Abstract Body (Do not enter title and authors here): Background: Heart failure management increasingly requires accurate, continuous assessment of left ventricular (LV) function, including parameters such as systolic and end-diastolic pressure (LVSP, LVEDP). These metrics are typically measured intermittently using invasive catheterization, providing only point-in-time values rather than continuous hemodynamic monitoring. Currently, no established method exists for direct, long-term monitoring of LV pressures, limiting timely detection of decompensation and individualized therapy adjustment.

Objective: To develop and validate an intramyocardial cardiac monitoring system enabling real-time estimation of LV function.
Methods: A micro-electro-mechanical systems (MEMS) pressure sensor array was implanted into the LV free wall and integrated with a machine learning model trained to estimate intracardiac pressures from local myocardial dynamics. Validation was performed in acute swine experiments (n=3) under dobutamine stress, using intraventricular pressure-volume catheterization as the reference standard.

Results: The system estimated LVSP and LVEDP with mean differences of 0.06 ± 5.3 mmHg and 0.15 ± 3.6 mmHg, respectively. Across all markers, the average coefficient of determination exceeded 0.85, indicating strong agreement.

Conclusion: This MEMS-machine learning platform enables accurate, continuous estimation of LV function without the need for intra-cardiac instrumentation. Its compact design and real-time capabilities support its potential for even less invasive epicardial implementation, remote heart failure monitoring, and personalized therapy adjustment.