Incidence and Survival of Head & Neck Cancers in Estonia, 1996–2016

Background and Material Changing patterns of alcohol and tobacco consumption and human papillomavirus (HPV) infection have affected the epidemiology of head & neck cancers. The aim of the study was to examine 20-year trends in the incidence and survival of head & neck cancers in Estonia by site, sex, morphology and stage. Methods Data on all adult cases of invasive head & neck cancers diagnosed in Estonia in 1996–2016 were obtained from the population-based cancer registry. TNM stage was available for 2010–2016. Incidence trends were modelled with joinpoint regression and ve-year relative survival ratios (RSR) were calculated.


Background
Head & neck cancers (including cancers of lip, mouth & pharynx, salivary glands, nasal cavity and sinuses, larynx, and thyroid) comprise around 8% of all incident cancers diagnosed in the world each year [1]. The etiology of these cancers varies and globally, the incidence and mortality trends of head & neck cancers observed recently are rather diverse, due to differences in the prevalence of risk factors and diagnostic activities. Decreasing use of tobacco has brought along declining trends of oral and laryngeal cancer, whereas growing burden of HPV infection has caused increasing rates of oropharyngeal cancers [2,3]. Furthermore, there have been reports of rising incidence of late-stage oral and pharyngeal cancers [4]. Thyroid cancer trends are deeply affected by wide-spread diagnostic activities and the detection of indolent tumors in many highly developed countries [5][6][7].
According to GLOBOCAN 2018 [1], the incidence of head & neck cancers in Estonia is not high, lower than the very high rates of several Eastern European countries. However, the survival of these tumors has been among the lowest in Europe both in EUROCARE-3 [8] and in EUROCARE-5 [9], particularly in men.
A detailed examination of incidence and survival of head & neck cancers in Estonia was thus warranted, in view of the decreasing burden of smoking-related cancers in Estonia [10] but the very high incidence of cervical cancer [11], which suggests high prevalence of HPV infection in the population. Also, previous studies have shown a large effect of diagnostic activities on cancer incidence during the transition of the health-care system and increasing availability of modern diagnostic methods [12,13].
The aim of the study was to examine 20-year trends in the incidence and survival of head & neck cancers in Estonia by site; separate analyses were done by sex, morphology and stage.

Methods
The Estonian Cancer Registry provided data on all adult ( Death certi cate only and autopsy cases were excluded from survival analyses. Patients who were diagnosed and died on the same day were included with one day of survival time. Relative survival ratio (RSR) with 95% con dence interval (CI) was calculated by dividing the observed survival in the study cohort by the expected survival, derived from age-, sex-and calendar-period-speci c life tables of the Estonian general population, using the Ederer II method [14]. Cohort method was used to calculate RSRs for cases diagnosed in 1996-2002 and 2003-2009, and period method for 2010-2016. For stagespeci c survival analysis for 2010-2016, complete method was used. International standards were used for age-standardization of RSRs [15]. Two-sided p-values were used to compare proportions. Survival analyses were done with strs algorithm in STATA 14 [16].

Results
Overall, 6769 cases of head & neck cancers were diagnosed in Estonia in 1996-2016. The most common site was thyroid, followed by larynx and oropharynx (Table 1). In general, %MV was high and %DCO and %Autopsy were low. The proportion of women was 36% among all incident cases, while it ranged from 4% for hypopharynx to 82% for thyroid. In site group mouth & pharynx, women comprised 21% of the cases.
Median age of all patients was 62 years; it was the lowest for thyroid and nasopharynx (59 years) and the highest for lip cancer (73 years). The age distribution shifted signi cantly towards older age groups as the proportion of cases age 70 years and over increased from 23-31% from 1996-2002 to 2010-2016 (p < 0.001).

Incidence
A signi cant decline was seen in the incidence of lip and laryngeal cancer (Fig. 1). Starting from early 2000s, a signi cant increase was seen for oropharyngeal and hypopharyngeal cancer. For tongue cancer, an increase started in 2010. A slight but nonsigni cant increase was seen for thyroid cancer. The incidence of papillary thyroid cancers increased signi cantly at a rate of 4.9% (95% CI 2.5; 7.4) per year until 2007 and levelled off thereafter. The incidence of follicular thyroid cancer decreased signi cantly throughout the study period (APC − 5.3, 95% CI -8.4;-2.0).
The incidence increases of oropharyngeal and hypopharyngeal cancer were signi cant only for ages 60 years and over, while the modest increases seen for younger people were not signi cant. Similar ndings were observed for thyroid cancer.
Within the group of mouth & pharyngeal cancers, the proportion of oropharyngeal and hypopharyngeal cancers increased signi cantly over the study period, in parallel with the signi cant drop in the proportion of oral cavity cancers (Supplementary Table 1).
The majority of cases among mouth & pharyngeal cancers were histologically squamous cell carcinomas and this proportion did not change over the study period (Supplementary Table 2). The proportion of squamous cell carcinomas among laryngeal cancers increased signi cantly on the account of other and NOS tumors. For thyroid, there was a signi cant increase in the proportion of papillary cancers, while the proportion of follicular tumors and medullary tumors decreased.
Stage was available for cases diagnosed from 2010. For mouth & pharyngeal cancers, the majority of cases (62%) were diagnosed at stage IV (Fig. 2). For hypopharynx, the proportion of stage IV tumors was close to 80%. Thyroid and laryngeal cancer had the lowest proportion of stage IV cancers. Over 40% of thyroid cancer cases were diagnosed at stage I.

Survival
Overall, the age-standardized ve-year RSR for mouth & pharyngeal cancer increased signi cantly over the study period, from 21% to 33% (Table 2). An even larger increase (from 20-35%) was seen for squamous cell carcinomas of the same site group. By site, the largest survival increases were seen for oral cavity, tongue and laryngeal cancers, while modest changes were seen for oropharyngeal and hypopharyngeal cancers. The RSR for thyroid cancer increased by 7 percentage units.

Incidence
The slightly declining trend in oral cavity cancer rate, together with the steep decrease in lip and laryngeal cancer incidence is in concordance with previously observed decreasing rates of lung cancer in Estonia [10]. Smoking is one of the strongest risk factors for most head & neck cancers. Daily smoking prevalence among working-age men in Estonia has dropped from 45% in 1996 to 23% in 2018; the respective change among women was from 21-13%; however, the change in women has been too recent to in uence cancer occurrence [17]. There have been no particular changes in alcohol consumption as the proportion of men and women who consumed alcohol at least a few times a week uctuated from 20% in 1996 to 27% in 2008 and to 24% in 2018 [17]. As only 4-6% of oral cancers are known to be HPV-positive, papillomavirus is not expected to play important role in oral cavity cancer pathogenesis [18]. Inconsistent patterns in the incidence of oral cavity cancers has been observed worldwide with trends in some countries contrasting those of lung cancer, suggesting the role of other factors or their interaction [3]. The recent signi cant increase of tongue cancer incidence in Estonia since 2010 needs further monitoring to con rm. Laryngeal cancer, also mainly caused by smoking and alcohol consumption [19], demonstrated a signi cant decline, which correlates with prior research from other European countries [20][21][22].
The oropharyngeal cancer trend has turned into a steep rise since 2004, which is consistent with similar trends in Europe, US and South-East Asia [2,3,20]. These trends have been driven by increasing incidence of HPV-positive cancers, and the role of HPV infection has been con rmed by the increase in the proportion of HPV-positive tumors, shown to reach 70% in the US and Sweden [23,24]. In the US, the rise in HPV-positive tumors was accompanied by a decrease in HPV-negative tumors [23]. HPV-infections play higher role in economically more-developed countries, whilst in economically less-developed countries, the disease is still mainly caused by tobacco [25][26][27]. Worldwide, approximately 30% of oropharyngeal cancers are due to HPV infection [28]. The trends observed in this study regarding HPV-related cancer sites, together with the continuously increasing risk of cervical cancer in successive birth cohorts [11] and rising incidence of anal cancer among younger women in Estonia [29] suggest an important role of HPV infection. Recent studies have also shown an increase in the prevalence of genital warts in Estonia [30]. HPV vaccination, expected to have a protective role against HPV-associated cancers, is available for girls in the national immunization program in Estonia only since 2018. The vaccination coverage reached 61% for the rst birth cohort in 2019. Even so, vaccination is unlikely to play a role in incidence rates anytime soon, and unfortunately, boys are not covered by HPV vaccinations in Estonia.
The incidence for hypopharyngeal cancers has shown a signi cant increase since 2002. Similar trends have been observed in other European countries [2,20,21], regardless of decreases in alcohol consumption, the main known risk factor for hypopharyngeal cancer. Alcohol consumption is also not a likely explanation for our ndings as the trend resembled that of oropharyngeal cancer rather than that of oral cavity cancer, and therefore partial role of HPV can be suggested. The predominance of men was particularly evident for hypopharyngeal cancer as only 4% of the cases were seen in women, while a fth of all mouth & pharyngeal cancers were diagnosed in women. The overall proportion of women was similar to that seen in other countries [31]. The excess of men is probably explained by differences in health behavior.
Thyroid cancer was the most common head & neck tumor in females, who constituted 82% of the cases.
Recent studies from Europe, USA, Canada, and Australia have reported a steady increase in thyroid cancer incidence over time with a particular increase in papillary thyroid cancer [7,32]. These trends may be explained by increases in diagnostic activity and changed histological criteria [33] or linked to changes in risk factors such as iodine supplementation, radiation exposure, Hashimoto thyroiditis, and hormonal or reproductive factors [34][35][36]. In addition, a link between obesity and thyroid malignancies has been demonstrated [37]. The latter explanations are supported by reports of increasing incidence of advancedstage thyroid cancers [5]. At the same time, other countries have not reproduced these ndings [38]. A recent report suggested a substantial contribution of overdiagnosis to the rising incidence in high-income, but also in less a uent countries [39]. The rise in thyroid cancer incidence in Estonia was signi cant only in women age 60 years and over (data not shown). A larger increase may have occurred before the start of our study as the availability of new diagnostic procedures increased rapidly in the beginning of the 1990s. According to Cancer in Five Continents, the ASR for women increased from 2.8 to 5.5 from 1983-1987 to 1998-2002 and from 0.6 to 1.2 in men [40] and stabilized thereafter. Trends in thyroid cancer warrant further monitoring, particularly among younger age groups, as there have been reports of increasing incidence of papillary tumors among young adults [41].

Survival
The survival of mouth & pharyngeal cancer has increased in Estonia, but a large de cit compared to other European countries persists. For these sites, the RS estimates for Estonia in 2010-2016 have not yet reached those observed in Europe in 2000-2007 [9]. The largest survival increase was seen for oral cavity cancers, but the latest estimates for both men and women still remain well below those observed for the Nordic countries [42]. Relative survival of oropharyngeal cancers was 58% in Denmark in 2010-2014 [21] and only 24% in our study. Stage at presentation is the most important prognostic factor for squamous cell head & neck cancers [43] and patients with advanced tumors show the shortest survival [44]. The majority (62%) of mouth & pharyngeal cancers in Estonia in 2010-2016 presented in stage IV. Roughly, only ¼ of the patients in Estonia presented with an early disease, which is similar to United States, where approximately 29% of head & neck tumors were diagnosed as localized [45]. We observed a huge gender gap in survival for all mouth & pharyngeal cancers, but particularly for oropharyngeal and salivary gland cancers. Signi cantly later stage at diagnosis for mouth & pharyngeal cancers among men compared to women (proportion of stage IV tumors 65% in men and 52% in women, p < 0.001, data not shown) is one potential explanation. Male sex has been found to be a predictor of late-stage head & neck cancer, together with increased age, black race, absence of health insurance and tumor site [4]. A signi cant female survival advantage was found in Estonia for ve of the nine studied common solid tumors, including mouth & pharyngeal tumors, even after adjusting for age, stage and subsite, suggesting the role of less co-morbidities, higher treatment compliance and better health behavior among women [46].
Smoking status has an impact on survival even in long-term [47,48], and differences in smoking prevalence and intensity (both before and after diagnosis) might partially explain sex differences. Recent studies elsewhere have shown a change in the pro le of oropharyngeal cancer patients towards nonsmokers, younger age at diagnosis with better performance status, less co-morbidities and HPV-positivity [49][50][51]. HPV-positive patients have displayed better overall survival and progression-free survival in comparison to HPV-negative cancer patients indicating separate risk factors, treatment response and prognosis for the disease [48,50]. The persistently low survival observed in our study, however, suggests the predominance of tobacco-and alcohol-related cancers in Estonia.
Patients with head & neck cancer are known to have a higher comorbidity burden in comparison to general population [52], which may in uence the choice of treatment modality and has a signi cant negative impact on their survival outcome [53,54]. Estonian cancer patients have been shown to have more comorbidities than their European counterparts [55] and this may be particularly true for patients with health behavior-related cancers. Moreover, we also observed a large age difference in mouth & pharyngeal cancer survival, over 20 percentage points in favor of the youngest age group (49% for age < 50, 27% for age ≥ 70 years, data not shown).
Treatment possibilities in terms of timely access to radiotherapy, use of concomitant chemoradiation, adequate palliative care, social support and rehabilitation, utilization of PET-scan for better diagnostics, and centralization of head and neck cancer patients into specialized units are key clinical factors for ensuring the best possible outcomes. Head & neck cancers are mainly treated at one comprehensive cancer center in Estonia. However, the availability of radiotherapy has been severely hindered due to low number of radiotherapy treatment machines up until 2012 [56]. This de cit has caused prolonged waiting times, which have been shown to cause progress of head & neck cancers to next stage [57]. In 2012, less than 60% of patients in Estonia who would have required at least one radiotherapy course were actually treated [58]. The total number of linear accelerators per million inhabitants increased to 4.6 only in 2016, and this recent improvement had no effect on our results.
In EUROCARE-5, the survival de cit of Estonian laryngeal cancer patients compared to those in Central or Northern Europe exceeded 10 percent units [9]. There has been a steady increase and the most recent survival estimates in this study for both men and women are close to those observed in Finland for 2012-2016 [42]. In contrast to Estonia and Finland, men have higher survival for laryngeal cancer than women in other Nordic countries [42].
Thyroid cancer survival in Estonia is good and higher than in Denmark in 2010-2014, where the ve-year RS estimates were 82% in women and 74% in men, and 91% for papillary and 80% for follicular cancer [6].
Our results for both sexes are well comparable with NORDCAN estimates for all Nordic countries [42]. Nevertheless, the gender survival gap of 12 percentage units was larger in Estonia than in the Nordic countries. One possible explanation for women's survival advantage is that they have higher proportion of papillary tumors than men (67% vs 55%, data not shown), and women have their tumors diagnosed at a signi cantly earlier stage (stage I proportion 44% for women, 26% for men, data not shown). Also, overdiagnosis may be more common among women, in ating survival estimates [39].
The main strength of our study was the use of high-quality population-based cancer registry data over a 20-year period. Another strength was the availability of relatively complete TNM-stage information for recent years, as the proportion of unknown stage did not exceed 11% for any site.
The main limitations of the current study were the lack of individual data on major behavioral, socioeconomic and other risk factors, most importantly HPV-status, smoking and alcohol consumption habits, and comorbidities, as well as data on diagnostic and treatment delays and treatment compliance. Small numbers prevented more speci c analysis of incidence and survival trends (e.g. by site/morphology and age).

Conclusions
Discordant trends were observed in the incidence of head & neck cancers in Estonia, with declines seen for oral and laryngeal cancers, but recent sharp increases for sites with persistently dismal prognosis, such as oropharyngeal and hypopharyngeal cancers. Primary prevention targeting health behavior but also HPV vaccination have a central role in reducing mortality from these cancers. A shift in the diagnosis towards earlier stage is crucial for improving the survival and the quality of life of head & neck cancer patients. Increasing awareness of these cancers and educating the public, as well as physicians, are critical in achieving this goal. Further analyses should focus on ascertaining the effect of HPV-infection on the incidence and survival trends, examining the availability of optimal treatment for all patient groups and identifying factors that affect treatment choice and survival outcome.

Declarations
Ethics approval and consent to participate: Ethical approval for the research has been con rmed.
Availability of data and materials: All data generated or analysed during this study are included in this published article and its supplementary les.
Authors' contributions: The authors declare that they have no competing interests Consent for publication: Not applicable