# Critical views for safe surgical phase progression in endoscopic endonasal transsphenoidal pituitary adenoma resection: modified Delphi consensus

## TL;DR

Experts developed a framework called CVPPs to guide safe steps during pituitary tumor surgery, which could also help train surgeons and support AI tools.

## Contribution

The paper introduces the first international consensus on Critical Views for Phase Progression (CVPPs) in pituitary surgery, standardizing visual benchmarks for safe surgical steps.

## Key findings

- Consensus identified essential and desirable CVPPs for three surgical phases in both micro- and macroadenoma surgeries.
- Validation showed high agreement between expert ratings and the CVPP framework, with discrepancies only in intentionally incomplete views.
- CVPPs can improve training and serve as a foundation for AI-driven surgical guidance systems.

## Abstract

Endonasal transsphenoidal surgery is the gold-standard for pituitary adenoma resection, yet no intraoperative framework exists to confirm safe phase progression. Inspired by the Critical View of Safety in laparoscopic cholecystectomy and engineering “phase-gate” process, we propose the Critical Views for Phase Progression (CVPPs) – a set of visual cues confirming phase objectives and safe phase progression. Designed to be clinically relevant and machine-readable, CVPPs aim to support training and future AI-driven guidance systems.

A three-round modified Delphi process was conducted involving 15 pituitary surgery experts from 13 centres across Europe and North America. CVPPs for the naso-sphenoid, sellar, and closure phases were classified as “Essential”, “Desirable” or “Not Necessary”. Consensus required ≥ 70% agreement. A local validation study was subsequently performed involving six experts who reviewed 15 intraoperative video clips and rated their confidence to proceed, which was compared against the predefined reference derived from the finalised CVPPs.

Consensus identified essential and desirable CVPPs across all three phases for both micro- and macroadenoma variants, reflecting differences in exposure goals and surgical risk. Validation demonstrated high concordance between participant ratings and predefined references. Discrepancies arose only in a minority of intentionally incomplete (“unsafe”) views and were attributable to contextual misinterpretation of short video segments, rather than disagreement with the CVPP framework.

This international, multicentre consensus is the first to define CVPPs. By standardising intraoperative visual benchmarks, CVPPs can enhance training, mitigate risks, and provide a foundation for future AI-driven guidance systems capable of real-time anatomical annotation and decision support.