Artificial intelligence for the detection of local signs of infections around catheters.
Presentation of the project
DeepCath aims to develop the first therapeutic decision support device for the risk of infection around a catheter coupling an intuitive interface and an image recognition algorithm.
In the current context, in Europe, the incidence density of infections linked to central venous catheters (CVC) varies from 1 to 3.1 per 1000 patients per day. In the United States, the National Nosocomial Infections Surveillance System (NNIS) estimates that there are 80,000 CVC-related bloodstream infections per day. These infections are associated with increased mortality, increased lengths of stay and increased hospitalization costs.
The implementation of an automatic system facilitating the monitoring of patients with a catheter, regardless of their place of care (hospitals, clinics and home) would make it possible to detect the risks of infections very early, reduce complications and would thus participate in all the monitoring strategies, beneficial for the patient.
DeepCath aims to become a simple digital tool, at the service of healthcare professionals and patients in hospitals and in the city, to diagnose local signs compatible with the diagnosis of catheter-related infections in patients. Its integration into current clinical practice will respond to the problems encountered related to the difficulty of early diagnosis, and will thus contribute to improving the management of patients suspected of having an infectious complication.
The project team
PU-PH Intensive care, Bichat Hospital APHP, OUTCOMEREA association
Avicenna Hospital APHP, OUTCOMEREA association
Doctor researcher, Geneva Hospitals
PhD. Computer sciences, ICUREsearch
Biostatistician, OUTCOMEREA association, ICUREsearch
Master’s degree in clinical study management, ICUREsearch
Web developer, ICUREsearch
20.01.2023 – Launch of the Léon Bérard center in Lyon
This Friday, January 20, we had an appointment at the Léon Bérard Center in Lyon (CLB) with Doctor Hervé Rosay, who has supported the DeepCath project since its beginnings!!! The CLB is a hospital specializing in oncology, it is a member of the UNICANCER network (national federation of centers for the fight against cancer). We met motivated and dynamic teams. A good omen for the completion of the catheter photo database!
02.11.2022 – MASTER II INTERNSHIP OFFER
An internship is available for a Research / Artificial Intelligence Engineering position.
Title: Development of an artificial intelligence algorithm for the recognition of local signs of infection around catheters from pictures.
19.10.2022 – The first phase of the DeepCath project is officially launched!
Are you a healthcare professional? Are you in contact with patients with implanted catheters? Do you want to advance medical research? So, REGISTER on https://deepcath.com to PARTICIPATE ACTIVELY in this ambitious project by helping us for this first phase where the objective is to collect a MAXIMUM of pictures of catheters!!!
18.10.2022 – DeepCath is displayed at the SPIADI Congress!
This year, the 4th SPIADI National Mission Day took place at the Palais des Congrès in Tours. A huge thank you to the organizers, in particular to Doctor Nathalie VAN DER MEE and Doctor Anne-Sophie Valentin, for their welcome. We took advantage of the forum offered to our team to present and share with the more than 500 participants, the DeepCath project. The program of these days was very rich with in particular interventions by high-level researchers (Information: https://www.spiadi.fr/missionday)
17.06.2022 – The DeepCath clinical study has obtained official ethical approval!
DeepCath has obtained the agreement of the OUEST V Personal Protection Committee! This is a prospective multicenter observational study. The study is also listed on ClinicalTrials.gov (NCT05440396).
– Main objective: To train an artificial intelligence model of image recognition through images of intravascular catheters from hospitalized and outpatients to automatically identify, from a digital photograph, the presence of local signs associated with an infection .
– Expected result: Validation of the deep learning algorithm for detecting local signs from a picture.