Together with Meander Medical Center in Amersfoort, we shape the future of digital surgery. We have various projects around the topic of surgical AI, video analysis, and data science. The projects are carried out at the Meander AI & Data Science Center, where clinicians, technical physicians, and computer scientists are connected closely. Below, you can find our three major research directions.
For the most recent updates, you can follow our LinkedIn:
https://www.linkedin.com/showcase/meander-ai-data-science-center/. Or feel free to reach out!
Surgical Data Science in the OR: Benchmarking Surgical Performance
An aging population and a growing burden of chronic diseases and comorbidities pose increasing demands on hospitals. At the same time, healthcare systems are constantly pressured to improve patient outcomes with limited resources. Innovations in the digitization of healthcare offer opportunities to effectively address these challenges. AI is one of the most promising technologies for the digitization of healthcare. The unique synergy of surgeons, technical physicians, informaticians, and other colleagues enables the development of innovative AI tools to enhance surgical outcomes and performance. The main goal is the intraoperative detection and statistical analysis of surgical data in surgical videos. Additionally, we utilize pre-/post-/intra-operative patient information, offering many benefits such as additional support for streamlined surgical planning, improved surgical training, and a benchmark for surgical practices. To achieve this we focus on high-volume procedures such as laparoscopic gallbladder surgery and hernia repair.
Scene Understanding in Operating Room Videos
With the introduction of electronic systems in the operating room, such as patient monitoring, laparoscopic surgery, and robot assistance, more and more data are recorded during surgical procedures. This trend gives data-driven systems, such as machine learning models, the opportunity to gain a more prominent role in the surgical environment. We explore the development of deep learning based algorithms for scene understanding in operating room videos. For example, by detecting medical staff and identifying their role or by recognizing clinical phases in surgical procedures. Automatic scene understanding can form the basis for higher-level algorithms that assist the surgical team, help to evaluate completed procedures or bring insights for surgery planning. We approach the development of new algorithms from the perspective of geometric deep learning, with a focus on end-to-end differentiable methods and graphical models. In our work, we pay explicit attention to the privacy of the medical staff and the patients. Our goal is to find practical and effective solutions that respect the privacy of the people who enter the observed environment thereby contributing to the digital trend that makes the operating room safer, more efficient, and more pleasant.
Patient-Specific Prediction of Surgical Outcomes
During patient treatment, numerous data sources are gathered (i.e. Electronic Health Record data, diagnostic imaging or surgical videos). This data can be used to establish a more patient-tailored treatment. Therefore, we focus on the development and use of predictive models. Based on pre- and intra-operative data, we aim to predict how certain surgeries will unfold. For example, predicting the duration of the surgery, the likelihood of complications, and the expected recovery time. Such a model can lead to a more individualized and data-driven treatment when implemented in the clinic. Predictive models can impact healthcare in various ways. The model can support the decision-making of the practitioner by objectively analyzing the data. Additionally, during treatment, various sources of information (e.g., surgical video or electro-surgical energy consumption) can be automatically integrated, resulting in an adaptive prediction of postoperative outcomes (e.g., length of hospital stay). Furthermore, an accurate prediction of procedure duration and hospital admission can lead to a more efficient use of our healthcare resources.