Exploring clinician interpretability of AI decision support systems for surgical anatomy recognition
Year: 2026
Type: article
Study design: Observational Study
Abstract: Artificial intelligence (AI) decision support systems for surgery hold promise but face barriers to adoption, particularly around the interpretability of their outputs. We conducted an international cross-sectional survey of 47 neurosurgeons to evaluate perspectives on literature-derived explanation techniques for AI-generated anatomical segmentations, using endoscopic pituitary surgery as a high-risk exemplar. Participants ranked certainty scores, certainty maps, saliency maps, scene similarity scores, and nearest-neighbour illustrations, and rated them using a modified Explanation Satisfaction Scale alongside free-text feedback. Certainty-based techniques were consistently ranked and rated highest for interpretability — valued for aligning with surgical decision-making by conveying confidence (via scores) and anatomical boundaries (via maps). Saliency- and similarity-based methods were judged less clinically relevant and better suited to educational settings. Certainty-based explanations, therefore, appear most acceptable to surgeons for clinical integration of decision support systems, though their impact on AI acceptability, trust calibration, and performance requires prospective evaluation across surgical domains.
Source: Scientific Reports
Authors: Danyal Zaman Khan, Tobias Joseph Adams, Anjana Wijekoon, Roxana Ramirez Herrera, Sophia Bano, Peter McCulloch, Danail Stoyanov, Matthew J Clarkson, Enrico Costanza, Ann Blandford, Hani Joseph Marcus, CARES Evaluation Group, Mohammed Mustafa Adawi, Joao Paulo Almeida, Jensen Ang, Anouk Borg, Roman Bošnjak, Alexandros Boukas, Aswin Chari, Lekhaj C Daggubati, Neil L Dorward, Samuel Emerson, Laurence Johann Glancz, Paul Gwammache, John Gerrard Hanrahan, Joachim Starup-Hansen, Lauren Harris, Mark Hughes, Abhiney Jain, Adham M Khalafallah, Taufiq Khan, Muhammad Sohaib Khan, Ramez Wadie Kirollos, Angelos G Kolias, Jacek Kunicki, Ruth Lau, Juan Carlos Fernandez-Miranda, Ramesh Nair, Nicola Newall, Amr Nimer, Fabio Massimo Oddi, Piero Andrea Oppido, Piotr Paździora, Hafiz Muhammad Irfan Razzaq, Luis Alejandro Flores Salazar, Olabisi Sanusi, Thomas Santarius, Sebastian Senger, Jonathan Shapey, Nathan A Shlobin, Nicholas Wake Macmeikan Thomas, Sebastian Miguel Toescu, Georgios Tsermoulas, Juan Francisco Villalonga, Adam Williams, Nina Yoh, Brett E Youngerman, Gabriel Zada
Institutions: National Hospital for Neurology and Neurosurgery, UCL Queen Square Institute of Neurology, University College London, Manchester University NHS Foundation Trust, University of Oxford, Indiana University Bloomington, National Neuroscience Institute, Ljubljana University Medical Centre, Oxford University Hospitals NHS Trust, University of Washington, Queen's Medical Centre, Harvard University, NHS Lothian, University of Miami, Queen Elizabeth Hospital Birmingham, Lady Reading Hospital, University of Cambridge, The Maria Sklodowska-Curie National Research Institute of Oncology, Hospital Universitari Joan XXIII de Tarragona, Stanford University, Charing Cross Hospital, Imperial College London, University of Rome Tor Vergata, Sapienza University of Rome, Medical University of Silesia, Punjab Institute of Neurosciences, Hospital Barros Luco-Trudeau, Oregon Health & Science University, Cambridge University Hospitals NHS Foundation Trust, Sozialstiftung Bamberg, King's College London, Columbia University Irving Medical Center, National University of Tucumán, North Bristol NHS Trust, University of Southern California
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,690
OpenAlex ID: W7223446601