[Paper Review] Complicated relationships between tissue T2 relaxation time and in vivo tissue diffusion measures, depending on the ranges of T2 value

Yì Wáng|arXiv (Cornell University)|Jun 19, 2023

MRI in cancer diagnosisMedicine3 citations

TL;DR

This paper argues that apparent diffusion coefficient (ADC) values in MRI are significantly influenced by T2 relaxation time, especially in tissues with long T2, leading to misleading diffusion measurements. The study demonstrates that high T2 values can dominate signal intensity on b=0 and high b-value images, causing ADC maps to reflect T2 effects rather than true diffusion, particularly in tumors with cystic or myxoid degeneration.

ABSTRACT

Apparent diffusion coefficient (ADC) is a measure of the magnitude of diffusion of water molecules within tissues. We argue that ADC value contains information of both diffusion and T2 relaxation. In this letter, we list literature evidence to support our argument. Firstly, we list uterine myometrium tumors as examples. Myometrium has a T2 relaxation time of 79 ms at 3T. Literature shows, when the myometrium tumors are T2 weighted signal intensity hypertensive (relative to myometrium), these tumors likely show diffusion restriction. This is similar to that, while the spleen is T2 weighted signal intensity hypertensive relative to the liver, the spleen demonstrates diffusion restriction relative to the liver. On the other hand, when the myometrium tumors are T2 weighted signal intensity hypotensive (relative to myometrium), these tumors likely do not show diffusion restriction. However, when the myometrium tumors are very highly hypertensive such as the cases of leiomyoma cystic degeneration and myxoid degeneration, the relationship between T2 weighted signal intensity and diffusion is similar to that of a normal gallbladder, i.e., T2 weighted signal highly hypertensive without diffusion restriction. For most of the tumors, much longer T2 may suggest a tumor is waterier. On b=0 images, the signal of tumor (relative to other tissues) can be dominated by T2 effect. On a high b-value images such as b=1000 images, it is possible that tumor regions are dominated by noises. In the end, the ADCs, which are determined by the slope between the signal at b=0 and the signal at b=1000, are also dominated by the T2 effect, with longer T2 being associated with higher ADC measure. In fact, since there is no diffusion gradient with b=0 images, and there are only noises on b=1000 (or b=800), ADC map is devoid of diffusion information.

Motivation & Objective

• To investigate how T2 relaxation time influences apparent diffusion coefficient (ADC) measurements in in vivo tissues.
• To clarify the confounding role of T2 effects in diffusion-weighted imaging (DWI), particularly in tumors with abnormal T2 values.
• To challenge the assumption that ADC exclusively reflects water diffusion, showing instead that T2 can dominate signal behavior.
• To explain discrepancies in diffusion restriction assessment when comparing tissues with varying T2 relaxation times.
• To highlight clinical implications in interpreting DWI in tumors with high or very high T2, such as leiomyoma cystic degeneration.

Proposed method

• Review of existing literature on T2 relaxation times and diffusion measures in specific tissues, including uterine myometrium and spleen.
• Analysis of signal intensity patterns on b=0 and high b-value (e.g., b=1000) images in relation to T2 and ADC values.
• Use of signal intensity relationships between b=0 and high b-value images to infer ADC calculation bias due to T2 effects.
• Comparison of clinical cases with T2-hypertensive and T2-hypotensive tumors to assess diffusion restriction patterns.
• Evaluation of noise dominance on high b-value images, particularly in tissues with long T2, affecting ADC quantification.
• Application of the principle that ADC is derived from the slope between b=0 and high b-value signals, which can be skewed by T2 effects when b=0 is T2-dominated and high b-value is noisy.

Experimental results

Research questions

• RQ1 How does T2 relaxation time affect the apparent diffusion coefficient (ADC) in tissues with varying T2 values?
• RQ2 Why do some T2-hypertensive tumors show no diffusion restriction despite high signal intensity on T2-weighted images?
• RQ3 What explains the paradoxical lack of diffusion restriction in tumors with very high T2, such as those with cystic or myxoid degeneration?
• RQ4 To what extent does T2 relaxation dominate signal intensity on b=0 and high b-value images, affecting ADC map interpretation?
• RQ5 How does noise on high b-value images (e.g., b=1000) impact the accuracy of ADC measurements in tissues with long T2?

Key findings

• In tissues with long T2, such as uterine myometrium tumors with cystic or myxoid degeneration, high T2 leads to T2-weighted signal hyperintensity without corresponding diffusion restriction.
• ADC values are significantly influenced by T2 relaxation, with longer T2 values associated with higher ADC measurements, even in the absence of true diffusion changes.
• On b=0 images, signal intensity is primarily governed by T2 effects, making tumor signal appear elevated regardless of diffusion status.
• On high b-value images (e.g., b=1000), signal is often dominated by noise, especially in tissues with long T2, undermining the reliability of ADC calculation.
• ADC maps derived from the slope between b=0 and high b-value signals may reflect T2 effects more than true diffusion, particularly when b=0 is T2-dominated and high b-value is noisy.
• In cases like normal gallbladder or leiomyoma with cystic degeneration, high T2 leads to high signal on T2-weighted images but no diffusion restriction, demonstrating that T2 hyperintensity does not imply restricted diffusion.