The feasibility of utilising a cloud-based storage and analysis software in multicentre clinical electrophysiology research

FUNDING ACKNOWLEDGEMENTS: Type of funding sources: Private grant(s) and/or Sponsorship. Main funding source(s): BioSense Webster Investigator Initiated Study (IIS) Grant. INTRODUCTION: The COVID-19 pandemic enforced long-term changes in healthcare towards remote working, implementing digital systems that support the safe online sharing of clinical information. A cloud-based storage and analysis software has been developed that uploads 3D mapping cases within the CARTO-3 system for remote review. In addition, it uses artificial intelligence/ machine learning to analyse large procedural datasets. We tested the feasibility of this technology as a future research tool in analysing the first 8 cases recruited to the Ripple AT PLUS study. METHODS: The Ripple AT PLUS trial is a planned multicentre randomised trial comparing ablation outcomes in scar based atrial tachycardia (AT) using different CARTO mapping approaches. The study recently commenced recruitment (September 2022). The CARTO files for the first 8 cases recruited into this study from different hospitals were securely transferred from CARTO 3 workstations to the Siemens teamplay gateway. Patient identifiers were removed, and anonymized datasets were successfully uploaded onto CARTONET Microsoft Azure cloud. A secure Web-site (https://eu.cartonet.net/login) was accessed for remote review by the research team. RESULTS: Uploaded study cases were readily separated from other clinical cases using a research filter. Figure 1 summarises graphical datasets for these study cases presented by CARTONET, (including total procedural duration, mapping vs ablation times, ablation lesions). Their corresponding numerical values were exportable as a Microsoft Excel spreadsheet for individualised review. More detailed mapping and ablation analytics within each case was feasible – this is exemplified in figure 2A, which presents a histogram of the proportion of acquired electrogram bipolar voltages below a modifiable cut-off (0.20mV in this example). Individual case data was easily extracted and combined from all study cases to test different hypothesis – again exemplified in figure 2B, which ascertains how tissue voltage might affect the amount of ablation delivered. CONCLUSION: CARTONET is a secure, simple to use, and efficient cloud-based AI/machine learning system that offers access to clinically relevant data-sets from the CARTO system remotely from the hospital premises. This allows for a variety of rapid analytics, with the potential to improve the efficiency of multicentre clinical electrophysiology research collaboration. [Figure: see text] [Figure: see text]