Cancer of the tongue is an uncommon cancer site, with only 31,378 cases in the SEER 1975-2017 database, fewer than 1% of all reported cancers. This article updates trends in incidence, prevalence, short and long-term survival and mortality of tongue carcinoma.
Background & Importance.—Cancer of the tongue is infrequent, and although it accounts for 25% (31,378) of all oropharyngeal cancers (126,888) in the National Cancer Institute’s SEER frequency database, 1975-2017,1 it comprises only 0.64% of all cancers (4,879,410) for the same period, and a total of 24,805 microscopically confirmed malignant cases in the SEER survival database2 considered in this article. Tongue cancer is most frequently diagnosed among people ages 55-64 years (28%) and the mean age by sex in males and females was 61 and 64 years, respectively; by race in whites and blacks, 63 and 59 years respectively; mean age at death is 67 years. Lingual cancer is more common in males than females (male-to-female ratio 2.1:1) and by race, far more common in whites than in blacks (white-to-black ratio 10.6:1). In the United States, there are estimated to be 17,660 new cases in 2020 (1% of all new cancers) and 2830 estimated deaths (0.5% of all cancer deaths).3 While the incidence, frequency and deaths are low in the United States, they are exceptionally important from a clinical and surgical standpoint. The morphological and functional changes require a multidisciplinary approach to treatment including surgical, medical and radiation oncology, speech and physical rehabilitation therapy, and emotional support from family, friends, psychologists or social workers. The base of the tongue is the site of origin of slightly less than 50% of lingual cancers, most of the neoplasms that arise on the anterior two thirds of the tongue begin on the lateral surfaces. Greater than 96% of these neoplasms are squamous cell (TSCC) characterized by rapid local invasion and early lymph node metastasis and most of the remainder are adenocarcinomas or sarcomas. Risk factors include heavy alcohol use, tobacco use and infection with human papillomavirus (HPV), especially HPV-16.
Objective.—To update trends in incidence, prevalence, short and long-term survival and mortality of tongue carcinoma using the statistical database of SEER*Stat 8.3.6.14 for diagnosis years 1975-2017 employing multiple case selection variables.
Methods.—This study was both a retrospective and prospective population-based cohort study in design using nationally representative data from the National Cancer Institute’s (NCI) Surveillance, Epidemiology, and End Results (SEER) program (www.seer.cancer.gov) to evaluate 31,378 cases for diagnosis years 1975-2017 comparing multiple variables of age, sex, race, stage, cohort entry time-period, and disease duration. Frequencies/age-specific incidence rates, median age, extent of spread at diagnosis/stage, survival and mortality were examined. Relative survival statistics were analyzed in two cohorts: 1975-1996 and 1997-2017. Survival statistics were derived from: SEER*Stat 8.3.6.1 Database: Incidence – SEER 9 Regs Research Data, (1975-2017) released April 2020, based on the November 2019 submission. General methods and standard life table methodologies for converting SEER survival data to comparative mortality and explanations of cancer staging and grading procedures are described in previous Journal of Insurance Medicine articles5,6 and other publications.7,8 Excluded were all death certificate only and those alive with no survival time. The percentage of microscopically confirmed malignant behavior cancers for case selections was 100%. Incidence and survival rates were obtained from the most current SEER Cancer Statistics Review (CSR)9 publication and prevalence counts are based on 2017 population estimates from the US Bureau of the Census.
Statistical significance.—Standard errors are shown for survival rates in the SEER survival tables. Actuarial method: Ederer II method is used for cumulative expected survival. Ederer II method calculates the expected survival rates for patients under observation at each point of follow-up so the matched individuals are considered to be at risk until the corresponding cancer patient dies or is censored.10 Confidence interval: Log (-Log ()) Transformation; the level is 95%. Poisson confidence intervals at the 95% level based on the number of observed deaths are used in this study but not displayed here to conserve space on the mortality tables.
Results.—Incidence rates and trends, absolute & relative frequency distributions, survival & mortality by age, sex, stage of tongue carcinoma in two entrant time-periods as recorded in the SEER Program of the National Cancer Institute for diagnosis years 1975-2017 (SEER*Stat 8.3.6.1) are summarized. Shifts in trends over time are identified, and the findings are correlated with prognosis, including short and long-term observed (actual), expected & relative survival, median observed and relative survival, mortality rates & excess death rates per 1000 people.
Conclusions.—Trends in SEER incidence, prevalence, survival & mortality by age, sex, race, stage, cohort entry time-period, relative frequency & percent distribution, were examined to provide a current epidemiologic and medical-actuarial risk assessment framework for tongue carcinoma in the 1975-2017 time-frame.
Incidence:Table 1 by sex, indicates that tongue cancer AAPC incidence changes from 2013 to 2017 are higher in females than males, 3.4 vs 2.3per 100,000, respectively. By race, AAPC incidence trends are lower in black males than females, -2.5 vs -1.6 per 100,000, respectively, and lower in blacks than whites.
Limited available data in Table 2 indicates that incidence rates increase with age and vary by sex and race in the United States. Incidence is higher in males than in females, higher in whites than in blacks, and higher in the United States than the average elsewhere in the world.
SEER Case Statistics: The total of 31,378 cases of tongue carcinoma in the 1975-2017 SEER database1 was distributed by age/sex/race/stage/grade and cohort entry period as shown in Table 3. Tongue carcinoma diagnostic frequency compared by age and sex is illustrated in the Figure 1. Five percent of cases were unstaged, and 23% were of unknown grade. Parenthetically but not displayed are 1020 cases of in-situ carcinoma of the tongue. Regional and distant stages comprised 60% of staged cases. The zenith of diagnostic frequency (see Figure 1) by attained age in men occurred at quinquennial ages 60-64 (16.9%), and declining thereafter. From ages 75 up, 13.7% of male cases were diagnosed; male mean age at diagnosis, 61.4 years. In women the zenith of diagnostic frequency by attained age also occurred at quinquennial ages 60-64 years, then tended to plateau to ages 75-79 years, declining to 7.3% at 80-84 years, then slightly increased to 7.6% at ages 85+ years; from ages 75 up, 25.0% of female cases were diagnosed (see Figure 1). Female mean age at diagnosis was 63.9 years. There was a 2-to-1 male-to-female ratio by diagnostic frequency and males accounted for 68% of tongue cancers while females accounted for only 32% of cases. In the United States, tongue cancer is a disease of the white race (85%) and 15% or less in black, other and unknown races. Mean ages in whites and blacks are 62.3 years and 58.5 years, respectively.
Cancer Case Characteristics: Men made up 68% of the total cases of cancer of the tongue, and women 32%; there were few patients (8%) under age 45. Cases were excluded from the survival analysis if the diagnosis was made by death certificate or at autopsy. The SEER historic code is used for staging (local, regional or distant) in Tables 5-8. SEER ‘Combined Summary Stage 2000 (2004+)’ is used in Table 5 to select regional stage patterns of invasion (POI) by direct extension only, lymph nodes involved only, and by both direct extension and lymph node involvement to examine risk outcomes by methods of tumor spread. Tables 5-9 show both durational observed input data and calculated indices of comparative mortality including the mortality ratio (MR), excess death rate (EDR); and for comparative survival, observed and expected survival, P and P’, respectively, survival ratio (SR), and observed and relative median survival.
Follow-up (FU): Standard procedures utilized by the SEER registries were so effective that few cases were lost to FU. According to the 1975-2015 Cancer Statistics Review, aggregate FU of cases diagnosed 1975 and followed to 2014 was 98.9% complete for males and 97.9% complete for females.
Results:Table 5 shows results for local cancer of the tongue in the entire 1975-2017 database, by sex and for age groups <65, 65 up, and all ages combined. In comparison with major sites such as breast in women and prostate in men, excess mortality is very high in localized cancer of the tongue in both sexes and age groups. Survival is correspondingly low, with few survivors, less than 10%, after 15 years in patients age 65 and up. Peak mortality occurs at duration 1-2 years, with MR over 450% and EDR about 80 to 92 per 1000 per year, all ages combined, males and females, respectively. MR decreases and EDR increases with advancing age. MR tends to be higher in women because of their lower expected mortality, but sex differences in EDR appear to be small and irregular. Excess mortality persists to 15-20 years in both men and women, and total 5-year SR is only 74% in men and 77% in women.
Table 6 displays results for regional and distant carcinoma of the tongue. Excess mortality is very high, with 5-year SR of 54% in regional, and 32% in distant stages, respectively. Maximum mortality occurred in the second year in patients under 65 (MR 1922% and EDR 166 per 1000), and in the first year in patients 65 up (MR 778% and EDR 288 per 1000). When all ages were combined for the regional stage MR and EDR values tended to be higher in females. EDR decreases below 100 per 1000 after 5-year duration. A similar pattern of decreasing excess mortality is seen in the distant stage. Highest values are in the first year, with MR of 1808% and EDR of 365 per 1000 per year (1229 of 3181 patients [39%] died in the 1st duration).
Table 7 results compare subgroups matched by a combination of stage and grade, all ages combined in 2 cohorts, 1975-1996 and 1997-2017. Results are for durations 0-5 and 5-10 years for both sexes in the upper part of the table, and by sex for local stage only in the lower part. At 0-5 years, excess mortality appears to be somewhat lower and SR higher in the localized stage with grades 1&2 in the 1997-2017 cohort, and this is also true for the regional stage. For the more severe grades 3&4, mortality is lower and survival increased in the 1997-2017 cohort than in the earlier cohort as a result of improvements in treatment. Most cancers were graded, and the proportion of cancers graded increased in the more recent cohort. Decrease in mortality and increase in survival as a result of improvements in treatment in the 1997-2017 cohort are modest with carcinoma of the tongue, in contrast to results in some sites such as cancers of the breast and the prostate. Grading at the 3&4 level is associated with a worse 5-year prognosis than grades 1&2 in local cancer of the tongue, but not in regional cancer.
Table 8 summary results for the 1997-2017 cohort are given by race and stage. Excess mortality, both MR and EDR, in all stages, was substantially greater in black-other-unknown races than in the white race. For example, in the 1st duration of local stage cancer, MR was 357% for black-other-unknown patients and 280% for whites; EDR values were 43 and 40 per 1000 per year, respectively.
Table 9 illustrates specific risk results for differing patterns of invasion (POI) in patients with regional stage squamous cell carcinoma of the tongue and details differences in the prognostic impact with direct tissue invasion alone (DTIA), nodal involvement alone (NIA) or a combination of both by race, all ages combined using the SEER Combined Summary Stage 2000 (2004+) selection choice. Table 9 shows that in black, other and unknown patients, MR and EDR risk values are approximately twice that for regional stage white patients – both combined tissue extension and nodal invasion – and are associated with much diminished median survival times. However, in the 0-5-year durational interval for all races combined summarized in Table 4, differences in MR, EDR, and relative cumulative survival (SR) are not large. Nevertheless, a pattern of invasion with nodal involvement alone (NIA) compared to direct tissue invasion alone (DTIA), all races combined, 5-year mortality is notably less and survival improved in patients with NIA. Median survival time with NIA is more than twice that with DTIA, and this applies also to both black and white patient-populations listed in Table 9.
Overall, Table 9 prognostic indices by race are comparatively better in the white population. Results for local and distant stages adjusted for cohort entry time-period are proportionally similar to those in the previous tables for all ages, durations 0-5 and 5-10 years.
Comment: Cancer of the tongue is an uncommon cancer site, with only 31,378 cases in the SEER 1975-2017 database, fewer than 1% of all reported cancers. It is characterized by a male predominance, a mean age of 62 years, and a high excess mortality in cases staged as localized. We speculate that staging is more difficult for tongue cancer than for many other sites, because of the highly vascular soft tissue stroma characteristic of the tongue. It is possible that many cancers classified as local have already spread undetected to lymph nodes or adjacent tissues.
Tables 5 and 6 provide detailed results for local and regional cancer by a combination of sex and two age groups, <65 and 65 up, plus all ages combined in a 20-year FU. Highest EDR values are in the second year, about 92 per 1000, all ages in local, and 175 in regional cases, with little sex difference. EDR decreases after 2 years, and excess mortality appears to persist up to 15 years, although there are few survivors after 15 years. EDR values are even higher, over 350 per 1000 in the 1st duration in distant cancer of the tongue. Results in Table 7 do show improvement in the 1997-2017 cohort over results in the matching subgroup of the 1975-1996 cohort (subgroups are for interpolations of stage and grades 1&2, 3&4, unknown grade, or all grades (G&U), at durations 0-5 and 5-10 years, all ages combined). Grading is also less effective in prognosis than it is in many other cancer sites. In Table 8, EDR and MR at 0-5 years duration in black-other-unknown races of patients in all stages appears to be higher, and SR is lower than in white patients. Table 9 illustrates racial differences in tongue carcinoma mortality and survival results by pattern of invasion in the regional stage. However, prognostic indices using the SEER Combined Summary Stage 2000 are proportionally similar in the most current cohort time-period.
Profound appreciation to my esteemed colleague Dr. Richard B. Singer, MD (deceased) for his friendship, invaluable collaboration and scholarly expertise, encouragement and assistance in the initial (unpublished) iterations of long-term comparative mortality and survival in patients with tongue carcinoma in 2003.
