Integrating deep proteomics into precision oncology for multi-omics characterization of metastatic cancer.

Abstract

e15065 Background: Precision oncology is enhancing cancer treatment by tailoring therapies to specific genetic and molecular profiles. However, identifying suitable treatment targets and resistance mechanisms remains challenging in many cases. In Denmark, Rigshospitalet's Phase 1 Unit is conducting the Copenhagen Prospective Personalized Oncology (CoPPO) study to provide comprehensive sequencing for metastatic cancer patients, helping with targeted treatments [1]. Our study enables the integration of multi-omics information from the same biopsy, facilitating a comprehensive molecular understanding of metastatic cancer and advancing multi-omics-based precision oncology, with the aid of mass spectrometry (MS)-based proteomics. Methods: Patients with advanced solid tumors and exhausted treatment options have been enrolled in CoPPO, receiving genomic profiling through whole genome and RNA sequencing. These profiles are evaluated in national tumor board meetings to determine appropriate targeted therapies. Around 3,000 samples collected from 2016 to 2023, were further processed with the automatic system KingFisher and analyzed using the automated high throughput Evosep One system coupled with a high-sensitive Orbitrap Astral mass spectrometer. Artificial intelligence (AI) was employed to develop cancer classifiers and aid in the interpretation of high-dimensional quantitative protein profiles. Results: The refined workflow yielded high protein coverage and robust proteomic profiles, demonstrating resilience to up to 7 years of sample storage. Across around 3,000 samples, 15,755 unique protein groups were quantified at a 21-minute gradient, providing comprehensive proteomic data. Machine learning analysis accurately predicted primary cancer origin from metastatic lesions, with an area under the curve (AUC) of 0.84–0.93. Subgroup analysis of patients resistant to BRAF-targeted therapy, revealed resistance mechanisms and potential new drug targets, complementing DNA and RNA-based profiling. Further details from the ongoing in-depth analysis will be presented at the meeting. Conclusions: The integration of deep proteomics into precision oncology has revealed the proteomic landscape of metastatic cancers with high sensitivity and rapid turnaround times. This approach provides valuable insights into tumor behavior, resistance mechanisms, and potential therapeutic targets. Combining proteomics with AI-driven molecular tumor profiling offers a powerful tool for advancing precision oncology. Prospective clinical trials are needed to further validate the clinical utility of this integrated approach in refining cancer prognosis and personalizing treatment strategies. Ultimately, this methodology supports real-time personalized therapeutic strategies for cancer patients, aiming to overcome drug resistance and improve outcomes.