Molecular imaging logo for artificial intelligence.

AI in Molecular Imaging

AI-powered algorithms to drive MI further.

Although intelligent algorithms have been used for some time in segments of medical imaging, new methods of machine learning, based particularly on “deep learning,” are much more powerful. AI-powered algorithms may offer concrete prospects for quantification, standardized imaging, and reporting.

 

Robust AI-powered algorithms can be used to drive clinically relevant and actionable information in a way that is quantitative, accurate, reproducible, and outcome-oriented clinical decision-making. Additionally, AI-powered algorithms may increase standardization that provides efficient automation and more operator-independent actions.

 

Thanks to already implemented innovative and forward-thinking technologies and the now available increased computing power, we can bring practical AI implementations to both scanners and reading solutions.

 


Molecular imaging PET/CT scanner Biograph Vision.

AI-powered algorithms are integrated seamlessly on our Biograph™ PET/CT scanners with the new, intelligent AIDAN1 platform. Several AI-powered features are available such as


Molecular imaging 3D lung image with color lobes.
Data courtesy of University Hospital Basel, Basel Switzerland.

Our syngo.via reading solution features AI-powered applications to support automated clinical analysis of PET and SPECT images such as:

  • Automated Image Registration
  • RT Image Suite
  • Spine and Rib Labeling
  • Auto Views
  • Auto Ranges
  • Auto Lung 3D1
  • Auto Identifier1

AI infographic showing AI, machine learning, and deep learning.

Artificial intelligence (AI) is transforming care delivery and expanding precision medicine. We have served as a pioneer in AI development for more than 20 years and the new deep learning technology now enables us to automate complex diagnostics and support optimal treatment.

Siemens Healthineers has been involved in the field of machine learning since the 1990s. We own more than 500 patent families related to machine learning; more than 125 patent families thereof are related to deep learning. More than 40 offerings on the market are AI-powered.

1