Incidence patterns and trends of chordoma in the United States: a population-based analysis of over 6,000 cases Authors: Hyuna Sung, Chenxi Jiang, Hela Koka, Jiwei Bai, Alisa M Goldstein, Xiaohong R Yang Published: 2026 ABSTRACT AbstractIntroductionChordoma is a rare malignant bone tumor arising from notochordal remnants, with unknown etiology and limited descriptive characterization.MethodsData from the North American Association of Central Cancer Registries, covering 93% of the U.S. population, were used to estimate age-standardized incidence rates and trends from 2003 to 2022 by demographic and clinical factors. Trends were quantified using Joinpoint regression. Data from 21 Surveillance, Epidemiology, and End Results (SEER) registries were used to estimate relative survival for cases diagnosed from 2003 to 2021.ResultsA total of 6260 chordoma cases were diagnosed during 2003–2022. Cranial tumors were most frequent (38.7%), followed by sacral (32.7%) and spinal (25.4%) sites. The age-standardized incidence rate was 0.097 per 100,000, increasing with age and peaking at 80–84 years. Overall, incidence rates were higher in males than females (0.1173 versus 0.0800), with the divergence emerging after age 35. Incidence rates were comparable across racial and ethnic populations, except among non-Hispanic Black individuals, whose rates were 56% lower than those of non-Hispanic White individuals (95% CI = 0.39 to 0.49). Incidence rates were highest in the Northeast and metropolitan areas. From 2003 to 2022, incidence rose 1.47% annually (95% CI = 0.61–2.39), with steeper increases among younger individuals, women, and localized/cranial tumors. Among 2170 SEER cases, 5-year relative survival was 83.3%, varying by race/ethnicity and region. For distant-stage disease, survival ranged from 68.6% for cranial tumors to 38.3% for sacral tumors.ConclusionChordoma incidence and survival vary by demographic and clinical factors, suggesting differences in detection, healthcare access, and underlying etiology.Keywords: Chordoma, Tumor subtype, Incidence, Trend, Survival, Population-based cancer registry Introduction Chordoma is a rare malignant bone tumor with an incidence rate of less than 0.1 per 100,000 people in most parts of the world [1]. It is thought to arise from remnants of the embryonic notochord. During embryogenesis, the notochord serves as the main axial structure that induces formation of the neural tube and axial skeleton. Although the notochord normally regresses at birth, persistent remnants can give rise to chordoma [2]. Although chordoma accounts for only ∼3% of primary bone tumors, it is clinically significant because of its locally aggressive nature, high recurrence rate, and the complexity of its management. The tumor most frequently occurs along the axial skeleton, particularly at the skull base, spine, and sacrum, where its proximity to critical neurovascular structures makes complete resection challenging. Despite advances in imaging, pathology, and molecular characterization, the etiology of chordoma remains poorly understood. Population-based studies of chordoma are limited due to its rarity. For example, a recent study in Canada identified only 208 patients in the Ontario Cancer Registry between 2003 and 2019 [3], while a nationwide study in the Netherlands included 420 patients diagnosed between 1991 and 2020 [4]. The largest published study to date used United States Cancer Statistics (USCS) data and reported fewer than 4000 patients diagnosed between 2001 and 2014 in the United States [5]. These investigations provided important insights into demographic and clinical features but were constrained by modest sample sizes. Consequently, key questions remain regarding population-level incidence patterns, variation by demographic, socioeconomic, and geographic factors, and differences in incidence and survival by tumor location. To address these gaps, we conducted the largest descriptive epidemiologic study of chordoma to date, using the most up-to-date high-quality population-based cancer registry data from the North American Association of Central Cancer Registries (NAACCR). This dataset includes over 6000 pathologically confirmed chordoma cases—nearly one and a half times the number of cases in a prior study, allowing evaluation of incidence across diverse demographic and geographic strata. In addition, we compared clinical features and survival across tumor sites (skull base, spine, and sacrum) and assessed temporal trends in incidence by age, sex, stage, and tumor location. Methods Data sources and study population For incidence burden and trend analyses, we used the population-based cancer registry data from the Cancer in North America (CiNA) Dataset for Researcher produced by the North American Association of Central Cancer Registries (NAACCR) [6]. We identified individuals diagnosed with microscopically confirmed invasive chordoma (International Classification of Diseases for Oncology, 3rd edition (ICD-O-3), histology codes 9370/3, 9371/3, 9372/3) between 2003 and 2022 in 46 states and the District of Columbia that cover approximately 93% of the United States (U.S.) population. Arkansas, Indiana, Mississippi, and New Mexico were excluded because multiple diagnosis years within the study period did not meet the data quality standards [7]. Individuals of all ages were included. For relative survival analysis, we used data from the 21 Surveillance, Epidemiology, and End Results (SEER) registries, covering approximately 46% of the U.S. population [8]. Individuals diagnosed at age 15 and older from 2003 to 2021 were included with vital status followed through 2022. Only microscopically confirmed cases were included and those with unknown follow-up status were excluded. From both data sources, all races and ethnicities were included. Race and ethnicity were categorized as Hispanic, non-Hispanic American Indian or Alaska Native (AIAN), non-Hispanic Asian or Pacific Islander, non-Hispanic Black, non-Hispanic White, and unknown. To reduce racial misclassification for AIAN, incidence estimates were restricted to counties within Purchased/Referred Care Delivery Area. Primary tumor sites were categorized as cranial, sacral, spinal, or other based on ICD-O-3 SEER Site/Histology codes (Table S1). The study used de-identified publicly available data, which is considered non-human subjects research under the Common Rule (US federal regulation 45 CFR §46); thus, institutional review board approval and informed consent were not needed. Statistical analysis For cases diagnosed in NAACCR, age-standardized incidence rates (ASIRs; per 100,000 person-years, standardized to the 2000 U.S. population, 20 age groups) were calculated overall and by primary tumor sites for age at diagnosis (0–14, 15–39,40–64, 65+), diagnosis period (2003–2007, 2008–2012, 2013–2017, 2018–2022), race/ethnicity, stage (SEER Summary Stage: localized, regional, distant, and unknown), histology, geographic region (U.S. Census regions: Midwest, Northeast, South, and West), and rural/urban continuum code based on the counties of diagnosis address (Data Item #3312; developed by the United States Department of Agriculture using data from the Office of Management and Budget and the U.S. Census Bureau) [9]. Temporal trends in chordoma ASIRs were quantified using the National Cancer Institute (NCI)’s Joinpoint software (version 5.4.0.0), which uses a segmented line regression model to fit a series of joined straight lines to a logarithmic scale of the incidence rates. We estimated annual percent changes (APCs) and 20-year average annual percent changes (AAPCs) overall and by sex, age at diagnosis, race/ethnicity, primary tumor site, and stage at diagnosis. Rates in 2020 were excluded from trend analyses to account for the abrupt decline in cancer incidence reported during the COVID-19 pandemic, which could influence Joinpoint location and APC calculation [10]. As a complementary analysis, we applied age-period-cohort modeling to quantify differences in incidence rates by period of diagnosis and birth cohorts for cases diagnosed between ages 20 and 84, excluding younger individuals to minimize statistical uncertainties resulting from small case counts. Using the NCI's web tool for age-period-cohort analysis [11], we estimated period rate ratios and cohort rate ratios, which quantify the risk associated with a specific period or birth cohort relative to a reference group, while adjusting for the other factor (cohort effect for period rate ratios and period effect for cohort rate ratios) and age effects. We constructed 16 overlapping 10-year birth cohorts, referenced by mid-year of birth (1923 [1919–1927], 1928 [1924–1932]…,1993 [1989–1997],1998 [1994–2002]), by subtracting the mid-year age at diagnosis (20–24 [22] …, 80–84 [82]) from the mid-year period of diagnosis (2003–2007 [2005], …, 2018–2022 [2020]). The 1958 birth cohort, representing individuals born in 1954–1962, was selected as the reference, as it is the midpoint among the birth cohorts examined. The statistical significance of period and cohort deviations were determined by the Wald test [11]. Using SEER, we estimated age-standardized five-year relative survival for cases diagnosed from 2003 to 2021 and followed until 2022 (International Cancer Survival Standard 2- Age 15+) [12] by patient and tumor characteristics. Relative survival is a net survival measure representing cancer survival in the absence of other causes of death and is defined as the ratio of the proportion of observed survivors in a cohort of cancer patients to the proportion of expected survivors in a comparable set of cancer-free individuals [13]. Expected survival was estimated with the Ederer II method [14] and the Socioeconomic status, Geographic, and Race Annual Life Tables [15]. Relative survival was calculated with the actuarial method. All analyses, except Joinpoint and age-period-cohort modeling, were performed using the NCI's SEER*Stat (version 9.0.41.4). Two-sided p < 0.05 was considered statistically significant. Results Between 2003 and 2022, 6260 chordoma cases were diagnosed in 46 states and the District of Columbia (Table 1), corresponding to an age-standardized rate (ASR) of 0.0971 per 100,000. Men accounted for 57.4% of all cases; 40.6% were aged 40–64 years and 36.9% were ≥ 65 years. Approximately half of the cases were diagnosed at the local stage, followed by regional (34.9%) and distant (7%) stages. Consistent with the recent World Health Organization classification, which groups conventional and chondroid chordoma into a single histologic category, most cases were coded as chordoma NOS. Only 1.2% of cases were classified as dedifferentiated histology. Table 1Age-standardized incidence rates of chordoma and incidence rate ratios by demographic and clinical characteristics in the US, 2003–2022, the North American Association of Central Cancer Registries.TotalCase numberASR (SE)Incidence rate ratio (95% CI)Total, N62600.0971 (0.0012)Year of diagnosis2003–200719.7%0.086 (0.0025)Reference2008–201223.0%0.0925 (0.0025)1.08 (1–1.16)2013–201728.0%0.1055 (0.0026)1.23 (1.14–1.32)2018–202229.3%0.1032 (0.0025)1.2 (1.11–1.29) Age at diagnosis0–14 years3.8%0.0206 (0.0013)Reference15–39 years18.6%0.0583 (0.0017)2.82 (2.46–3.26)40–64 years40.6%0.1229 (0.0025)5.95 (5.21–6.82)65+ years36.9%0.2738 (0.0057)13.26 (11.61–15.21) SexMale57.4%0.1173 (0.002)ReferenceFemale42.6%0.08 (0.0016)0.68 (0.65–0.72) Race and ethnicityNon-Hispanic White73.4%0.1012 (0.0015)ReferenceNon-Hispanic Black5.2%0.0444 (0.0025)0.44 (0.39–0.49)Non-Hispanic American Indian and Alaska Native0.5%0.113 (0.0217)1.12 (0.74–1.62)Non-Hispanic Asian or Pacific Islander5.8%0.1089 (0.0058)1.08 (0.96–1.2)Hispanic, All Races14.1%0.1064 (0.0038)1.05 (0.97–1.14)Unknown Ethnicity1.8%0.2053 (0.0206)2.03 (1.65–2.48) Census RegionNortheast23.3%0.1171 (0.0031)ReferenceMidwest19.3%0.0913 (0.0027)0.78 (0.72–0.84)South33.0%0.0862 (0.0019)0.74 (0.69–0.79)West24.4%0.103 (0.0027)0.88 (0.82–0.95) Metro/non-metro (2013 Beale)Metropolitan area87.6%0.0996 (0.0014)ReferenceNon-metropolitan area12.4%0.0823 (0.0031)0.83 (0.76–0.89)Unknown0.0%–– StageLocalized49.6%0.0484 (0.0009)ReferenceRegional34.9%0.0337 (0.0007)0.7 (0.66–0.74)Distant7.0%0.0068 (0.0003)0.14 (0.13–0.16)Unknown8.5%0.0082 (0.0004)0.17 (0.15–0.19) Histology9370/3: Chordoma, not otherwise specified92.5%0.0898 (0.0012)Reference9371/3: Chondroid chordoma6.3%0.0061 (0.0003)0.07 (0.06–0.08)9372/3: Dedifferentiated chordoma1.2%0.0012 (0.0001)0.01 (0.01–0.02) Primary SitesCranial38.7%0.0389 (0.0008)ReferenceSacral32.7%0.0309 (0.0007)0.8 (0.75–0.84)Spinal25.4%0.0242 (0.0006)0.62 (0.58–0.66)Othera2.1%0.002 (0.0002)0.05 (0.04–0.06)aNot included in the subsequent analysis of site-specific rate calculations and survival analysis. Age-standardized incidence rates of chordoma and incidence rate ratios by demographic and clinical characteristics in the US, 2003–2022, the North American Association of Central Cancer Registries. Not included in the subsequent analysis of site-specific rate calculations and survival analysis. Incidence rates were similar across racial/ethnic groups (ASR = 0.10 to 0.11 per 100,000), except among non-Hispanic Black individuals, whose rate was 56% lower than that of non-Hispanic White individuals (ASR = 0.044 versus 0.11 per 100,000; 95% confidence interval (CI) = 0.39 to 0.49) (Table 1). Incidence was highest in the Northeast region (0.12 per 100,000) and lowest in the South (0.09 per 100,000), and higher in metropolitan areas compared to non-metropolitan areas (0.10 versus 0.082 per 100,000). Cranial chordoma was the most frequent site (38.7%, 0.039 per 100,000), followed by sacral (32.7%; 0.031 per 100,000) and spinal (25.4%, 0.0242 per 100,000) (Table 1). Demographic characteristics were generally similar across primary sites (Table 2); however, cranial cases were less likely to occur in older individuals (≥65 years: 22.1% vs. 50.1% for sacral and 42.7% for spinal) and showed less pronounced male (53.5% vs. 60.6% for sacral and 59.0% for spinal) and White (64.7% vs. 77.5% for sacral and 81.7% for spinal) predominance compared with sacral and spinal cases. Table 2Age-standardized incidence rates of site-specific chordoma and incidence rate ratios by demographic and clinical characteristics in the US, 2003–2022, the North American Association of Central Cancer Registries.CranialSacralSpinalCase numberASR (SE)Incidence rate ratio (95% CI)Case numberASR (SE)Incidence rate ratio (95% CI)Case numberASR (SE)Incidence rate ratio (95% CI)Total, N24250.0389 (0.0008)20490.0309 (0.0007)15870.0242 (0.0006)Year of diagnosis2003–200719.7%0.0334 (0.0015)Reference19.3%0.0274 (0.0014)Reference20.5%0.0228 (0.0013)Reference2008–201222.7%0.0364 (0.0016)1.09 (0.96–1.23)23.1%0.0298 (0.0014)1.08 (0.95–1.24)23.1%0.0233 (0.0012)1.02 (0.88–1.19)2013–201728.0%0.0428 (0.0017)1.28 (1.14–1.44)27.9%0.033 (0.0014)1.2 (1.06–1.37)27.9%0.0262 (0.0013)1.15 (0.99–1.34)2018–202229.6%0.0424 (0.0016)1.27 (1.13–1.43)29.7%0.0328 (0.0014)1.19 (1.05–1.36)28.5%0.0247 (0.0012)1.09 (0.94–1.26) Age at diagnosis0–14 years7.3%0.0152 (0.0011)Reference0.3%0.0006 (0.0002)Reference2.9%0.0039 (0.0006)Reference15–39 years28.6%0.0345 (0.0013)2.27 (1.93–2.7)9.3%0.0098 (0.0007)16.33 (7.76–41.17)15.4%0.0121 (0.0008)3.08 (2.24–4.32)40–64 years42.1%0.0509 (0.0016)3.35 (2.86–3.96)40.3%0.0389 (0.0014)65.08 (31.41–162.43)39.0%0.0293 (0.0012)7.48 (5.54–10.34)65+ years22.1%0.0624 (0.0027)4.11 (3.46–4.9)50.1%0.1228 (0.0039)205.69 (99.36–512.96)42.7%0.0806 (0.0031)20.58 (15.25–28.39) SexMale53.5%0.0426 (0.0012)Reference60.6%0.0405 (0.0012)Reference59.0%0.0305 (0.001)ReferenceFemale46.5%0.0355 (0.0011)0.83 (0.77–0.9)39.4%0.0229 (0.0008)0.57 (0.52–0.62)41.0%0.0191 (0.0008)0.63 (0.57–0.7) Race and ethnicityNon-Hispanic White64.7%0.0375 (0.001)Reference77.5%0.0328 (0.0008)Reference81.7%0.0277 (0.0008)ReferenceNon-Hispanic Black6.8%0.022 (0.0017)0.59 (0.49–0.69)5.0%0.0144 (0.0015)0.44 (0.35–0.54)2.7%0.006 (0.0009)0.22 (0.16–0.3)Non-Hispanic American Indian and Alaska Native0.6%0.0511 (0.0134)1.36 (0.75–2.29)0.6%0.0536 (0.016)1.63 (0.82–2.87)–––Non-Hispanic Asian or Pacific Islander6.8%0.0467 (0.0037)1.24 (1.05–1.46)6.6%0.0425 (0.0037)1.3 (1.07–1.55)3.4%0.0164 (0.0023)0.59 (0.44–0.78)Hispanic, All Races19.8%0.0528 (0.0025)1.41 (1.26–1.57)9.9%0.0281 (0.0021)0.86 (0.73–1)10.8%0.0219 (0.0018)0.79 (0.66–0.94)Unknown Ethnicity2.5%0.1051 (0.0143)2.8 (2.09–3.68)0.9%0.0404 (0.0098)1.23 (0.72–1.97)1.9%0.0534 (0.0105)1.93 (1.25–2.85) Census RegionNortheast23.1%0.047 (0.002)Reference24.1%0.0382 (0.0018)Reference23.3%0.0292 (0.0015)ReferenceMidwest19.1%0.0369 (0.0017)0.79 (0.69–0.89)20.4%0.0303 (0.0015)0.79 (0.69–0.91)18.7%0.0217 (0.0013)0.74 (0.63–0.87)South32.8%0.0342 (0.0012)0.73 (0.65–0.81)33.4%0.0278 (0.0011)0.73 (0.64–0.82)31.4%0.0206 (0.0009)0.71 (0.61–0.81)West25.0%0.0414 (0.0017)0.88 (0.78–0.99)22.2%0.0303 (0.0014)0.79 (0.69–0.9)26.5%0.0283 (0.0014)0.97 (0.84–1.12) Metro/non-metro (2013 Beale)Metropolitan area88.1%0.0398 (0.0009)Reference87.2%0.0318 (0.0008)Reference87.1%0.0249 (0.0007)ReferenceNon-metropolitan area11.8%0.0333 (0.0020)0.84 (0.73–0.95)12.8%0.0256 (0.0016)0.80 (0.70–0.92)12.9%0.0207 (0.0015)0.83 (0.71–0.97)Unknown0.1%––0.0%––0.0%–– StageLocalized53.2%0.0207 (0.0006)Reference43.3%0.0134 (0.0005)Reference53.1%0.0129 (0.0005)ReferenceRegional34.1%0.0133 (0.0005)0.64 (0.59–0.7)40.8%0.0125 (0.0004)0.93 (0.84–1.03)29.7%0.0072 (0.0003)0.55 (0.49–0.62)Distant5.3%0.0021 (0.0002)0.1 (0.08–0.12)8.6%0.0027 (0.0002)0.2 (0.17–0.24)6.8%0.0016 (0.0002)0.13 (0.1–0.15)Unknown7.3%0.0028 (0.0002)0.14 (0.11–0.16)7.3%0.0023 (0.0002)0.17 (0.14–0.2)10.3%0.0025 (0.0002)0.19 (0.16–0.23) Histology9370/3: Chordoma, not otherwise specified88.5%0.0344 (0.0008)Reference95.2%0.0295 (0.0007)Reference95.6%0.0232 (0.0006)Reference9371/3: Chondroid chordoma10.8%0.0041 (0.0003)0.12 (0.11–0.14)2.6%0.0008 (0.0001)0.03 (0.02–0.04)3.8%0.0009 (0.0001)0.04 (0.03–0.05)9372/3: Dedifferentiated chordoma0.7%0.0003 (0.0001)0.01 (0–0.01)2.1%0.0007 (0.0001)0.02 (0.02–0.03)0.6%0.0002 (0.0001)0.01 (0–0.01) Age-standardized incidence rates of site-specific chordoma and incidence rate ratios by demographic and clinical characteristics in the US, 2003–2022, the North American Association of Central Cancer Registries. Fig. 1 shows age-specific frequencies and incidence rates by sex and tumor site. Incidence rates increased with age, peaking at ages 80–84 years for males and both sexes combined versus ages 75–79 years for females. A divergence by sex appeared at ages 35–39, where incidence rates among men began to exceed those among women. Up to ages 55–59, cranial tumors predominated, after which sacral tumors had the highest rates. Spinal tumor rates also exceeded cranial tumor rates starting at ages 70–74. Age-specific incidence rates of chordoma by (A) sex and (B) primary tumor site in the U.S., 2003–2022, the North American Association of Central Cancer Registries. Fig. 2 shows incidence trends from 2003 to 2022 by demographic and tumor characteristics (APCs, 10-year AAPCs, and 20-year AAPCs in Table 3). Over the 20-year period, chordoma incidence rates increased by 1.47% per year (95% CI = 0.61 to 2.39) overall, by 1.36% per year (95% CI = 0.27 to 2.97) in men, and by 2.03% per year (95% CI = 1.04 to 3.15) in women (Table 3). Among men, the increase was only seen before 2010, after which the rate stabilized, whereas rates among women rose steadily. Although not statistically significant, increasing trends were observed in all groups except non-Hispanic Black individuals whose rates remained stable over the 20-year period (0.20% per year; 95% CI = -2.48 to 3.18). Increases were confined to younger individuals (<65 years; 20-year AAPC, 1.84% to 2.52% per year) and localized disease (3.46% per year). Although incidence rose across all primary sites, the increase was greatest for cranial chordomas (1.93% per year), followed by sacral (1.33% per year) and spinal (0.9% per year). Trends in incidence rates of chordoma and 20-year average annual percent changes (95% CI) by (A) sex, (B) age at diagnosis, (C) race and ethnicity, (D) primary tumor site, and (E) stage at diagnosis in the U.S., 2003–2022, the North American Association of Central Cancer Registries. Table 3Annual percent change (APC) and average annual percent change (AAPC) in incidence rates of chordoma in the NAACCR between 2003 and 2022.Segment 1APC (95% CI)Segment 2APC (95% CI)Segment 3APC (95% CI)Segment 4APC (95% CI)10-year AAPC, 2013–2022 (95% CI)20-year AAPC, 2003–2022 (95% CI)Total2003–20221.47 (0.61 to 2.39)1.47 (0.61 to 2.39)1.47 (0.61 to 2.39)SexMale2003–20094.58 (1.62 to 23.11)2009–2022−0.09 (−3.62 to 0.87)−0.09 (−2.02 to 0.89)1.36 (0.27 to 2.97)Female2003–20222.03 (1.04 to 3.15)2.03 (1.04 to 3.15)2.03 (1.04 to 3.15) Age at diagnosis0–14 years2003–2012−2.75 (−9.30 to 1.10)2012–201532.51 (−4.15 to 43.41)2015–2018−14.09 (−21.49 to 21.17)2018–20228.72 (−1.11 to 23.26)5.03 (1.85 to 7.47)2.52 (1.02 to 3.9)15–39 years2003–20221.89 (0.79 to 3.10)1.89 (0.79 to 3.1)1.89 (0.79 to 3.1)40–64 years2003–20221.84 (0.69 to 3.11)1.84 (0.69 to 3.11)1.84 (0.69 to 3.11)65+ years2003–20220.46 (−0.67 to 1.74)0.46 (−0.67 to 1.74)0.46 (−0.67 to 1.74) Race and ethnicityNon-Hispanic White2003–20162.32 (0.38 to 19.65)2016–2022−1.37 (−9.50 to 1.96)−0.16 (−2.66 to 1.89)1.14 (−0.11 to 2.98)Non-Hispanic Black2003–20220.20 (−2.48 to 3.18)0.2 (−2.48 to 3.18)0.2 (−2.48 to 3.18)Non-Hispanic Asian or Pacific Islander2003–20221.75 (−0.09 to 4.13)1.75 (−0.09 to 4.13)1.75 (−0.09 to 4.13)Hispanic, All Races2003–20221.88 (0.50 to 3.60)1.88 (0.5 to 3.6)1.88 (0.5 to 3.6) Primary SitesCranial2003–20221.93 (1.05 to 2.90)1.93 (1.05 to 2.9)1.93 (1.05 to 2.9)Sacral2003–20221.33 (0.30 to 2.42)1.33 (0.3 to 2.42)1.33 (0.3 to 2.42)Spinal2003–20220.90 (0.05 to 1.81)0.9 (0.05 to 1.81)0.9 (0.05 to 1.81)Other2003–20222.74 (−1.42 to 7.86)2.74 (−1.42 to 7.86)2.74 (−1.42 to 7.86) StageLocalized2003–20194.71 (4.17 to 7.36)2019–2022−2.95 (−7.49 to 3.19)2.09 (0.53 to 3.34)3.46 (2.74 to 4.56)Regional2003–2022−0.36 (−1.98 to 1.33)−0.36 (−1.98 to 1.33)−0.36 (−1.98 to 1.33)Distant2003–2022−1.36 (−3.81 to 1.10)−1.36 (−3.81 to 1.1)−1.36 (−3.81 to 1.1)Unknown2003–2005−21.81 (−30.29 to −3.13)2005–2022−1.32 (−2.97 to 3.34)−1.32 (−2.93 to 1.87)−3.7 (−4.9 to −1.61) Annual percent change (APC) and average annual percent change (AAPC) in incidence rates of chordoma in the NAACCR between 2003 and 2022. Fig. S1 shows trends in relative incidence rates of chordoma by period of diagnosis and birth cohorts. Relative risks increased over time and across cohorts, but the trends remained linear, with no significant deviations from linearity (Pperiod deviation = 0.08; Pcohort deviation = 0.13). Table 4 presents 5-year age-standardized relative survival for 2170 cases diagnosed in 21 SEER areas. Overall, the 5-year survival was 83.3%, ranging from 88.2% for localized disease to 55.7% for distant-stage disease. Survival estimates for localized disease were comparable across primary sites (82.0% to 83.7%) but varied widely for distant-stage disease, ranging from 68.6% for cranial chordoma to 38.3% for sacral chordoma. Survival estimates also varied by race/ethnicity, with notably lower rates among AIAN (51%) and non-Hispanic Black (79.3%) individuals compared to White individuals (84.4%). A modest but noticeable gradient was observed by rurality, with lower survival in counties with populations <250,000 (80.3%) vs. >1 million (84.1%). Table 4Age-standardized five-year relative survival for chordoma diagnosed in 21 Surveillance, Epidemiology, and End Results areas between 2003 and 2021.TotalaCranialSacralSpinal%Relative survival (95% CI)%Relative survival (95% CI)%Relative survival (95% CI)%Relative survival (95% CI)Total217083.3% (81.1% - 85.3%)88683.7% (79.6% - 87.0%)67682.0% (77.8% - 85.4%)53082.4% (77.8% - 86.2%)Year of diagnosis2003–200720.4%79.9% (74.9% - 84.0%)21.3%76.5% (67.3% - 83.5%)18.0%76.2% (65.8% - 83.9%)22.8%85.4% (75.9% - 91.4%)2008–201224.9%81.5% (77.3% - 85.0%)25.2%85.7% (78.1% - 90.8%)25.6%79.9% (71.9% - 85.8%)23.8%75.9% (66.8% - 82.8%)2013–201729.4%83.4% (79.3% - 86.7%)28.6%81.8% (74.1% - 87.3%)32.5%82.9% (75.7% - 88.2%)26.4%81.4% (72.2% - 87.8%)2018–202125.3%–24.9%–23.8%–27.0%– SexMale58.4%84.3% (81.3% - 86.8%)54.0%85.1% (78.8% - 89.7%)62.7%85.0% (79.8% - 89.0%)60.9%82.1% (76.2% - 86.8%)Female41.6%82.3% (78.9% - 85.2%)46.0%82.3% (76.6% - 86.7%)37.3%77.1% (70.1% - 82.8%)39.1%83.5% (76.1% - 88.8%) Race and ethnicityNon-Hispanic White63.9%84.4% (81.6% - 86.7%)56.1%84.6% (79.4% - 88.6%)66.9%83.5% (78.3% - 87.5%)73.4%83.7% (78.4% - 87.8%)Non-Hispanic Black5.0%79.3% (64.7% - 88.3%)5.4%82.3% (54.4% - 93.9%)5.6%71.5% (49.7% - 85.2%)3.0%–Non-Hispanic American Indian and Alaska Native0.6%51.0% (22.5% - 73.8%)0.5%–0.6%–0.8%–Non-Hispanic Asian or Pacific Islander9.9%81.6% (74.4% - 86.9%)10.8%84.2% (71.9% - 91.4%)11.5%77.3% (63.0% - 86.7%)7.0%73.6% (52.6% - 86.4%)Hispanic, All Races20.6%79.3% (73.3% - 84.1%)27.2%79.0% (69.0% - 86.0%)15.4%79.1% (66.9% - 87.2%)15.8%74.1% (58.5% - 84.5%) Registries by Census RegionNortheast29.8%83.9% (79.7% - 87.3%)28.4%85.8% (78.5% - 90.7%)33.4%80.1% (72.5% - 85.8%)28.1%84.6% (75.7% - 90.4%)Midwest2.4%81.5% (65.6% - 90.6%)2.4%-b1.8%-b3.4%-bSouth28.5%83.3% (79.0% - 86.8%)29.3%82.0% (73.6% - 87.9%)27.4%83.7% (75.3% - 89.5%)26.6%80.3% (71.1% - 86.8%)West39.2%81.9% (78.2% - 85.1%)39.8%82.1% (74.7% - 87.5%)37.4%81.2% (74.2% - 86.6%)41.9%80.2% (72.7% - 85.9%) Urban/rural continuum (code 2023)Counties in metropolitan areas ≥1 million population68.2%84.1% (81.4% - 86.4%)70.0%83.7% (78.8% - 87.5%)68.3%82.5% (77.5% - 86.5%)64.0%84.7% (78.8% - 89.0%)Counties in metropolitan areas of 250,000 to 1 million population17.9%81.8% (76.3% - 86.2%)17.8%81.5% (70.1% - 88.9%)18.6%82.6% (71.6% - 89.7%)17.5%70.9% (58.4% - 80.3%)Counties in metropolitan areas of less than 250,000 pop and non-metropolitan areas13.9%80.3% (74.0% - 85.2%)12.2%82.4% (70.5% - 89.9%)12.9%72.3% (58.0% - 82.5%)18.5%83.6% (72.8% - 90.4%)Unknown0.0%–0.0%–0.1%–0.0%– StageLocalized47.6%88.2% (85.0% - 90.8%)51.8%89.8% (83.5% - 93.8%)39.1%88.6% (82.1% - 92.9%)52.3%85.4% (78.6% - 90.2%)Regional34.9%83.3% (79.4% - 86.3%)33.5%77.2% (70.4% - 82.6%)42.0%85.8% (79.3% - 90.4%)29.8%82.8% (73.8% - 88.9%)Distant7.1%55.7% (46.3% - 64.1%)5.5%68.6% (51.7% - 80.6%)9.0%38.3% (23.8% - 52.8%)6.6%60.9% (42.3% - 75.2%)Unknown10.4%80.0% (72.9% - 85.4%)9.1%87.1% (72.5% - 94.3%)9.9%73.2% (57.3% - 84.0%)11.3%77.1% (62.3% - 86.8%) Histology9370/3: Chordoma, not otherwise specified92.4%83.6% (81.3% - 85.6%)87.4%83.5% (79.0% - 87.1%)96.3%83.0% (78.8% - 86.4%)95.8%82.8% (78.1% - 86.5%)9371/3: Chondroid chordoma6.9%82.7% (73.0% - 89.1%)12.4%84.7% (73.6% - 91.4%)2.4%76.9% (54.0% - 89.4%)3.6%73.1% (42.7% - 89.1%)9372/3: Dedifferentiated chordoma0.7%–0.2%–1.3%–0.6%–aTotal includes tumors classified as “other” (N = 133). See Table S1 for ICD-O-3 codes for subsite classification, including “other”.bNot presented due to limited case numbers that preclude stable estimates. Age-standardized five-year relative survival for chordoma diagnosed in 21 Surveillance, Epidemiology, and End Results areas between 2003 and 2021. Total includes tumors classified as “other” (N = 133). See Table S1 for ICD-O-3 codes for subsite classification, including “other”. Not presented due to limited case numbers that preclude stable estimates. Discussion In this large, population-based analysis of chordoma, we provide a comprehensive assessment of incidence patterns, temporal trends, tumor characteristics, and survival outcomes among more than 6000 patients diagnosed in the United States over two decades. Leveraging near-complete national cancer registry coverage and high-quality SEER survival data, this study represents the most extensive and up-to-date descriptive evaluation of chordoma burden to date. Beyond confirming established demographic features of this rare malignancy, our findings reveal clinically and epidemiologically meaningful heterogeneity by age, sex, race and ethnicity, geographic region, and anatomic tumor site, with important implications for diagnosis, referral patterns, and care delivery. Consistent with prior studies, chordoma incidence increased markedly with age and was higher overall in males. Notably, however, sex differences emerged only after age 35, with comparable incidence rates in younger individuals. This pattern appears to be driven in part by the higher proportion of cranial tumors among women and younger patients, as cranial chordomas tend to present earlier in life and show less pronounced male predominance. These findings raise the possibility that early-onset chordoma, particularly at cranial sites, may represent a biologically distinct subgroup. Such heterogeneity underscores the importance of age- and site-specific analyses in future clinical and molecular studies, especially in pediatric and young adult populations where data remain sparse. We also found striking differences by race, ethnicity, and geographic region. Incidence was lowest among non-Hispanic Black individuals, consistent with observations from previous smaller studies [3], [4], [5], [16]. Genetic susceptibility differences may play a role, e.g., a common genetic variant in TBXT (a chordoma susceptibility gene encoding a transcription factor playing an important role in notochord development) that was associated with increased chordoma risk is more prevalent among European individuals (50%) than African/African American individuals (26%) [17]. Disparities in diagnostic access, imaging utilization, or referral to specialized centers are also likely to contribute. Similarly, higher incidence in the Northeast and metropolitan (vs non-metropolitan) areas may reflect greater access to advanced imaging, subspecialty care, and academic referral centers where chordoma is more likely to be recognized and accurately diagnosed. Together, these findings suggest that observed incidence patterns may partially reflect healthcare system factors, highlighting the potential impact of care centralization and diagnostic pathways on reported disease burden. The relatively large sample size in this study also allowed for comparisons by primary tumor location. Cranial tumors were more common in younger and non-White patients, less male-dominant, and more often diagnosed at a localized stage compared with spinal and sacral tumors. These site-specific differences align with emerging molecular evidence that chordomas from different anatomic locations may have distinct biological features. For example, our previous tumor sequencing analysis of chordoma patients residing in the U.S. and Canada showed that cranial tumors were less likely to carry driver gene mutations and copy number alterations compared to sacral and spinal tumors [18]. Clinically, patients with cranial tumors had the highest relative survival, which may reflect earlier detection, advancements in surgical techniques, and more favorable disease biology, although the challenges of surgical management in this region remain considerable. Our findings reinforce the need for site-specific management strategies and highlight the potential benefit of referral to high-volume centers with multidisciplinary expertise. A key finding of this study is the steady increase in chordoma incidence over the past two decades, averaging approximately 1.5% annually, particularly for cranial and localized tumors. This rise likely reflects improved diagnostic imaging, better case ascertainment, and reclassification of tumors that might previously have been misdiagnosed. This is particularly true for cranial tumors, for which the advancement of the endoscopic endonasal surgery technique has facilitated transnasal biopsy of clival lesions [2], [19] and improved prognosis [20], while routine brachyury immunohistochemistry in the recent decade has significantly improved the accuracy of diagnosis and survival [21]. The steeper increase among women compared with men (2.03% vs 1.36% per year) may mirror the rise of cranial tumors, which show less pronounced male predominance. Additionally, we found the increase in chordoma incidence was limited to younger individuals (<65 years), whereas rates remained stable among older adults. While the increase in incidence rates may suggest a rising prevalence of risk factors, the specific risk factors underlying this trend remain unknown. Importantly, the increase was confined to localized disease, whereas rates of regional and distant diseases were stable or declined, supporting that earlier detection plays a major role rather than a true increase in underlying risk. Consistent with this interpretation, age–period-cohort modeling analysis showed that the increase in incidence was gradual and linear rather than driven by a specific birth cohort, suggesting that the observed rise is more likely to be attributable to period-related improvements in cancer diagnosis than to changes in risk factor exposure. Our study has several limitations. As with all registry-based analyses, we lacked detailed clinical information on treatment, recurrence, and comorbidities. Misclassification of tumor site or histology is possible, although minimized by modern registry coding practices. Survival analyses were descriptive and based on net survival using 5-year relative survival estimates. Therefore, the observed heterogeneity in survival by subtype, race/ethnicity, and rurality should be interpreted with caution and warrants further investigation. Multivariable analyses using data supplemented with information on access to care (e.g., health insurance status) and receipt of guideline-concordant treatment information may help identify the mechanisms underlying the observed disparity in survival. Lastly, although derived from the largest available population-based data in the U.S., small case counts led to age-standardized rates and 5-year relative survival estimates with wide confidence intervals, particularly when stratified simultaneously by primary site and patient characteristics and should be interpreted with caution. Nonetheless, the use of 20 consecutive years of high-quality population-based registry data with over 90% coverage of the U.S. population alongside with SEER data for survival analysis strengthens the robustness and generalizability of our findings. These comprehensive datasets allow for a more accurate assessment of the overall disease burden, as well as detailed examination of variations in incidence, trends, and survival outcomes across key demographic subgroups. In conclusion, this comprehensive analysis demonstrates that chordoma incidence is rising in the U.S., and varies substantially by age, sex, race, geography, and tumor site. While some of these patterns likely reflect differences in diagnostic practices and healthcare access, the observed heterogeneity across regions, racial and ethnic groups, and tumor sites may also provide important clues to underlying etiologic factors. These findings have important implications for healthcare access, referral patterns, and centralization of care for this rare malignancy. Furthermore, our findings offer updated population-based benchmarks for this rare disease and highlight the need for integrated epidemiologic and molecular studies to clarify disease mechanisms and guide strategies for earlier detection and improved outcomes. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work, the author(s) used Codex for producing the graphical abstract. The author(s) reviewed and edited the output as needed and take full responsibility for the content of the published article. CRediT authorship contribution statement Hyuna Sung: Writing – review & editing, Writing – original draft, Funding acquisition, Formal analysis, Conceptualization. Chenxi Jiang: Writing – review & editing, Formal analysis. Hela Koka: Writing – review & editing, Conceptualization. Jiwei Bai: Writing – review & editing, Validation. Alisa M. Goldstein: Writing – review & editing, Conceptualization. Xiaohong R. Yang: Writing – review & editing, Writing – original draft, Conceptualization. Ethics This study utilized de-identified, publicly available data and is therefore classified as non-human subjects research under the Common Rule (45 CFR §46). Accordingly, institutional review board approval and informed consent were not required. Role of the funder/sponsor The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Funding This work was supported by the Intramural Research Department of the American Cancer Society (HS and CJ) and in part by the Intramural Research Program of the National Institutes of Health (NIH) (HK, AMG, and XRY). The contributions of the NIH author(s) are considered Works of the United States Government. The findings and conclusions presented in this paper are those of the author(s) and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services. Declaration of competing interest We declare no competing interests. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.jbo.2026.100774. Supplementary data Supplementary Table S1-S2 Data availability Two databases used in this study (SEER*Stat Database: NAACCR Incidence Data - CiNA Research Data, 1995–2022, Public Use [https://www.naaccr.org/cina-data-products-overview/] & SEER Research Plus Limited-Field Data, 21 Registries [www.seer.cancer.gov]) are publicly available. References Associated Data Supplementary Materials Supplementary Table S1-S2 Data Availability Statement Two databases used in this study (SEER*Stat Database: NAACCR Incidence Data - CiNA Research Data, 1995–2022, Public Use [https://www.naaccr.org/cina-data-products-overview/] & SEER Research Plus Limited-Field Data, 21 Registries [www.seer.cancer.gov]) are publicly available.