Wearable AI device may facilitate HF monitoring

A wearable, noninvasive device that estimates pulmonary capillary wedge pressure (PCWP) using artificial intel­ligence (AI) technology may facilitate heart failure (HF) monitoring with an ac­curacy comparable to right heart cath­eterization (RHC).

The device, known as CardioTag, captures electrocardiography, seis­mocardiography and photoplethys­mography data from patients with HF for PCWP estimation using a machine learning algorithm. [Klein L, et al, AHA 2024]

In the SEISMIC-HF I study, Cardio­Tag signals and RHC pressure tracing data were collected simultaneously from 943 patients scheduled to under­go RHC at 15 centres across the US (mean age, 63 years; male, 55 percent; White, 27 percent; with HF diagnosis, 88 percent; reduced left ventricular ejection fraction [ie, ≤40 percent], 39 percent; New York Heart Associa­tion function class II–IV, 90 percent). Blinded core-lab adjudicated RHC PCWP tracings were used as gold standard to evaluate the AI algorithm’s performance.

With RHC measurement, mean PCWP was 15.8 +/- 9.1 mm Hg, while mean pulmonary artery pressure was 42.6 +/- 18.0 mm Hg systolic and 18.4 +/- 9.4 mm Hg diastolic. “Model perfor­mance comparing the hold-out dataset to gold standard, simulating perfor­mance validation, showed a mean error of 1.04 +/- 5.57 mm Hg,” reported the investigators.

“Analysis of the machine learning al­gorithm’s performance in a racially and geographically diverse population sug­gests that this noninvasive technology may offer comparable accuracy to exist­ing invasive methods,” they concluded. “This technology could become a novel adjunctive tool for haemodynamic-guided clinical management of HF patients.”