Wearable sensors closely match arterial lines in initial ICU tests

An arterial catheter sits inside an artery, tracking every rise and fall in a critically ill patient’s blood pressure. At Johns Hopkins Hospital, researchers are testing a less invasive alternative: two wearable sensors and an AI model produced continuous blood-pressure waveforms in 28 ICU patients, closely matching the catheter readings.

The system, called MOSAIC, uses one sensor on the chest and another on a finger. Together they capture the heart’s electrical activity and the flow of blood through the body. A deep-learning model then turns those signals into a waveform showing how blood pressure changes over time, rather than the occasional number delivered by a standard arm cuff.

That matters because blood pressure in intensive care can swing quickly. If it rises too high, it can contribute to stroke, heart attacks and kidney damage. If it falls too low, the brain and vital organs may not receive enough blood. Arterial lines provide the continuous view clinicians need, but inserting them carries risks of bleeding, clotting and infection and limits a patient’s mobility.

And so what, concretely? If larger tests confirm the first results, hospitals could monitor arterial blood pressure without putting a catheter into an artery. The researchers say the approach might eventually extend beyond intensive care to regular hospital wards and the home, including daily monitoring for people with hypertension.

The next step is already under way: Johns Hopkins is validating the sensors and algorithm in a larger group of ICU patients. The initial study shows the system closely matched arterial catheter recordings, but it does not yet establish that wearable monitoring can replace arterial lines in routine care. The results come from an initial study of patients at Johns Hopkins Hospital.