RACOON is delighted to announce its extension! The Radiological CooperativeNetwork is the imaging data platform of the Network of University Medicine (NUM). It consists of decentralised RACOON nodes at all academic medical centers and a RACOON central unit. This network makes it possible to record image and clinical data in a structured manner and to analyse it in a federated and centralised manner. Modern platforms are used to realise AI-supported segmentation and radiomics workflows.
RACOON thus creates the basis for
- the provision of quantitatively valid image data
- the development of AI tools for diagnostics and prognosis
- comprehensive research on nationwide cohorts
- transfer of results into clinical routine and early warning systems
The forthcoming extension is recognition of our joint work - and an incentive to continue on our successful path.
With this in mind, we look forward to sharing some insights and success stories from our collaboration to date with you shortly.
Progress in the sub-projects:
RACOON FADEN - Early detection of Adenomyosis and Endometriosis
Although both diseases are common, diagnosis remains difficult. The FADEN project uses MRI-based imaging to detect structural features of the uterus and correlates these with clinical symptoms. So far, 339 MRIs have been performed on 145 patients and 78 volunteers. Initial findings indicate that the contractility of the uterus could be a more meaningful biomarker than the thickness of the junctional zone, which has often been used to date.
RACOON-RESCUE - Radiological care for childhood NHL
To improve the diagnosis of non-Hodgkin's lymphoma (a group of cancers of the lymphatic system) in children, a Germany-wide, pseudonymised image and clinical dataset with up to 980 cases is being created. An AI-supported workflow, new segmentation tools and a structured reporting scheme have been successfully implemented. The first publications are in preparation.
RACOON CORE-PE - AI for pulmonary embolism prognosis
The aim is to improve AI-supported individual risk assessment in acute pulmonary embolism. Over 1,600 patients were included in a multi-site setting. In addition to structured reporting, AI-based image segmentation is currently being carried out. Initial results are expected shortly.
RACOON PDAC & PRECISE-MD - Patient-specific therapy decisions in pancreatic cancer and hepatocellular carcinoma
The aim is to use extensive real-world data to enable precise prediction of therapy response in pancreatic adenocarcinoma and hepatocellular carcinoma. The study will combine image and clinical data to identify predictive biomarkers. These will serve as the basis for well-founded, patient-specific therapy decisions in clinical practice. The statistical analysis phase has now begun and the first promising results are expected shortly.