LUMICARE

Laser-Based Ultra-Resolution System for Multimodal Imaging of Tumors and Reactive Oxygen Singlets in Photodynamic Therapy
(BMFTR 2026-2028)

The project LUMICARE (Laserbasiertes Ultraresolution-System für das multimodale Imaging von Tumoren und reaktiven Sauerstoff-Singuletts in photodynamischer Therapie) develops an innovative theragnostic method for photodynamic therapy (PDT) of head and neck tumors. PDT uses photosensitizers (PS) that selectively accumulate in tumor cells and are activated by laser. This produces reactive oxygen species (ROS) that destroy tumor cells. However, there is currently no method that directly quantifies these short-lived molecules during treatment and controls the therapy in real time. LUMICARE therefore pursues the approach of developing a multimodal diagnostic and therapy system with closed-loop control that continuously monitors the effect and dosage. The aim is to achieve maximum therapeutic efficiency with minimal damage to healthy tissue. The core innovation is the combination of fluorescence lifetime imaging (FLIM) and infrared-sensitive quantum detection to simultaneously, contactlessly, and in real time detect PS concentrations, metabolic changes, and ROS formation. State-of-the-art detector technologies and modulated pulse lasers are applied.

A novel laser system is used both for PDT irradiation and for stimulating autofluorescence for monitoring purposes. Particular emphasis is placed on the development of a 3D in vitro tissue model that realistically replicates tumor and tissue conditions. It enables the parameterization of life imaging, the investigation of PS degradation processes, and the validation of monitoring parameters. Based on this, an AI-based control algorithm is developed that integrates sensor data, recognizes patterns in ROS formation, and dynamically controls light dosage. This makes PDT adaptive and customizable. The end result is a fully integrated, endoscopic PDT system that makes therapy effects measurable and controllable for the first time — a step from experience-based to data-driven standard practice in oncology.

As part of this project, the AILab develops AI-supported methods for analyzing photonic image data with a focus on FLIM. This includes data structures, neural models for image reconstruction with low photon counts, domain transfer for endoscopic images, performance optimization for real-time applications, and the development of an AI model for predicting therapy success while taking uncertainties into account.

Project Partners

References