Reducing Imaging-Scan Times for MRI Systems

Statement of Government Interest

This work was supported by NIH grants R01 AR4504, AR051041 , and performed at a NIH supported resource center (NIH RR02305). The government may have certain rights in this invention.

Field of the Invention

The present invention relates to magnetic resonance imaging (MRI), and more particularly, a magnetic resonance (MR) pulse sequence for reducing imaging-scan times for MRI systems.

Background of the Invention

MRI, or Magnetic Resonance Imaging, (including spectroscopy, conventional, and fast imaging techniques) is viewed as a conventional medical procedure having acceptable risks and certain concerns regarding bio-effects and patient safety. Of these concerns, electromagnetic energy adsorption may result in a host of undesired effects such as tissue or cellular damage. Absorption of electromagnetic energy by the tissue is described in terms of Specific Absorption Rate (SAR) ₁ which is expressed in Watts/kg. SAR in MRI is a function of many variables including pulse sequence and coil parameters and the weight of the region exposed. In the United States, for example, the recommended SAR level for head imaging is 8 Watts/kg. Ti _p is commonly referred to as the longitudinal relaxation time constant in the rotating frame. Ti _p MRI produces images with contrast different from conventional T _r or T ₂ - weighted images. Ti _p relaxation is obtained by spin-locking the magnetization in the transverse plane with the application of a low power RF pulse(s). Ti _p relaxation is influenced by molecular processes that occur with a correlation time, τ _c , that is proportional to the frequency of the spin-lock pulse (γB-ι/2π). This frequency typically ranges from zero to a few kilohertz. In biological tissues, T-, _p is approximately T ₂ , the

spin-spin relaxation time constant, for very low amplitude spin-lock pulses and increases with higher intensity Bi fields. The sensitivity of Ti _p to low-frequency interactions facilitates the study of biological tissues in a manner that is unattainable by other MR methods. MRI using Ti _p -weighted contrast has been used to investigate and assess the condition of a variety of tissues such as breast, brain, and cartilage. Contrast in magnetic resonance (MR) images derives from the magnetic relaxation properties of tissues. Variations in tissue relaxation times help to distinguish the healthy and the pathological states. An unconventional contrast mechanism called Ti p imaging shows sensitivity to the breast cancers, early acute cerebral ischemia, knee cartilage degeneration during osteoarthritis, posttraumatic cartilage injury, and the intervertebral discs among people with nonspecific lower back pain. In addition, functional Ti p imaging shows an augmented signal to brain activation and oxygen consumption (metabolism), and other applications.

Time constraints during an MR clinical examination place certain restrictions on Ti p imaging sequences. For example, to diagnose a patient presenting chronic knee joint pain requires a pulse sequence with full volume coverage of the articular cartilage of the patella, femoral condyle and tibial plateau. Present, pulse sequences are insufficient, however, for a standard clinical examination, because of either incomplete anatomical coverage, or prohibitively-long scan durations. That is, present, single-slice, 2D TSE-based acquisition schemes require an acquisition time on the order of a couple of minutes per slice. This time quickly increases if multiple slices are required. Compounding the time issue is the fact that multiple acquisitions are required to generate Ti _p maps of the tissue. Spin-locked EPI (SLEPI) has a much briefer scanning time for single slice imaging, but the nonselective spin-lock pulse used does not allow for 3D data acquisition. A multi-slice 2D sequence with an equivalent adjacent slice-spacing to a 3D acquisition would result in cross-talk between slices due to imperfect excitation pulse slice profiles and thin slices are not achievable. Since Ti _p mapping involves collection of at least four 3D data sets at varying SL times, it is inherently inefficient.

Conventional 3D FGRE, multi-slice and 2D EPI-based sequences typically require 20-25 minutes for gathering a single Ti _p map. Still further, conventional 3D Ti _p

maps are typically collected with 2-4 mm slice thickness since it is too time consuming to collect 3D maps with isotropic voxel sizes. Ti _p -weighted volume sets in clinical MRI studies examining pathologies in extended regions, such as, the articular surfaces of the knee joint, brain and heart, cannot be obtained under the time constraints of a viable clinical exam. Therefore, at least two views, e.g. sagittal and axial, are required to properly visualize anatomical structures in 3D Ti _p maps, which presently requires a prohibitively long duration.

Summary of the Invention

To solve these and other problems, the present invention described herein, introduces a method of, and system for, rapid MRI imaging-scanning that provides 2D or 3D coverage, high precision, and high-temporal efficiency without exceeding SAR limits.

In one embodiment, this is accomplished by using a pulse sequence process that includes a Ti p preparation period followed by a very rapid image acquisition process, which acquires multiple lines of kspace data. The combination of Tip preparation and acquisition of multiple lines of kspace, allows scan times to be shortened by as much as 3 or 4-fold or more, over conventional scanning methods.

In one embodiment of the invention, the Ti p pulse sequence includes five stages: a pre-preparation, Ti p preparation, post-preparation, and image acquisition stages, and post image acquisition period which facilitates clinical imaging.

In addition to the decreased scan duration, there are other aspects of the invention. For instance, in yet another embodiment, the Ti p preparation period is insensitive to magnetic field inhomogeniety. This is valuable, since magnetic field inhomogenieties can cause image artifacts and prevent accurate diagnoses. In another embodiment, methods to reduce Ti p image blurring and incorrect measurement of Ti p relaxation times or enhance signal to noise, are performed by a post-processing filter, a variable-flip-angle acquisition method, and/or a Half-Fourier acquisition method. Additionally, scan time can still be further shortened by reducing the T1- relaxation delay as part of the post image acquisition period, while compensating for magnetization saturation.

The Ti _p pulse sequence of the invention may be adapted for use with a wide array of clinical assessments including but not limited to:, intervertebral disk pathology, tumors, to study Alzheimer's disease, neuro-degeneration, myocardial abnormalities, arthritis, joint injuries and abnormalities, heart disease, and scanning cartilage pathology.

Additional objects, advantages and novel features of the invention will be set forth in part in the description, examples and figures which follow, all of which are intended to be for illustrative purposes only, and not intended in any way to limit the invention, and in part will become apparent to those skilled in the art on examination of the following, or may be learned by practice of the invention. Brief Description of the Drawings

The detailed description is explained with reference to the accompanying figures. In the Figures, the left-most digit(s) of the reference number identifies the figure in which the reference first appears.

Fig. 1 illustrates an MRI system 100 within which the present invention can be either fully or partially implemented.

Fig. 2 is an exemplary method 200 for performing rapid MRI imaging- scanning through the use of an MRI system, such as system 100 of Fig. 1 Figure 2 needs to include post-image acquisition period.

Fig. 3 shows Ti p images of the brain at 3 Tesla using the methodology of the present invention.

Figs. 4a and 4b illustrate two alternate Ti p preparation embodiments.

Fig. 5 shows an exemplary pulse sequence for T1 p-prepared balanced steady-state free precession for rapid 3D imaging.

Fig. 6 shows an exemplary k-space filter used to reduce blurring during image acquisition.

Fig. 7 shows another exemplary pulse sequence for Ti _p -weighted MRI, a spin-locking pulse cluster, consisting of two anti-phase spin-locking lobes surrounded by alternate phase 90° RF pulses, including a pre-preparation (fat saturation), post- preparation (α/2), image acquisition (bSSFP) and post image acquisition (relaxation delay) periods.

Fig. 8 shows calculated T _1p -values obtained with the pulse sequence on an agarose gel phantom as a function of delay_time parameter. Also shown are the R ² values of the exponential fits per Equation 1. This is an example of a parametric mapping technique.

Fig. 9 shows images of the knee joint of a healthy volunteer acquired with the sequence and illustrates another example of the parametric mapping technique. The anatomy of the knee joint (bright signal from cartilage and synovial fluid, dark signal from bones such as the femur, tibia and patella) is clearly visible in the early spin lock duration images in both the sagittal and axial views. The long spin lock duration images show decreased signal in the cartilage of interest while signal from fluid remains strong

Fig. 10 shows images of the calculated parametric Ti _p value maps for the articular cartilage overlaid onto a T ₂ -weighted image..

Detailed Description of the Invention

Reference herein to "one embodiment", "an embodiment", or similar formulations herein, means that a particular feature, structure, operation, or characteristic described in connection with the embodiment, is included in at least one embodiment of the present invention. Thus, the appearances of such phrases or formulations herein are not necessarily all referring to the same embodiment. Furthermore, various particular features, structures, operations, or characteristics may be combined in any suitable manner in one or more embodiments. Fig. 1 illustrates an MRI system 100 within which the present invention can be either fully or partially implemented. As appreciated by those skilled in the art, there are various ways to implement an MRI system 100. In one possible embodiment, MRI system 100 includes hardware components 102, and a control system 104. As is well known by those skilled in the art, typical hardware components 102 include: a magnet 106 for producing a stable and very intense magnetic field, gradient coils 108 for creating a variable field, and radio frequency (RF) coils 1 10, which are used to transmit energy and to encode spatial positioning.

Control system 104 controls hardware components 102, such as the scanning sequencing operations, and processes information obtained from scanning. Control system 104 may be implemented as a computer or control device, which includes at least one processor 112, and memory 114. Memory 114 may include volatile memory (e.g., RAM) and/or non-volatile memory (e.g., ROM). It is also possible for other memory mediums (not shown) having various physical properties to be included as part of control system 104. Control system 104 may also include code 116 stored in memory 114, such as software and/or firmware that causes MRI system 100 to perform scanning, and processing of images.

Much of the discussion below will focus on embodiments for performing operations of control system 104 — that may be embodied as code 116 — used to control MRI system 100. In particular, the Tip sequence used for issuing RF and gradient pulses, and image acquisition stages.

Fig. 2 is an exemplary method 200 for performing rapid MRI imaging- scanning through the use of an MRI system, such as system 100 of Fig. 1. Method 200 includes blocks 202, 204, 206, 208, and 210. (each of the blocks represents one or more operational acts). The order in which the method is described is not to be construed as a limitation, and any number of the described method blocks may be combined in any order to implement the method. Furthermore, the method can be implemented in any suitable hardware, software, firmware, or combination thereof. Additionally, although each module in Fig. 2 is shown as a single block, it is understood that when actually implemented in the form of computer-executable instructions, logic, firmware, and/or hardware, that the functionality described with reference to it may not exist as separate identifiable block.

In block 202, pre-preparation is performed. In one embodiment, pre- preparation involves several RF, gradient pulses and delays that may be activated at any time during a sequence to modify T1 p contrast. It is appreciated by those skilled in the art after having the benefit of this disclosure that preparation periods may be used to complement Ti p imaging in order to reduce blurring, artifacts, etc. These are not necessarily mutually exclusive from the image acquisition period. Examples of pre-

preparation pulses include Inversion, Gradient Tagging, Diffusion-Weighting, and Spectral Excitation / Saturation. Each will be described with greater detail below as follows:

Inversion: An magnetization inversion block (typically an RF pulse with a flip angle of 180°) may be used to null signal from a certain tissue. One application is to reduce the signal from joint space fluid in the knee or fluid in the ventricles, which may wash-out or blur the Ti p image contrast. Variations of this block can be used for saturation recovery (an RF pulse with a flip angle of 90 or any general flip angle).

Gradient Tagging: The use of gradients to tag the spatial MR signal to form grids to track a region over time.

Diffusion-Weighting: The use of gradients to yield sensitivity to diffusion processes. Spectral Excitation / Saturation: Used to enhance or diminish sensitivity to magnetic nuclei precessing at different rates. This is commonly used to eliminate the signal from fatty tissues.

In block 204, Ti p preparation is performed. Ti p preparation involves instructing MRI system 100 to issue a series of RF pulses used to obtain Ti p contrast. There are several variations of RF pulses used. A novel variation of one embodiment is the δB0 and B1 insensitive sequence Figure 3, which refocuses dephasing caused by magnetic field inhomogenieties. Both external magnetic field inhomogenieties δB0 and RF field inhomogenieties B1 produce similar banding or shading artifacts in magnetic resonance images. Some of these artifacts are shown in Figure 3 on the human brain, which shows Ti p images of the brain at 3 Tesla. Low spin lock RF fields cause banding artifacts in traditional Ti p sequences (listed B1 Compensation and BO compensation above). To eliminate these artifacts a B1-and-B0-T1p-preparatory sequence may be employed which is immune to both kinds of artifacts. More details of these variations are given in Witschey, et al. Artifacts in Ti p-weighted imaging: Compensation for B1 and BO field imperfections. JMR 186:75-85 (2007), incorporated herein by reference.

Figs. 4a and 4b illustrate two alternate spin-acquisition embodiments. In the embodiment of Fig. 4a magnetization is flipped along the y-axis, where it nutates

about the effective field (z'-axis) back along the y-axis. In the embodiment of Fig. 4b the magnetization follows a similar path, but with two differences: (1 ) the excitation flip angle does not need to be 90° and (2) B1 insensitivity is maintained by flipping the magnetization along the -z-axis. The embodiment of Fig. 4b may be preferred over 4a, because it is insensitive to both external magnetic field inhomogenieties and RF field inhomogeniety.

Referring now to Fig. 5, is an exemplary pulse sequence for Ti p-prepared balanced-steady-state free precession for rapid 3D imaging. Ti p-weighted MRI uses a balanced gradient echo as illustrated in Figure 5. "Balanced" means the transverse phase of magnetic nuclei due to gradient pulses (i.e., gradient moment) is zero at the end of each repetition time (TR).

One way to achieve Tip contrast is to apply a 90° pulse with an arbitrary initial phase flips the initial magnetization into the transverse plane where it is spin- locked by a pair of rotary echo pulses (phase ±90 degrees to the initial pulse) which provide Ti p-weighting to the initial magnetization MO. The duration or amplitude of the spin-locking pulse determines the final Ti p contrast in the image. Following the spin- lock period, the magnetization is flipped longitudinally by another 90° pulse (phase 180 degrees to the initial pulse). The Ti p prepared magnetization is stored for image acquisition using the balanced gradient echo (balanced Steady-State Free Precession bSSFP) sequence. Those skilled in the art, will appreciate that the angle of the pulse is not restricted to a 90 pulse for the flip angle after having the benefit of this disclosure. An initial b-SSFP preparation period is used prior to image acquisition to reduce artifacts caused by blurring caused by the transient echo amplitudes during the initial bSSFP image acquisition. The initial bSSFP preparation period consists of an α/2 pulse (phase 0) used to prevent oscillations of the transient echo amplitudes during the image acquisition period. Following this α/2 pulse, any number of dummy pulses of phase alternating (± 0) α pulses is applied to pulse the magnetization toward the steady- state. The loss of the initial T1 p prepared magnetization depends on the flip angle of the α pulses, the repetition time (TR) and spin-lattice T1 and spin-spin T2 relaxation times. To achieve optimal Ti p-weighting, there are no dummy pulses following the initial Ti p preparation period.

Next phase alternating (±0 )α pulses are used to acquire the Ti p prepared magnetization. The image acquisition gradients consists of both frequency and phase encoding gradients to acquire the kspace data. In this example, the echo amplitudes are recorded using a rectilinear kspace acquisition with frequency encoding performed in the x-direction and phase encoding performed in the y- and z-directions.

Finally, the magnetization is restored to MO by a T1 relaxation delay period where no RF pulsing occurs. While Figure 5 shows a specific example of rapid Ti p image acquisition, it should be understood by those skilled in the art, after having the benefit of this disclosure that the combination of a Ti p preparation period with ANY rapid image acquisition technique is a part of the present invention. Thus, many modifications to the pulse sequence shown in Figure 5 achieve the same result of rapid Ti p image acquisition. Some examples of generalizations of the pulse sequence are explained here.

Gradient Echo: The balanced gradient echo (bSSFP) shown in Figure 5 is only one example of the more general gradient echo. Gradient echoes, both spoiled and unspoiled, which spoil or refocus the gradient moment are equally valid techniques for rapid image acquisition. Any rapid gradient echo sequence consists of a series of RF pulses during which the so-called 'gradient echo' is used to acquire kspace data. This technique has multiple names: SPGR, FFE, FLASH, FISP, SSFP, FIESTA, CISS, DESS ₁ as well other names. By definition, the gradient echo uses a gradient to dephase the magnetization in the transverse plane with a gradient dephaser pulse, followed by a gradient rephaser during which kspace data is acquired with the MR hardware. The technique for gradient echo image acquisition consists of both phase and frequency encoding to 'encode' the spatial magnetization in kspace. The encoding is reconstructed using a technique such as the fast Fourier Transform to create an image. The RF pulses in the gradient echo sequence can be phase cycled to reduce artifacts or adjust the image contrast. A specific example of gradient echo RF pulses is a phase cycling routine which causes RF spoiling of transverse magnetization. By rotating the phase of the RF pulse each acquisition period, the transverse magnetization

accumulates an arbitrary phase each period and can cancel the transverse magnetization from a previous period. This technique is useful for preventing steady- state artifacts.

Following the acquisition of the kspace data during a repetition period, the gradient moment can be refocused (balanced) or spoiled. Either technique can be used, however, in the exemplary embodiment, the balanced-gradient-echo technique is used, because it achieves higher signal than spoiled gradient echo techniques. Balanced gradients refocus the gradient moment of transverse magnetization each repetition time, while spoiled gradients further dephase the transverse magnetization by a large gradient.

Spin Echo: Rapid spin echo acquisition is obtained by modifying the gradient echo sequence above to refocus magnetic field inhomogeniety by the use of a refocusing RF pulse. This technique has multiple names: Carr-Purcell Meiboom Gill Spin Echo, Fast Spin Echo, Turbo Spin Echo, Specifically, an RF pulse is used to generate transverse magnetization. Because of local magnetic field inhomogenieties, the transverse magnetization is dephased, however, if a refocusing pulse is used at a time TE/2, at a later time TE, an echo is created by rephased magnetization. Half-Fourier Acquisition: Half-Fourier acquisition is used to acquire a partial set of kspace data. This can also be used to increase the SNR and reduce blurring by simultaneously increasing the flip angle during image acquisition.

Variable Flip Angle Image Acquisition: The flip angle of the gradient or spin echo image acquisition train may be varied to maintain Tip contrast and prevent imaging artifacts. One such application is to reduce blurring caused by the approach to the steady-state by repeated RF pulsing. Having explained the Tip preparation of block 204, it is now possible to discuss blocks 206, and 208.

Referring back to Fig. 2, in block 206 post preparation is performed. Post preparation involves generally the same pulsing as performed with pre-preparation (see the discussion above with respect to block 202 above).

In block 208, image acquisition is performed. Image acquisition is obtained using a very rapid gradient echo or spin echo acquisition technique. Instead of acquiring only a single line of kspace data, a rapid image acquisition technique can acquire any

number of lines of kspace data following the initial Ti p preparation of block 204 (Fig. 2). This allows the scan time to be substantially shortened. If, for example, 128 lines of kspace are acquired immediately following the T1 p preparation, then the scan time is shortened by 128-fold. This technique is especially suitable for clinical imaging, where patient motion or comfort is prohibitive.

In block 210 is a post image acquisition period. An example of post-image acquisition periods are T1 relaxation delays or storage pulses. See also the discussion below. Example of block 210 is perform the post image acquisition delay. It is possible to further accelerate the image acquisition by shortening the T1 -delay in Figure 6. T1- delays are important to fully recover the magnetization to return to its equilibrium distribution. This can take as long as 2-4 s, depending on the tissue. It is possible to substantially shorten the T1 -delay (for example, to 0.3-0.4 s), shortening the scan time by another eight-to-ten fold, however, a model for magnetization saturation must be used to obtain the corresponding Ti p contrast. One such model can be calculated for Ti p magnetization in the steady-state after repeated Ti p preparation periods. The model depends only on the T1 , T2, flip angle and image acquisition techniques, but has been implemented for 3D Ti p imaging To reduce blurring a k-space filter may be used to correct a non-constant echo amplitude during image acquisition as shown in the embodiment of Fig. 6. The filter design of Fig. 6, compensates for the transient signal during b-SSFP acquisition of Ti p magnetization.

As a result of using method 200, time constraints during an MR clinical examination are eliminated. For example, prior to the invention, Tip imaging in a patient presenting chronic knee joint pain required a pulse sequence with full volume coverage of the articular cartilage of the patella, femoral condyle and tibial plateau. Two conventional choices are superior to others, a Ti p prepared 2D multislice fast spin echo sequence or a Ti p prepared 3D gradient echo Ti p imaging sequence. Still both sequences are insufficient for a standard clinical examination with incomplete volume coverage or unreasonably long scan times.

The sequence of method 200 obviates these slower conventional sequences. To be more specific, conventional sequences acquire only a single line of kspace data

after the initial Ti p preparation period. This is because only a single gradient echo or spin echo is acquired following the initial Tip preparation. In accordance with the present invention, multiple RF pulses (2 or more and likely 128 or 256 pulses), gradient echoes or spin echoes are used to acquire the kspace data as in method 208. There are numerous ways to acquire kspace space data including but not limited to Cartesian, radial or spiral acquisitions. This technique rapidly accelerates the time for image acquisition and the scan time is shortened proportional to the number of kspace lines that are acquired following the initial T1 p preparation.

It is possible to further accelerate the image acquisition by shortening the T1- delay in Figure 6. T1 -delays are important to fully recover the magnetization to return to its equilibrium distribution. This can take as long as 2-4 s, depending on the tissue. It is possible to substantially shorten the T1 -delay (for example, to 0.3-0.4 s), shortening the scan time by another eight-to-ten fold, however, a model for magnetization saturation must be used to obtain the corresponding Tip contrast. One such model can be calculated for Ti p magnetization in the steady-state after repeated Ti p preparation periods. The model depends only on the T1 , T2, flip angle and image acquisition techniques, but has been implemented for 3D Ti p imaging. Another feature of the T1 p acquisition is the use of multiple spin lock amplitudes or durations to generate parametric maps. Two such examples are a Ti p map measuring the spatial Ti p relaxation times or the spatial distribution of the signal obtained at a ratio of different spin lock amplitudes.

Another feature of the T1 p acquisition is the use of exogenous contrast agents to complement or enhance the T1 p contrast. Two such examples are inhaled magnetic molecular oxygen ( ¹⁷ O ₂ ) or paramagnetic contrast agents such as Gd-DTPA. The foregoing can also be surmised as follows: A balanced Steady-State Free Precession (b-SSFP) technique of rapid image acquisition of single-slice, multi-slice, or three-dimensional images and has been found to be an exceptional pulse sequence candidate for imaging articular cartilage, especially in patients with osteoarthritis. This sequence is also commercially named true fast imaging with steady precession (TrueFISP), balanced fast field encoding (b-FFE), and fast imaging employing steady- state excitation (FIESTA). In its conventional version, the b-SSFP pulse sequence

consists of a series of excitation pulses of alternating phase, each followed by a gradient-echo readout, and is capable of generating images with contrast based on the ratio T2/T1. Except, here a method for acquiring Ti _p -weighted three-dimensional volumes in a time-efficient manner by using spin-lock pulses in conjunction with the b- SSFP technique in a new pulse sequence called SLIPS (Spin-Locked Imaging with Precession in the Steady-state) is described. The signal expression of the new sequence was simulated and actual Ti _p measurements were performed in a homogeneous phantom of known T _1p as well as in vivo in the human knee joint to map Ti _p in cartilage. Pulse sequence design

Fig. 7 shows a pulse sequence in accordance with another embodiment of the invention. A fat saturation pulse was applied in each segment to attenuate signal from fat in the marrow of the knee joint during in vivo experiments. T _1p contrast was generated by applying a four-pulse cluster during each segment after fat-saturation. In this pulse cluster, a non-selective ττ/2 pulse excites spins into the transverse plane that are then spin-locked by the application of two phase-alternating (±90° phase-shifted from the phase of the first ττ/2 pulse) SL pulses. Phase-alternating SL pulses have been previously demonstrated to reduce image artifacts resulting from Bi inhomogeneity. The duration of the SL pulses is denoted as TSL and typically ranges from a millisecond to ~50ms for in vivo imaging. A delay of 20μs was maintained between the SL pulse segments hardware switching between RF pulse excitations. A second non-selective ττ/2 pulse then returns the magnetization to the longitudinal axis. For imaging the Ti _p -prepared signal, a sequence of excitation pulses and gradient echo sampling are used to acquire the segment of k-space. The first excitation pulse has angle α/2, and the following pulses have angle α, with alternating phase, until the final pulse, which has angle α/2. Each of these excitation pulses is separated by a time defined as "short repetition time" (Short TR). Once the whole segment of k-space has been acquired, the magnetization is allowed to relax toward thermal equilibrium for a time defined as "long repetition time" (Long TR). This acquisition method is then repeated for each subsequent segment in k-space until the whole volume has been acquired.

The b-SSFP acquisition of Spin-Lock prepared magnetization signal greatly reduces scan time, but also increases the complexity of the weighting of the signal as compared to TSE-based spin-locking sequences. To begin, the short TR approximation of the steady-state signal generated by b-SSFP is given by the equation:

The equation shows a Ti/T ₂ -weighting that is typical in b-SSFP images. The addition of SL pulses does not result in an additional multiplicative factor to this equation (as it does with the single-slice TSE-based and EPI-based methods). A complicated signal expression arises from the fact that the prepared magnetization and steady-state magnetization are not directly related to each other. The reason for this is that the preparatory RF pulses (e.g. fat-saturation, Ti _p preparation, etc.) have the greatest effect on the magnetization immediately after their application, while the steady-state magnetization is produced only after a long period of repetitive pulsing. For this reason, the magnetization gradually reduces from a T _1p -prepared to a steady-state value in the SLIPS pulse sequence. Materials And Methods

A MRI "phantom" and two healthy male volunteers were imaged on a 1.5T Sonata Siemens clinical MRI scanner (Siemens Medical Solutions, Erlangen, Germany) using an eight-channel knee coil (MRI Devices Corp., Muskego, Wl). The phantom consisted of gel of 2% (w/v) agarose in phosphate-buffered saline (Sigma-Aldrich, St. Louis, MO) doped with 0.2 mM MnCI ₂ to reduce T-i. For this particular study, we only studied healthy subjects without any clinically meaningful acute or chronic medical problems. Estimation of T _{1 p} in agarose phantoms

The ability of pulse sequence to estimate accurate Ti _p values was evaluated using two agarose bottle phantoms. A series of Ti _p -weighted images were acquired with the pulse sequence at seven spin-lock durations (TSL) (1 , 5, 10, 20, 30, 35, and 40 ms). Other imaging parameters were TE = 2.5 ms, short TR = 5 ms, FOV = 180mm, 256x128 matrix size with 4mm-thick sections and spin-locking frequency at 400 Hz.

The parameter long TR was varied to determine its dependence on the resulting calculated Ti _p values. Circular regions of interest (ROI) were manually selected by a single user in each phantom. Identical ROIs were applied for all scans of the same phantom. Mean intensity values were calculated within the ROI and the results were fit to Equation 1 to generate an exponential decay rate with respect to TSL. Estimation of Ti _p in the human knee articular cartilage

The utility of the pulse sequence to generate meaningful and accurate Ti _p maps in vivo was evaluated. Each subject's left knee was imaged by placing the knee in the coil, and padding was placed to restrict motion during the scan. A series of Tip- weighted images were acquired with pulse sequence at five spin lock durations (1 , 10, 20, 30, and 40 ms). The TE and short TR parameter values used were calculated for the minimum possible values under SAR limitations. Therefore, TE varied between 2.5 and 3.0 ms, and short TR was exactly twice the TE value. However, within each series, the two parameters were kept constant. Other imaging parameters were long TR = 1 s, FOV = 140mm, 256x128 matrix size with 4mm-thick slices, and a spin-locking frequency fixed at 400Hz. Each data set (one per TSL) was smoothed using a 3x3x3 averaging matrix. These data sets were then used to generate Ti _p "maps" by fitting signal intensities as a function of TSL by linear regression to Eq. 1. In the fitting routine, pixels whose intensities correlated poorly (i.e., R ² < 0.98) with the equation were set to zero.

The calculated Ti _p values from these maps were verified by comparing them to single-slice Ti _p maps of the center slice of the acquisition volume obtained with a 2D TSE-based T _1p pulse sequence. These images were acquired with the same FOV, slice thickness, and image dimensions as the Ti p images images. In all, each MRI exam was conducted in less than thirty minutes, including a scout image and full collection of two views (sagittal and axial) with a set of five TSLs per view and imaging matrix of 256x128. Data Processing The images were transferred to a Dell lnspiron computer (Dell Inc., Round

Rock, TX) for processing. Phantom and human knee images were processed in custom-written software in the IDL programming language (RSI Corp., Boulder, CO). Measurements of T _1p relaxation times were performed on the entire FOV of the images.

Measured values of T _1p in the articular cartilage made with both the SLIPS and SL-TSE sequences were overlaid onto T ₂ -weighted images (Figure 4). To increase consistency of comparison of scans between the same patient, IDL code was used to co-register and realign images. Results

Figure 8 shows plots of calculated T _1p values for the agarose phantoms and resultant R- squared values of the exponential fits, respectively. As can be seen, as long TR is increased, calculated T _1p values for the phantom increase, and the R-squared value of the fit approaches the ideal value of 1.

Rapid T1 p -acquired images of the knee joint of a healthy volunteer are shown in Fig. 9. Fat-suppressed anatomy is clearly shown in the low TSL images.

Higher spin lock duration images are reduced in intensity, but fluid intensity remains very high. The cartilage also retains much of its intensity as it has a higher T _1p value than the surrounding tissue. The Ti _p map generated from the Tip-weighted images (Figure 10, left) show similar values to those obtained by the Ti _p -TSE sequence (Figure 10, center).

Application of the pulse sequence to measure T _1p in knee cartilage was demonstrated in this work. The pulse sequence has the advantage of rapid three- dimensional acquisition of T _1p data over the conventional Ti _p -prepared TSE sequence. Studies that examine T _1p of all articular surfaces of cartilage in the knee joint can potentially be performed clinically by using pulse sequence. The pulse sequence scheme used also allows for the addition of more slices in the acquisition volume, without significantly adding to the scan time, since the actual active scanning acquisition time is proportional to the number of slices, but very short in comparison to the delay time during which no acquisition is taking place. Each additional slice will add an additional time NPE X (Short TR) per volume acquisition where NPE is the number of phase encoding lines per slice. The disclosure of each patent, patent application and publication cited or described in this document is hereby incorporated herein by reference, in its entirety.

While the foregoing specification has been described with regard to certain preferred embodiments, and many details have been set forth for the purpose of

illustration, it will be apparent to those skilled in the art without departing from the spirit and scope of the invention, that the invention may be subject to various modifications and additional embodiments, and that certain of the details described herein can be varied considerably without departing from the basic principles of the invention. Such modifications and additional embodiments are also intended to fall within the scope of the appended claims.