28.1 Magnetic Relaxation Times

Magnetic relaxation times (T1, T2, etc.) govern the processes that dictate the decay of MRI signal amplitude during an MRI acquisition and can be exploited to manipulate the tissue contrast in the resulting image.

T2: The spin-spin or transverse relaxation time T2 refers to the rate of the decay of the MRI signal over time resulting from the loss of phase coherence from neighboring nuclear magnetic moments after excitation has occurred, hence the term “spin-spin relaxation.” When the magnetic moments have phase coherence immediately after they have been excited by a radiofrequency pulse, they create a detectable net magnetization vector (the MRI signal). As the individual magnetic dipole moments influence one another, they lose phase coherence and no longer contribute toward the MRI signal (the signal strength decays to zero over time). This process is quite rapid, with soft tissue T2 times typically ranging from 40 ms and greater. Images that are obtained in order to primarily focus on this process are said to be “T2-weighted.”

T2*: In practice, localized inhomogeneities of the main magnetic field also contribute toward the rate of loss of phase coherence from neighboring spins. Combining this effect with the spin-spin dephasing governed by T2 relaxation gives a combined dephasing rate T2*. Images may specifically be acquired to be either T2- or “T2*-weighted.”

T1: The spin-lattice or longitudinal relaxation time T1 refers to the rate at which excited magnetic moments realign back to their unexcited equilibrium orientation over time. The term “spin-lattice relaxation” refers to the fact that this relaxation process is the loss of energy from the excited magnetic moments to their surroundings (typically large macromolecules). Due to the energy exchange nature of this process, T1 is typically an order of magnitude longer than T2 for most tissues, and it increases with magnetic field strength. Images that are obtained in order to primarily focus on this process are said to be “T1-weighted.”

T1⍴: Through the application of a continuous radiofrequency pulse to lock the spins in a single orientation, loss of phase coherence can be prevented and instead the rate of decay of the MRI signal can be made proportional to the low-frequency thermal interactions between the excited hydrogen molecules and local macromolecules. This process is governed by the relaxation time T1⍴. Through application of specific acquisition sequences that involve longer applications of continuous radiofrequency pulses, images may be acquired that are strongly “T1⍴-weighted.”

The number of protons available within a unit volume of tissue will also affect the amount of signal that tissue can contribute to MRI. If acquisition parameters are chosen such that there is no notable contrast in the image from either T1 or T2 decay, the image is said to be “proton density (or spin density) weighted.”

There are an incredible variety of additional types of MRI contrast available, as well as the ability to mix contributions of contrast in order to optimally focus on tissue and biological processes with MRI.

MRI may be acquired in either two or three dimensions in any orientation, with the static nature of most joints making them ideal targets for high-resolution three-dimensional imaging, which tends to require longer acquisition times. Three-dimensional MRI also benefits from the ability to reconstruct the data in multiple planes to observe a variety of structures within a joint. MRI is now widely held as a clinical gold standard for assessment of cartilage and joints. 1, 2

28.2 Imaging Coils

Most clinical MRI systems are whole-body imaging systems with a range of radiofrequency coils used to acquire images of specific body areas. Knee imaging, for example, is commonly acquired using a dedicated knee coil (see Fig. 28.2), although dedicated extremity systems are also available for rapid throughput of extremity imaging (see Fig. 28.3).

Extremity coils (such as knee coils) tend to have multiple receiver elements and benefit from a broadly cylindrical design to give excellent signal-to-noise ratio (SNR) and homogeneity of transmit field (B₁). Image quality and spatial resolution on these coils therefore tend to be some of the best available on clinical imaging systems. The majority of quantitative MRI cartilage research is therefore performed on whole-body MRI systems with dedicated body area–specific coils.

Dedicated extremity coils also benefit from a more reproducible placement of the body part than flexible, general purpose coils. Dedicated extremity coils are generally of solid construction and therefore fixate the volume of interest more reproducibly than flexible coils. When imaging subjects longitudinally (before and after surgical repair, for example), reproducibility of patient positioning within the imaging system is essential to assess subtle changes in joint morphology or geometry. A reproducible and rapid patient positioning protocol is also essential for studies involving large populations.

Due to their rigid construction, dedicated extremity coils do, however, have a finite bore size, so larger subjects may need to be scanned using flexible coils.

28.3 Multinuclear Coils

Hydrogen is the nucleus most commonly used in MRI due to the natural abundance of ¹H in the body as well as the high nuclear magnetic resonance sensitivity of ¹H. Most clinical imaging systems therefore use imaging coils tuned to the resonant frequency of ¹H. To obtain data from other nuclei such as in applications using ³¹P magnetic resonance spectroscopy (MRS) or ²³Na imaging, for example, coils must be used that are tuned to the specific resonant frequency of those nuclei. Multinuclear coils may be purchase or constructed that are either tuned to one specific frequency, or are switchable between frequencies (known as “dual-tuned”). For example, a ³¹P coil may be used that is “dual-tuned” to allow acquisition of ¹H data as well as ³¹P data. The much lower relative nuclear magnetic resonance sensitivity of non-¹H nuclei means that SNR of this type of data is much lower and that acquisition times may often therefore be much longer and data acquisition more challenging (see Table 28.1).