and 19.0% in first primary laryngeal cancer patients. We also predicted the cumulative risk for a 64-year-old patient who was diagnosed in 1983 (the median age and calendar year of diagnosis of the study population) to have comparable curves across sites. There were no major differences in the curves for the overall compared to the predicted curves for the 64-year old.

Figure 1.

Selected second primary cancers after mouth and tongue (a), pharyngeal (b) and laryngeal (c). There were no major differences in the curves for the observed versus predicted curves for a 64-year-old patient diagnosed at 1983.

Table II shows the SIRs of SPCs following head and neck cancers by age at diagnosis of the first primary. Second primary head and neck cancers were associated with first primary head and neck cancers, especially for those who had their first primaries diagnosed at younger than 56 years old (SIR = 14.9, 95% CI = 13.6–16.3). The SIR estimates declined at older ages (p for heterogeneity <0.0001 using Poisson regression). Similar associations were also observed for second primary esophageal and lung cancers.

Table II. Numbers of Cases (N) and Standardized Incidence Ratio (SIR) of Selected Second Cancers after Head and Neck Cancer by Age at Diagnosis of First Cancer
Site of second cancer (ICD-9)<5656–6566–7475+Total
NSIR95% CINSIR95% CINSIR95% CINSIR95% CINSIR95% CI
  • Other tobacco-related cancers include the following: nasopharynx (ICD-9: 147), nose and nasal cavity (160), stomach (151), liver (155), pancreas (157), cervix (180), bladder (188), kidney (189) and leukemia (204–208). Other alcohol-related cancers include the following: colorectal (153, 154), liver (155) and female breast (174).

  • *

    p < 0.0001 compared to the youngest group using Poisson Regression.

All malignant (140–208)2,5352.45(2.35–2.55)4,0171.99(1.93–2.05)*2,9631.59(1.53–1.65)*1,3111.47(1.39–1.55)*10,8261.86(1.83–1.90)
Head and neck (141–149, 161, excl. 142, 147)46914.9(13.6–16.3)46810.3(9.37–11.3)*2719.50(8.40–10.7)*868.37(6.69–10.3)*1,29411.2(10.6–11.8)
 Tongue (141)11820.8(17.3–25.0)12414.9(12.4–17.7)7013.0(10.1–16.4)3014.7(9.91–21.0)34216.0(14.3–17.7)
 Mouth (143–145)10414.5(11.9–17.6)12812.0(10.1–14.3)8011.4(9.06–14.2)3312.0(8.23–16.8)34512.5(11.2–13.9)
 Pharynx, excl. nasopharynx (146–149, excl. 147)18117.0(14.6–19.7)14310.2(8.59–12.0)*819.82(7.80–12.2)*134.78(2.54–8.17)*41113.5(12.2–14.9)
 Larynx (161)668.26(6.39–10.5)735.83(4.57–7.33)405.07(3.62–6.90)103.61(1.73–6.64)1896.06(5.23–6.99)
Lip (140)252.10(1.36–3.10)572.66(2.01–3.44)422.36(1.70–3.19)232.83(1.79–4.23)1472.48(2.09–2.91)
Oesophagus (150)17110.3(8.78–11.9)2296.64(5.80–7.55)*1414.59(3.86–5.41)*312.06(1.40–2.92)*5725.90(5.43–6.41)
Lung (162)8664.46(4.17–4.77)1,3963.56(3.37–3.75)*8512.69(2.51–2.87)*2372.10(1.84–2.39)*3,3503.30(3.19–3.41)
Other tobacco-related cancers3571.50(1.35–1.66)6581.40(1.30–1.52)5391.22(1.11–1.32)2991.36(1.21–1.52)1,8531.35(1.29–1.42)
Other alcohol-related cancers2751.29(1.15–1.46)4841.29(1.18–1.41)4201.21(1.09–1.33)2341.25(1.10–1.43)1,4131.26(1.19–1.32)
All the other cancers4021.17(1.06–1.29)7761.10(1.02–1.18)7441.06(0.99–1.14)4171.20(1.08–1.32)2,1971.08(1.04–1.13)

Table III shows the SIRs of SPCs following head and neck cancers by follow-up time. Increased risks of SPCs persisted 10 years after diagnosis of the first primary mainly for SPCs in the head and neck, esophagus, lung other tobacco-related cancers and other alcohol-related cancers. The SIRs for second primary head and neck (SIR = 11.5, 95% CI = 10.1–13.1) and esophageal cancers (SIR = 7.22, 95% CI = 5.96–8.67) were very high especially within the first year of diagnosis of the first primaries head and neck cancer. The SIRs decreased with follow-up time but remained elevated. For second lung and other tobacco-related cancers, there was an increased SIR 1–4 years after the diagnosis.

Table III. Numbers of Cases (N) and Standardized Incidence Ratios (SIR) of Selected Second Cancers after Head and Neck Cancer by Follow-Up Years
Site of second cancer (ICD-9)<12 months1–4 years5–9 years≥10 years
NSIR95% CINSIR95% CINSIR95% CINSIR95% CI
  • Other tobacco-related cancers include the following: nasopharynx (ICD-9: 147), nose and nasal cavity (160), stomach (151), liver (155), pancreas (157), cervix (180), bladder (188), kidney (189) and leukemia (204–208). Other alcohol-related cancers include the following: colorectal (153, 154), liver (155) and female breast (174).

  • *

    p < 0.0001 compared to the <12 months group using Poisson Regression.

All Malignant (140–208)1,6341.89(1.80–1.99)3,9792.01(1.95–2.07)2,8091.94(1.87–2.01)24041.59*(1.52–1.65)
Head and neck (141–149, 161, excl. 142, 147)23611.5(10.1–13.1)50311.7(10.7–12.8)34212.03(10.8–13.4)2138.86(7.71–10.1)
 Tongue (141)6418.2(14.0–23.3)13216.8(14.1–19.9)9517.53(14.2–21.4)5111(8.17–14.4)
 Mouth (143–145)449.71(7.05–13.0)15615.6(13.2–18.2)9313.46(10.9–16.5)528.52(6.37–11.2)
 Pharynx, excl. nasopharynx (146–149, excl. 147)8714.9(12.0–18.4)1309.86(8.24–11.7)11012.18(10.0–14.7)9112(9.66–14.7)
 Larynx (161)416.23(4.47–8.45)857.2(5.75–8.91)446.21(4.51–8.34)193.33(2.01–5.21)
Lip (140)192.00(1.21–3.13)632.98(2.29–3.81)362.47(1.73–3.42)292.06(1.38–2.96)
Oesophagus (150)1157.22(5.96–8.67)2426.98(6.13–7.91)1355.62(4.71–6.65)803.58(2.84–4.46)*
Lung (162)4312.73(2.48–3.00)1,3013.62(3.42–3.82)*9243.63(3.40–3.87)*6942.85(2.64–3.07)
Other tobacco-related cancers2821.34(1.19–1.51)6471.37(1.27–1.48)4831.43(1.30–1.56)4411.26(1.14–1.38)
Other alcohol-related cancers1971.18(1.02–1.35)4941.29(1.18–1.41)3771.35(1.22–1.49)3451.18(1.06–1.31)
All other cancers3821.45(1.31–1.61)7831.13(1.05–1.21)5461.04(0.95–1.13)6291.07(0.99–1.16)

The SIRs of head and neck cancer as a SPC following a first primary cancer of various sites are presented in Table IV. Head and neck cancers were closely related to each other (SIR = 11.2, 95% CI = 10.6–11.8) as well as with esophageal (SIR = 5.25, 95% CI= 4.32–6.32) and lung cancers (SIR = 2.43, 95% CI = 2.24–2.64). Among patients with a first diagnoses of lymphohaematopoietic malignancies, the risk of second primary head and neck cancer increased with increasing duration of follow-up, with a SIR of 2.29 (95% CI = 1.84–2.82) 10 years after the diagnosis. Particularly high 10-year SIR estimates were found for second tongue cancer (SIR = 3.72, 95% CI = 2.36–5.58, data not shown) following a lymphohaematopoietic tumor.

Table IV. Numbers of Cases (N) and Standardized Incidence Ratios (SIR) of Head and Neck Cancers after Selected First Cancers by Follow-Up Years
Site of primary cancer (ICD-9)<12 months1–4 years5–9 years≥10 yearsTotal
NSIR95% CINSIR95% CINSIR95% CINSIR95% CINSIR95% CI
Head and neck (141–149, 161, excl. 142, 147)23611.6(10.1–13.1)50311.7(10.7–12.8)34212.0(10.8–13.4)2138.85(7.70–10.1)1,29411.2(10.6–11.8)
 Tongue (141)4212.5(9.03–17.0)12319.7(16.4–23.5)6917.6(13.7–22.3)4814.2(10.4–18.8)28216.7(14.8–18.8)
 Mouth (143–145)7616.6(13.1–20.8)12913.6(11.3–16.1)10917.9(14.7–21.6)408.16(5.83–11.1)35414.1(12.7–15.7)
 Pharynx, excl. nasopharynx (146–149, excl. 147)6913.4(10.4–17.0)11615.2(12.6–18.3)6616.4(12.7–20.8)248.42(5.39–12.5)27514.0(12.4–15.8)
 Larynx (161)496.67(4.93–8.81)1356.93(5.81–8.20)986.82(5.53–8.31)1017.84(6.39–9.53)3837.08(6.39–7.83)
Lip (140)332.97(2.05–4.17)942.53(2.04–3.09)662.02(1.56–2.57)741.70(1.34–2.14)2672.15(1.90–2.42)
Salivary gland (142)84.10(1.77–8.08)81.64(0.71–3.23)82.08(0.89–4.09)182.65(1.57–4.20)422.41(1.73–3.25)
Oesophagus (150)394.26(3.03–5.82)497.57(5.60–10.0)175.76(3.36–9.23)62.33(0.85–5.06)1115.25(4.32–6.32)
Lung (162)1822.05(1.76–2.37)2383.02(2.65–3.43)1102.92(2.40–3.52)611.62(1.24–2.08)5912.43(2.24–2.64)
Female breast (174)511.16(0.87–1.53)1471.13(0.95–1.33)1291.33(1.11–1.58)1331.16(0.97–1.37)4601.19(1.09–1.31)
Cervix uteri (180)193.45(2.08–5.40)271.71(1.13–2.48)352.15(1.65–3.21)921.74(1.40–2.13)1731.91(1.64–2.22)
Prostate (185)1080.78(0.64–0.94)2890.92(0.81–1.03)1270.96(0.80–1.15)541.06(0.79–1.38)5780.91(0.84–0.98)
Bladder (188)791.52(1.20–1.89)1791.36(1.17–1.58)1271.44(1.20–1.72)901.25(1.01–1.54)4751.38(1.26–1.51)
Kidney (189) (−189.3, 189.4)130.70(0.37–1.19)431.05(0.76–1.41)240.91(0.58–1.35)230.98(0.62–1.47)1030.94(0.77–1.14)
Lymphohaematopoietic (200–208)420.89(0.64–1.20)1301.34(1.12–1.59)681.34(1.04–1.70)882.29(1.84–2.82)3281.41(1.26–1.57)
 Lymphoma (200–202)210.98(0.61–1.50)811.72(1.36–2.13)451.54(1.12–2.06)642.32(1.79–2.97)2111.69(1.47–1.93)
 Leukemia (204–208)161.00(0.57–1.63)341.06(0.73–1.48)181.17(0.70–1.86)151.85(1.03–3.05)831.16(0.92–1.44)

Discussion

This study confirms that patients with a primary head and neck cancer have an excess risk of developing a SPC when compared to the general population. Lung cancer contributed to the highest proportion of the SPCs with a 20-year cumulative risk of 13%. High second primary lung cancer incidence may contribute to the poor prognosis for head and neck cancer patients with a SPC.5, 19, 20

The most frequent SPC sites reported in previous studies after a first primary head and neck cancer were head and neck (35%–73%), lung (15%–32%) and esophagus (9%) over a 10-year follow up period.5, 19, 21, 22 Most of these studies were based on hospital registries where surveillance bias could have resulted in more SPCs being diagnosed. The anatomical proximity of these cancer sites might have led to an earlier detection of second primary head and neck cancers during follow up examinations or a misdiagnosis of local spread of the primary head and neck cancer to other sites as a new primary. On the other hand, our study suggests that surveillance bias is not entirely responsible for the observed associations between head and neck cancers and second primaries in the aerodigestive tract, as the associations were persistent after 10 years of follow up when surveillance bias would be expected to be minimized.

An excess of SPCs was observed for tobacco- and alcohol-related cancers. In 1953, Slaughter et al.23 proposed the concept of field cancerization, where carcinogenic effects from tobacco and alcohol may simultaneously act on multiple parts of the aerodigestive tract mucosa, triggering the development of multiple primary cancers that are independent of each other. Studies have reported significantly higher SPC occurrences in active smokers than never smokers for head and neck cancer patients, with intermediate risks reported for former smokers.3, 24, 25 The occurrence of SPCs in the UADT suggests that the smoking and drinking habits of patients before the first primary diagnosis play a role in the SPC occurrence. It will be important to inform cancer patients who continued tobacco or tobacco-related habits can increase their risk of SPCs, including lung cancer, which subsequently leads to very poorer survival.

We observed that the SIR for head and neck cancer as a second primary was higher in younger onset cases (<56 years) than older onset cases (≥75 years). A possible explanation is that the increasing rates with age in the general population14 results in higher SIRs in the youngest group. Another possible reason is that factors such as tobacco, alcohol, HPV infection, treatment or genetics may contribute to the differences in SIRs between young and old patients. Previous studies have reported that young patients had minimal use of tobacco and alcohol.26 However, other studies27 have indicated that there were young patients with heavy smoking and alcohol drinking who might have been indulged in the habit for more than 20 years. Cases who are diagnosed at younger ages may have increased genetic susceptibility, which also predispose them to multiple tumors. However, common environmental exposures may contribute to an apparent increase in multiple primary tumors due to family history of cancer. Young patients tend to be treated with combined therapy of surgery and radiation or radiation and chemotherapy28 which may lead to the development of SPCs.

An important common risk factor that may be more important in younger patients is HPV infection. It has been noted in several epidemiological studies that HPV-related head and neck cancers tend to be presented at a younger age, with less tobacco and alcohol exposure, and for the orapharyngeal site.27, 29, 30 In our analysis restricted to women (data not shown), elevated risks were observed for both cervical cancer as a SPC following head and neck cancer (SIR = 1.71, 95% CI = 1.28–2.24) and head and neck cancer as a SPC following cervical cancer (SIR = 1.91, 95% CI = 1.64–2.22), especially among the youngest population. Similar analyses with the US SEER data also showed associations between first primary cervical cancer and second primary head and neck cancer,31 and between first primary genital cancers (squamous cell carcinoma of cervix, cervical carcinoma in situ, vulva/vagina and anus) and second primary oral squamous cell carcinoma.32

Although we observed various associations for SPCs in the head and neck, there are several limitations in our study. The definition for a SPC is that the SPC and the first primary cancer should be independent of each other.5 Recent molecular studies have shown that SPCs can share some or even all genetic markers with the index tumor, indicating that both tumors have arisen from a common cell clone.33 We cannot exclude the possibility of misclassification of a metastasis as the second primary, especially for those identified within 1 year of the first primary cancer.34 To address this concern, we stratified our analysis on follow-up periods and examined the patterns of SPC risk and found that the SIRs for cancers of the head and neck region, esophagus and lung were high during the first year and decreased with follow-up years indicating that some misclassification may have influenced our results.

In conclusion, our study suggests that patients with head and neck cancers are at increased risk of developing a SPC of the oral cavity and pharynx, esophagus, larynx and lung. The major advantage of this study is that the results in our study have higher precision than most previous studies due to the large number of first and second primary cases. These patterns suggest that head and neck SPCs are largely due to shared common risk factors such as tobacco and alcohol.

Acknowledgements

Drs. Ghislaine Scelo and Drs. Shu-Chun Chuang worked on this study under the tenure of a Special Training Award from the International Agency for Research on Cancer.

Ancillary