Applications in Oncology

Aims

The medical objectives are to enhance diagnostic and treatment strategies for tumors using advanced MRI techniques. This involves developing methods to assess tumor heterogeneity and identify markers of treatment resistance, thereby improving prognosis. A key goal is to implement a non-invasive digital biopsy to evaluate crucial tumor characteristics and oxygenation levels without the need for contrast agents. Additionally, the aim is to apply these MRI advancements to tailor adaptive radiotherapy treatments based on individual tumor biology, improving precision and effectiveness in clinical settings. The approach also seeks to facilitate MRI-only radiotherapy planning, offering a reliable and innovative alternative for personalized cancer therapy. 

MRI

....

Diagnosis and prognosis from routine clinical images

The richness of data from existing protocols in clinical practice is still underexploited. We are therefore exploring:

1. increasing the robustness and reliability of current measurements extracted from conventional imaging systems;
2. integrating complex descriptors using representative learning solutions;
3. exploiting advanced statistical models to improve diagnosis and prognosis.

The entirety of our imaging protocols is designed to be conducted within clinical protocols involving patients, with sessions not exceeding 30 minutes. This setup allows us to validate the diagnostic and prognostic values of our approaches on patients and cohorts expertly managed at the Léon Bérard Center.

Optical guided cancer surgery

Interventional imaging is crucial in surgery guiding and impacts the clinical decision-making regarding surgical interventions. The shift towards personalized treatment planning relies on the ability to predict the accuracy and reliability of intraoperative biomarker imaging. The development and clinical implementation of predictive digital twins' models mark a shift towards data-driven, personalized neurosurgery. Traditionally, decisions regarding use of intraoperative imaging to guide have been based primarily on general guidelines and the surgeon’s experience. We develop predictive multimodal (MRI-Optics) biomarker modeling to support these decisions by quantitative data specific to each patient's pathology characteristics, enhancing the precision of surgical interventions and supporting clinical decision-making.
Our work focuses on Glioma neurosurgery addressing clinical issues of tumour margin delineation trough the proof of concept clinical studies GLIOSPECT ref. NCT02473380 and FLUOSPECT with HCL.
We proposed original approaches to identifying biomarkers using machine learning methods (Leclerc2020) and original model for fluorescence spectroscopy quantification without a-priori on parasitic fluorophores based on a multi-excitation fluorescence spectroscopy approach (Gautheron2023).