AI Wearable Enables Continuous Noninvasive Arterial Blood Pressure Monitoring

Continuous, beat-to-beat blood pressure monitoring in intensive care units often relies on arterial lines, which are invasive and carry risks such as bleeding, clotting, and infection. Intermittent cuff readings are safer but cannot capture rapid hemodynamic changes that threaten organ perfusion. Clinicians need noninvasive, continuous monitoring that approaches invasive accuracy to guide timely interventions. Researchers have now created a wearable sensor and AI system intended to deliver continuous blood pressure waveforms without arterial catheterization.

Rsearchers from Johns Hopkins University and Johns Hopkins Medicine have developed MOSAIC, a two-sensor platform for continuous arterial blood pressure estimation. The system is positioned as a potential alternative to arterial catheters used in intensive care units (ICUs) and operating rooms. Initial patient use indicates it could extend continuous blood pressure monitoring beyond critical care settings.

MOSAIC uses a chest-mounted sensor and a finger sensor to capture the heart’s electrical activity and peripheral blood flow. These synchronized signals are processed by a deep learning model that reconstructs a continuous arterial pressure waveform. The approach aims to provide real-time, noninvasive monitoring comparable to invasive lines while preserving patient mobility.

In an initial study of 28 ICU patients at Johns Hopkins Hospital, MOSAIC generated blood pressure waveforms that closely matched those recorded by arterial catheters. The findings have been published in Computers in Biology and Medicine on August 15, 2026. The team is now validating the sensors and algorithm in a larger cohort of intensive care patients.

Because standard cuffs provide only intermittent measurements, a reliable noninvasive waveform could help clinicians detect instability earlier and tailor therapy. The investigators note potential applications on general wards and at home, including for people with hypertension who may benefit from continuous tracking similar to glucose monitoring in diabetes. The sensors might also enable studies of daily blood pressure patterns in healthy individuals.

"Patients in the ICU need continuous blood pressure monitoring to catch problems early, but it means an arterial line, which comes with a risk of bleeding, clotting and infection. We wanted to find a better way," said Carl Harris, Ph.D. student in biomedical engineering.

"We reconstruct waveform data in a way that's meaningful, accurate, reliable and, most importantly, non-invasive. It's a possible solution for avoiding the current standard of care for measuring blood pressure, arterial lines, a very invasive procedure with a risk of many complications," said Robert Stevens, chief of the Division of Informatics, Integration, and Innovation at Johns Hopkins Medicine.

Related Links
Johns Hopkins University