Recent studies reveal a “dual evolutionary” pattern in pancreatic cancer similar to that of ovarian cancer: the HRD pathway is primarily characterized by DNA deletions, while the WGD pathway drives tumor malignancy through chromosomal fragmentation and ecDNA amplification. Key findings include that mitochondrial mutations significantly affect prognosis (with a hazard ratio of 3.1), and CDK12 mutations can serve as independent adverse prognostic indicators. These findings provide a molecular basis for developing targeted metabolic interventions and synthetic lethality therapies, especially for WGD-type pancreatic cancer patients who are resistant to conventional chemotherapy. This study was published in Nature Communications (14.7).
High-grade serous ovarian cancer (HGSOC) is the deadliest gynecological malignancy. Although PARP inhibitors significantly improve the prognosis of patients with homologous recombination deficiency (HRD), approximately half of HRD-negative patients still face significant treatment challenges. Behind this clinical dilemma lies the extremely complex landscape of genomic structural variations in HGSOC, which has not been fully elucidated—from large-scale chromosomal rearrangements to mitochondrial genome mutations. How these variations collectively shape tumor evolutionary trajectories and clinical outcomes remains a core scientific question that needs to be addressed in the field.
An international team led by the MRC Human Genetics Unit at the University of Edinburgh published this landmark study in Nature Communications. By integrating whole-genome sequencing (WGS) data from 324 cases of HGSOC from Scotland, Australia, Canada, and the United States, the researchers have for the first time mapped a panoramic view of structural variations in HGSOC, revealing two distinctly different tumor evolutionary pathways: the HRD-dominated genome is characterized by deletion variations, while the WGD-dominated genome gains amplification advantages through complex events such as chromosomal fragmentation, breakage-fusion-bridge cycles (BFB), and extrachromosomal DNA (ecDNA).
Key technical methods include:
- Deep whole-genome sequencing of 324 HGSOC samples from multiple centers (median coverage 71X) and integrated analysis of RNA-seq data
- Using ShatterSeek to detect chromothripsis, AmpliconArchitect to predict ecDNA, and gGnome to analyze BFB and other eight types of complex structural variations (cSV)
- Identification of copy number variations (CNA) and structural variation (SV) hotspots based on GISTIC and FishHook algorithms
- Mitochondrial genome mutation analysis using the RtN! algorithm to eliminate nuclear-mitochondrial pseudogene (NUMT) interference
Main research findings:Extreme structural diversity generates “collateral damage” across the genomeGenomic analysis shows that 89% of gene disruptions are driven by repetitive variations, but most high-frequency variant genes (such as NAALADL2) are located at fragile sites and are considered “passenger mutations.” The true driving events are concentrated in key genes such as CCNE1, whose amplification is achieved through complex mechanisms like BFB, resulting in a 2.8-fold increase in expression.
Dual evolutionary pathways driven by HRD and WGDHRD samples are enriched in deletion variations and chromosomal interlocks (chromoplexy), while WGD samples exhibit a co-occurrence of chromosomal fragmentation (49%), BFB (27%), and ecDNA (19%). Notably, severe chromosomal fragmentation involving more than two chromosomes is associated with better prognosis.
Prognostic value of mitochondrial mutation burden26% of samples carry pathogenic mutations that disrupt mitochondrial complex I (CI) function, with high heterogeneity levels increasing the risk of death by 3.1 times (HR=3.1, p=0.0002). This effect is particularly pronounced in WGD tumors, highlighting the importance of nuclear-mitochondrial genome interactions.
Multi-omics driven clinical translation modelElastic net regression models confirm that CDK12 mutations (HR=2.1) and mitochondrial mutations are independent prognostic factors, while CEP89 amplification and severe chromosomal fragmentation indicate poor and good prognosis, respectively. This model provides a new standard for precise stratification of HRD-negative patients.
This study fundamentally changes the understanding of the genomic complexity of HGSOC:
- For the first time, it systematically demonstrates the causal relationship between structural variation hotspots and expression changes, identifying eight novel candidate driver genes such as WRN.
- It reveals that WGD is both an engine of genomic instability and a “buffer” that tolerates extreme variations, explaining the heterogeneity of WGD prognosis.
- It establishes mitochondrial mutations as the first cross-omics prognostic biomarker, providing a theoretical basis for metabolic interventions.
- The proposed “dual evolutionary” model points the way for developing synthetic lethality strategies targeting HRD-negative tumors, particularly treatment combinations targeting CCNE1 amplification and CDK12 loss.
These findings not only address significant challenges in the molecular classification of HGSOC but also provide important insights into the genomic evolutionary patterns relevant to other highly heterogeneous cancers (such as triple-negative breast cancer and pancreatic cancer). As WGS technology enters clinical practice, the integrated analytical framework established by this study will accelerate the transition to precision medicine in ovarian cancer.