Multiparametric mapping in the brain from conventional contrast‐weighted images using deep learning
Shihan Qiu, Yuhua Chen, Sen Ma, Zhaoyang Fan, Franklin G. Moser, M. Marcel Maya, Anthony Christodoulou, Yibin Xie, Debiao Li
Abstract
Purpose To develop a deep‐learning–based method to quantify multiple parameters in the brain from conventional contrast‐weighted images. Methods Eighteen subjects were imaged using an MR Multitasking sequence to generate reference T 1 and T 2 maps in the brain. Conventional contrast‐weighted images consisting of T 1 MPRAGE, T 1 GRE, and T 2 FLAIR were acquired as input images. A U‐Net–based neural network was trained to estimate T 1 and T 2 maps simultaneously from the contrast‐weighted images. Six‐fold cross‐validation was performed to compare the network outputs with the MR Multitasking references. Results The deep‐learning T 1 /T 2 maps were comparable with the references, and brain tissue structures and image contrasts were well preserved. A peak signal‐to‐noise ratio >32 dB and a structural similarity index >0.97 were achieved for both parameter maps. Calculated on brain parenchyma (excluding CSF), the mean absolute errors (and mean percentage errors) for T 1 and T 2 maps were 52.7 ms (5.1%) and 5.4 ms (7.1%), respectively. ROI measurements on four tissue compartments (cortical gray matter, white matter, putamen, and thalamus) showed that T 1 and T 2 values provided by the network outputs were in agreement with the MR Multitasking reference maps. The mean differences were smaller than 1%, and limits of agreement were within 5% for T 1 and within 10% for T 2 after taking the mean differences into account. Conclusion A deep‐learning–based technique was developed to estimate T 1 and T 2 maps from conventional contrast‐weighted images in the brain, enabling simultaneous qualitative and quantitative MRI without modifying clinical protocols.