Toward ubiquitous surgical assistance services via private 5G infrastructure

PURPOSE: Advanced data processing methods, including artificial intelligence (AI), are increasingly integrated into clinical practice, improving treatment quality and reducing practitioner strain. These methods entail elevated computational demands, necessitating secure and efficient resource provision. This study investigates a wireless infrastructure based on private 5G networking capable of provisioning scalable computational resources, exemplified by surgical assistance services (SAS) requiring live feedback. METHODS: To this end, five exemplary SAS covering medical imaging, speech, and device data processing were deployed as virtualized applications on a private 5G-connected edge computing cluster. Latency and concurrent stream capacity were evaluated against a local processing baseline. RESULTS: The private 5G configuration supports up to five concurrent 4K or 41 concurrent 720p video streams with on-site-like responsiveness and guaranteed latencies. A 720p end-to-end endoscopic SAS, covering local image capture, remote segmentation, and return, consistently achieves capture-to-display latencies below 290 ms. CONCLUSION: The findings demonstrate how medical systems can wirelessly offload diverse computational workloads, satisfying responsiveness requirements for surgical applications despite a preliminary private 5G configuration. Compared to Wi-Fi, private 5G requires greater infrastructure effort but provides dedicated spectrum with reduced interference susceptibility, enhanced security, and temporal determinism, properties that may be of particular relevance for surgical navigation and robotic systems.