CoCross: An ICT Platform Enabling Monitoring Recording and Fusion of Clinical Information Chest Sounds and Imaging of COVID-19 ICU Patients

Vassilis Kilintzis, Nikolaos Beredimas, Evangelos Kaimakamis, Leandros Stefanopoulos, Evangelos Chatzis, Edison Jahaj, Militsa Bitzani, Αnastasia Kotanidou, Aggelos K. Katsaggelos, Nicos MaglaverasHealthcare 2022 · DOI

TL;DR

This paper introduces CoCross, an ICT platform designed to collect, fuse, and analyze multimodal data (chest sounds, imaging, and clinical parameters) from COVID-19 patients in the ICU. By creating a dynamic, annotated database, CoCross aims to improve understanding of the disease's pathophysiology and support clinical decision-making, offering a much-needed tool for managing critically ill patients.

What Problem Does This Paper Address?

The COVID-19 pandemic highlighted significant challenges in monitoring and treating severely ill patients in Intensive Care Units (ICUs). Understanding the complex pathophysiology of the disease and its progression requires high-quality, multi-parameter, and annotated databases. Traditional monitoring methods are often cumbersome in the ICU setting due to the need for personal protective equipment, and existing data collection systems lack the integration of diverse clinical information like chest sounds and imaging. This gap hinders comprehensive analysis and the development of advanced diagnostic tools. The paper addresses the need for an integrated, user-friendly system that can capture and fuse various data streams from ICU patients, facilitating real-time assessment and long-term research.

Methods

CoCross is a three-component system:

  1. CoCross4Pros Android Application: A modular app that manages the interaction with portable medical devices (e.g., Bluetooth stethoscopes, pulse oximeters) within the ICU. It handles user authentication, recording session management, and local data storage.
  2. Cloud-based Data Management Services: This backend infrastructure utilizes HL7 FHIR (Fast Healthcare Interoperability Resources) and ontologies to store and manage the collected data. It ensures secure data handling, authorization, and provides a RESTful API for client applications. The data is stored as RDF graphs serialized in Turtle format.
  3. Web-based Application for Intensivists: This application allows healthcare professionals to access and analyze the acquired data in real-time from outside the ICU. It provides a dashboard view of patient information, allows playback of auscultations, and viewing of DICOM imaging series.

The system emphasizes ease of deployment, scalability, and interoperability, using technologies like Docker for containerization and Keycloak for authentication.

Datasets

The paper describes the creation of a new, dynamic, and annotated database of multisource data from COVID-19 ICU patients.

  • CoCross Database: This database integrates clinical information, chest sounds (lung and heart), and diagnostic imaging (X-rays, lung ultrasound).
  • Data Acquired: As of the paper's publication, data from 176 ICU patients was collected, representing an average monitoring period of five days. This includes a dataset with 3477 distinct 20-second auscultations.

The paper does not explicitly mention pre-existing datasets used for training or evaluation, but it references the ICBHI respiratory database as a precedent for creating annotated lung sound datasets.

Key Insights and Ideas

  • The Power of Data Fusion: The core idea is that combining diverse data streams (sounds, images, vital signs) provides a more holistic and accurate picture of a patient's condition than any single modality alone. This is crucial for understanding complex diseases like COVID-19.
  • Bridging the ICU Gap: CoCross effectively overcomes the physical barriers and logistical challenges of data collection within an ICU, especially during a pandemic, by enabling remote access and analysis for clinicians.
  • Real-time Clinical Support: The platform's ability to provide near real-time data access and analysis empowers intensivists to make more informed and timely decisions, potentially leading to better patient outcomes.
  • Foundation for Future Research: The annotated, multisource database generated by CoCross is a valuable resource for developing advanced AI/ML models for COVID-19 diagnosis, prognosis, and treatment optimization.
  • User-Centric Design: The platform's design prioritizes user experience, aiming for ease of use and minimal technical support, which is critical for adoption in demanding clinical environments.

Key Results

  • Successful Deployment: CoCross was successfully deployed in two COVID-19 reference hospitals in Greece (Thessaloniki and Athens) starting in June 2020.
  • Data Acquisition: Data from 176 COVID-19 ICU patients was collected, with an average monitoring period of 5 days per patient. This resulted in a dataset of 3477 distinct 20-second auscultations.
  • Technical Robustness: The system experienced no technical issues or downtime during the first nine months of deployment, demonstrating its reliability.
  • Data Accessibility: Recorded data was accessible via the web application for intensivists within seconds of acquisition. Auscultation recordings were uploaded to the backend within approximately 40 seconds.
  • User Acceptance: The platform was well-accepted and positively rated by users, with qualitative feedback and quantitative survey results (UEQ-S score of 1.725) indicating it is an inventive, easy-to-use system that efficiently addresses user needs.

Limitations and Future Work

  • Limitations:
    • The current system supports only one stethoscope model.
  • Tools for monitoring the expansion of the generated database are not yet integrated into the web application for intensivists.
  • Future Work:
    • Aggregation and preliminary analysis of the data to assess database quality and completeness.
  • Adoption of new devices for auscultation that offer better signal quality.
  • Expansion of the platform to include additional ventilator parameters and recording of interventional events (e.g., intubation, extubation).
  • Leveraging the CoCross database for machine learning and clinical research protocols to advance the Integration Information Initiative (III).

Summary updated

Abstract

Monitoring and treatment of severely ill COVID-19 patients in the ICU poses many challenges. The effort to understand the pathophysiology and progress of the disease requires high-quality annotated multi-parameter databases. We present CoCross, a platform that enables the monitoring and fusion of clinical information from in-ICU COVID-19 patients into an annotated database. CoCross consists of three components: (1) The CoCross4Pros native android application, a modular application, managing the interaction with portable medical devices, (2) the cloud-based data management services built-upon HL7 FHIR and ontologies, (3) the web-based application for intensivists, providing real-time review and analytics of the acquired measurements and auscultations. The platform has been successfully deployed since June 2020 in two ICUs in Greece resulting in a dynamic unified annotated database integrating clinical information with chest sounds and diagnostic imaging. Until today multisource data from 176 ICU patients were acquired and imported in the CoCross database, corresponding to a five-day average monitoring period including a dataset with 3477 distinct auscultations. The platform is well accepted and positively rated by the users regarding the overall experience.
BibTeX
@inproceedings{kilintzis2022cocross,
  author = {Vassilis Kilintzis and Nikolaos Beredimas and Evangelos Kaimakamis and Leandros Stefanopoulos and Evangelos Chatzis and Edison Jahaj and Militsa Bitzani and Αnastasia Kotanidou and Aggelos K. Katsaggelos and Nicos Maglaveras},
  title = {{CoCross: An ICT Platform Enabling Monitoring Recording and Fusion of Clinical Information Chest Sounds and Imaging of COVID-19 ICU Patients}},
  booktitle = {Healthcare},
  year = {2022},
  doi = {10.3390/healthcare10020276},
}
Figure