Source-linked AI summary
Surgical Data Science: Enabling Next-Generation Surgery
Lena Maier-Hein, Swaroop Vedula, Stefanie Speidel, Nassir Navab, Ron Kikinis, Adrian Park, Matthias Eisenmann, Hubertus Feussner, Germain Forestier, Stamatia Giannarou, Makoto Hashizume, Darko Katic, Hannes Kenngott, Michael Kranzfelder, Anand Malpani, Keno März, Thomas Neumuth, Nicolas Padoy, Carla Pugh, Nicolai Schoch, Danail Stoyanov, Russell Taylor, Martin Wagner, Gregory D. Hager, Pierre Jannin
TL;DR
Large-scale data science has been slow to enter interventional medicine, where only a fraction of patient-related data is digitized and standardized. The paper presents a consensus definition of Surgical Data Science as an emerging field aimed at improving interventional healthcare quality and value through data capture, organization, analysis, and modeling.
Problem
Large-scale data science has been slow to enter interventional medicine, and only a fraction of patient-related data is digitized and stored in a structured, standardized manner.
Source: §21 · p. 2Method
The paper develops a consensus definition of Surgical Data Science and identifies associated challenges, opportunities, and a roadmap for advancing the field.
Source: What is Surgical Data Science? · p. 4Results
Surgical Data Science is defined as an emerging field aimed at improving interventional healthcare quality and value through capturing, organizing, analyzing, and modeling data.
Source: What is Surgical Data Science? · p. 4Takeaways & Limitations
The paper envisions Surgical Data Science providing surgeons quantitative decision support and linking decisions to patient outcomes.
Source: Evolution of Surgical Practice · p. 3High-quality large-scale databases documenting surgical care processes and patient outcomes are lacking.
Source: Key Challenges · p. 6
Abstract
from arXiv · showhide
This paper introduces Surgical Data Science as an emerging scientific discipline. Key perspectives are based on discussions during an intensive two-day international interactive workshop that brought together leading researchers working in the related field of computer and robot assisted interventions. Our consensus opinion is that increasing access to large amounts of complex data, at scale, throughout the patient care process, complemented by advances in data science and machine learning techniques, has set the stage for a new generation of analytics that will support decision-making and quality improvement in interventional medicine. In this article, we provide a consensus definition for Surgical Data Science, identify associated challenges and opportunities and provide a roadmap for advancing the field.
6 Department of Computer Science, University of Bremen, Bremen, Germany … 21 INSERM, Rennes, France
The paper argues that personalized interventional care should move toward objective decisions grounded in large-scale, heterogeneous data and a close partnership among caregivers, patients, technology, and information systems. It identifies limited digitization and variation in perioperative practice as barriers to data science in interventional medicine.
21 INSERM, Rennes, France
- Interventional care should shift from physicians’ individual experiences and traditions toward objective decisions based on large-scale data from heterogeneous sources.The paper frames this change as part of personalized medicine and calls for close partnership among caregivers, patients, technology, and information systems.Source: §21 · p. 2
- Only a fraction of patient-related data is digitized and stored in structured, standardized forms such as registries (1,2), delaying large-scale data science in interventional medicine.Source: §21 · p. 2
- Variation in caregiver training, experience, and institutional routines can produce disparities in outcomes when data do not reveal actual perioperative practice.Source: §21 · p. 2
Evolution of Surgical Practice
Surgery has advanced from experience-based practice to imaging- and computer-assisted interventions, but data and computer aids remain poorly integrated; Surgical Data Science envisions quantitative, patient-centered decision support linked to outcomes.
Evolution of Surgical Practice
- 9 million of an estimated 300 million surgical procedures worldwide per year encounter major complications, despite rising expectations for outcomes and safety (4, 5).Surgery is described as a profession defined by its authority to cure through bodily invasion (3).Source: Evolution of Surgical Practice · p. 3
- Surgical practice evolved from anesthesia and antiseptics to outcome measurement, minimally invasive access, multimodal imaging (6), and computer- and robot-assisted interventions (7).Computer aids are not yet seamlessly integrated into surgical environments, while electronic patient records are often unstructured, weakly quality-controlled, and disconnected from those systems.Source: Evolution of Surgical Practice · p. 3
- Surgical Data Science aims to shift surgery toward explicit models, objective decisions, and quantitative assessment, supporting personalized care and linking surgical decisions to patient outcomes.By analyzing data at scales beyond individual perception, it could support surgical actions and reduce variability attributable to patient characteristics rather than surgeon or facility choice.Source: Evolution of Surgical Practice · p. 3
What is Surgical Data Science?
Surgical Data Science is an emerging field that uses data capture, organization, analysis, and modeling to improve the quality and value of interventional healthcare. It encompasses the patient-care process, its people and technologies, and relevant domain knowledge.
What is Surgical Data Science?
- The field focuses on procedural data about patients, clinical and technical effectors, sensors, and domain knowledge, and builds on data-intensive disciplines while complementing robotics, smart operating rooms, and electronic patient records.Relevant sensors capture images, vital signs, medical-device data, and motion; domain knowledge includes standards, study findings, guidelines, and practical knowledge from prior procedures.Source: What is Surgical Data Science? · pp. 3, 5
- Surgical Data Science aims to improve interventional healthcare quality and value across disciplines that manipulate anatomical structures for diagnostic, prognostic, or therapeutic goals.These include surgery, interventional radiology, radiotherapy, and interventional gastroenterology.Source: What is Surgical Data Science? · p. 4
- Its data may span initial presentation through long-term outcomes and concern patients, caregivers, and care-delivery technology, interpreted using evidence, guidelines, practice patterns, caregiver experience, and patient preferences.Data can come from medical records, imaging, devices, and sensors positioned on patients or caregivers or integrated into care-delivery instruments and technology.Source: What is Surgical Data Science? · pp. 3–4
- It seeks improvement by understanding care processes, predicting events and outcomes, supporting decisions and execution, optimizing ergonomics, controlling devices, and advancing prevention, training, simulation, and assessment.Source: What is Surgical Data Science? · p. 4
Key Clinical Applications
Surgical Data Science offers applications across the patient care pathway, including decision support, context-aware assistance, and surgical training.
i) Decision Support
- Decision support can evolve from preoperative, cross-sectional registries with missing pathway data (9,10) to continuously updated analytics using comprehensive, unconventional, multimodal patient-and-population data integrated into user-friendly workflows (11–15).This approach models individual patients in the context of population-level data and other multimodal sources (14,15).Source: i) Decision Support · p. 5
ii) Context-aware Assistance
- Context-aware assistance spans pathway monitoring for procedure duration and resource needs (16), phase recognition (17,18), patient-specific simulations (19), and collaborative robots (20), augmenting care safety, quality, efficiency, and provider performance.These applications can provide timely information and assistance when integrated into surgical care pathways.Source: ii) Context-aware Assistance · p. 5
iii) Surgical Training
- Training can use OCASE (24), robot-assisted active learning (25), tailored simulation and coaching (26,27), and analytics for targeted feedback (28,29), addressing skill-related risks to care outcomes (21–23).Surgical Data Science represents a new frontier for training in complex care environments with limited resources.Source: iii) Surgical Training · pp. 5–6
Key Challenges
The paper identifies two immediate challenges for Surgical Data Science: securing high-quality, consistently documented surgical data and analyzing highly heterogeneous, multimodal data. These challenges span resource creation, reliable modeling of team-based interventions, and integration of procedural data with other patient information.
Data availability
- Surgical Data Science lacks large-scale, high-quality databases documenting care processes and patient outcomes.Privacy and confidentiality concerns, inconsistent capture and annotation, missing ontologies for care activities, and resource-intensive annotation impede data creation; the authors call for consistent longitudinal collection integrated into clinical workflows and short-term wins that build stakeholder trust (30, 31, 32).Source: Key Challenges · p. 6
Analysis of heterogeneous data
- Surgical analysis must model coordinated actions by multiple care-team members and patient responses, while ensuring robust and reliable methods because errors can cause serious complications or death (33).The head surgeon, anesthetists, assistant surgeons, circulators, and nurses contribute at different workflow steps.Source: Key Challenges · p. 6
- Surgical processes vary by procedure, patient, and surgeon, while hardware, imaging protocols, context, training, guidelines, and physicians add heterogeneity that complicates method development and validation (34).Source: Key Challenges · pp. 6–7
- Procedural data must be analyzed holistically with genetics, biomarkers, demographics, imaging, and pre- and intraoperative data to support knowledge-based, data-driven medicine.Shared tools for discovery and researcher training could advance the field (35).Source: Key Challenges · p. 7
Dissemination and Impact
Surgical Data Science is expected to reshape medical education and training while translating discoveries into clinical workflows and products. Its broader aim is data-driven interventional healthcare supported by continuous measurement, assessment, and improvement.
Dissemination and Impact
- Surgical Data Science could transform how medical students learn from complex data, while creating training pathways for Surgical Data Scientists and embedding data science in curricula.The paper also anticipates distinct career pathways and data science in undergraduate and medical school curricula.Source: Dissemination and Impact · p. 7
- Effective translation into patient-care workflows depends on academic and commercial stakeholders collaborating from discovery through development of data products and services.Potential outputs include decision support systems, smart instrumentation, intelligent technologies, and surgical training; large-scale data can also support usability studies and product optimization.Source: Dissemination and Impact · p. 7
- Surgical Data Science could disseminate its impact across medical training, surgical imaging, instrumentation, user interfaces, and patient information systems updated through analysis of dynamic data.The paper describes a broad range of potential applications, including next-generation advanced patient information systems.Source: Dissemination and Impact · p. 7
Towards next-generation surgery
- Surgical Data Science aims to move interventional healthcare from artisanal practice to data-driven care, improving quality and efficiency through continuous measurement, assessment, and improvement.The proposed approach makes evidence from data a core component of care.Source: Dissemination and Impact · p. 8
- Professional societies can advance this shift through best practices, comprehensive data registries, and active engagement and oversight, while hospitals establish data science in education and career pathways for future interventionalists.The paper identifies these measures as an actionable path toward next-generation surgery.Source: Dissemination and Impact · p. 8