Luna Maris

Doctoral fellow – Medical Imaging and Signal Processing (MEDISIP) research group – UGent; Faculty of Engineering and Architecture
Research Engineer – XEOS Medical NV
Principal investigators: prof. Christian Vanhove (PhD) and prof. Vincent Keereman (PhD)
Research focus
In curative breast cancer surgery, complete tumor removal is vital to ensure no tumor cells remain within the breast. Unfortunately, in up to 20% of breast cancer patients, pathological assessment of the resected breast cancer specimen indicates that the tumor removal was incomplete. As pathology takes several days, these patients must undergo repeated surgery to remove remaining tumor cells. Recently, micro-PET-CT imaging of resected breast cancer specimens emerged as a promising technique for fast intraoperative visualization and verification of the resection margins. This technique requires preoperative administration of the radiotracer 18F-FDG to the patient. As 18F-FDG may show nonspecific uptake in metabolically active benign tissue, like inflamed tissue, interpreting micro-PET-CT images can be challenging in some cases. Hence, we are developing a decision-support system that can aid physicians to interpret micro-PET-CT images of resected breast cancer specimens. To reach this goal, we are developing techniques to co-register micro-PET-CT specimen images with the corresponding ground truth whole-slide histopathology images. These co-registered images provide insight into the radiotracer uptake patterns of different breast tissue types. In addition, we use these co-registered images to develop a deep learning model that can automatically segment breast tumors in micro-PET-CT images of resected breast cancer specimens. We hope that this technology will allow physicians to treat patients with more confidence and will avoid re-surgeries.
Biography
Luna Maris graduated as a Biomedical Engineer from Ghent University in 2021, where her master dissertation was focused on non-invasive prognosis of prostate cancer using deep learning on MRI. Since 2021, she works as a Research Engineer at XEOS Medical NV, where she conducts research on how we can use deep learning to facilitate the interpretation of micro-PET-CT images of breast cancer specimens and guide image-based decision making. In 2022 she started her PhD, focused on using deep learning to link micro-PET-CT images with the histopathological assessment of breast cancer specimens, in a collaboration between XEOS and the MEDISIP research group of UGent.
Key publications
- ‘Direct co-registration of [18F]FDG uptake and histopathology in surgically excised malignancies of the head and neck: a feasibility study’. European Journal of Nuclear Medicine and Molecular Imaging (EJNMMI), 2023. (PMID: 36854863)
Contact & links
- Lab address: UZ campus entrance 37a; Corneel Heymanslaan 10; 9000 Gent
- Medical Imaging and Signal Processing, MEDISIP
- XEOS Medical NV
- Google Scholar
- Luna Maris is interested to receive invitations for presentations or talks