Since its start in September 2023, MR-O-MICS has progressed ahead of schedule in clinical integration and data generation. Following ethical approval, patient recruitment began in December 2023. To date, over 100 women with high-grade serous ovarian carcinoma have been included, and more than 500 tissue samples and matched MRI datasets have been acquired.
At the Institut du Cancer de Montpellier, a new research-dedicated 3-tesla MRI workflow has been implemented in routine care. Every eligible patient now undergoes MRI the day before surgery, creating a seamless collaboration between radiologists and surgeons. These MRI “roadmaps” are already used to guide cytoreductive surgery, marking one of the earliest examples of immediate clinical translation within an ERC Starting Grant.
The project also developed a high-field 9.4-tesla MRI protocol for resected tumour specimens. Using custom 3D-printed moulds, the team can precisely align imaging slices with histology and digital pathology. This innovation enables voxel-to-cell correspondence and provides an unprecedented multiscale view of tumour heterogeneity.
Radiomic and pathomic analyses have been launched using IBSI-compliant Python pipelines. Radiomic features are extracted from both 3-tesla and 9.4-tesla datasets, while digital histology is analysed with open-source tools such as QuPath. These data are now used to train explainable machine-learning models predicting stromal proportion and immune infiltration, validated against genomic homologous recombination deficiency (HRD) profiles.
The molecular component of MR-O-MICS has evolved toward spatial transcriptomics and single-cell sequencing, offering higher biological resolution and preserving tissue architecture. Eighteen spatial transcriptomic datasets from ovarian, omental, and peritoneal samples have already been generated, linking microscopic gene-expression patterns to MRI features.
A unique interdisciplinary ecosystem has emerged: radiologists, physicists, biologists, surgeons, and data scientists now work together in the newly established PINKCC laboratory (Precision Imaging as a New Key in Cancer Care). The ERC project has also fostered strategic partnerships with Siemens Healthineers (MRI sequence development), NVIDIA and Scaleway (cloud-based AI infrastructure), and AstraZeneca (support for molecular analyses).
In 2025, the team organised the first PINKCC Ovarian Cancer Challenge — a European data-science competition that attracted more than 350 researchers to develop AI algorithms for lesion segmentation. This initiative reinforced open science, strengthened the European AI-in-oncology network, and directly improved the project’s radiomics pipeline.
Beyond research, MR-O-MICS has catalysed institutional transformation. The Principal Investigator, now Full Professor of Radiology and AI Scientific Chair at the Institut du Cancer de Montpellier, leads clinical AI implementation and chairs the European Society of Urogenital Radiology (ESUR) Female Pelvic Imaging Working Group. These roles ensure direct dissemination of ERC-funded knowledge into clinical standards and European guidelines.
