Ethnic Background as a Determinant of Thyroid Cancer Risk: A Decade-Long National Cohort Study Reveals Divergent Epidemics in Saudi Arabia

Summary

Objectives:

To analyze and compare temporal trends, demographic characteristics, and stage at diagnosis for thyroid cancer (TC) between Saudi and non-Saudi populations in order to identify potential population-specific etiological drivers.

Methods:

We conducted a national retrospective cohort study of TC cases in Saudi Arabia from 2013–2022. To assess population-level trends, we computed age-standardized incidence rates (ASIRs) and multivariable Poisson regression alongside piecewise regression models for comparative analysis.

Results:

Among 14,159 total cases, TC incidence increased disproportionately in Saudi nationals, rising from 884 to 1,677 (90%), significantly outpacing the 184 to 275 (49%) increase observed in non-Saudis. The incidence rate ratio for Saudis versus non-Saudis was 5.67 (95% CI [5.41, 5.94]). Saudis presented with a significantly younger median age at diagnosis (5–10 years earlier, p < 0.001) and demonstrated a steeper annual increase in ASIRs (Interaction β = 0.0307, p < 0.001). A marked stage migration was concurrently observed among Saudi nationals, with localized diagnoses increasing from 55% to 71%. The recent period of the study (post-pandemic) revealed an acceleration in the Saudi cohort (+316.50 cases annually), a pattern absent in non-Saudis.

Conclusion:

The profound divergence in incidence trends provides compelling evidence consistent with distinct, population-specific etiological drivers. This suggests a need to investigate genetic and environmental factors unique to the Saudi population. However, the heterogeneity of the non-Saudi comparator group necessitates cautious interpretation of the observed incidence rates.

Introduction

Thyroid cancer (TC) has emerged as the most prevalent endocrine malignancy worldwide, with incidence rates demonstrating a steep, sustained upward trajectory over the past several decades [1,2,3]. This epidemiological shift is frequently ascribed to a complex synergy of heightened diagnostic sensitivity and genuine increases in TC disease burden, influenced by environmental and genetic determinants. While often characterized by a favorable prognosis, this escalating incidence imposes a substantial clinical and economic burden on healthcare systems globally. The dominant hypothesis largely attributes this observed ‘epidemic’ to enhanced diagnostic sensitivity, specifically citing the widespread use of neck ultrasonography and fine-needle aspiration. This has led to the detection of indolent subclinical microcarcinomas, a phenomenon commonly termed as overdiagnosis [4,5]. However, solely focusing on overdiagnosis may be reductive given the observed population-specific trends. A complex interplay of genuine risk factors, including environmental exposures, lifestyle changes, and genetic predispositions, is also believed to contribute to the rising disease burden, though their relative contributions remain poorly defined [6,7].

Saudi Arabia's singular demographic and environmental context offers a unique opportunity to disentangle this complex etiology. The nation has undergone rapid socioeconomic transformation and extensive healthcare expansion, which has led to a pronounced increase in diagnostic capacity and a corresponding shift in disease detection patterns. More critically, its population structure is uniquely bifurcated: a stable, young national population coexists with a large, transient, and heterogeneous non-national resident population. These 2 populations, while sharing the same geographic space and being subjected to similar contemporary diagnostic practices, differ profoundly in their genetic ancestry, early-life environmental exposures, and lifestyle risk factors. This divergence establishes a powerful, real-world comparative setting. If rising incidence were solely attributable to enhanced diagnostic intensity, parallel trends would be expected across both cohorts. Conversely, divergent trends provide compelling evidence for the operation of population-specific etiological drivers.

Within Saudi Arabia, prior investigations have indeed documented an increasing incidence of thyroid malignancy, particularly noting a strong female predominance. However, these studies have been constrained by limited temporal scope or, crucially, have not leveraged the nation's demographic structure to conduct a comparative analysis. The absence of a longitudinal, population-based study that rigorously compares incidence trends, demographic patterns, and staging characteristics between national and non-national residents constitutes a critical gap in the literature. Determining whether rising TC incidence represents a uniform phenomenon or is disproportionately driven by specific subpopulations is essential for developing targeted public health strategies and generating novel etiological hypotheses.

To address this critical gap, we conducted a comprehensive national cohort analysis of all TC cases reported in Saudi Arabia from 2013 to 2022. We hypothesized that incidence trends, demographic profiles, and patterns of stage at diagnosis would diverge significantly between Saudi nationals and non-Saudi residents, thereby indicating distinct, population-specific etiological drivers. Our specific objectives were to: i) characterize and compare temporal trends in incidence and age-standardized incidence rates (ASIRs) between the 2 cohorts, identifying significant temporal breakpoints; ii) investigate disparities in demographic patterns, including gender distribution and median age at diagnosis, and stage at diagnosis; iii) determine, using multivariable regression analysis, whether the temporal trends in incidence are statistically divergent between populations; and iv) generate a short-term forecast of the expected case burden to inform healthcare resource planning and public health policy.

Method

A national retrospective population-based cohort study of all incident TC cases in Saudi Arabia reported to the Saudi Cancer Registry (SCR) between January 2013 and December 2022 was conducted. The SCR is a comprehensive nationwide surveillance system that collects data on all newly diagnosed malignant tumors from healthcare facilities across Saudi Arabia, achieving high coverage and data completeness. The aggregate data used in this study are publicly available through the SCR's official annual reports hosted by the Saudi Health Council. This study leverages the unique demographic contrast between the stable Saudi national population, which has universal access to government healthcare services, and the transient, heterogeneous non-Saudi resident population. A key limitation of this comparative design is that the ‘non-Saudi’ cohort is highly heterogeneous, encompassing diverse nationalities with differing genetic backgrounds, risk profiles, and durations of residence, which is treated as a single group due to data availability constraints.

The study population included all individuals, regardless of nationality, with a histologically confirmed primary malignant thyroid carcinoma recorded in the SCR during the 10-year study period. Cases were identified using the International Classification of Diseases for Oncology, 3^rd Edition (ICD-O-3) code C73.9. We excluded benign tumors, non-thyroid malignancies, and cases with missing or invalid data for nationality, gender, or year of diagnosis.

The primary outcome was the incidence of TC, measured as both the annual frequency of cases and the ASIRs. The ASIRs were calculated per 100,000 person-years using the Segi world standard population to allow for international comparisons. Population denominator data for calculating ASIRs was obtained from the official annual population statistics published by the Saudi General Authority for Statistics. The primary explanatory variable was nationality, categorized as ‘Saudi’ or ‘Non-Saudi’. This variable serves as a proxy for divergent genetic backgrounds, early-life exposures, and potentially lifestyle factors. Other variables included: i) Demographic factors: gender (male/female) and age at diagnosis (continuous and categorized for ASIR calculation). ii) Temporal factor: calendar year of diagnosis (2013-2022). iii) Tumor characteristic: SEER summary stage at diagnosis, categorized as localized (confined to the primary site), regional (extended to regional lymph nodes or adjacent organs), distant (metastatic disease), and unknown (missing or incomplete data). The significant reduction in the “unknown category” over the study period was specifically monitored as a proxy for improved diagnostic completeness.

The study utilized de-identified aggregated public health surveillance data that is routinely collected for public health purposes, along with publicly available population statistics. As such, the data did not contain any personally identifiable information. Consistent with standard public health research practices for de-identified secondary data analysis, this study was determined to be exempt from formal ethical review by an institutional review board, as it did not involve human subjects research as defined by relevant ethical guidelines.

Statistical analysis

A suite of statistical analyses was employed to characterize trends and test for disparities. I) Descriptive statistics: we computed annual case frequencies, percentages, and medians with interquartile ranges (IQR). The Compound Annual Growth Rate (CAGR) was calculated to summarize the average annual growth over the decade. All changes over the study period 2013 to 2022 were quantified using 2 metrics: absolute change, defined as the difference in percentage points between the final and initial values, and relative change (or percentage increase/decrease), calculated as the absolute change divided by the initial value and multiplied by 100%. These metrics were used to summarize decadal changes in incidence rates, case counts, and stage distribution. II) Incidence rates and trends: ASIRs were computed by direct standardization. Temporal trends in annual case counts and ASIRs were initially assessed using linear regression models, with the coefficient of determination (R²) and p-value reported. III) Comparative analysis: the Mann-Whitney U test was used to compare non-normally distributed continuous variables such as age at diagnosis between groups. The Chi-square test of independence was used for categorical variable associations such as gender proportion and stage distribution. Spearman's rank correlation (ρ) was used to assess the relationship between the proportions of localized and unknown-stage cases over time. IV) Modeling incidence trend divergence: to formally test for statistical divergence in the temporal trends between the Saudi and non-Saudi populations, we employed a multivariable Poisson regression model. This model analyzed the annual number of cases (the count data), Y_it, for each nationality i in year t. The model specified the log of the expected annual case count, E[Y_it], as the outcome variable: log(E[Y_it] = β ₀ + β₁.Nationality + β₂.Year_Scaled + β ₃.(Nationality × Year_Scaled predictors included nationality (coded as a binary factor), calendar year (treated as a continuous scaled variable), and the interaction term between nationality and year. This model did not include an offset term, as the objective was to test for divergence in the trends of absolute case counts between populations, rather than to model incidence rates. A statistically significant interaction term (β₃, with p < 0.05) indicates that the temporal trends in case counts were statistically different between the Saudi and non-Saudi cohorts. V) Identifying temporal breakpoints: we employed piecewise regression to identify significant changes in the slope of incidence trends over time. This analysis objectively identifies the years when the trend significantly accelerated or decelerated. VI) Stage migration analysis: to quantify the trend towards earlier stage at diagnosis, we performed a logistic regression analysis limited to cases where the stage was known (localized, regional, or distant). The outcome was a binary dependent variable: ‘late-Stage Diagnosis’ (coded as 1, encompassing regional and distant stages) versus ‘early-stage diagnosis’ (coded as 0, encompassing localized stage), with calendar year as the continuous predictor. Separately, the Chi-square test of independence was performed on the aggregated 2013 and 2022 stage proportions (including the unknown category) to formally assess the overall significance of the shift in stage distribution over the decade. This analysis was limited to the Saudi cohort, and the unavailability of staging data for the non-Saudi population precludes a crucial comparative analysis of disease severity at presentation. VII) Forecasting: an autoregressive integrated moving average (ARIMA) model was selected based on the time series characteristics of the Saudi national data to generate short-term forecasts for the expected number of new cases for 2023-2025. All analyses were performed using Python, version 3.9, with the pandas, statsmodels, scipy, and scikit-learn libraries. A 2-sided p-value <0.05 was considered statistically significant.

Results

In burden and divergent incidence trends, a total of 14,159 TC cases were recorded in Saudi Arabia from 2013 to 2022. Analysis revealed profoundly divergent epidemiological landscapes between national and non-national populations. The total number of cases among Saudis increased from 884 to 1,677 (90%) over the decade far exceeding the 184 to 275 (49%) growth observed in non-Saudis in Table 1, Fig. 1A.

Table 1.

Key epidemiological metrics for thyroid cancer in Saudi Arabia 2013-2022.

Variable	Population	2013	2014	2015	2016	2017	2018	2019	2020	2021	2022
Total Cases n	All Saudis	884	959	1,020	1,076	1,107	1,323	1,331	1,044	1,626	1,677
Percentage of All Cancers	All Saudis	7.60%	8.20%	8.50%	8.20%	8.00%	8.40%	8.20%	7.40%	9.10%	8.70%
Cases n	Saudi F	702	734	793	854	873	1,045	1,024	816	1,304	1,288
	Saudi M	182	225	227	222	234	278	307	228	322	389
Ranking (All Ages)	Saudi F	2nd	2nd	2nd	2nd	2nd	2nd	2nd	2nd	2nd	2nd
	Saudi M	12th	9th	8th	9th	8th	10th	9th	10th	8th	9th
ASIR (per 100,000)	Saudi F	8.1	7.8	8.4	8.9	9	10.6	10.3	8.1	17.1	16.5
	Saudi M	2.2	2.5	2.5	2.5	2.5	2.9	3.3	2.4	4.7	5.4
Median Age (Years)	Saudi F	40	40	39	41	40	40	40.5	42	42	43
	Saudi M	41	44	44	48	42	45	48	46	45	45
Female to Male Ratio	All Saudis	385:100	326:100	349:100	385:100	373:100	375:100	334:100	358:100	405:100	331:100
Total Cases n	All non-Saudis	184	182	223	214	223	246	205	171	283	275
Percentage of All Cancers	All non-Saudis	5.50%	5.20%	6.40%	5.80%	5.80%	6.00%	5.60%	3.30%	6.70%	5.70%
Cases n	Non-Saudi F	118	114	151	146	149	162	140	116	197	195
	Non-Saudi M	66	68	72	68	74	84	65	55	86	80
Ranking (All Ages)	Non-Saudi F	2nd	3rd	2nd	2nd	2nd	2nd	3rd	N/A	2nd	3rd
	Non-Saudi M	9th	9th	8th	9th	N/A	8th	9th	N/A	9th	N/A
ASIR (per 100,000)	Non-Saudi F	2.4	2.1	3.3	2.2	1.5	1.5	2.5	8.1	8.7	9.7
	Non-Saudi M	1.8	1.4	1.7	1.3	1.3	1.1	1	2.4	4.4	4.3
Median Age	Non-Saudi F	48	47	47	49	49	49	49	49	48	47
	Non-Saudi M	52	53	53	54	54	53	53	55	53	52

Saudi F: Saudi females; Saudi M: Saudi males; ASIR: Age-standardized incidence rates.

Fig. 1.

Comparative trends of thyroid cancer in Saudi versus non-Saudi populations 2013-2022. This figure provides a detailed, 6-panel comparison of thyroid cancer trends between the Saudi and non-Saudi populations. Panels A and B display the annual trends for the number of cases and the ASIRs. C and D illustrate the median age at diagnosis, stratified by gender. Panels E and F present a linear regression analysis for the total number of cases and ASIR, respectively, providing a visual representation of long-term trends.

This disparity was confirmed by ASIRs, with Saudi female ASIRs rising from 8.1 to 16.5 per 100,000 and male ASIRs from 2.2 to 5.4 per 100,000 (Table 1, Fig. 1B). Saudi nationals bore a disproportionately high burden, with 12,037 total cases compared to 2,122 among non-Saudis, a 5.7-fold difference (incidence rate ratio: 5.67). The Poisson regression model, which analyzed annual case counts without an offset to test for divergence in trends, estimated an annual increase of 6.76% (95% CI: 6.00-7.54%) for Saudis and 3.53% (95% CI: 2.02-5.02%) for non-Saudis (Table 2). The significant interaction term (β = 0.0307, 95% CI: 0.015-0.047, p < 0.001) formally confirmed that the temporal trend in case counts was statistically steeper in the Saudi population.

Table 2.

Advanced statistical analyses of thyroid cancer patterns 2013-2022.

Analysis Type	Comparison Group	Primary Result	95% CI	p-value	Effect Size/Model Fit
Demographic Comparisons	Saudi Females vs. Males	W = 5.5	-	0.0008*	rrb = 0.890
	Saudi vs. non-Saudi	W = 0.0	-	0.0002*	rrb = 0.920
	Stage Distribution 2013 vs. 2022	χ2 = 157.44	-	<0.001*	V = 0.248
Temporal Age Trends	Saudi Females	β = 0.33 years/year	[0.166, 0.494]	0.004*	R2 = 0.661
	Saudi Males	β = 0.339 years/year	[-0.119, 0.798]	0.185	R2 = 0.208
	Non-Saudi Females	β = 0.048 years/year	[-0.159, 0.256]	0.659	R2 = 0.026
	Non-Saudi Males	β = 0.036 years/year	[-0.172, 0.245]	0.742	R2 = 0.014
Stage Migration	Saudi Patients	OR = 0.9614	[0.948, 0.975]	<0.001*	-
**Poisson Regression Model for Case Count Trends	Intercept (non-Saudi reference)	β 0 = 5.391	[5.349, 5.433]	<0.001*	-
	Nationality (Saudi vs. Non-Saudi)	β 1 = 1.685	[1.639, 1.731]	<0.001*	-
	Year Effect (Non-Saudi trend)	β 2 = 0.0347	[0.020, 0.049]	<0.001*	-
	Year×Nationality Interaction	β3 = 0.031	[0.015, 0.047]	<0.001*	-
Derived Annual Percentage Increases	Saudi Population	6.76%	[6.00%, 7.54%]	<0.001*	-
	Non-Saudi Population	3.53%	[2.02%, 5.02%]	<0.001*	-

P < 0.05 indicates statistical significance; W: Mann-Whitney U test; χ²: Pearson's chi-square test; β: linear regression coefficient; OR: logistic regression odds ratio; r_rb: Rank-biserial correlation; V: Cramer's V; R²: Coefficient of determination; CI: Confidence interval. The Poisson regression model analyzed annual case counts without an offset to test for divergence in trends between populations.

In distinct demographic and clinical presentation patterns, stark differences in age at diagnosis characterized the 2 populations. Saudi patients were consistently diagnosed 5-10 years earlier than non-Saudis (p < 0.001), with Saudi females having a median age of 39-43 years versus 47-49 years for non-Saudi females (Table 3, Fig. 1C-D).

Table 3.

Comparative epidemiological characteristics of thyroid cancer in Saudi Arabia 2013-2022.

Metric	Saudi Population	Non-Saudi Population	Absolute Difference	Relative Difference	P-value
Cases
Total 2013-2022	12,037	2,122	+9,915	5.67x higher	-
Average Annual Cases	1,204	212	+992	5.68x higher	-
Percentage of All Cancers	8.20%	5.60%	+2.60%	46% higher	-
Gender Distribution
Female Cases n (%)	9,239 (76.8%)	1,489 (70.2%)	+7,750	6.21x higher	<0.001*
Male Cases n (%)	2,798 (23.2%)	632 (29.8%)	+2,166	4.43x higher	<0.001*
Female-to-Male Ratio†	330:100	236:100	+94	40% higher	<0.001*
Age at Diagnosis
Median Age - Female	43 years	47 years	-4 years	9% younger	<0.001‡
Median Age - Male	45 years	52 years	-7 years	16% younger	<0.001‡
Incidence Rates
ASIR (2022) - Female	16.5/100,000	9.7/100,000	+6.8	70% higher	-
ASIR (2022) - Male	5.4/100,000	4.3/100,000	+1.1	26% higher	-
Cancer Ranking
Female	2nd	2nd-3rd			-
Male	8th-12th	8th-9th			-
Decadal Change (2013-2022)
Total Cases Increase	+793 (90%)	+91 (49%)	+702	8.7x greater increase	<0.001**
Female ASIR Increase	+8.4 (104%)	+7.3 (304%) *	+1.1	-	-
Male ASIR Increase	+3.2 (145%)	+2.5 (139%)	+0.7	-	-
Stage Distribution Change (Saudi Only) ††
% Localized Stage	+15 pts	-	+15 pts	+28.2%	<0.001**
% Regional Stage	-3.4% pts	-	-3.4% pts	-12.6%	<0.001**
% Distant Stage	+0.1% pts	-	+0.1% pts	+2.5%	<0.001**
% Unknown Stage	-12.3% pts	-	-12.3% pts	-87.9%	<0.001**

Chi-square test for proportion comparison (χ² = 41.82, p < 0.001).

^‡Mann-Whitney U test for median age comparison. Multivariable Poisson regression interaction term (Year * Nationality). Female-to-male ratio presented as number of female cases per 100 male cases.

Stage Distribution change metrics reflect the shift in proportion from 2013 to 2022 among the Saudi population only. Non-Saudi data was unavailable for stage assessment. P-value is derived from the Chi-square test comparing the stage proportions in 2013 versus 2022, confirming a statistically significant stage migration over the study period. The very high percentage increase in Non-Saudi Female ASIR is influenced by a low baseline rate and potential fluctuations in the population denominator, particularly during the COVID-19 period, and should be interpreted with caution.

A significant stage migration was observed over the study period (χ² = 157.44, df = 3, p < 0.001). The proportion of localized-stage diagnoses increased markedly, while unknown-stage cases saw a near-complete reduction (Table 3). Logistic regression quantified this trend, showing each passing year was associated with a 3.86% reduction in the odds of late-stage diagnosis (OR = 0.9614, 95% CI: 0.948-0.975, p < 0.001; Table 2). Spearman's rank correlation confirmed a strong inverse relationship between localized and unknown-stage proportions (ρ = -0.795, p = 0.006), indicating major improvements in diagnostic precision and data completeness within the Saudi healthcare system. Critically, this analysis was necessarily restricted to the Saudi population due to the unavailability of staging data for non-Saudi cases, which precludes a comparative assessment of stage at diagnosis between the 2 groups.

In temporal dynamics and the post-pandemic surge, piecewise regression identified distinct temporal breakpoints, revealing fundamentally different epidemic dynamics (Table 4). The most striking divergence emerged in the post-2020 period. Among Saudis, case numbers surged at an unprecedented rate of 316.50 additional cases annually, a 4-fold acceleration over pre-pandemic growth rates (Fig. 1E). This dramatic rebound was not mirrored in the non-Saudi population, which exhibited different trajectory patterns throughout the decade, potentially reflecting the out-migration of a portion of this transient population during the pandemic, which may have artificially inflated subsequent incidence rates among those who remained.

Table 4.

Piecewise regression analysis of thyroid cancer incidence trends with statistical significance 2013-2022.

Dataset	Break-point 1	Break-point 2	Segment 1 Period	Slope (95% CI)	P-value	Segment 2 Period	Slope (95% CI)	P-value	Segment 3 Period	Slope (95% CI)	P-value	Model R2
Saudi Total Cases	2019	2020	2013-2019	77.00 (40.70, 113.30)	0.006**	2019-2020	-198.50 (-417.32, 20.32)	0.126	2020-2022	316.50 (180.67, 452.33)	0.004**	0.915
Saudi Female ASIR	2020	2021	2013-2020	0.23 (-0.06, 0.52)	0.17	2020-2021	7.40 (5.19, 9.61)	0.001***	2021-2022	-0.60 (-3.23, 2.03)	0.671	0.95
Saudi Male ASIR	2020	2021	2013-2020	0.08 (-0.01, 0.17)	0.147	2020-2021	1.82 (1.11, 2.54)	0.002**	2021-2022	0.70 (-0.15, 1.55)	0.157	0.947
Non-Saudi Total Cases	2018	2020	2013-2018	11.49 (1.34, 21.65)	0.068	2018-2020	-26.32 (-51.76, -0.88)	0.089	2020-2022	54.17 (24.57, 83.78)	0.012**	0.769
Non-Saudi Female ASIR	2019	2020	2013-2019	-0.10 (-0.32, 0.12)	0.421	2019-2020	6.11 (4.79, 7.43)	0.001***	2020-2022	0.80 (-0.02, 1.62)	0.104	0.978
Non-Saudi Male ASIR	2019	2021	2013-2019	-0.13 (-0.19, -0.07)	0.005**	2019-2021	1.67 (1.47, 1.86)	0.001***	2021-2022	0.01 (-0.43, 0.45)	0.962	0.988

^*p < 0.01,

^**p < 0.05, and

^*** p < 0.001 are considered statistically significance.

The ARIMA (1,1,1) model forecasted a potential plateau in annual new Saudi cases for 2023–2025 (point estimates ~1,677 cases/year; Fig. 2). Critically, the 95% prediction intervals were substantial, ranging from 889 to 2,463 cases by 2025. This wide range underscores the considerable uncertainty in the forecast and highlights the necessity for flexible public health planning that can accommodate this potential variability.

Fig. 2.

Autoregressive integrated moving average model forecast for Saudi thyroid cancer cases 2023–2025. ARIMA: Autoregressive integrated moving average. The blue line represents historical data, the red dashed line represents the point forecast, and the shaded red area represents the 95% confidence interval. The forecast suggests a plateau in incidence but with high uncertainty in later years.

Discussion

This comprehensive national study reveals a rapidly evolving TC landscape in Saudi Arabia (2013-2022), characterized by 4 principal findings: i) a dramatic 90% incidence surge among Saudi nationals far exceeding the 49% increase in non-Saudis; ii) persistent female predominance alongside unexpectedly rapid ASIR acceleration in Saudi males (CAGR = 10.49%); iii) substantial staging improvements with localized diagnoses rising from 55% to 71%; iv) and most significantly, stark divergence in epidemiological trends suggesting population-specific etiological drivers.

The magnitude of increase among Saudis, equating to a significant 6.76% annual rise, aligns with global TC detection epidemics largely attributed to heightened diagnostic scrutiny [8,9]. This study, however, challenges a purely diagnostic explanation; the persistent female-to-male ratio of 3.3:1 within the Saudi cohort notably exceeds those reported in Western epidemiological studies, thereby strongly implicating the involvement of non-diagnostic population-specific risk factors [10]. This distinctive national profile is corroborated by a recent regional study from Aljouf, where Saudis constituted 91.2% of thyroid pathology cases, reporting a nearly identical female-to-male ratio (3.3:1) and identifying papillary carcinoma (75.8%) as the dominant histologic subtype [11]. The profound and statistically significant disparity between national and non-national residents, who share the same geographic space and healthcare system, suggests that factors tied to ethnic background and ancestral determinants, rather than diagnostic intensity alone, are required to explain the divergence. The rapid acceleration among Saudi males despite lower baseline risk warrants investigation into gender-specific environmental or hormonal exposures.

Coronavirus disease 2019 (COVID-19) pandemic had a clear and complex impact on TC trends, characterized by an initial sharp decline in diagnoses in 2020 due to disruptions in non-essential healthcare services, followed by a substantial rebound. Our findings of a dramatic post-2020 surge in Saudi cases are consistent with global patterns of delayed diagnosis during the pandemic waves. A recent longitudinal study of US SEER data found that TC incidence declined markedly during COVID-19 variant waves, with an estimated 10,200 cases remaining undiagnosed by the end of 2021, potentially creating a future pool of more advanced presentations [12]. This aligns with evidence from China showing that post-lockdown TC patients presented with significantly more aggressive features, including higher rates of lymph node metastasis and extra thyroidal extension [13].

In our cohort, the steeper post-2020 acceleration among Saudis is therefore most parsimoniously explained by this large-scale catch-up of delayed diagnoses, compounded by possible increases in diagnostic capacity. The intensification of underlying population-specific risk factors remains a secondary, non-exclusive hypothesis. Crucially, this dramatic rebound was not mirrored in the non-Saudi population, where trends were further complicated by pandemic-era demographic shifts. This differential recovery pattern underscores how pre-existing disparities in healthcare system integration can be amplified during public health crises, potentially exacerbating long-term cancer outcome disparities between population groups.

Marked staging improvements reflect substantial diagnostic advancements, with the strong inverse correlation between localized and unknown-stage cases (ρ = -0.795) indicating major data quality progress. This “stage migration” signifies a positive public health development, as earlier detection associates with superior outcomes [14]. The non-Saudi cohort provides a crucial epidemiological counterpoint. Despite significant ASIR increases, their absolute burden remained 5.7-fold lower, with a consistently later diagnosis (approximately 5-10 years older). This pattern, observed within the same national borders and healthcare environment, is unlikely to be explained by healthcare access differentials alone. It may reflect the ‘healthy migrant effect’, distinct genetic/environmental exposures, or delayed adoption of nation-specific risk factors. Most critically, our study design and advanced regression modeling formally confirmed significantly steeper incidence increases among Saudis (interaction p < 0.001). The post-pandemic acceleration phase saw unprecedented surges, Saudi total cases increased by 316.50 annually, unmirrored in non-Saudis. This robust statistical divergence, within our natural experiment framework, suggests the epidemic drivers are potentiated by factors endemic to the national population.

The robust statistical divergence confirmed by our regression modeling, which separates the Saudi and non-Saudi cohorts, strongly supports the presence of distinct disease etiologies. While our analysis is rooted in the administrative variable of nationality (Saudi versus non-Saudi) as recorded by the national registry, we interpret the observed divergent trends as reflective of differences rooted in ethnic background and associated risk factors. Nationality, in this natural experiment framework, serves as a crucial proxy measure for ancestral genetic profiles, region-specific environmental exposures such as historical iodine intake patterns in countries of origin, and varying pre-migration screening histories. Therefore, the significant difference in the acceleration of TC incidence is unlikely to be solely due to local differences in diagnostic access but rather points to underlying ethnic determinants of TC susceptibility and burden.

We hypothesize several non-mutually exclusive explanations for this divergence that warrant future investigation. First, genetic predispositions, including susceptibility gene polymorphisms prevalent in Middle Eastern populations or v-raf murine sarcoma viral oncogene homolog B (BRAF) V600E variation, may be explored as potential risk factors [15,16,17,18]. Second, unique environmental/lifestyle factors accompanying rapid modernization may disproportionately affect Saudis [19,20]. Third, and critically, disparities in health-seeking behavior and cultural awareness may also contribute to the observed differences. The Saudi population, with universal access to the healthcare system and increasing health awareness driven by national public health campaigns may be more likely to undergo screening and present earlier with symptomatic or even subclinical disease. This is supported by recent national surveys indicating that engagement with the healthcare system such as undergoing thyroid function tests or ultrasound is strongly associated with higher TC awareness among Saudis, and that knowledge, while variable, is actively developing within the population, particularly among educated demographics and those with medical connections [17,21]. In contrast, the non-Saudi cohort, which includes a large transient workforce, might face barriers such as language, health insurance complexities, or a focus on acute rather than preventive care, potentially delaying diagnosis. Future research should integrate qualitative methods to directly explore these cultural and behavioral determinants, providing deeper insights into the drivers of healthcare utilization and diagnostic timing between these populations.

Study limitations

Our registry-based analysis, while offering significant national-level insights, is constrained by several limitations that necessitate cautious interpretation of the findings. The primary constraint lies in the absence of crucial individual-level risk factor data such as radiation exposure history, genetic predisposition, and iodine status within the registry. This fundamental gap precludes a direct quantitative assessment of their contributions to TC incidence and consequently restricts the scope of mechanism-driven etiological inferences that can be credibly derived from this study. Furthermore, the unavailability of cancer stage data for the non-Saudi cohort presents a significant analytical barrier, preventing direct comparisons of disease progression between the nationality groups and thereby restricting detailed etiological inferences concerning tumor staging and survival to the Saudi population exclusively.

The analytical interpretation of the non-Saudi cohort is subject to substantial methodological constraints rooted in its complex composition. A principal limitation is the inherent heterogeneity of this population, an aggregate of individuals from numerous countries across distinct geographic and epidemiological regions such as Asia, Africa, and the West. Pooling these diverse source populations likely dilutes specific disease risk signals and compromises the precision required for subgroup-specific comparisons. A further challenge is the absence of granular surveillance data, specifically on country of origin, duration of residence, or ethnic background, which precludes a nuanced epidemiological assessment of risk mechanisms within this transient group. Finally, the cohort's inherent transience, compounded by a measurable pandemic-era exodus, likely led to an artificial inflation of the ASIR in the latter study years. This artifact occurs when the denominator (the total estimated population) decreases disproportionately relative to the number of incident cancer cases, thereby necessitating rigorous caution in interpreting recent temporal trends for this cohort.

In summary, our national study documents a rapidly escalating burden of TC in Saudi Arabia (2013–2022), characterized by a 90% increase in cases among Saudi nationals—a surge far exceeding the 49% growth observed in the non-Saudi resident population. The stark divergence in incidence trends, age distribution, and temporal dynamics between these populations, empirically confirmed through advanced regression modeling, is strongly consistent with the presence of distinct population-specific etiological drivers. Firm mechanism-driven conclusions are precluded by critical limitations, notably the lack of individual-level data on key risk factors such as genetic predisposition, radiation exposure, and iodine status and the aggregation of a highly heterogeneous non-Saudi population into a single comparator group. Nonetheless, the findings strongly suggest that factors beyond differential diagnostic intensity alone are propelling the escalating incidence among Saudi nationals. This evidence warrants immediate public health planning for the growing clinical burden and mandates targeted etiological investigation. Therefore, future research must adopt a multidisciplinary approach to investigate genetic predispositions such as (BRAF), rearranged during transfection (RET) mutations, environmental exposures, and modifiable lifestyle factors specific to the Saudi population. This will require designing targeted prospective studies and fostering cross-disciplinary collaborations to move beyond documenting epidemiological trends and toward definitively elucidating the drivers of this national health challenge.

Funding information

This study was funded by the Deanship of Graduate Studies and Scientific Research, Jazan University, Saudi Arabia, through Project number: JU- 202503321 -DGSSR- RP -2025.

Acknowledgmemnt

The author expresses sincere appreciation to LANGVARA (https://langvara.com/en/) for English language editing.

Disclosure statement

During the preparation of this manuscript, the authors used Google Gemini solely to improve the readability and language of the text. All AI-generated suggestions were critically reviewed, edited, and approved by the authors, who take full responsibility for the final content. This work has not been presented at any conference.

Disclosure

The authors have no conflicts of interest to declare. We confirm that this article has not been previously published in any preprint server. This study was funded by the Deanship of Graduate Studies and Scientific Research, Jazan University, Saudi Arabia, through Project number: JU- 202503321 -DGSSR- RP -2025.