Surgical Data Science Initiatives

17 Jun 2019
10:30 - 12:30

Surgical Data Science Initiatives

Chair: Stefanie Speidel (National Center for Tumor Diseases (NCT) Dresden, Germany)

Smart Image
Adrian Park (Anne Arundel Medical Center, Johns Hopkins University, United States)

The advent of laparoscopic cholecystectomy almost 30 years ago would change forever the way surgeons visualize and interact with target anatomy Patients continue to benefit from different yet related image guided therapies that also allow access to pathology by minimally invasive means.  As we continue to depend upon images to guide and inform patient interventions it is instructive to review the advances made in surgical visualization over its recent history and look forward to issues that will need to be addressed toward optimization of interventional visualization.  These issues will be reviewed from the perspective of a clinician, and not a computer scientist nor physicist with attention also paid to the often neglected topics of ergonomics and human factors considerations in surgical visualization.

CONDOR – Connected Optimized Network & Data in Operating Rooms
Pietro Mascagni and Tong Yu (CAMMA, IHU Strasbourg & University of Strasbourg, France)

CONDOR is a French multi-institutional project funded by BPI France aimed at developing and deploying standards, infrastructures and methods for operating rooms (OR) data.

Modern ORs are data-intensive, complex thus error-prone environments. In France alone, of the roughly 6 million surgical procedures performed annually, more than 9000 complicates with adverse events leading to a high burden on patients, health systems and the society as a whole.

Similarly to aeronautics, OR black boxes and surgical control towers capable of real-time data capturing, transmission, recording and analysis could translate into increased patients’ safety. For instance, OR and endoscopic vigilance could alert the surgeon in case of dangerous deviation from planned workflows, eventually preventing errors. Furthermore, the correlation of vast amounts of OR data with clinical information could offer new insights into surgical practices leading to favorable outcomes. Finally, such systems could optimize surgical resource allocations to decrease costs.

CONDOR is delivering a new worldwide communication standard for surgical videos, infrastructures allowing ultra-low latency high-definition video compression and deep learning methods for context-aware video analysis. Results will be implemented in the innovative platforms for image-guided surgery at IHU-Strasbourg in order to advance research, re-think training and impact patients.

OR Black Box: using data to improve surgical safety
Teodor Grantcharov (St. Michael’s Hospital, University of Toronto, Canada)

NCT Surgical Data Science Initiative
Keno März (German Cancer Research Center (DKFZ), Germany)

The National Center for Tumor Diseases (NCT) in Heidelberg was founded in 2004 as an alliance between the German Cancer Research Center (DKFZ), Heidelberg University Hospital, and the German Cancer Aid. The NCT strives to offer each patient a personalized treatment option specifically tailored for his or her individual needs. The mission of the NCT Surgical Oncology program is to establish the field of surgical data science in the complex research and clinical environment of the NCT. A key aim is therefore to develop the future infrastructure, tools and workflow concepts for enabling data-driven oncologic surgery. This talk will present the status of the project with a particular emphasis on the intra-operative communication architecture.

AI4OR, A Network Initiative for Context-Aware Operating Theaters
Duygu Sarikaya (Inserm, University of Rennes 1, France)

Artificial Intelligence (AI) is projected to deeply transform the practice of medicine and bring socioeconomic changes along with these transformations. This is largely due to the recent advances in machine learning, especially deep learning, which demonstrated a breakthrough performance in human perception tasks such as object recognition and natural language processing. Achieving human-like competence, AI is already shaping the future of medical imaging diagnostics, personalized medicine and clinical decision support systems. Due to the unique environment and the complex nature of surgical procedures, however, AI has taken longer to integrate with operating theaters. We believe that AI has tremendous potential to shape the future of interventional healthcare by increasing safety, efficiency, precision and reproducibility in operating theaters. The main aim of this proposed Action is to create innovative solutions to context-awareness in operating theaters, with a particular focus on surgical process modeling, analysis and recognition of surgical processes, and cognitive surgical robotics, primarily using AI algorithms. This proposed Action aims to advance this emerging field via a bottom-up, interdisciplinary network of academics, clinicians, and industry partners with a common goal of defining the technologies of the future AI powered, context-aware operating theaters. This network will help democratize knowledge and expertise, and close the divide between renowned research centers and small communities, train and mobilize young talents. It will help create a market of specific skills by bringing academia, healthcare institutions and industry together, and help attract and keep young talents in Europe. Moreover, it will give solid ground to multi-stakeholder research collaborations.