A Multi-Sensorization Approach to Improve Safety in Transesophageal Echocardiography

Giovanni Faoro, Izadyar Tamadon, Selene Tognarelli, Arianna Menciassi

Research output: Contribution to journalArticleAcademicpeer-review

81 Downloads (Pure)

Abstract

Real-time 3D transesophageal echocardiography (RT-3D TEE) allows 3D visualization of patient heart and catheters without exposing patient and operators to ionizing radiations. Nonetheless, during such procedures esophageal injuries occur due to improper probe manipulation and probe overheating. To tackle these problems, we propose a multisensorization approach to provide information on probe pose and temperature throughout the procedure. Electromagnetic (EM) tracking is fused with inertial sensing thanks to a finite state machine integrating Extended and Incremental Kalman filters. This approach allows for a statistically significant improvement in static tracking with respect to standard EM, as reported by the Mann-Withney test. A novel sensor fault detection based on angular velocities discrepancy allows for robust tracking under different electromagnetic interferences, such as the one provided by ferro-, dia- and paramagnetic materials occupying the interventional room. Fiber optic technology is exploited for temperature estimation, taking advantage of its immunity to EM fields and the possibility of distributed sensing. Performances are compared with a commercial thermistor to guarantee feasibility and a root mean square error of 1.59 ∘ C is finally reported. We believe that these results demonstrate how sensing technologies can be integrated in TEE-guided surgical procedures to improve overall outcome and safety.
Original languageEnglish
Pages (from-to)829-838
Number of pages10
JournalIEEE Transactions on Medical Robotics and Bionics
Volume6
Issue number3
DOIs
Publication statusPublished - Aug 2024

Fingerprint

Dive into the research topics of 'A Multi-Sensorization Approach to Improve Safety in Transesophageal Echocardiography'. Together they form a unique fingerprint.

Cite this