<p>Enhanced External Counterpulsation (EECP) is a proven effective treatment for cardiovascular diseases, yet its use is limited in patients with arrhythmias due to the therapy’s reliance on precise cardiac cycle synchronization. This paper proposes a novel framework to broaden the accessibility of EECP therapy by integrating artificial intelligence (AI) with private 5G network infrastructure. Private 5G provides the secure, low-latency data transmission necessary for real-time monitoring. We introduce a two-step deep learning approach using Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) architectures. The first step involves multi-label classification of cardiac rhythms from ECG data. The second step utilizes this rhythm analysis to inform a binary classification model that triggers a preemptive shutdown of the EECP machine upon detecting potentially hazardous arrhythmias. This serves as a critical safety measure. While our models show promising performance, we acknowledge limitations in detecting rare arrhythmia types due to dataset imbalances. Nonetheless, this study demonstrates the feasibility of an AI- and 5G-enabled system to enhance the safety of EECP, laying the groundwork for future development and extending its therapeutic potential to a wider patient population previously excluded due to arrhythmia risks.</p>