<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=2248836&amp;fmt=gif">

What Are the Differences between AI Auto-Contouring and Atlas-Based Methods?

What are the real-world expectations for AI auto-contouring results?

Video Summary

Mild Case: 
High-Quality Contrast Liver CT

Atlas Contouring: In the first example, the atlas-based approach led to a reasonably accurate result for a high-quality contrast CT.

AI Auto-Contouring: Almost every slice is accurately contoured and ready to go, right out of the box with Contour ProtégéAI™. Some slices may need minor edits.


More Challenging Case: 
Y90 Therapy Patient on a Non-Contrast Liver CT with Artifacts on Slices

Atlas Contouring: Every slice needs editing in this instance, requiring a significant time commitment.

AI Auto-Contouring: When running Contour ProtégéAI, the difference between AI auto-contouring and atlas-based contouring is significant. The AI auto-contouring results are much more accurate, with minimal editing needed overall.


Difficult Case: 

Low-Contrast CT on a Patient with a Small Body Size and an Increased Liver

Atlas Contouring: There isn't a close-match atlas subject to use in this instance, and the atlas results are inaccurate and unusable. For this atlas example, it would be the most practical to contour from scratch, which requires a significant time dedication.

AI Auto-Contouring: Overall, the auto-segmented example using Contour ProtégéAI handled this complex case well and requires minimal editing after review.

As demonstrated in the video, AI auto-contouring offers significant improvements in quality and time savings in cases of varying complexity. Contour ProtégéAI is clinically implemented in hospitals worldwide and is available now.

Schedule a Demo
Learn More about Contour ProtégéAI

Find Out How MIM Can Improve Your Clinical Workflow