Modeling and performance analysis of image transfer in PACS

Area of Science:

• Medical Imaging Informatics
• Computer Science
• Systems Engineering

Background:

• Performance simulation is vital for designing effective Picture Archiving and Communication Systems (PACS).
• Understanding clinical workflow is essential for accurate PACS performance evaluation.
• Existing PACS designs require optimization for efficient image retrieval in outpatient settings.

Purpose of the Study:

• To present simulation models for PACS within an outpatient clinical workflow context.
• To evaluate the performance of PACS components at both workflow and physical levels.
• To determine the impact of system resources and loads on image transfer times.

Main Methods:

• Developed two-level simulation models (workflow and physical) using Petri Nets.
• Workflow level modeled patients, medical staff, and images as key resources.
• Physical level modeled PACS components like networks and hard disks.
• Conducted performance evaluations by varying data transfer rates, network speeds, and system loads.

Main Results:

• Achieving a 5-second image retrieval time for physicians requires high-performance hard disks/arrays and 100 Mbps networks.
• High-speed networks are particularly effective during the image acquisition phase.
• Using standard Ethernet and workstations necessitates limiting PACS to a single clinic for optimal performance.

Conclusions:

• The proposed Petri Net modeling method provides a simple yet effective approach for PACS performance evaluation.
• Simulation is highly useful in the early stages of PACS design to identify performance bottlenecks.
• Optimizing network and storage components is critical for meeting clinical demands for rapid image access.