What makes salivary gland cancer so difficult to compare internationally is its hidden biological diversity. The category pools multiple histologic entities that differ sharply in anatomic predilection, molecular features, recurrence behavior and prognosis, and population-based data from the United States show wide incidence variation by subtype, sex, age, race and anatomic site. International estimation systems aggregate all of these tumors under the ICD-10 C07–C08 top site, covering both major and minor glands, which makes cross-country comparison possible but masks subtype-specific variation. Most existing knowledge comes from institutional and national series whose heterogeneous case mix and referral contexts limit comparison, and population-based studies from the United Kingdom, Denmark, the Netherlands and Spain have exposed differences in ascertainment, classification and observation periods. Whether the burden varies along a national development gradient had mostly been examined in single national series rather than comparatively, leaving rare head-and-neck sites underrepresented in global oncology assessments.

Methodologically, the study is a descriptive, population-based, cross-national analysis built entirely on the GLOBOCAN 2024 estimates published by the Global Cancer Observatory of the International Agency for Research on Cancer. The researchers extracted incident cases, deaths, 5-year prevalent cases, age-standardized incidence rates (ASIR) and age-standardized mortality rates (ASMR) for each of the 19 sovereign member economies and four global HDI aggregate groups, with all rates standardized to the 1966 Segi–Doll World Standard Population and expressed per 100,000. They then computed three all-age derived ratios: the mortality-to-incidence ratio (MIR) as a case-fatality proxy, the prevalence-to-incidence ratio (PIR) as a burden proxy, and the prevalence-to-mortality ratio (PMR) as an exploratory survival-pool indicator. Associations with HDI were tested using two-sided Spearman rank correlations across the 19 economies, with 95 percent confidence intervals obtained by non-parametric percentile bootstrap with 20,000 resamples, since GLOBOCAN publishes no uncertainty intervals for its case counts.

The geographic picture is strikingly uneven. Age-standardized incidence ranged from 0.30 per 100,000 in Saudi Arabia to 0.86 in Canada, with the three highest rates in Canada, Italy and the United States (0.86, 0.84 and 0.84) and the three lowest in Saudi Arabia, South Africa and India (0.30, 0.43 and 0.45). Mortality told a different story: Indonesia recorded the highest age-standardized mortality rate at 0.40 per 100,000, followed by the Russian Federation (0.25) and Italy (0.20), while Mexico sat lowest at 0.12. Canada also led on age-standardized 5-year prevalence (3.19), followed by the United States (3.09), Italy (3.08) and Australia (3.05). The MIR spanned from 0.22 in the United States to 0.54 in Indonesia, meaning Indonesia combines moderate incidence with the highest case-fatality proxy and only fourth place in absolute case count, a profile the authors read as consistent with limited diagnostic capacity and later presentation.

The study’s central paradox emerges around development level. Across the four global HDI aggregates, the burden clearly followed development: incidence was highest in the very-high-HDI group while mortality was highest in the low-HDI group, where MIR peaked at 0.53 against 0.34 in the very-high-HDI group, and PMR varied 3.9-fold, from 2.04 in the low-HDI group to 8.05 in the high-HDI group. Yet when the same relationship was tested across the 19 individual member economies, the country-level HDI–MIR correlation was null (r = −0.04, 95 percent CI −0.48 to 0.42; P = 0.88), and it remained non-significant with the sign reversed after a prespecified sensitivity analysis excluded four economies with reconstructed both-sexes counts. Meanwhile, ASIR rose significantly with HDI at country level (r = 0.50, P = 0.028), an association driven by the female analysis (r = 0.59, P = 0.007), and PIR was strongly correlated with HDI (r = 0.91, P < 0.001). The authors stress that the aggregate rows are global pools covering all 186 GLOBOCAN populations, not subsets of the G20, so an aggregate gradient cannot be read as a property of member economies.

Sex disparities are pervasive and consistent. Male age-standardized incidence exceeded female incidence in all 19 economies, with the widest absolute gaps in Indonesia (1.05 versus 0.50 per 100,000), Canada (0.99 versus 0.76) and the United States (0.97 versus 0.74). The male excess held for mortality in 18 of 19 countries, most starkly in Indonesia (0.6 versus 0.2) and the Russian Federation (0.4 versus 0.1). The MIR behaved slightly differently: male MIR exceeded female MIR in 16 of the 17 economies where the ratio was computable, the sole exception being Türkiye, where female MIR was marginally higher, and the widest male excesses appeared in Saudi Arabia (0.48 versus 0.21) and the United Kingdom (0.54 versus 0.28). In absolute deaths, the male disadvantage was greatest in China (2,453 versus 1,282), India (1,359 versus 674) and Indonesia (854 versus 424). Male and female MIR could not be computed for Australia and Brazil because the export does not publish both required sex-specific cells.

Age structure dominates every outcome. Age-standardized incidence rose roughly 70-fold from the 0–19-year group to the ≥75-year group (0.05 to 3.6 per 100,000), mortality rose from 0.01 to 2.0, and age-specific 5-year prevalence climbed from 0.1 to 10.5. Of the 20,707 world all-ages deaths in 2024, a global total rather than the G20 subtotal, the 40–64, 65–74 and ≥75-year groups accounted for 38.0 percent, 24.5 percent and 31.4 percent respectively. The older-age share varied enormously by economy: persons aged 59 years and above represented 78 percent of cases in Germany and 73 percent in Canada and Italy, against just 49 percent in South Africa, 39 percent in India and 26 percent in Saudi Arabia. In Canada the incidence curve rises steeply and monotonically to a peak in the 85-plus band, while Indonesia’s mortality curve is bimodal, with peaks at 55–59 and 65–69 years, and the male curve lies above the female curve across most of the age range.

The projections to 2050 are where the analysis turns genuinely alarming. Holding 2024 age-specific rates constant under the GLOBOCAN Cancer Tomorrow default demographic scenario, which applies 2024 rates to United Nations population projections and captures population growth and aging only, deaths at ages 75 and above would rise by 152.8 percent by 2050, from 6,499 to 16,429 worldwide. Growth would be largest in the low- and medium-HDI aggregates (+134.4 percent and +96.1 percent) and smallest in Japan (+8.4 percent). Country-level mortality increases were steepest for Mexico (+121.8 percent) and Brazil (+105.1 percent), with Australia and Saudi Arabia both at +100.0 percent. Projected incidence by 2050 reaches 13,568 cases in China and 11,413 in India, followed by the United States (7,735), Indonesia (3,968), Brazil (2,704) and Japan (1,823), and Japan is the only economy projected to decline in incidence (−4.7 percent). A separate annual-percentage-change scenario analysis for the six highest-incidence economies, run from −3 to +3 percent in 1 percent steps, offers sensitivity envelopes rather than forecasts. Notably, the projected percentage increase in deaths was inversely associated with baseline mortality (r = −0.51, P = 0.044), meaning the settings with the lowest current burden face the steepest relative growth.

Interpreting these ratios demands caution, and the authors are explicit about the limits. Because the design is ecological, with the country as the unit of analysis, a difference between national MIR values licenses no inference about any individual patient’s survival probability, and the ratios inherit whatever error the underlying registry and modelling steps carry. Three non-mutually-exclusive mechanisms could explain elevated MIR in lower-resource settings: diagnostic delay and late presentation shifting the stage distribution toward advanced disease; restricted access to the specialized surgical resection, adjuvant radiotherapy and systemic therapy that salivary gland cancer primarily requires, with services concentrated in a small number of centers; and registry under-ascertainment, where missed incident cases inflate the calculated ratio while deaths are captured through vital registration. A fourth explanation, variation in histologic mix across populations, cannot be separated using top-site C07–C08 estimates. Within-country evidence does support the care-pathway reading: US cohorts link insurance status to stage and survival with racial disparities, London data show ethnicity-specific incidence differences, and German and US cohorts identified socioeconomic disadvantage and comorbidity as adverse prognostic factors.

The policy implications are structural rather than opportunistic. Rare tumors serve as an informative stress test for health systems, because small patient numbers mean the fixed costs of maintaining pathology, imaging and specialist surgical capacity cannot be distributed across a large volume, so a system that delivers good care for common cancers may still underperform for rare ones. Three interventions follow directly: expanding radiotherapy and surgical oncology capacity in lower-resource G20 economies, since resection completeness and adjuvant radiotherapy delivery both substantially influence outcomes; strengthening pathology and diagnostic infrastructure, including immunohistochemistry and molecular testing for rarer histologic subtypes, which also enriches registry data; and establishing multidisciplinary head-and-neck tumour boards with supporting referral pathways at relatively low cost. Measurement is a constraint in its own right: population-based cancer registration covers roughly one-quarter to one-third of the global population, with coverage lowest precisely where under-ascertainment is most severe, particularly across Africa and Asia. The authors also caution that steepest relative growth in Mexico and Brazil should not justify moving resources away from the largest absolute contributors, China and India, and that these demographic-only projections are descriptive modelling results, not predictions of what will occur or guidance for any individual patient’s care.

Subject of Research: Global burden, geographic and sex disparities, and projections to 2050 for salivary gland cancer across G20 countries: a GLOBOCAN 2024 analysis

Article Title: Global burden, geographic and sex disparities, and projections to 2050 for salivary gland cancer across G20 countries: a GLOBOCAN 2024 analysis

Article References: Yang, W., Ma, Y., & Shu, C. (2026). Global burden, geographic and sex disparities, and projections to 2050 for salivary gland cancer across G20 countries: a GLOBOCAN 2024 analysis. Frontiers in Oncology, 16, Article 1995714. https://doi.org/10.3389/fonc.2026.1995714

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1995714

Keywords: Global, burden, geographic, disparities, projections, salivary, gland, cancer, across, countries, GLOBOCAN, analysis