54P - Integrative omics workflow identifies credible druggable targets in high-grade serous cancer

Background

The ovarian high-grade serous carcinoma (HGSC) lags in the exploitation of personalized biological therapies, while targeted inhibition of cancer driver events has improved the patient care in other solid tumor types. This ows to the copy-number driven HGSC pathogenesis, leading to few short driver mutations. We introduce a systematic multi-omic workflow to integrate evidence from genomic and transcriptomic data to identify credible, recurring, and targetable driver events.

Methods

The multi-omics workflow was piloted with data from 831 cancer samples of 290 patients with HGSC, enrolled to the ongoing prospective and longitudinal DECIDER study at the Turku University Hospital. We assembled short mutations, copy number changes, large indels, and translocations, and classified them as potential loss or gain of function. We evaluated the events by their heterogeneity and direct and downstream consequence on the RNA level, using prediction algorithms and databases as a support. The credible oncogenic events were overlayed with targets and biomarkers for cancer medicines listed in the Open targets and OncoKB databases, to identify promising treatment modalities for testing in patient-derived organoids.

Results

The most frequent oncogenic aberrations affecting the MAPK and PI3K patways included NF1 loss and AKT2 amplification, affecting 19% and 9% of cases, followed by PTK2 or KRAS amplifications or PTEN loss, each affecting about 5% of cases. At least one of these aberrations was detected in 47% cancers. The frequency of the aberrations varied by the cancer homologous recombination (HR) status, so that NF1 losses were enriched in HR-deficient cancers, but the AKT2 amplifications were detected almost exclusively in HR-proficient cancers. In a drug sensitization screen for carboplatin with different targeted inhibitors of the MAPK pathway, the pan-RAS inhibitor RMC-7977 was effective for a patient-derived organoid with NF1 loss.

Conclusions

We developed a systematic and comprehensive omics workflow for analysis of a copy number-driven cancer. We demonstrated the utility of the integrated genomic and transcriptomic analysis in identifying druggable driver aberrations and validated the predicted sensitivity in a patient-derived organoid.

Editorial acknowledgement

Clinical trial identification

NCT04846933.

Legal entity responsible for the study

University of Helsinki.

Funding

European Union.

Disclosure

All authors have declared no conflicts of interest.