Disease burden, risk factors, and temporal trends of eye cancer: A global analysis of cancer registries

This study aims to investigate the global disease burden, risk factors, and temporal trends of eye cancer by sex and age group.

incidence trend and a decreasing mortality trend.Notably, all countries reporting decreasing trend in mortality were in the Asian or European region.Conclusions: Although higher incidence was observed in both African and European regions, only the Sub-Saharan Africa region reported high mortality, indicating inequity in the access of healthcare and treatment resource.Higher prevalence of UV exposure was associated with both higher incidence and mortality.Education should be provided to increase the awareness of eye protection.An overall declining mortality trend was found, but it was limited to only Asian and European countries.
disease burden, eye cancer, risk factor, temporal trend

| INTRODUCTION
2][3] Ocular cancers may affect the inside of the eye (intraocular) or the outer parts (extraocular), with the most common of primary intraocular cancers in adults being melanoma and lymphoma, and retinoblastoma in children. 4 Other ocular malignancies are secondary cancers that metastasize from other areas of the body, such as the breast or lungs, which often manifest in the uvea of the eye.Early detection and treatment are essential for avoiding vision impairment, vision loss, endangerment of eyeball integrity, and fatality. 5The 3-year survival rate is approximately 90%, whilst the 5-year is around 80%, but may plummet to a 5-year survival rate of 15% if the cancer spreads to distant parts of the body. 6,7lthough ocular cancer has indeterminable causes, different eye cancers may be more prevalent in certain groups of the population.Risk factors for intraocular melanoma include older age, being white, having a fair skin tone, light coloured eyes, atypical moles, exposure to ultraviolet (UV) radiation, and inherited cancer condition caused by the BAP1 gene. 8,94][15] Retinoblastoma in children has been associated with genetic changes if they inherited the retinoblastoma gene (RB1) mutation from their parents; delayed detection may lead to retinoblastoma-related life loss. 16,17owever, previous research suggests the detection of minimal dissemination can be used as a tool to assess retinoblastoma to aid in treatment and reduce morbidity in retinoblastoma-related life loss. 18,19he rarity of eye cancer cases has led to a scarcity of studies on the epidemiological trends of ocular cancers.The present study aims to provide an updated assessment of the global disease burden.An evaluation of the distribution and temporal trends across age groups, genders, and geographical regions, and risk factors associated with ocular cancer will be conducted.

| Ethics
This study was approved by the Survey and Behavioural Research Ethics Committee, The Chinese University of Hong Kong (No. SBRE-20-332).

| Data sources
The global incidence and mortality rates of eye cancers were examined with specific reference to malignant neoplasms of the eye and adnexa (C69) [conjunctiva (C69.0),cornea (C69.1),retina (C69.2),choroid (C69.3),ciliary body (C69.4), and other unspecified parts of the eye (C69.9)].However, the eyelid cancer (C43.1) was classified as malignant neoplasms of skin based on the ICD-10 and therefore was not included in this study. 20The Human Development Index (HDI) for each country and region was retrieved from World Bank and United Nations, with HDI rates of <0.550, 0.550-0.699,0.700-0.700and ≥0.800 categorised as low, medium, high, and very high.2][23] The CI5 database holds cancer incidence-related data obtained from global, regional, and national cancer registries.This includes proportion of cases registered, rate of cases recorded microscopically, and cancer incidence by age, primary tumour year, and region.Cancer-related statistics from the Nordic region and United States was accessed through the NORD-CAN database and SEER programme.The incidence of eye cancer in 2020 was from GLOBOCAN for the descriptive analysis.For trend analysis, we used yearly incidence data available from 1980 to 2012 from CI5, SEER, and NORD-CAN.For the mortality trend analysis, cancer-related death data for each country and region was obtained from the WHO IARC mortality database. 24National civil cancer registries, on a local and national level, are responsible for logging verified cancer deaths and their causes and compiling annual reports to be submitted to WHO.Only figured with medium quality or above were published by the WHO mortality database to ensure data accuracy and comprehensiveness.Data for the age-adjusted prevalence of tobacco use, alcohol consumption, physical inactivity, obesity, and chronic health conditions for each country was extracted from the WHO Global Health Observatory data repository. 25Global incidence and mortality of eye cancer was estimated from CI5 and GLOBOCAN using skin melanoma as the proportionate estimator as it shares similar risk factors with eye cancer.
The Segi-Doll world reference population was used to transform all cancer incidence and mortality figures to generate a weighted age-standardised rate (ASR). 26,27eighting proportionate to the number of people in each corresponding age group within the standard population.The Appendix S1 contains detailed information on the data sources used for the analyses.

| Statistical analysis
Choropleth maps were constructed on the global incidence and mortality of eye cancers in 2020.A linear regression analysis examining the relationships between HDI, risk factors, and eye cancer incidence and mortality was carried out.The beta coefficients (β) and the corresponding 95% confidence intervals (CI) were generated using Stata 16.0.β estimates indicated the degree of change in ASR for eye cancer incidence and mortality rates.A logarithm transformation was performed prior to conducting the Joinpoint regression analysis to assess the temporal trend of incidence and mortality rates for all ages.This was categorised according to gender (male and female) and geographical regions (Asia, Oceania, Northern America, Southern America, Northern Europe, Western Europe, Southern Europe, Eastern Europe).Joinpoint regression analysis software (Version 4.8.0.1-April 2020; Statistical Methodology and Applications Branch, Surveillance Research Program, National Cancer Institute) was used to calculate the average annual percentage change (AAPC) and its 95% CI.For analysis on transition trends, AAPC is preferred over annual percentage change (APC) as it does not assume linearity and takes into consideration the length of time. 28Previous studies have utilised this method to determine the epidemiological trend of other cancers, with a positive or negative AAPC indicating an increasing or decrease trend in incidence or mortality. 29All p values less than 0.05 were considered statistically significant.

| Global incidence of eye cancer
In 2020, there was an estimated 46 680 new cases of eye cancer with an ASR of 0.49 new cases per 100 000 persons (Figure 1).The highest incidence was found in Sub-Saharan Africa (ASR = 4.06), followed by Western Europe (ASR = 0.89) and Northern Europe (ASR = 0.84), whilst lower incidence was observed in Asian regions, particularly Eastern Asia (ASR = 0.09) and South-Central Asia (ASR = 0.14).Males had a slightly higher incidence (ASR = 0.58 vs. 0.42 in females), and such disparity was consistent throughout all regions.Countries with the highest incidence included Côte d'Ivoire (ASR = 33.63),Namibia (ASR = 18.11), and La Réunion, France (ASR = 12.19), all in the Sub-Saharan African region.The overall and sex-specific ASR of incidence and number of new cases of all countries and regions can be found in Table S1a.

| Global mortality of eye cancer
There were approximately 8202 eye cancer-related deaths in 2020 (ASR = 0.08).Sub-Saharan Africa (ASR = 1.59) had a distinctively higher mortality than other regions (ASR = 0.05-0.16)(Figure 2).Despite a high incidence found in Western Europe and Northern Europe, their respective mortality remained low (ASR Western Europe : 0.07; ASR Northern Europe : 0.09).Likewise, the mortality was higher in males (ASR = 0.11) than females (ASR = 0.06) overall and in all regions.Similar to the incidence, countries with higher mortality were all Sub-Saharan African countries, the highest mortality was found in Côte d'Ivoire (ASR = 8.87), followed by Malawi (ASR = 4.06), Angola (ASR = 2.73), and the Republic of Congo (ASR = 2.73).The national and regional, overall and sex-specific ASR of mortality and number of eye-cancer related deaths of all are listed in Table S1b.

| Risk factors associated with higher eye cancer mortality
In men, higher eye cancer mortality was associated with higher UV exposure (β = 0.243; 95% CI = 0.135,  cancer was found in Sub-Saharan Africa, and Western and Northern Europe, but higher mortality was found only in the Sub-Saharan African region.(2) Higher incidence and mortality were associated with lower HDI, higher prevalence of UV exposure and several other factors; (3) a relatively stable trend in incidence of eye cancer was observed overall, despite highly remarkable increases were found in some populations; (4) there was an overall decreasing trend in mortality, all countries reporting decreasing mortality were from Asia or Europe.

| Comparison with previous literature
It was found that eye cancer incidence was associated with HDI in a U-shaped relationship.HDI was significantly associated with improved drinking-water sources (β = 83.93,95% CI = 64.71,103.15) and improved sanitation facilities (β = 199.90,95% CI: 174.39, 225.42). 30A study conducted in the United States has shown that eye cancer incidence was inversely correlated with the availability of fluoridated water. 313][34] It is suspected that the inequity in sanitation might have contributed to the association.Meanwhile, higher incidence found in higher HDI regions in Europe was likely a result of the white race being a risk factor. 8On the other hand, the inverse association between eye cancer mortality and HDI was probably a result of delayed diagnosis due to a lack of medical resource, which contributed to a worse prognosis and reduced rates of patient survival, 35 this was supported by the disparity in mortality among regions with high incidence.In addition, higher ASR in the incidence of eye cancer was observed in some countries in Africa, especially in the sub-Saharan Africa region, which may be attributed to the local prevalence of conjunctival squamous cell carcinoma (SCC).The immune suppression caused by HIV and UV exposure has been confirmed to be associated with the incidence of conjunctival SCC. 36herefore, the concentration of HIV-infected individuals in sub-Saharan Africa and the prevalence of HIV may have influenced the prevalence and increased incidence of eye cancers, especially conjunctival SCC. 37n this study, UV exposure has been found significantly associated with increased mortality and incidence.From a genetic perspective, a previous study looked at the DNA signature of iris melanoma and found that the anterior uveal tract was affected by UV, resulting in DNA damage, 38 and genetic mutations and gene chain changes have been found in uveal melanoma, which may be related to UV exposure. 39,40Nevertheless, previous epidemiological findings on the effect of UV exposure on eye cancer incidence had been inconsistent. 41,42everal lifestyle habits and metabolic syndromes including smoking, obesity, and diabetes mellitus were identified as possible risk factors of both eye cancer incidence and mortality.So far, they had not been linked to risk of eye cancer, despite links between various eye diseases and smoking, 43 obesity, 44 and diabetes 45 suggested in previous studies.The effects found in this study might be a result of the adoption of an ecological study design, where risk factors were examined on a population level, and that these factors were partly correlated with HDI, which was a main risk factor identified.
In the current study, we have found an overall decreasing trend in eye cancer mortality, particular among Asian and European countries, which accorded with previous studies concerning the trend in Europe. 46t is speculated that the decline in mortality was attributable to the better eye protection from UV radiation and improvement in the treatment of the neoplasms. 46As the development of diagnostic and treatment tool continues to progress 47 and the increased awareness in eye protection, 48 it is forecasted that there will be a further decline in mortality but a relatively stable or slightly increasing trend incidence.In addition, countries such as the Kingdom, Israel, and have observed significant increase in the incidence of eye cancer, especially in the male population and in the age group <50 years.The factors responsible for this increase are not fully described.However, the increased incidence eye cancer may the prevalence of its associated lifestyle risk factors in these countries.A report from the United Kingdom pointed out that the number of overweight and obese population in the United Kingdom continues to increase.In 2021, about 64% of the population will be defined as obese or overweight, and the population between 35 and 54 years old is the most serious. 49Similar situations exist in Israel and Canada. 50,51In addition, according to the WHO report, 52 in the past 10 years, the per capita alcohol consumption in the United Kingdom and Canada has declined slightly, while the per capita alcohol consumption in Israel has increased.This may be one of the reasons that explains the particularly high rising trend in the incidence of eye cancer in Israel.In addition, the alcohol consumption habits and frequency of the younger population are higher than those of the older population, especially those after the age of 65, which may be used to explain the significant upward trend in the younger group. 53owever, the drivers behind specific trends in individual countries are often multifaceted and may require more in-depth research and investigation.

| Limitations
The current study has a few limitations.Incidence rates may have been over-estimated as cancer diagnoses were predicted to decline because of the pandemic, whilst mortality rates may have been under-estimated due to the late diagnoses of eye cancer cases or coinfection of COVID-19.In developing countries, there is the possibility of under-reporting of eye cancers due to a lack of well-established infrastructure and mechanisms for cancer screening and reporting.As each country may utilise differing methods of registering cancer cases, the inconsistency that arises may lead to difficulties when directly comparing rates across different regions.Lastly, due to the database utilised in this study did not classify the subtype of eye cancer, the risk factors or trends of different subtypes were not analysed and discussed in this study.Future research can use more available databases to classify and analyse different subtypes.

| Conclusion
Lower HDI has been identified as an associated risk factor of eye cancer incidence and mortality, indicating potential inequity of access to sanitation and healthcare service, and the need of intervention.Higher prevalence of UV exposure has been associated with higher cancer incidence and mortality, the biological mechanism behind the link should be further explored.There has been an encouraging overall decline in mortality from eye cancer possibly due to the advancement in treatment and increased awareness on eye protection.Nevertheless, a few populations had reported significant increases in incidence.More research should be done to explore the causes behind the temporal epidemiological trends of different subtypes.Subsequent trend analysis can also be done to capture the trends of eye cancer in the post COVID-19 era to evaluate the impact of the COVID-19 pandemic, especially after cancer diagnosis has got back to normal.

1
Global incidence of eye cancer overall and by sex in 2020.

4
Associations between risk factors and eye cancer incidence.