DCE-FORMER: A Transformer-based Model With Mutual Information And Frequency-based Loss Functions For Early And Late Response Prediction In Prostate DCE-MRI

Dynamic Contrast Enhanced Magnetic Resonance Imaging aids in the detection and assessment of tumor aggressiveness by using a Gadolinium-based contrast agent (GBCA). However, GBCA is known to have potential toxic effects. This risk can be avoided if we obtain DCE-MRI images without using GBCA. We propose, DCE-former, a transformer-based neural network to generate early and late response prostate DCE-MRI images from non-contrast multimodal inputs (T2 weighted, Apparent Diffusion Coefficient, and T1 pre-contrast MRI). Additionally, we introduce (i) a mutual information loss function to capture the complementary information about contrast uptake, and (ii) a frequency-based loss function in the pixel and Fourier space to learn local and global hyper-intensity patterns in DCE-MRI. Extensive experiments show that DCE-former outperforms other methods with improvement margins of +1.39 dB and +1.19 db in PSNR, +0.068 and +0.055 in SSIM, and -0.012 and -0.013 in Mean Absolute Error for early and late response DCE-MRI, respectively.

SFT-KD-Recon: Learning a Student-friendly Teacher for Knowledge Distillation in Magnetic Resonance Image Reconstruction

Deep cascaded architectures for magnetic resonance imaging (MRI) acceleration have shown remarkable success in providing high-quality reconstruction. However, as the number of cascades increases, the improvements in reconstruction tend to become marginal, indicating possible excess model capacity. Knowledge distillation (KD) is an emerging technique to compress these models, in which a trained deep teacher network is used to distill knowledge to a smaller student network such that the student learns to mimic the behavior of the teacher. Most KD methods focus on effectively training the student with a pre-trained teacher unaware of the student model. We propose SFT-KD-Recon, a student-friendly teacher training approach along with the student as a prior step to KD to make the teacher aware of the structure and capacity of the student and enable aligning the representations of the teacher with the student. In SFT, the teacher is jointly trained with the unfolded branch configurations of the student blocks using three loss terms - teacher-reconstruction loss, student-reconstruction loss, and teacher-student imitation loss, followed by KD of the student. We perform extensive experiments for MRI acceleration in 4x and 5x under-sampling on the brain and cardiac datasets on five KD methods using the proposed approach as a prior step. We consider the DC-CNN architecture and setup teacher as D5C5 (141765 parameters), and student as D3C5 (49285 parameters), denoting a compression of 2.87:1. Results show that (i) our approach consistently improves the KD methods with improved reconstruction performance and image quality, and (ii) the student distilled using our approach is competitive with the teacher, with the performance gap reduced from 0.53 dB to 0.03 dB.