Annabelle Mürb, Working Student at Flatiron Health, Doctoral Candiate at the Universitätsklinikum Heidelberg (UKHD), shared on LinkedIn:
“Excited to have presented my first poster at the AI and Digital Oncology Congress 2026 by ESMO – European Society for Medical Oncology!
Our work: ‘LLM-based Analysis of Longitudinal Patient Narratives following Loco-regional Breast Cancer Therapy. A secondary analysis of the international, randomized CINDERELLA trial.’
Patients’ free-text narratives capture experiences that structured PROs often miss, but they’re hard to analyze at scale. We tested whether open-weight LLMs can turn them into clinically meaningful insights. Both models reproduced clinically meaningful conclusions, with Llama 3.3 70B performing significantly better overall.
Huge thanks to my supervisor André Pfob for his guidance and support, to the Cinderella Project team and all co-authors across Europe, and to Universitätsklinikum Heidelberg (UKHD) and the Institute for Medical AI. I’m looking forward to taking this research further!”
About AI and Digital Oncology Congress 2026
AI and Digital Oncology Congress 2026, organised by the European Society for Medical Oncology (ESMO), is an international scientific event focused on the role of artificial intelligence and digital technologies in cancer research and clinical practice.
- Organiser: European Society for Medical Oncology (ESMO).
- Focus: Artificial intelligence, digital health and innovation in oncology.
- Key topics: AI-powered diagnostics, digital pathology, precision oncology, data analysis and clinical decision support.
- Research: Presents new studies and developments exploring how AI can improve cancer research and patient care.
- Audience: Medical oncologists, researchers, healthcare professionals, data scientists and technology experts.
- Purpose: To encourage collaboration and explore how digital innovation can advance cancer prevention, diagnosis, treatment and care.
You may also be interested in:
Humaid Al-Shamsi: Nexomic Brings the Largest Team Representation to ESMO AI 2026