A deep learning method for image‐based subject‐specific local SAR assessment

• Published in: Magnetic resonance in medicine Vol. 83; no. 2; pp. 695 - 711
• Main Authors: Meliadò, E.F., Raaijmakers, A.J.E, Sbrizzi, A., Steensma, B.R., Maspero, M., Savenije, M.H.F., Luijten, P.R., Berg, C.A.T.
• Format: Journal Article
• Language: English
• Published: United States Wiley Subscription Services, Inc 01.02.2020; John Wiley and Sons Inc
• Subjects: Algorithms; Artificial neural networks; Computer Simulation; convolutional neural network; Databases, Factual; Deep Learning; Full Papers—Computer Processing and Modeling; Healthy Volunteers; Humans; Image Processing, Computer-Assisted - methods; In vivo methods and tests; Machine learning; Magnetic Resonance Imaging; Male; Mathematical models; Models, Statistical; Neural networks; Neural Networks, Computer; parallel transmit; Phantoms, Imaging; Prostate; Prostate - diagnostic imaging; Prostatic Neoplasms - diagnostic imaging; Qualitative analysis; Reproducibility of Results; Safety factors; Safety margins; Signal-To-Noise Ratio; specific absorption rate; subject‐specific SAR assessment; Teaching methods; ultrahigh‐field MRI; deep learning; ultrahigh-field MRI; specific absorption rate; parallel transmit; subject-specific SAR assessment; convolutional neural network
• ISSN: 0740-3194; 1522-2594; 1522-2594
• DOI: 10.1002/mrm.27948

Purpose Local specific absorption rate (SAR) cannot be measured and is usually evaluated by offline numerical simulations using generic body models that of course will differ from the patient's anatomy. An additional safety margin is needed to include this intersubject variability. In this work, we present a deep learning–based method for image‐based subject‐specific local SAR assessment. We propose to train a convolutional neural network to learn a “surrogate SAR model” to map the relation between subject‐specific B1+ maps and the corresponding local SAR. Method Our database of 23 subject‐specific models with an 8–transmit channel body array for prostate imaging at 7 T was used to build 5750 training samples. These synthetic complex B1+ maps and local SAR distributions were used to train a conditional generative adversarial network. Extra penalization for local SAR underestimation errors was included in the loss function. In silico and in vivo validation were performed. Results In silico cross‐validation shows a good qualitative and quantitative match between predicted and ground‐truth local SAR distributions. The peak local SAR estimation error distribution shows a mean overestimation error of 15% with 13% probability of underestimation. The higher accuracy of the proposed method allows the use of less conservative safety factors compared with standard procedures. In vivo validation shows that the method is applicable with realistic measurement data with impressively good qualitative and quantitative agreement to simulations. Conclusion The proposed deep learning method allows online image‐based subject‐specific local SAR assessment. It greatly reduces the uncertainty in current state‐of‐the‐art SAR assessment methods, reducing the time in the examination protocol by almost 25%.