Shaping the future of oncology: Key insights from ESMO 2024

340,000—that's my estimate of how many coffees were served to the 34,000 participants at this year’s ESMO Congress in Barcelona. Over five intense days of presentations, discussions, and networking, attendees relied on caffeine to keep up with the pace of new developments! 

This year's event was packed with cutting-edge breakthroughs and a clear consensus on the importance of using artificial intelligence (AI) to shape the future of precision oncology. From novel therapeutic strategies like bispecific antibodies and optimized combination therapies to the emergence of multimodal AI in cancer research, the 2024 ESMO edition reflected the rapidly evolving landscape of oncology and the increasing complexity of treating cancer effectively. During his opening speech, ESMO President Pr. Andres Cervantes,  emphasized the need for specific treatment approaches and for embracing emerging trends and technologies, such as AI, to drive cancer research forward. In line with this approach, ESMO has announced the launch of a new congress dedicated to AI and digital oncology, scheduled to take place in Berlin in November 2025.

We were especially pleased this year to attend the congress and contribute to the exciting new developments in cancer research by introducing MOSAIC Window: an incredible opportunity for researchers worldwide to look inside our MOSAIC initiative, the largest spatial omics and multimodal dataset in oncology. After reaching the 1,000-patient milestone, we are now opening a 60-patient subset accessible through the European Genome-Phenome Archive (EGA). This subset includes data from 5 cancer indications, allowing researchers to work with cutting-edge modalities like spatial and single-cell omics to conduct proof-of-concept studies before scaling to larger projects. Additionally, we presented a scientific poster showcasing our work on the use of deep learning to detect tertiary lymphoid structures on digitized histology slides and their prognostic value in sarcoma.

One of the standout themes of the conference was the progress being made in the field of bispecific antibodies. These engineered molecules can bind to two antigens on cancer and/or immune cells simultaneously. This makes them interesting candidates for combining the benefits of two medicines in one, offering a higher level of precision that could make treatments more effective and potentially less toxic. Bispecific antibodies were highlighted as a promising tool to increase the immune system's ability to identify and kill cancer cells, improve tumor targeting, and overcome the limitations of single-target therapies. 

Another primary focus in multiple presentations was the necessity for optimized combination therapies. Many key opinion leaders at ESMO emphasized the importance of drug combinations to keep pace with the complex and adaptive nature of cancer that single therapies often struggle to address. Combinations are proving to be essential in overcoming resistance mechanisms and extending the benefits of treatment to more patients. However, with more sophisticated therapeutic approaches becoming available, designing the best combinations is not trivial. This is where AI could make a significant difference. We are exploring this through our AI drug positioning engine, DrugMATCH.

While bispecifics and combination therapies drew significant attention, the conference also showcased several growing areas of interest that could make an impact in the years to come. Radioligand therapies, for instance, offer a novel approach to targeting tumors. By attaching radioactive particles to molecules that specifically target cancer cells, radioligand therapies deliver a localized dose of radiation while limiting impact on surrounding healthy tissues. Another approach discussed was synthetic lethality, which takes advantage of cancer cells' unique genetic vulnerabilities. By exploiting specific genetic interactions, synthetic lethality allows for targeted killing of cancer cells without affecting normal cells—opening new possibilities for patients, especially those with otherwise limited therapeutic options. PARP (poly (ADP-ribose) polymerase) inhibitors were the first example of a treatment exploiting synthetic lethality in BRCA-mutated cancers. Professors Andrew Tutt, Institute of Cancer Research (London, UK), and Alan d’Andrea, Dana Faber Cancer Institute (Boston, US), specifically discussed understanding and overcoming PARP inhibitor resistance, also highlighting the need for drug combinations. 

Throughout the conference, one of the overarching messages was the critical role of biomarkers and the need to better understand resistance mechanisms in cancer. Treatments are only as effective as our ability to predict which patients will benefit from them. Numerous talks emphasized the growing need for reliable biomarkers to guide therapeutic decisions, track treatment response, and anticipate resistance. This is another area where AI and multimodal data can make a difference. According to Pr. Jakob Nikolas Kather, Technical University Dresden (Dresden, Germany), during his keynote lecture: we should use AI to improve decision-making in precision oncology. At Owkin, our work is aligned with this mission. We are deeply invested in developing AI methods to integrate and analyze large, multimodal datasets that include cutting-edge data modalities like single-cell and spatial omics. By doing this, we aim to uncover patterns and insights that are impossible to detect with traditional and single-modality approaches and ultimately contribute to developing more effective and personalized treatments. 

We live in an era where we can generate more data and use technologies that are more advanced than ever before. Several presenters emphasized the need for regulatory frameworks to evolve and keep pace with these innovations. Adapting regulatory pathways and approval mechanisms is critical to ensure that new treatments reach patients faster to avoid unnecessarily prolonged approval times. As regulators begin to see the value of incorporating AI across the drug discovery and development process, Amy Abernethy, a former FDA official, pointed out that for clinicians to overcome skepticism and embrace AI, algorithms must be reliable and robust at the individual patient level. 

In conclusion, this year’s ESMO Congress was another great success. From the rise of bispecific antibodies to the growing interest in selecting optimal combination therapies and the use of AI and data-driven approaches, the field is moving towards cancer treatments that are becoming increasingly sophisticated. 

We are proud to be at the forefront of AI research in oncology, and with a fresh cup of coffee in hand, we look forward to continuing to drive new AI-driven discoveries in the years to come!

Ariane Peyret, Discovery Solutions Manager, Pharma BD