Artificial intelligence enables whole-body positron emission tomography scans with minimal radiation exposure.

Yan-Ran Joyce WangLucia BarattoK Elizabeth HawkAshok J TheruvathAllison PribnowAvnesh S ThakorSergios GatidisRong LuSantosh E GummidipundiJordi Garcia-DiazDaniel RubinHeike Elisabeth Daldrup-Link

Published in: European journal of nuclear medicine and molecular imaging (2021)

Our CNN model could generate simulated clinical standard 18F-FDG PET images from ultra-low-dose inputs, while maintaining clinically relevant information in terms of diagnostic accuracy and quantitative SUV measurements.

Keyphrases

positron emission tomography
artificial intelligence
computed tomography
deep learning
low dose
convolutional neural network
big data
machine learning
pet imaging
high resolution
pet ct
high dose
magnetic resonance imaging
optical coherence tomography
dual energy