News & Articles

Clinical validation started for the Cholecystectomy OR Assistant. This pioneering study, led by Prof. Silvana Perretta, tests real-time AI guidance to improve implementation of the Critical View of Safety.

Institutional Review Board approves comparative trial for the Cholecystectomy OR Assistant. This US-based study, led by Prof. Adnan Alseidi, will demonstrate how AI-powered guidance improves surgical safety and patient outcomes.

Official FDA feedback (Q-Sub) confirms streamlined pathway for first AI-powered SaMD for real-time assistance during laparoscopic cholecystectomy. This major milestone accelerates our path to real-world impact.

Scialytics team helps uncover key relationship between surgical technique variations and adverse events in bariatric procedures. Published in Surgical Endoscopy, this multicenter video analysis led by Dr. Joël Lavanchy identifies specific high-risk phases in laparoscopic Roux-en-Y gastric bypass, with 77% of adverse events occurring during gastric pouch creation and anastomosis formation. By correlating technical approaches with complication patterns, this research provides surgeons with actionable insights to enhance safety in one of the most commonly performed bariatric procedures.

Read the article: Surgical Endoscopy

Scialytics' CTO Luca Sestini breaks new ground with fully unsupervised approaches to instrument segmentation. By achieving results comparable with fully-supervised methods, these innovative techniques unlock the potential to leverage vast amounts of unlabeled surgical video for AI development. This breakthrough represents a significant step toward more practical and scalable computer-assisted surgery applications.

Read the articles: Medical Image Analysis (MedIA) | Medical Image Analysis (MedIA)

Scialytics’ team co-authored SurgPrOD, an AI system that assesses laparoscopic cholecystectomy difficulty from just the initial minutes of surgery. This groundbreaking work led by Saurav Sharma enables early identification of challenging cases, allowing for timely expert intervention and optimized resource allocation. By predicting operative difficulty before complications arise, this system has the potential to significantly improve patient outcomes and operating room efficiency.

Read the article: International Journal of Computer Assisted Radiology and Surgery

Scialytics team releases comprehensive dataset for advancing surgical safety in laparoscopic cholecystectomy. Published in Scientific Data, Endoscapes 2023 includes 201 expertly annotated surgical videos with precise segmentation masks and critical view of safety assessments from three trained surgeons. This groundbreaking resource enables developers to build more reliable and reproducible AI models for anatomy recognition and Critical View of Safety assessment, directly contributing to safer cholecystectomy procedures worldwide.

Read the article: Scientific Data

Scialytics collaborates in piloting the transformative "Marginal Gains" approach for laparoscopic right hemicolectomy training. Published in the European Journal of Surgical Oncology, this groundbreaking study led by Dr. Morales-Conde demonstrates how AI-powered video assessment supports a structured, step-based framework for surgical education. By combining practical training with intelligent performance analysis, this paradigm-shifting approach establishes a new standard for surgical quality assurance and skills development in complex procedures.

Read the article: European Journal of Surgical Oncology

Scialytics co-founders organize the first-ever AI challenge led by a surgical society. The SAGES CVS challenge generated a dataset of more than 1000 surgical videos to support the development of AI models for assessing the critical view of safety during cholecystectomy. This groundbreaking initiative demonstrates growing surgical society support of AI-powered surgical safety tools.

Visit the Challenge website

Scialytics leads groundbreaking research on surgical foundation models through extensive data composition analysis. Presented at MICCAI 2024, this work demonstrates that strategic pre-training dataset composition outperforms existing. Through 300+ experiments across 20 pre-training datasets, 9 procedures, and 7 hospitals, we establish new standards for building surgical AI that works across diverse clinical settings.

Read the article: MICCAI 2024

Scialytics’ co-founder Dr. Pietro Mascagni contributes to a groundbreaking quality assessment framework for preclinical AI studies led by Dr. Giulio Antonelli. Published in Gut, QUAIDE establishes consensus recommendations from 32 international experts across medicine and computer science to standardize AI development methodology. Covering critical domains from data acquisition to result interpretation, these guidelines accelerate the translation of AI innovations into clinical practice by ensuring reproducibility and generalizability.

Read the article: Gut

Scialytics outlines a comprehensive framework for bringing surgical AI from research to clinical practice. Published in Revista Española de Cirugía, this paper uses colorectal cancer care as a model to address critical technical, clinical, and governance challenges in surgical AI implementation. By examining the entire patient journey, this work bridges the gap between AI's theoretical capabilities and practical value in improving surgical outcomes for cancer patients.

Read the article: Revista Española de Cirugía

Scialytics team pioneers advances in Self-Supervised Learning (SSL) - learning from unlabeled data - that dramatically accelerate surgical AI development. Published in Elsevier Medical Image Analysis (MedIA), this comprehensive study benchmarks SSL methods for surgical computer vision, validating their ability to generalize across hospitals, tasks, and datasets. This approach reduces dependency on surgical expertise for dataset generation while maintaining high performance in critical applications.

Read the article: Medical Image Analysis (MedIA)

Scialytics team contributes to develop an innovative approach to surgical action modeling through structured triplet detection. This work, led by Chinedu Nwoye, has culminated in the organization of international challenges, scientific publications and datasets release , advancing surgical AI from basic recognition to precise detection of instrument-action-anatomy relationships, creating foundations for more sophisticated AI assistance in image-guided surgery.

Read the articles formalizing and modelling surgical action triplets: MICCAI 2020 | Medical Image Analysis (MedIA)‍
Read about the CholeTriplet Challenges: MedIA | MedIA

Scialytics helps Dr. Joël Lavanchy develop and validate innovative deep learning solutions to protect patient and OR staff privacy in surgical videos. This technology automatically identifies and filters out-of-body scenes in endoscopic recordings, ensuring sensitive information remains protected during video-based surgical analysis. By addressing critical privacy concerns, this advancement enables broader adoption of video analysis for quality improvement while maintaining strict ethical standards for privacy protection.

Read the article: Scientific Reports

Scialytics co-founder Tong Yu pioneers first-ever real-time surgical video retrieval system. Published in MEDIA, this groundbreaking technology enables surgeons to instantly find similar procedures from video archives during live operations. By implementing compressed uncertainty learning, the system efficiently matches current surgical scenarios with relevant historical examples, allowing surgeons to see how experts previously handled comparable situations. This "Google for surgery" represents a significant advancement in bringing just-in-time knowledge to the operating room.

Read the article: Medical Image Analysis (MedIA)

Scialytics participates in a comprehensive systematic review of AI applications for technical skill evaluation. Published in Surgical Endoscopy, this analysis led by Dr. Romina Pedrett reveals that while AI shows promising performance in assessing minimally invasive surgical skills, most models lack critical external validation and generalizability. The review establishes clear recommendations for benchmarking using standardized metrics and emphasizes the need for clinical implementation studies before widespread adoption.

Read the article: Surgical Endoscopy

Scialytics team leads a perspective on surgical AI's journey to clinical impact. Published in npj Digital Medicine, this invited review assembles key opinion leaders to map AI's current methods, discuss enablers, and address ethical, cultural, and educational considerations for AI adoption in surgery.

Read the article: npj Digital Medicine

Scialytics’ co-founder Pietro Mascagni exploreAI's transformative potential in colorectal cancer procedures. Published in Cancers, this comprehensive review led by Dr. Giuseppe Quero provides surgeons with an accessible introduction to AI fundamentals and explores how computer vision can enhance surgical care across the colorectal cancer treatment pathway. This work establishes a roadmap for applying surgical computer vision to more complex procedures to expand its clinical value.

Read the article: Cancers

Scialytics contributes to landmark surgical society consensuses on best practices video annotation and usage. Published in Surgical Endoscopy, these SAGES recommendations establish crucial frameworks for surgical AI and Computer Vision research led by Dr. Ozanan Meireles and Dr. Daniel Hashimoto. By bringing together diverse stakeholders from both clinical and computational domains, these consensus documents provide essential guidance for video annotation, data governance, and appropriate data usage across AI research, education, and quality improvement. This foundation enables standardized datasets, reliable performance benchmarks, and improved collaboration throughout the surgical AI ecosystem.

Read the Surgical Endoscopy articles: SAGES Annotation Framework | SAGES Data Use Recommendations

Scialytics’ team introduces novel FedCy framework to address data related challenges in surgical AI development. Published in TMI, this work demonstrates how federated semi-supervised learning can overcome annotation bottlenecks while protecting hospital data privacy. The breakthrough enables AI models to learn from multiple institutions while keeping sensitive patient data secure within each center.

Read the article: IEEE Transactions on Medical Imaging (TMI)

Scialytics participates in a perspective article led by Dr. Lorenzo Cobianchi on ethical dilemmas in surgical AI adoption. Published in the Journal of the American College of Surgeons, this work identifies eight key ethical principles, including human agency, accountability, technical robustness, and algorithmic fairness. The paper advocates for structured digital health education as essential preparation for surgeons entering the AI era.

Read the article: Journal of the American College of Surgeons (JACS)

Scialytics co-founder Nicolas Padoy and Pietro Mascagni participate in this landmark MEDIA mapping the future of surgical AI implementation. This comprehensive work led by Prof. Lena Maier-Hein defines the challenges and solutions for translating surgical data science research into real-world clinical applications.

Read the article: Medical Image Analysis (MedIA)

Scialytics co-founder Pietro Mascagni joins an expert group led by Dr. Kyle Lam to define Digital Surgery. Published in npj Digital Medicine, this Delphi study unites 38 international experts across surgery, AI, law, and ethics to establish standards for digital surgical technologies. The framework addresses privacy, consent, and governance challenges while outlining future research priorities.

Read the article: npj Digital Medicine

In this invited contribution to the Journal of Visceral Surgery, Scialytics co-founders explain their vision of an AI-augmented operating room, a framework where AI could provide real-time support to surgical teams, similar to how air traffic controllers support pilots.

Read the article: Journal of Visceral Surgery

Breakthrough AI platform documents surgery 5x better than standard reports. This Annals of Surgery publication lead by Scialytics’ team introduced EndoDigest, which automatically locates critical safety events in surgical videos. Multi-center validation confirmed that AI-produced short video clips provide superior documentation compared to traditional dictated reports.

Read the original article: Annals of Surgery
Read the validation study: Surgical Endoscopy

70,000 surgeons witness history at World Congress of Endoscopic Surgery. In partnership with NVIDIA, Scialytics’ team demonstrated real-time AI assistance during laparoscopic cholecystectomy, proving that intraoperative AI support is possible.

Read the article: British Journal of Surgery (BJS)
Read the blog: BJS Academy

Scialytics’ team proves a simple cognitive aid can triple the implementation of critical safety measures. Published in the Journal of the American College of Surgeons, our video-based assessment of 343 procedures revealed that timing out to recall best practices before irreversible steps improves guideline compliance.
Dr. L. Michael Brunt praised it as "one of the highest benefit-to-risk ratio interventions" in recent years.

Read the article: Journal of the American College of Surgeons (JACS)
Read Dr. Brunt's commentary: JACS Commentary

Scialytics co-founder invited to examine transformative potential of visual AI for surgical care in high-impact review. Published in Surgery, this work led by Thomas Ward chronicles how deep learning breakthroughs have enabled computers to accurately understand surgical imagery and video. The paper details remarkable progress in surgical phase recognition and instrument identification across multiple procedures, with some AI systems achieving surgeon-level accuracy. Looking ahead, this foundation will expand through larger datasets and improved algorithms to create clinically useful models that augment surgical capabilities and enhance patient safety globally.

Read the article: Surgery

Scialytics co-founders Pietro Mascagni and Nicolas Padoy helo examine how data science and AI are revolutionizing surgical training. Published in Journal of Surgical Oncology, this review led by Thomas Ward examines how the capture and analysis of procedural data enables new educational approaches in surgery. By leveraging artificial intelligence for automated coaching, feedback, and assessment, surgical data science creates unprecedented opportunities to enhance technical skills development. This work provides a comprehensive overview of how these technologies can be specifically applied to improve education in surgical oncology procedures.

Read the article: Journal of Surgical Oncology

Scialytics team invited to contribute essential guide on deep learning in prestigious textbook “Artificial Intelligence in Surgery" edited by Dr. Daniel Hashimoto. This work provides surgeons with accessible insights into neural networks and deep learning accompanied by practical hands-on materials requiring no prior coding experience. By demystifying AI concepts specifically for surgical applications, Scialytics empowers clinicians to become active participants in developing the next generation of AI-powered surgical tools.

Read the chapter pre-print: arXiv
Try the tutorial: GitHub

Scialytics’ team publishes groundbreaking research on the automatic assessment of the Critical View of Safety using deep learning. Our study demonstrates how AI can evaluate key safety steps during cholecystectomy, highlighting its potential to scale surgical expertise and improve patient outcomes.

Read the article: Annals of Surgery

Scialytics co-founder Nicolas Padoy co-authors revolutionary approach for surgical tool tracking with minimal supervision. Published in IJCARS, the work led by Chinedu Nwoye proposed an innovative method requiring only binary tool presence labels rather than precise spatial annotations to effectively track instruments in laparoscopic videos, dramatically reducing the labeling burden. This advancement represents a critical step toward intelligent operating rooms by making surgical tool tracking more practical to implement at scale.

Read the article: International Journal of Computer Assisted Radiology and Surgery (IJCARS)

Scialytics co-founder Nicolas Padoy advances context-aware OR systems through innovative early recognition of surgical procedures. Published in TMI, this work led by Siddharth Kannan introduces a novel LSTM-based approach that can identify the type of laparoscopic surgery from just the initial phases of the video. This breakthrough enables "smart" operating rooms to provide automated, procedure-specific assistance from the earliest stages of surgery, while also facilitating rapid database indexing for efficient surgical video management.

Read the article: IEEE Transactions on Medical Imaging (TMI)

Scialytics co-founder Dr. Pietro Mascagni explores how emerging imaging technologies revolutionize surgical visualization and precision. Published in the Journal of Surgical Oncology, this work examines how AI-powered interpretation transforms raw data into actionable guidance, enabling surgeons to leverage advanced visualization technologies for enhanced surgical accuracy across specialties.

Read the article: Journal of Surgical Oncology

Scialytics co-founder Nicolas Padoy advances operating room efficiency with innovative surgery duration prediction system. Published in TMI, this work led by Andru Twinanda introduces RSDNet, a deep learning pipeline that automatically estimates remaining procedure time using only laparoscopic video. By eliminating the need for manual annotations during training, this groundbreaking approach enables seamless implementation across various surgical procedures, significantly improving OR scheduling, patient safety, and resource optimization.

Read the article: IEEE Transactions on Medical Imaging (TMI)

Scialytics co-founder Nicolas Padoy landmark contribution to surgical AI with EndoNet. Published in TMI, this pioneering work led by Andru Twinanda established the foundation for modern surgical workflow recognition. This publication also introduced Cholec80, now the most widely used public dataset in surgical AI research, accelerating progress throughout the field.

Read the article: IEEE Transactions on Medical Imaging (TMI)

Scialytics co-founder Nicolas Padoy established foundational principles for surgical workflow analysis in landmark paper. Published in Elsevier Medical Image Analysis (MedIA), this pioneering pre-deep learning research introduced statistical modeling approaches to analyze surgical procedures. By demonstrating the viability of computational methods for inferring surgical workflows, this seminal work laid the groundwork for modern surgical AI and set the stage for today's advanced computer vision approaches to surgical analysis.

Read the article: Medical Image Analysis (MedIA)