top of page

AI Autocontouring is just Atlas contouring with a few tech gimmicks – Myth 6

Updated: May 3, 2023





10 May 2021

AI Autocontouring is just Atlas contouring with a few tech gimmicks


Myth 6 – AI Auto contouring is just Atlas contouring with a few tech gimmicks

Today, Atlas technology is being replaced by superior AI-based Deep Learning auto contouring, such as DLCExpert. Both solutions deliver auto contouring, but AI superior to Atlas and is increasingly being shown to deliver more usable clinical contours.

In essence, the difference between Atlas and Deep Learning Contouring can be summarized as follows:

o Atlas-based auto contouring uses a combination of previously contoured anonymized patients, generally 20+, and a matching algorithm selects the patient or contour most similar to today’s patient before deformable image registration is used to transfer the contours from the Atlas onto the patient.

o Deep learning uses hundreds of past anonymized cases (which have been carefully selected to ensure high-quality, consistent input data) to train an artificial intelligence that learns how to predict the same structures onto new data. Deep learning contouring is a form of artificial intelligence based on neural networks mimicking the human brain.

More and more studies are showing that Deep Learning techniques outperform Atlas techniques. Atlas may be sufficient for some basic applications, but for most clinical contouring in radiation therapy, we strongly recommend choosing a Deep Learning option. Mirada’s automation platform, WorkflowBox, supports both.

The keystone to any good model is well-curated, robust data sets. Mirada now offers an AI library of contours covering the four anatomical regions, Head and Neck, Thorax, Prostate, and Breast.

With Mirada’s AI, you can test your data and see the output quality before your purchase. Mirada’s AI autocontouring has been the first in-clinic since 2018 and has been peer-reviewed.

Read more An evaluation of atlas selection methods for atlas-based automatic segmentation in radiotherapy. 2019, Schipaanboord B et al. https://ieeexplore.ieee.org/document/8684891 To help clinics experience the benefits of AI, we have launched an AI Accelerator program. A tailored package offers accelerated clinical implementation, rapid clinical optimization, and access to our world-class expertise for a zero-cost, no-obligation, introductory period.

Places on the program are limited, so please complete the form below to discuss your requirements for an initial engagement with our team.

Apply to be part of the Accelerator Program. [pardot-form width=”800″ height=”1000″ id=”3121″ title=”Accelerator form”]

Comments


bottom of page