Early-onset cancer signature (EOCS): Pan-cancer molecular alterations with prognostic significance in early-onset (EOC) and late-onset (LOC) cancer.

M Mahesh Iddawela (Monash University Faculty of Medicine, Nursing, and Health Sciences, Melbourne, VIC, Australia) M Mei Sim Lung (Peninsula Health, Frankston, VIC, Australia) R Ryan De Ruyter (Peninsula Clinical School, Monash University, Frankston, VIC, Australia) N Nandan Prasanna Kumar (Peninsula Clinical School, Monash University, Frankston, VIC, Australia) S Sally Heath (Peninsula Health, Frankston, VIC, Australia) B Babak Tamjid

Abstract

3126 Background: Early-onset cancer (EOC) is an emerging global epidemic with biological and molecular features distinct from late-onset cancer (LOC). Improved delineation of genomic pathway aberrations in EOC is essential to inform risk stratification and therapeutic decision-making. However, large-scale, unbiased analyses of biological drivers across tumour types remain limited. To address this gap, we developed a pan-cancer Early-Onset Cancer Signature (EOCS) to advance understanding of the genomic architecture underpinning EOC. Methods: The AACR GENIE v16 dataset comprising 241,551 tumour samples was analysed, including 22,408 early-onset cases (≤40 years) and 219,143 late-onset cases (≥41 years) from multiple international repositories. Differentially enriched genomic alterations and pathways between EOC and LOC were identified to define the EOCS. The EOCS was subsequently evaluated across TCGA pan-cancer datasets and validated in independent cohorts, including METABRIC (breast cancer), MSKCC colorectal cancer, and MSKCC prostate cancer datasets. Results: Analysis of the AACR GENIE cohort revealed that the most significantly enriched alterations in EOC compared with LOC were CCDC6 (20% vs 10%), SYNE1 (8% vs 6%), BRAF (9% vs 6%), FLI1 (19% vs 0.5%), and EML4 (8% vs 4%). In contrast, alterations significantly enriched in LOC included KRAS (15% vs 5%), TP53 (38% vs 29%), EGFR(15% vs 5%), STK11 (3% vs 1%), and CDKN2A (10% vs 6%). Gene set enrichment analysis of EOCS-associated genes identified DNA recombination, regulation of histone H3K9 trimethylation, and homologous recombination–mediated double-strand break repair as the top three enriched Gene Ontology pathways. The EOCS was altered in 30% (3,736/10,967) of TCGA pan-cancer samples, with the highest prevalence observed in endometrial, skin, carcinosarcoma, and gastric cancers. The most frequently altered genes included SYNE1(12%), CTNNB1 (4%), PPMD1 (3%), BCR (3%), ETV6 (3%), RAD52 (3%), and H3-3A (3%). Across TCGA pan-cancer and tumour-specific cohorts, EOCS alterations were consistently associated with inferior overall survival irrespective of age at diagnosis, including pan-cancer (68 vs 83 months; p = 9.1 × 10⁻³), breast cancer (142 vs 164 months; p < 0.04), colorectal cancer (45 vs 56 months; p = 1.75 × 10⁻⁴), and prostate cancer (56 vs 77 months; p = 2.0 × 10⁻⁴). Conclusions: This pan-cancer Early-Onset Cancer Signature identifies distinct genomic pathways driving tumourigenesis in EOC. The findings provide novel biological insights into disease pathogenesis and highlight potential diagnostic and therapeutic targets, supporting the clinical relevance of EOCS across multiple cancer types.

Article Details

Volume / Issue Vol. 44, Issue 16_suppl
Published June 01, 2026
Pages 3126-3126
ISSN 0732-183X
Publisher Lippincott Williams & Wilkins

Journal Info

Journal of Clinical Oncology

Lippincott Williams & Wilkins

ISSN: 0732-183X Health Sciences

Authors (6)

M

Mahesh Iddawela

Monash University Faculty of Medicine, Nursing, and Health Sciences, Melbourne, VIC, Australia

M

Mei Sim Lung

Peninsula Health, Frankston, VIC, Australia

R

Ryan De Ruyter

Peninsula Clinical School, Monash University, Frankston, VIC, Australia

N

Nandan Prasanna Kumar

Peninsula Clinical School, Monash University, Frankston, VIC, Australia

S

Sally Heath

Peninsula Health, Frankston, VIC, Australia

B

Babak Tamjid