Age-dependent chromosomal instability in osteosarcoma: younger patients exhibit paradoxically higher genomic damage
Tiara
Jamison
1✉
Emailtiara@annieguard.com
1
A
AnnieGuard Research
5520 Research Park Drive, Suite 100
21228
Baltimore
MD
USA
Tiara Jamison1*
1 AnnieGuard Research, 5520 Research Park Drive, Suite 100, Baltimore, MD 21228, USA
* Correspondence: tiara@annieguard.com
Abstract
Osteosarcoma is an aggressive bone malignancy with peak incidence during adolescence and a secondary peak in older adults. While the disease is characterized by extreme chromosomal instability, the relationship between patient age and genomic damage has not been systematically quantified. Here, using two independent cohorts totaling 589 osteosarcoma samples, we demonstrate a paradoxical inverse relationship between age and chromosomal instability. In the discovery cohort (n = 460, Foundation Medicine), younger patients (≤ 25 years) exhibited significantly higher loss of heterozygosity (LOH: 16.52% vs 12.70%, p = 0.0002) compared to older patients (> 40 years). This finding was validated in an independent cohort (n = 129, MSK-IMPACT) using fraction genome altered (FGA: 36.5% vs 26.5%, p = 0.028). The age-instability correlation was continuous and negative in both cohorts (r=-0.15 to -0.25, p < 0.01). Notably, chromosomal instability did not differ between primary and metastatic tumors, suggesting metastatic competence is established early rather than acquired through additional genomic damage. These findings challenge conventional cancer models where genomic damage accumulates with age and support the chromothripsis hypothesis for adolescent osteosarcoma, wherein catastrophic chromosomal shattering during rapid bone growth drives tumorigenesis. Our results suggest that adolescent and adult osteosarcomas may represent biologically distinct entities requiring different therapeutic strategies.
Keywords:
osteosarcoma
chromosomal instability
loss of heterozygosity
age
chromothripsis
pediatric cancer
Introduction
Osteosarcoma is the most common primary malignant bone tumor, with a bimodal age distribution showing peak incidence during adolescence (ages 15–25) and a secondary peak in adults over 60 years1,2. Despite decades of research, the 5-year survival rate has remained stagnant at approximately 60–70% for localized disease, with significantly poorer outcomes for metastatic cases3,4. This therapeutic plateau highlights the critical need to understand the fundamental biology driving osteosarcoma development across different age groups.
A
Genetically, osteosarcoma is characterized by extreme chromosomal instability rather than recurrent point mutations5,6. Unlike many pediatric cancers that harbor specific chromosomal translocations (e.g., EWSR1::FLI1 in Ewing sarcoma, PAX3::FOXO1 in alveolar rhabdomyosarcoma), osteosarcoma exhibits complex karyotypes with numerous structural and numerical aberrations affecting nearly every chromosome7,8. The most frequently altered genes include TP53 (28–40%), RB1 (8–21%), and ATRX (6–8%), which are involved in maintaining genomic stability9–11.Recent studies have demonstrated that osteosarcoma frequently harbors chromothripsis—a catastrophic genomic event involving the shattering and random reassembly of one or more chromosomes12–14. This phenomenon has been particularly associated with pediatric osteosarcoma and may explain the disease's propensity to arise during periods of rapid bone growth15. However, the quantitative relationship between patient age and chromosomal instability has not been systematically characterized across large cohorts.
Age-related differences in osteosarcoma have been reported at multiple levels. Younger patients typically present with tumors in the appendicular skeleton near growth plates, while older patients more frequently develop axial or secondary tumors16,17. Recent genomic profiling has revealed age-associated differences in specific driver mutations, with CCNE1 and MYC amplifications enriched in younger patients and MDM2 and CDKN2A/B alterations more common in older patients18,19. However, whether these qualitative differences in mutation spectra translate to quantitative differences in overall genomic instability remains unknown.
In conventional cancer models, genomic instability is expected to accumulate with age as cells acquire progressive DNA damage20,21. This paradigm has been validated in numerous adult cancers, where tumor mutational burden correlates positively with patient age22,23. Whether this relationship holds in osteosarcoma—a disease that predominantly affects adolescents and may arise through fundamentally different mechanisms—has not been established.
In this study, we systematically analyzed chromosomal instability metrics, specifically loss of heterozygosity (LOH) and fraction genome altered (FGA), across two independent osteosarcoma cohorts totaling 589 patients. We hypothesized that if adolescent osteosarcoma arises through catastrophic chromothripsis events during rapid bone growth, younger patients might paradoxically exhibit higher chromosomal instability than older patients. Our findings reveal a robust inverse relationship between age and chromosomal instability that challenges conventional cancer paradigms and has significant implications for understanding osteosarcoma biology and developing age-stratified therapeutic approaches.
Results
Cohort characteristics
We analyzed two independent osteosarcoma cohorts from publicly available data (cBioPortal). The discovery cohort (Bone Cancer, Foundation Medicine) comprised 460 osteosarcoma samples profiled by comprehensive genomic profiling. Patients ranged from 5 to 89 years of age (median 21, mean 30.2 ± 20.1) with a slight male predominance (252 male, 208 female). Loss of heterozygosity (LOH) was quantified across 320 samples. The validation cohort (Sarcoma, MSK-IMPACT 2022) included 129 osteosarcoma samples sequenced using the MSK-IMPACT platform. Patients ranged from 8 to 78 years (median 19, mean 26.1 ± 16.6) with 79 males and 50 females. Fraction genome altered (FGA) was available for all 129 samples (Table 3).
|
Characteristic
|
Discovery Cohort
|
Validation Cohort
|
|---|---|---|
|
Total samples
|
460
|
129
|
|
Age range (years)
|
5–89
|
8–78
|
|
Median age (years)
|
21
|
19
|
|
Mean age ± SD
|
30.2 ± 20.1
|
26.1 ± 16.6
|
|
Male
|
252 (54.8%)
|
79 (61.2%)
|
|
Female
|
208 (45.2%)
|
50 (38.8%)
|
|
Primary instability metric
|
LOH (%)
|
FGA (%)
|
|
Median instability
|
15.0%
|
31.7%
|
|
Median TMB (mut/Mb)
|
0.033
|
2.0
|
Driver gene frequencies were remarkably consistent between cohorts (Table 1). TP53 was altered in 28.5% (discovery) and 25.6% (validation) of samples. RB1 alterations were observed in 8.9% and 7.0%, ATRX in 6.1% and 6.2%, and other drivers including CDKN2A, PTEN, PIK3CA, and SETD2 at comparable frequencies. This consistency supports the validity of cross-cohort comparisons despite different sequencing platforms.
|
Gene
|
Discovery (n = 460)
|
Validation (n = 129)
|
|---|---|---|
|
TP53
|
28.5%
|
25.6%
|
|
RB1
|
8.9%
|
7.0%
|
|
ATRX
|
6.1%
|
6.2%
|
|
CDKN2A
|
2.0%
|
1.6%
|
|
PTEN
|
1.7%
|
1.6%
|
|
PIK3CA
|
1.7%
|
2.3%
|
|
SETD2
|
1.7%
|
3.9%
|
|
NF1
|
1.1%
|
1.6%
|
Younger patients exhibit higher chromosomal instability
Contrary to conventional expectations, we observed that younger patients had significantly higher chromosomal instability than older patients. In the discovery cohort, patients aged ≤ 25 years exhibited mean LOH of 16.52% ± 8.83% (n = 189), compared to 12.70% ± 8.80% in patients aged > 40 years (n = 89). This difference was highly significant (p = 0.000198, Welch's t-test) (Fig. 1A). The relationship between age and LOH was continuous, with Pearson correlation r=-0.147 (p = 0.0083), indicating a significant negative correlation across the entire age spectrum (Fig. 1B).
This paradoxical finding was robustly validated in the independent cohort using fraction genome altered (FGA) as the chromosomal instability metric. Younger patients (≤ 25 years) had mean FGA of 36.50% ± 19.38% (n = 82), compared to 26.53% ± 20.54% in older patients (> 40 years, n = 24). This difference was statistically significant (p = 0.028, Welch's t-test) (Fig. 1C). The negative correlation was stronger in the validation cohort (r=-0.251, p = 0.0041) (Fig. 1D).
Using an alternative measure of chromosomal instability (Fraction CNA) in the validation cohort further confirmed these findings. Younger patients exhibited significantly higher Fraction CNA (73.45% ± 18.30%) compared to older patients (56.92% ± 23.20%, p = 0.0024). The consistency across different instability metrics and sequencing platforms strengthens confidence in this finding (Table 2).
|
Cohort
|
Young (≤ 25)
|
Old (> 40)
|
p-value
|
Correlation r (p)
|
|---|---|---|---|---|
|
Discovery (LOH)
|
16.52% ± 8.83% (n = 189)
|
12.70% ± 8.80% (n = 89)
|
0.0002
|
-0.147 (0.008)
|
|
Validation (FGA)
|
36.50% ± 19.38% (n = 82)
|
26.53% ± 20.54% (n = 24)
|
0.028
|
-0.251 (0.004)
|
|
Validation (CNA)
|
73.45% ± 18.30% (n = 74)
|
56.92% ± 23.20% (n = 20)
|
0.002
|
-0.289 (0.005)
|
Fig. 1
Younger osteosarcoma patients exhibit higher chromosomal instability. (A) Bar plot showing mean LOH in young (≤ 25 years) versus old (> 40 years) patients in the discovery cohort. (B) Scatter plot showing negative correlation between age and LOH in the discovery cohort. (C) Bar plot showing mean FGA in young versus old patients in the validation cohort. (D) Scatter plot showing negative correlation between age and FGA in the validation cohort. Error bars represent standard deviation. ***p < 0.001, *p < 0.05.
Low TMB combined with high chromosomal instability signature
Both cohorts confirmed osteosarcoma's characteristic genomic profile: extremely low tumor mutational burden (TMB) combined with high chromosomal instability. In the discovery cohort, median TMB was 0.0333 mutations/Mb, among the lowest of any cancer type. In contrast, median LOH was 15.0%, indicating substantial structural genomic damage. The validation cohort showed similar patterns with median TMB of 2.0 mutations/Mb and median FGA of 31.7%. This signature—few point mutations but extensive chromosomal rearrangements—supports the concept that osteosarcoma is fundamentally a disease of genomic architecture rather than sequence-level mutations.
Metastasis does not increase chromosomal instability
We next examined whether metastatic progression was associated with increased chromosomal instability. In the discovery cohort, primary bone tumors had mean LOH of 14.66% ± 8.59% (n = 109), while metastatic samples (predominantly lung, n = 81) showed LOH of 15.97% ± 9.56%. This difference was not statistically significant (p = 0.324). Similarly, in the validation cohort, primary tumors had FGA of 34.38% ± 20.89% (n = 77) compared to 28.39% ± 19.12% in metastatic samples (n = 52, p = 0.111) (Fig. 2, Table 4).
These findings suggest that metastatic competence in osteosarcoma is an intrinsic property established early in tumorigenesis rather than acquired through additional chromosomal damage during disease progression. This has important clinical implications, suggesting that the genomic features driving metastasis are present from tumor initiation.
Fig. 2
Chromosomal instability does not differ between primary and metastatic tumors. (A) Comparison of LOH between primary bone tumors and metastatic (lung) samples in the discovery cohort. (B) Comparison of FGA between primary and metastatic samples in the validation cohort. Error bars represent standard deviation. ns = not significant.
|
Cohort
|
Primary
|
Metastatic
|
p-value
|
|---|---|---|---|
|
Discovery (LOH)
|
14.66% ± 8.59% (n = 109)
|
15.97% ± 9.56% (n = 81)
|
0.324 (ns)
|
|
Validation (FGA)
|
34.38% ± 20.89% (n = 77)
|
28.39% ± 19.12% (n = 52)
|
0.111 (ns)
|
Driver gene consistency across cohorts
Analysis of driver gene frequencies revealed remarkable consistency between the discovery and validation cohorts (Fig. 3). TP53 remained the most frequently altered gene in both cohorts (28.5% and 25.6%, respectively), followed by RB1 (8.9% and 7.0%) and ATRX (6.1% and 6.2%). This consistency across different sequencing platforms and patient populations reinforces the core driver architecture of osteosarcoma and validates cross-cohort comparisons.
Fig. 3
Driver gene frequencies are consistent across cohorts. Bar plot comparing alteration frequencies of major driver genes between the discovery cohort (Foundation Medicine, n = 460) and validation cohort (MSK-IMPACT, n = 129).
Discussion
Our findings reveal a paradoxical inverse relationship between patient age and chromosomal instability in osteosarcoma, with younger patients exhibiting significantly higher genomic damage than older patients. This finding, validated across two independent cohorts using different sequencing platforms and instability metrics, challenges conventional cancer models where genomic damage accumulates with age and has significant implications for understanding osteosarcoma biology.
The observed age-instability paradox supports the chromothripsis hypothesis for adolescent osteosarcoma. Chromothripsis—the catastrophic shattering and random reassembly of chromosomes—has been increasingly recognized as a major driver of pediatric osteosarcoma12–15. Our data suggest that tumors arising during periods of rapid bone growth (adolescence) may be more susceptible to such catastrophic events, resulting in higher chromosomal instability despite younger age. In contrast, osteosarcomas arising in older patients may develop through more gradual accumulation of genomic alterations, resulting in lower overall instability.
Our findings complement and extend recent work by Outani et al.18, who demonstrated age-associated differences in specific driver mutations in osteosarcoma. While their study focused on qualitative differences in mutation spectra (which genes are altered), our analysis reveals quantitative differences in overall genomic instability (how much total damage). The observation that CCNE1 and MYC amplifications are enriched in younger patients18,19 is consistent with chromothripsis-mediated oncogene amplification, providing mechanistic support for our findings.
The finding that metastatic tumors do not exhibit increased chromosomal instability compared to primary tumors has important clinical implications. This suggests that metastatic competence in osteosarcoma is an intrinsic property established early in tumorigenesis rather than acquired through progressive genomic damage. Clinically, this supports the importance of early intervention and suggests that genomic profiling of primary tumors may be predictive of metastatic potential.
Our results have implications for precision oncology approaches in osteosarcoma. The biological distinction between adolescent and adult osteosarcomas suggested by our data supports the development of age-stratified therapeutic strategies. Younger patients with high-instability tumors may benefit from therapies targeting chromosomal instability or DNA damage response pathways, while older patients may require different approaches targeting more conventional oncogenic mechanisms.
Several limitations should be acknowledged. Our analysis was retrospective and based on publicly available data, which may introduce selection biases. The different sequencing platforms used in the discovery and validation cohorts measure chromosomal instability differently (LOH vs FGA), though the consistency of findings across metrics strengthens our conclusions. Future studies incorporating whole-genome sequencing would enable more detailed characterization of chromothripsis patterns across age groups.
In conclusion, we demonstrate a robust inverse relationship between patient age and chromosomal instability in osteosarcoma, validated across two independent cohorts. This age-instability paradox challenges conventional cancer models and supports the hypothesis that adolescent and adult osteosarcomas may represent biologically distinct entities arising through different oncogenic mechanisms. These findings have implications for understanding osteosarcoma pathogenesis and developing age-stratified precision oncology approaches.
Methods
Data sources
Genomic and clinical data were obtained from the cBioPortal for Cancer Genomics (https://www.cbioportal.org/). The discovery cohort was derived from the Bone Cancer study (Foundation Medicine, bone_msk_2022) containing 460 osteosarcoma samples. The validation cohort was derived from the Sarcoma study (MSK, Nat Commun 2022, sarcoma_mskcc_2022) containing 129 osteosarcoma samples.
Chromosomal instability metrics
Loss of heterozygosity (LOH) was defined as the percentage of the genome exhibiting loss of one allele while retaining the other, as calculated by Foundation Medicine's comprehensive genomic profiling platform. Fraction genome altered (FGA) was defined as the fraction of the genome with copy number alterations, as calculated by the MSK-IMPACT platform. Fraction CNA was calculated as the proportion of assessed genes exhibiting copy number alterations.
Statistical analysis
Comparisons between age groups were performed using Welch's two-sample t-test. Correlations between continuous variables were assessed using Pearson correlation coefficients. Age cutoffs of ≤ 25 years (young) and > 40 years (old) were selected based on the bimodal age distribution of osteosarcoma and established clinical conventions. All statistical analyses were performed using Python (scipy, numpy, pandas) and R. P-values < 0.05 were considered statistically significant.
Data Availability
All data analyzed in this study are publicly available through the cBioPortal for Cancer Genomics (https://www.cbioportal.org/). The discovery cohort is available as 'Bone Cancer (Foundation Medicine)' and the validation cohort as 'Sarcoma (MSK, Nat Commun 2022)'.
Acknowledgments
We thank the patients and families who contributed samples to the Foundation Medicine and MSK-IMPACT databases. We acknowledge the cBioPortal team for making these data publicly accessible.
A
Funding
Declaration
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Competing Interests
The author declares no competing interests.
A
Author Contribution
T.J. conceived the study, performed all analyses, and wrote the manuscript.
A
Data Availability
All data analyzed in this study are publicly available through the cBioPortal for Cancer Genomics (https://www.cbioportal.org/). The discovery cohort is available as 'Bone Cancer (Foundation Medicine)' and the validation cohort as 'Sarcoma (MSK, Nat Commun 2022)'.
References
1.
Mirabello L, Troisi RJ, Savage SA. Osteosarcoma incidence and survival rates from 1973 to 2004: data from the surveillance, epidemiology, and end results program. Cancer 115, 1531–1543 (2009).
2.
Cole S, et al. Osteosarcoma: a surveillance, epidemiology, and end results program-based analysis from 1975 to 2017. Cancer 128, 2107–2118 (2022).
3.
Smeland S, et al. Survival and prognosis with osteosarcoma: outcomes in more than 2000 patients in the EURAMOS-1 cohort. Eur J Cancer 109, 36–50 (2019).
4.
Marina NM, et al. Comparison of MAPIE versus MAP in patients with a poor response to preoperative chemotherapy for newly diagnosed high-grade osteosarcoma. Lancet Oncol 17, 1396–1408 (2016).
5.
Chen X, et al. Recurrent somatic structural variations contribute to tumorigenesis in pediatric osteosarcoma. Cell Rep 7, 104–112 (2014).
6.
Perry JA, et al. Complementary genomic approaches highlight the PI3K/mTOR pathway as a common vulnerability in osteosarcoma. Proc Natl Acad Sci USA 111, E5564-E5573 (2014).
7.
Martin JW, Squire JA, Zielenska M. The genetics of osteosarcoma. Sarcoma 2012, 627254 (2012).
8.
Behjati S, et al. Recurrent mutation of IGF signalling genes and distinct patterns of genomic rearrangement in osteosarcoma. Nat Commun 8, 15936 (2017).
9.
Sayles LC, et al. Genome-informed targeted therapy for osteosarcoma. Cancer Discov 9, 46–63 (2019).
10.
Kovac M, et al. Exome sequencing of osteosarcoma reveals mutation signatures reminiscent of BRCA deficiency. Nat Commun 6, 8940 (2015).
11.
Nacev BA, et al. Clinical sequencing of soft tissue and bone sarcomas delineates diverse genomic landscapes and potential therapeutic targets. Nat Commun 13, 3405 (2022).
12.
Stephens PJ, et al. Massive genomic rearrangement acquired in a single catastrophic event during cancer development. Cell 144, 27–40 (2011).
13.
Cortes-Ciriano I, et al. Comprehensive analysis of chromothripsis in 2,658 human cancers using whole-genome sequencing. Nat Genet 52, 331–341 (2020).
14.
Valle-Inclan JE, et al. Ongoing chromothripsis underpins osteosarcoma genome complexity and clonal evolution. Cell 188, 1–18 (2025).
15.
Smida J, et al. Genomic alterations and allelic imbalances are strong prognostic predictors in osteosarcoma. Clin Cancer Res 16, 4256–4267 (2010).
16.
Joo MW, et al. Osteosarcoma in Asian populations over the age of 40 years: a multicenter study. Ann Surg Oncol 22, 3557–3564 (2015).
17.
Kim C, et al. Osteosarcoma in pediatric and adult populations: are adults just big kids? Cancers 15, 5044 (2023).
18.
Outani H, et al. Age-related genomic alterations and chemotherapy sensitivity in osteosarcoma: insights from cancer genome profiling analyses. Int J Clin Oncol 30, 397–406 (2024).
19.
Gounder MM, et al. Clinical genomic profiling in the management of patients with soft tissue and bone sarcoma. Nat Commun 13, 3406 (2022).
20.
Alexandrov LB, et al. Signatures of mutational processes in human cancer. Nature 500, 415–421 (2013).
21.
Tomasetti C, Vogelstein B. Cancer etiology: variation in cancer risk among tissues can be explained by the number of stem cell divisions. Science 347, 78–81 (2015).
22.
Wu S, et al. Substantial contribution of extrinsic risk factors to cancer development. Nature 529, 43–47 (2016).
23.
Stratton MR, Campbell PJ, Futreal PA. The cancer genome. Nature 458, 719–724 (2009).
Figure 1. Younger osteosarcoma patients exhibit higher chromosomal instability.
(A) Bar plot showing mean LOH in young (≤ 25 years) versus old (> 40 years) patients in the discovery cohort. (B) Scatter plot showing negative correlation between age and LOH in the discovery cohort. (C) Bar plot showing mean FGA in young versus old patients in the validation cohort. (D) Scatter plot showing negative correlation between age and FGA in the validation cohort. Error bars represent standard deviation. ***p < 0.001, *p < 0.05.
Figure 2. Chromosomal instability does not differ between primary and metastatic tumors.
(A) Comparison of LOH between primary bone tumors and metastatic (lung) samples in the discovery cohort. (B) Comparison of FGA between primary and metastatic samples in the validation cohort. Error bars represent standard deviation. ns = not significant.
Figure 3. Driver gene frequencies are consistent across cohorts.
Bar plot comparing alteration frequencies of major driver genes between the discovery cohort (Foundation Medicine, n = 460) and validation cohort (MSK-IMPACT, n = 129).
Table 3. Cohort characteristics
|
Characteristic
|
Discovery Cohort
|
Validation Cohort
|
|---|---|---|
|
Total samples
|
460
|
129
|
|
Age range (years)
|
5–89
|
8–78
|
|
Median age (years)
|
21
|
19
|
|
Mean age ± SD
|
30.2 ± 20.1
|
26.1 ± 16.6
|
|
Male
|
252 (54.8%)
|
79 (61.2%)
|
|
Female
|
208 (45.2%)
|
50 (38.8%)
|
|
Primary instability metric
|
LOH (%)
|
FGA (%)
|
|
Median instability
|
15.0%
|
31.7%
|
|
Median TMB (mut/Mb)
|
0.033
|
2.0
|
Table 4. Primary vs metastatic tumor comparison
|
Cohort
|
Primary
|
Metastatic
|
p-value
|
|---|---|---|---|
|
Discovery (LOH)
|
14.66% ± 8.59% (n = 109)
|
15.97% ± 9.56% (n = 81)
|
0.324 (ns)
|
|
Validation (FGA)
|
34.38% ± 20.89% (n = 77)
|
28.39% ± 19.12% (n = 52)
|
0.111 (ns)
|
ns = not significant
Table 1. Driver gene frequencies across cohorts
|
Gene
|
Discovery (n = 460)
|
Validation (n = 129)
|
|---|---|---|
|
TP53
|
28.5%
|
25.6%
|
|
RB1
|
8.9%
|
7.0%
|
|
ATRX
|
6.1%
|
6.2%
|
|
CDKN2A
|
2.0%
|
1.6%
|
|
PTEN
|
1.7%
|
1.6%
|
|
PIK3CA
|
1.7%
|
2.3%
|
|
SETD2
|
1.7%
|
3.9%
|
|
NF1
|
1.1%
|
1.6%
|
Table 2. Age-instability relationship across cohorts
|
Cohort
|
Young (≤ 25)
|
Old (> 40)
|
p-value
|
Correlation r (p)
|
|---|---|---|---|---|
|
Discovery (LOH)
|
16.52% ± 8.83% (n = 189)
|
12.70% ± 8.80% (n = 89)
|
0.0002
|
-0.147 (0.008)
|
|
Validation (FGA)
|
36.50% ± 19.38% (n = 82)
|
26.53% ± 20.54% (n = 24)
|
0.028
|
-0.251 (0.004)
|
|
Validation (CNA)
|
73.45% ± 18.30% (n = 74)
|
56.92% ± 23.20% (n = 20)
|
0.002
|
-0.289 (0.005)
|