Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Is Registering Raw Tagged-MR Enough for Strain Estimation in the Era of Deep Learning?

Zhangxing Bian, Ahmed Alshareef, Shuwen Wei, Junyu Chen, Yuli Wang, Jonghye Woo, Dzung L. Pham, Jiachen Zhuo, Aaron Carass, Jerry L. Prince

TL;DR

The paper addresses motion and strain estimation in tag MRI (tMRI) under tag fading, modeling fading through $T_1$ relaxation and repeated RF pulses. It compares deep learning-based registration on raw tMRI against Harmonic Phase (HARP) methods, using a VoxelMorph-based framework with multiple similarity losses and a dedicated tag-fading simulation, evaluated on 1,000 simulated movies and a motionless phantom. The main contributions are (1) a mathematical tag-fading model that accounts for steady-state dynamics, (2) a systematic evaluation of popular image similarity losses for DL registration on fading tMRI, and (3) a finding that HARP-based inputs, particularly with NCC loss, yield more accurate motion and strain estimates than raw-tMRI inputs. The work highlights cautions for intensity-based registration in time-evolving contrast scenarios and suggests directions such as deep similarity metric learning and CSPAMM to mitigate fading while preserving fast acquisition for SPAMM.)

Abstract

Magnetic Resonance Imaging with tagging (tMRI) has long been utilized for quantifying tissue motion and strain during deformation. However, a phenomenon known as tag fading, a gradual decrease in tag visibility over time, often complicates post-processing. The first contribution of this study is to model tag fading by considering the interplay between $T_1$ relaxation and the repeated application of radio frequency (RF) pulses during serial imaging sequences. This is a factor that has been overlooked in prior research on tMRI post-processing. Further, we have observed an emerging trend of utilizing raw tagged MRI within a deep learning-based (DL) registration framework for motion estimation. In this work, we evaluate and analyze the impact of commonly used image similarity objectives in training DL registrations on raw tMRI. This is then compared with the Harmonic Phase-based approach, a traditional approach which is claimed to be robust to tag fading. Our findings, derived from both simulated images and an actual phantom scan, reveal the limitations of various similarity losses in raw tMRI and emphasize caution in registration tasks where image intensity changes over time.

Is Registering Raw Tagged-MR Enough for Strain Estimation in the Era of Deep Learning?

TL;DR

The paper addresses motion and strain estimation in tag MRI (tMRI) under tag fading, modeling fading through relaxation and repeated RF pulses. It compares deep learning-based registration on raw tMRI against Harmonic Phase (HARP) methods, using a VoxelMorph-based framework with multiple similarity losses and a dedicated tag-fading simulation, evaluated on 1,000 simulated movies and a motionless phantom. The main contributions are (1) a mathematical tag-fading model that accounts for steady-state dynamics, (2) a systematic evaluation of popular image similarity losses for DL registration on fading tMRI, and (3) a finding that HARP-based inputs, particularly with NCC loss, yield more accurate motion and strain estimates than raw-tMRI inputs. The work highlights cautions for intensity-based registration in time-evolving contrast scenarios and suggests directions such as deep similarity metric learning and CSPAMM to mitigate fading while preserving fast acquisition for SPAMM.)

Abstract

Magnetic Resonance Imaging with tagging (tMRI) has long been utilized for quantifying tissue motion and strain during deformation. However, a phenomenon known as tag fading, a gradual decrease in tag visibility over time, often complicates post-processing. The first contribution of this study is to model tag fading by considering the interplay between relaxation and the repeated application of radio frequency (RF) pulses during serial imaging sequences. This is a factor that has been overlooked in prior research on tMRI post-processing. Further, we have observed an emerging trend of utilizing raw tagged MRI within a deep learning-based (DL) registration framework for motion estimation. In this work, we evaluate and analyze the impact of commonly used image similarity objectives in training DL registrations on raw tMRI. This is then compared with the Harmonic Phase-based approach, a traditional approach which is claimed to be robust to tag fading. Our findings, derived from both simulated images and an actual phantom scan, reveal the limitations of various similarity losses in raw tMRI and emphasize caution in registration tasks where image intensity changes over time.
Paper Structure (5 sections, 2 equations, 3 figures, 1 table)

This paper contains 5 sections, 2 equations, 3 figures, 1 table.

Table of Contents

  1. INTRODUCTION
  2. Method
  3. Experiments
  4. Materials & Training Details
  5. Conclusion

Figures (3)

  • Figure 1: tMRI imaging sequence. "TF" stands for time-frame.
  • Figure 2: Simulation pipeline.
  • Figure 3: Qualitative results on real motionless data. Left panel shows horizontal (h-tag) and vertical (v-tag) tagging for three timeframes. Right panel shows the color-coded MPS$^{95}$ map and zoomed-in motion field of four methods. The first and second row show the registration results of ref-to-8$^{\text{th}}$ and ref-to-40$^{\text{th}}$ frame.