SterilAIze: Identification and Optimization of Surgical Trays Based on Instrument Use Patterns

Name of participant: Yuhao Gao

Project’s name: SterilAIze: Identification and Optimization of Surgical Trays Based on Instrument Use Patterns

Project description:

SterilAIze investigates how surgical instruments can be automatically recognized using artificial intelligence and computer vision and how their usage patterns can be analyzed. The aim is to optimize the composition of surgical trays according to actual needs, increase the efficiency of sterile processing, and simultaneously ensure patient safety.

The correct provision of sterilized and fully functional instruments is a fundamental prerequisite for safe surgical procedures. At the same time, sterile processing accounts for a considerable share of the secondary process costs in the operating room. Surgical trays often contain instruments that are not used at all or are used only rarely during a procedure. Nevertheless, all instruments must be inspected, cleaned, sterilized, packaged, and stored after every operation. This results in additional expenditure of time, energy, materials, and personnel resources.

Within the project, image and video data of surgical instruments are collected and annotated in both laboratory and real clinical environments. Based on these data, deep-learning models for object detection and instance segmentation are developed. The models are intended first to locate and identify instrument categories and then to distinguish characteristic features such as shape, size, and model. By analyzing the relative positions of instruments and surgical trays, as well as recordings from multiple viewing angles, the frequency and duration of instrument use are to be derived.

The identified usage patterns form the basis for an optimized tray composition. Depending on their frequency of use, instruments are classified as frequently used, rarely used, or unused. Frequently required instruments remain in the standard tray, rarely required instruments can be provided separately and on demand, while instruments that remain unused over time are considered potentially redundant. The proposed adjustments are reviewed together with medical professionals and other stakeholders to ensure that the functionality of the surgical workflow and patient safety are maintained at all times.

In addition to the technical development, SterilAIze also evaluates the economic and environmental effects of the optimized trays. The analysis includes potential savings in processing time, personnel effort, energy and water consumption, material wear, inventory management, and procurement. The project thus combines artificial intelligence, medical image processing, and process optimization in a practical approach to achieving more transparent, efficient, and sustainable sterile processing in hospitals.

Software Campus Partner: Karlsruhe Institute of Technology & Carl Zeiss AG

Implementation period: 01.01.2025 – 31.12.2026