CPMG DATA

FIELD OF THE INVENTION

The present invention relates generally to NMR or MRI methods and systems, and particularly to a method for extracting multiple T2* values from a single set of CPMG data.

BACKGROUND OF THE INVENTION

It is known in the art of NMR (nuclear magnetic resonance) or MRI (magnetic resonance imaging) that the NMR signal in a so-called CPMG (Carr-Purcell-Meiboom- Gill) sequence decays at a rate that is strongly affected by the self-diffusion of the water in the tissue under examination. It is also well established that in the presence of a substantially uniform constant gradient in the Bo field (static magnetic field), the decay is exponential for a single value of D and T2 in the voxel under study. (As is known, D is the self-diffusion constant and T2 is the spin-spin decay constant.)

In the CLEAR SIGHT system of Clear Cut Medical Ltd., excised tissue is examined via a CPMG sequence in the presence of a strong magnetic field with a strong gradient. The RF coil and the hardware are described in US Patent 9310450. The sampled signal is converted to a set of spectra (one spectrum per echo) by performing the Fourier transform of the (complex) time domain signal. An example of a prior-art simulated signal and its spectrum are shown in Fig. 1A (which illustrates the simulated signal in the time domain) and Fig. IB (which illustrates the Fourier transform of the time-domain signal).

For each echo, a weighted average of all the points in the spectrum is calculated. Zero weights are given to points outside the spectrum from the tissue (i.e. noise points), in order to improve the signal-to-noise ratio (SNR), and weights that depend on the amplitude are given to points within the spectrum. For example, in the spectrum shown in Fig. IB, the non-zero weights are given to the points whose frequencies lie approximately in the range [-75, +50] KHz, with the largest weights given to the points approximately in the range [-25, 25] KHz. Using methods well known to those skilled in the art of mathematics, it is possible to calculate the weights that optimize the final SNR. The prior art version of the CLEAR SIGHT system uses such weights. The signal of weighted average vs. echo number is fitted to an exponential, known as the T2 value; high T2 values are an indication that the voxel contains cancerous tissue. The presence of a Bo gradient during signal sampling implies that the different points in the spectrum (i.e., the different frequencies) contain the contributions of different positions in space. Indeed, if the Bo field gradient were purely in the Z direction (that is, perpendicular to the tissue surface), the different points in the spectrum would correspond to different depths into the tissue. In practice, the gradient may have a X-Y component as well (i.e. parallel to the tissue surface), complicating the simple geometrical interpretation, but the essential concept remains the same; namely, different points in the spectrum (i.e., the different frequencies) contain the contributions of different positions in space.

SUMMARY OF THE INVENTION

The present invention seeks to provide improved methods for extracting multiple T2* values from a single set of CPMG data, as is describe in detail hereinbelow.

There is provided in accordance with a non-limiting embodiment of the invention a method for extracting multiple T2* values from a single set of CPMG data including using an MRI system that includes an RF coil to image tissue and generate NMR time domain signals via a CPMG sequence, converting the signals to a set of spectra, one spectrum per echo, by performing a Fourier transform of complex time domain signals associated with the signals, and dividing each of the spectra into a number of segments and calculating a T2* value for each segment, wherein the T2* values represent decay rates for a subset of frequencies in the spectrum of the signal.

The method may use equal frequency limits for all echoes in the CPMG sequence.

The segments may include overlapping segments.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be understood and appreciated more fully from the following detailed description, taken in conjunction with the drawings in which:

Fig. 1A is a graphical illustration of a simulated signal in the time domain of a prior art MRI system, the signal being generated by the MRI system imaging tissue via a CPMG sequence in the presence of a strong magnetic field with a strong gradient;

Fig. IB is a graphical illustration of the Fourier transform of the time-domain signal of Fig. 1A; and

Fig. 2 is a simplified flow chart of a method for extracting multiple T2* values from a single set of CPMG data, in accordance with a non-limiting embodiment of the present invention.

DETAILED DESCRIPTION OF EMBODIMENTS Reference is now made to Fig. 2, which illustrates a method for extracting multiple T2* values from a single set of CPMG data in accordance with a non-limiting embodiment of the present invention.

An MRI system is used to image tissue and generate NMR time domain signals via a CPMG sequence in the presence of a strong magnetic field with a strong gradient (step 1). The signals are converted to a set of spectra (one spectrum per echo) by performing a Fourier transform of the (complex) time domain signals (step 2). In accordance with a non-limiting embodiment of the present invention, instead of taking the weighted sum of all the (non-zero) points in the spectrum, the spectrum is divided into a number of segments, and a T2 ^* value is calculated for each segment (step 3). The same frequency limits may be used for all the echoes in the CPMG sequence (step 4). Accordingly, multiple T2 values are calculated for each voxel. For example, if the spectrum is split into two segments, two T2 ^* values, T2 ^* (l) and T ₂ ^* (2) are calculated. If the gradient is purely in the Z direction, then T2 (1) represents the decay rate for tissue closer to the RF coil, and

T2 (2) represents the decay rate for tissue further from the RF coil (i.e., deeper into the tissue). However, even in the general case where the gradient is in an arbitrary direction,

T2 (1) and T2 (2) represent the signals from different volumes in the tissue.

More than two segments may be chosen (as many as there are points in the spectrum), including overlapping segments (step 5); e.g., the first 50% of the frequencies, the middle 50% of the frequencies and the last 50% of the frequencies and many other variations. If any one of the segments exhibits a high T2 value, it can be displayed to the user as a possible indication that cancer is present in that segment of this voxel.

The advantages of splitting the spectrum into multiple segments is that (a) this provides knowledge of where the high T2 value comes from (i.e., how deep into the tissue) and (b) this effectively narrows the size of the effective voxel, narrowing the diluting effect of healthy tissue (i.e., the partial volume effect that occurs if the voxel contains a mixture of tumor and healthy tissue). The disadvantage of splitting the spectrum is that the signal-to-noise ratio (SNR) of each segment is lower than the full spectrum.