The European Commission is calling for startups and SMEs with technologies and innovations that could help in treating, testing, monitoring or other aspects of the Coronavirus outbreak to apply urgently to the next round of funding from the European Innovation Council.

As D/Vision Lab, we think that during a new and unknown pandemic there is always a lack of clinical knowledge, clinical research and scientific papers as well as clinical history. There is therefore a strong need for effective and fast communication among clinicians.

Closely monitoring the pandemic and actively communicating with all involved counterparts, using state of the art technologies, could help to tackle the pandemic. Clinicians and medical staff can share anonymized patient data and medical images all over the world in order to build a vast dataset of clinical cases, treatments and outcomes. A real-time Artificial Intelligence (AI), powered by data clustering and visualization techniques, can be used in order to classify a new patient on the basis of the existing dataset, thus helping clinicians to provide additional information about patient status, treatment and potential outcome in order to reduce hospitalization and sudden death. A risk assessment using AI and statistical models can be done in order to predict potential disease treatment, hospitalization and outcome.

This project will be developed in collaboration with AISent, Bergamo University, Bergamo Hospital and Mario Negri Institute

The idea consists in a cloud-based platform to share clinical data among clinicians on an intra- and inter-center basis. In particular, the final user will be a clinician involved in the treatment of covid-19 patients. The clinician will be able to upload a set of structured data (i.e. clinical data, Digital Imaging and Communications in Medicine (DICOM images, audio files) that will be stored in a fully anonymized format into the main database. The data could also be collected directly from the clinical center internal Picture archiving and communication system (PACS) and electronic health records, to reduce the effort by the doctor.

After analyzing the uploaded case, the system will show to the clinician similar cases from the database, along with their clinical data and attached files (DICOM, images, audio). The main goal is to provide a structured data exchange and analysis for clinicians all over the globe during the spreading of the pandemic. The user is also encouraged to upload the treatment chosen for the specific case and the follow-up, so that other users can benefit from previous experiences.

The main goal will be reached through the implementation of different components:

  • The user interface with the platform will be a web app, containing data entry forms, dicom viewer, data visualization tools (i.e. mainly charts) and a chat to interact with other users.
  • Real-time chat and rapid sharing of clinical experience and clinical outcome.
  • The platform architecture will be cloud-based and fully scalable to meet increasing storing and computational demands from users.
  • Clustering methods will be applied on the dataset composed by the uploaded data off all users, in order to provide a space of similarity between patients.
  • Machine Learning / deep learning algorithms will also be applied to the dataset in order to provide short-term and long-term outcome predictions.

As a plus, this platform would allow the creation of a structured, labelled and accessible dataset of medical records. The dataset collected through the platform could be queried by external organizations, for research and scientific purposes, using a REpresentational State Transfer (REST) Application programming interface (API). Moreover, the data can also remain stored on a single institution data warehouse and be exchanged with the users through an open protocol. This meets particular security issues on data treatment.

In conclusion, this project aims to solve the problem of sharing clinical expertise in a structured, fast, and global way during a fast-paced event as a pandemic, with the support of artificial intelligence to help choose the proper treatment for a given patient.