Truace
Subscribe

Neurosurgeons rated certainty scores and maps as the clearest way to explain surgical AI

Study title: Exploring clinician interpretability of AI decision support systems for surgical anatomy recognition

An international cross-sectional survey asked 47 neurosurgeons to evaluate five ways of explaining AI-generated anatomical segmentations in endoscopic pituitary surgery. Participants ranked and rated certainty scores, certainty maps, saliency maps, scene similarity scores and nearest-neighbour illustrations, and also gave free-text feedback.

Read the study abstract

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 explanatory 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,…

By the Truvace desk1 min read1 sourceEvidence: Emerging · 46/100 — Survey · n = 47 · 1 independent sourcePeer-reviewedTRV-2026-2256· cite & verify
Exploring clinician interpretability of AI decision support systems for surgical anatomy recognition
Photo by Piron Guillaume on Unsplash · https://unsplash.com/@gpiron?utm_source=truvace&utm_medium=referral

In brief

The findings matter because limited interpretability is a barrier to adopting surgical AI, and the survey indicates which explanation styles clinicians find usable. What remains unknown is whether these preferences actually improve trust calibration or surgical performance, since that has not yet been tested prospectively.

Main points

  1. The survey covered 47 neurosurgeons internationally, with endoscopic pituitary surgery used as a high-risk example.
  2. Certainty scores and certainty maps were ranked and rated highest by respondents.
  3. Certainty scores convey confidence, while certainty maps convey anatomical boundaries, which surgeons saw as fitting their decision-making.
  4. Saliency and similarity-based methods were judged less clinically relevant and better suited to educational use.

The gain

Surgeons ranked and rated certainty scores and certainty maps highest because they match surgical decision-making, suggesting a workable route to more interpretable AI-assisted anatomical navigation.

The rundown

Researchers used a modified Explanation Satisfaction Scale together with ranking tasks to compare literature-derived explanatory techniques for AI anatomical segmentation. Free-text comments supplemented the quantitative ratings.

Surgeons valued certainty-based approaches because they present confidence levels and anatomical boundaries in a form aligned with intraoperative decisions. Saliency maps and similarity-based methods were seen as better for teaching than for clinical use.

The authors conclude that certainty-based techniques are the most acceptable option for improving interpretability, while noting that their effect on trust and performance needs prospective evaluation across surgical domains.

Sources

  1. Peer-reviewedScientific Reports2026-10-24

The debate