SOURCE-LINKED INTELLIGENCE

LUng Modeling and INtelligence for Advanced care: A Digital Twin Approach to Lung Cancer Diagnosis and Treatment

CORDIS · observation · Publication date unknown

tions, reducing the need for invasive biopsies even when using incomplete patient datasets, such as when a full genetic profile is unavailable. This beyond–state-of-the-art approach combines advanced Generative AI methods with multimodal data fusion to model personalized disease trajectories based on nodules and tumors dynamics. The core of this innovation is the Physics-Informed Attention Network, which employs a dual-pathway mechanism to synergistically fuse data-driven learning, enabling the capture of dynamic patterns in lung cancer and the generation of future CT images. Additionally, it incorporates a pointer-linked graph to integrate uneven data samples. LUMINA takes an ambitious step forward by integrating longitudinal, multimodal lung cancer data from both Europe and Asia. An initial model will be developed for risk stratification (focused on non-small cell lung cancer) and immunotherapy response prediction using multi-center data augmented with synthetic samples, then generalized across lung cancer subtypes and treatment regimens, and finally validated to demonstrate clinic

Read original source ↗ Open in workspace

recordType

award

status

SIGNED

region

EU

value

3998943.75

unit

EUR