WO1996004844A1

WO1996004844A1 - A method for quantitation of boron concentration by magnetic resonance imaging

Info

Publication number: WO1996004844A1
Application number: PCT/US1995/010191
Authority: WO
Inventors: Kenneth M. Bradshaw; Gary H. Glover; J. Rock Hadley; Martin P. Schweizer
Original assignee: University Of Utah Research Foundation
Priority date: 1994-08-11
Filing date: 1995-08-11
Publication date: 1996-02-22
Also published as: AU3320995A

Abstract

A magnetic resonance sensitive (MRS) element, notably boron (B), is compounded in a carrier, such as borocaptate sodium (22). A subject, typically a patient (21) with a brain tumor (24), is positioned in the imaging field of a magnetic resonance imaging machine, along with a reference (38) containing a known concentration of the MRS element. The magnetic resonance imaging machine is operated to produce a signal encoded by a reconstruction method that takes data immediately following the transverse RF pulse. The data is reconstructed to provide, for example, a subject image and a reference image. The subject image is calibrated from the reference image to derive a data matrix relating the concentration of the MRS element in the subject to locations within the subject. The data matrix may then be displayed as an image or downloaded to a computer for controlling radiation dosage during neutron therapy.

Description

-1-

A METHOD FOR QUANTITATTON OF BORON CONCENTRATION BY MAGNETIC RESONANCE IMAGING

BACKGROUND OF THE INVENTION

This invention was made with Government support under Contract No.

R43NS 30746-01 awarded by the National Institute of Neuro Disorders and Stroke

(NINDS) and C85-110763 to 16 awarded by the Department of Energy. The

United States government has certain rights in the invention. Field of the Invention: This invention pertains to the field of Magnetic

Resonance Imaging and more particularly to a method of quantitation of Boron in vivo and in vitro by magnetic resonance imaging.

State of the Art: Boron Neutron Capture Therapy (BNCT) is a promising new treatment that is under development for individuals with incurable brain tumors. Accurate neutron dose control is needed to optimize BNCT treatment procedures. The term "dose control" means controlling the direction, intensity, and duration of a neutron beam irradiating a tumor in a subject.

BNCT is currently limited by the lack of technology capable of determining

(quantitating) the actual concentration of boron in a subject. Boron uptake (rate of increase in concentration of boron in tissue with respect to time) varies by tissue type, by species, age and physical condition of the subject, and many other factors. Concentration mapping (the preparation of a pictorial presentation of the concentration of Boron as a function of position in a subject) of boron currently relies upon a multiplicity of biopsies from sacrificial animals or limited samples such as blood samples from a human subject. Concentration mapping is thus inaccurate, slow, and not readily integrated into patient treatment planning. BNCT requires a "capture element," such as the chemical element boron (¹⁰B). The BNCT procedure involves the "selective assimilation" or preferential deposition of the capture element in the tumor tissue. A neutron beam is used to promote fission of the capture element. A fission decay product serves as a lethal agent for killing tumor cells.

Boron compounds used for therapy are nontoxic and inert until the boron nucleus absorbs (captures) a thermal neutron resulting in fission particles and the release of thermal energy. A "harmless" boron compound can be selectively deposited into tumor tissue; it is then exposed to a beam of neutrons. The release of fission products from neutron capture by the ¹⁰B nucleus destroys the tumor cells. Selective assimilation effectively enables treatment of tumor cells with little harm to nearby normal tissue. The blood-brain-barrier normally separates the main components of the central nervous system from the blood in healthy tissue, restricting large molecules from passing into the central nervous system. A growing tumor may damage this barrier as it "recruits" its own blood supply. The large molecule, BSH, will then pass readily into a tumor but not into adjacent healthy brain tissue. This selective assimilation of BSH by the tumor is the basis for subsequent selective destruction of the tumor.

A neutron beam of epithermal neutrons (as opposed to "fast," with excessive energy, or "thermal," with inadequate energy,) penetrate the scalp and skull to provide a lethal dose of radiation to a deep-seated tumor. The brain tissue acts as a moderator to slow epithermal neutrons to thermal levels, at which they can be readily captured by the ¹⁰B nucleus. The nucleus then releases decay products, such as Lithium and Helium nuclei. The ⁷Li and He particles travel only about 10 microns, providing a destruction volume approximately the size of an average cancer cell. Treatment planning for BNCT assumes knowledge of boron levels in the tissue of interest. Boron concentration in tissue may be determined empirically from invasive animal studies or from preclinical human pharmaco-kinetic studies, neither of which has been practical in connection with BNCT procedures. In actual BNCT practice, boron levels are typically estimated, the estimates being based upon the boron concentration measured in blood samples. Radiation dose distribution is then calculated based upon these estimates, rather than actual boron concentration data.

It is recognized that the correlation between the concentrations of boron in blood and brain tissue, respectively, is imperfect. Moreover, every person is unique as is every individual and species of animal. Thus the absoiption, migration and accumulation of boron varies widely. To ensure that boron accumulates in a tumor at levels required for treatment, a technique is needed that can accurately monitor actual boron concentrations immediately, repeatedly and noninvasively within a tumor of an individual patient.

DEFINITIONS U.S. Patent 4,818,937 (Ailian et al.) issued April 4, 1989 and U.S. Patent

4,583,044 (Case et al.) issued April 15, 1986 are incorporated herein by reference for their disclosures concerning common terminology and technology related to magnetic resonance imaging and nuclear magnetic resonance phenomena generally. Unless some other meaning is clearly intended by the context of this disclosure, the following terms should generally by understood as follows:

"Boron neutron capture therapy" or "BNCT" refers to a method of destruction of tumorous cells according to which boron is introduced in a manner such that it binds to a molecule in a cell, after which the cell is irradiated by a neutron beam to cause localized fission destroying the cell. This phenomenon is reported in The International Journal of Radiation Oncology. Biology. Phvsics. vol. 28, no. 5, pp. 1067-1217; Barth R.F., Soloway A.H., Fairchild R.G., "Boron neutron capture therapy of cancer," Cancer Res (1990): 50: 1061-1070; Kabalka G.W. , Davis M., Bendel P., "Boron-11 MRI and MRS of intact animals infused with boron neutron capture agent," Magn Res Med (1988): 8:231-237; and other articles. The foregoing references are incorporated herein by reference for their teachings concerning BNCT and its associated technologies. Locher G.L. , "Biological effects and therapeutic possibilities of neutrons," Am J Roentgenol Radium Ther (1936): 36:1-13; and Sweet W.H. , "The use of nuclear disintegrations in the diagnosis and treatment of brain tumor," N Engl J Med (1952): 245:875-878 are incorporated herein by reference for their teachings concerning radiation therapy in general.

"MRI" is an acronym for "magnetic resonance imaging." This technique is based upon a phenomenon known as "nuclear magnetic resonance, " reflecting the influence of a magnetic field on the nuclei of atoms of interest. The techniques of MRI are regularly reported in journals such as Magnetic Resonance Imaging. Journal of Magnetic Resonance Imaging, and Society of Magnetic Resonance in Medicine. The principles of MRI are disclosed in Bovey F. A. , "Nuclear Magnetic Resonance Spectroscopy," 2nd Ed. p. 3, Acad. Press (1988); Richards T.L., Bradshaw K.M., "¹⁰B versus ⁿB signal-to-noise comparison at 2 Tesla" (abstr), In: Book of abstracts: Society of Magnetic Resonance in Medicine (1989); Berkeley, CA: Society of Magnetic Resonance in Medicine (1989): 665; Bradshaw, K.M., Richards T.L. , "In vivo chemical shift imaging of canine brain tumors after injection of Na₂B₁₂H_nSH" (abstr), In: Book of abstracts: Society of Magnetic Resonance in Medicine (1990); Berkeley, CA: Society of Magnetic Resonance in Medicine (1990): 1166; Kabalka G.W., Cheng G., Bendel P., "In vivo boron-11 MRI and MRS using

in the rat" (abstr), In: Book of abstracts: Society of Ma netic Resonance in Medicine (1991); Berkeley, CA: Society of Magnetic Resonance in Medicine (1991): 519.

"Reconstruction Method" refers to known methodology for encoding and decoding the signals derived during operation of an MRI machine. A reconstruction method of particular interest is "three dimensional projection reconstruction" or

"3DPR." This method provides an encoding and decoding technique by which MRI data can be reconstructed for the time period immediately following the application of a transverse, radio-frequency, magnetic field to a subject as the encoding gradients are being applied. The 3DPR method provides the distinct advantage of reconstructing data to yield a high signal to noise ratio (S/N) for the time period before the signal due to boron spin has decayed substantially.

The method is described in Glover G.H., Pauly J.M. , Bradshaw K.M., "Boron- 11 imaging with a three-dimensional reconstruction method," JMRI (1992): 2:47-52; Hayes C.E., Edelstein W.A., Schenck J.F., Mueller O.M., Eash M. , "An efficient, highly homogeneous radiofrequency coil for whole-body NMR imaging at 1.5T," J Magn Reson (1985): 63:622-628, incorporated herein by reference for their teachings of the operational principles of 3DPR, MRI, boron neutron capture therapy, general image reconstruction methods and the encoding and decoding of signal data from MRI machines. "T2" is the time required for the free induction decay of nuclear spin effects in a material subject to nuclear magnetic resonance effects. For example, a target may be placed inside a field coil to be exposed to a magnetic field, giving rise to a spin vector. When power to the field coil is shut off, the field collapses, and the spin effects decay in a time, T2.

"Tl " is the relaxation time required for the free induction decay of the spin effects of nuclei of interest (such as boron compounds), creating a detectable signal. For example, a transverse magnetic field at a radio frequency (RF field) may be pulsed, stimulating spin alignment of nuclei of interest within the field. When the field collapses, the spins of the nuclei of interest return from an orientation aligned with the field to their natural, random orientation. The nuclei of interest create their own detectable effects until they return to the natural orientation. "Fast" refers to the speed of relaxation of spin effects or spin vectors of nuclei in the imaging field of an MRI machine. A "Fast T2," for example, means that the T2 relaxation time is of a short duration compared to the time that encoding gradients are applied or that the MRI signal can be acquired. T2 can also be considered fast relative to Tl , the relaxation of spin effects due to the BQ field or its modulated, encoding gradients in an MRI machine. Boron has a time constant of T2 of approximately 600 microseconds, while the magnetic fields required for imposition of encoding gradients on a magnetic field may exist for milliseconds. Thus, the S/N ratio is poor for boron and any other nucleus having a T2 value less than a millisecond. "Biological material" means any material derived from a living organism.

The biological materials of immediate interest are body members of animals and patients as well as blood serum albumin and tissue samples.

"Target" means any material to be irradiated by a source of neutrons or to be imaged by magnetic resonance imaging. For example, a target may be a biological tissue sample in a test tube or a region, such as a head or a tumor, of a live subject.

A "subject" is a living person or animal or a sample of tissue or other biological material for study. For example, a person having a tumor is a subject likely to be treated, the head being the target of MRI and neutron irradiation. "Infusion" means any dispersion of a material into a target. For example,

Sodium borocaptate may be mixed with a tissue sample in a test tube or injected into a muscle or blood vessel to disperse in a region of interest in a live subject. The infusion may be made so that a region of interest may be more effectively imaged by MRI or irradiated by a neutron source.

A "capture element" is a chemical element having a nucleus capable of capturing a neutron from a neutron beam and splitting (fissioning) in response. Boron is an example of a capture element. Boron may be compounded in borocaptate sodium (BSH) so that it will chemically bind to molecules in biological material creating sites for fission. The boron isotope ¹⁰B is preferable for capturing neutrons and may be present as 95 percent of the infused boron during treatment. "MRS element" is a "magnetic resonance sensitive element" that may be imaged effectively. The specific MRS element selected depends upon the selection of operational parameters (such as resonant frequency) of the MRI machine. For example ^UB is an isotope of boron that is most effective for MRI. It occurs in natural abundance as 80 percent of boron isotopes. Pharmaco-kinetically, (it behaves identically to ¹⁰B and is used in treatment planning, including imaging, although ¹⁰B will be actually used during treatment. In general, an MRS element is chosen for use in the method disclosed herein based upon its ability to be assimilated and imaged. Thus the method may result in the quantitative dynamic imaging of "uptake" of, or pharmaco-kinetic response to, the MRS element by a patient irifused with the MRS element. Images taken at 10 minute time intervals are needed for pharmaco-kinetic studies of drugs and BNCT capture elements. The method provides images for quantitative dynamic observation of rates of uptake.

"Location" means a region in space, or a region of an image, which may be associated with a point or with boundaries enclosing the region. For example, an MRI image may be made, resulting in display by pixels of elements in a two dimensional domain. An MRI machine operates by encoding signals identifiable by their volume region of origination in three-dimensional space. Corresponding to each of a selected number of these regions in space, a value representing the magnitude of a signal may be recorded. The size of a region may be selected to optimize parameters associated with signal processing, signal to noise ratio, or display.

"Calibration" means comparison of an unknown value of a parameter associated with an item, object or signal to a known value of the same parameter associated with another item, object or signal. For example, absolute intensity (light txansmissivity of a film, grey scale value of an image) of an image may have no meaning until referenced to an intensity corresponding to a known value of an imaged parameter at a known location. In the method herein disclosed, boron concentration is determined by imaging a subject containing an unknown concentration and quantity of boron at the same time and in the same imaging field as a reference having a known amount and concentration of boron. The images have intensities associated with each region or pixel. The intensities of the image of any region of unknown concentration in the subject is compared to the intensity of a region of known concentration of the reference, thus determining the concentration of boron in the subject.

"Data matrix" refers to accumulated data organized in any format suitable for use as input to a processing device or for conversion into a useful graphical or analog format, such as an image. For example, a matrix of data reflecting intensities may be converted into an intensity image; a matrix of data reflecting concentrations may be processed to produce a concentration image.

SUMMARY Generally, this invention involves infusing a first material containing a magnetic-resonance-sensitive (MRS) element into a subject, positioning the subject within an imaging field region of a magnetic resonance imaging machine, positioning also within the same imaging field region a reference comprising a second material containing a known concentration of the MRS element, operating the magnetic resonance imaging machine to produce a signal having a plurality of amplitudes, each of which corresponds to a location in the field region, at least one such location being occupied by the reference, converting the plurality of amplitudes to an intensity matrix comprising a plurality of intensities corresponding to the plurality of amplitudes and calibrating the plurality of intensities with respect to the known concentration of the MRS element in the reference, thereby obtaining a concentration matrix reflective of the concentration of the MRS element in the subject as a function of position within the subject.

As applied to the quantification of boron uptake in a subject tissue, the method comprising infusing a quantity of boron into the subject tissue by conventional means. Intravenous drip procedures are generally preferred, but intramuscular injections, intraperitoneal injections, and intraarterial injections may also be used to introduce an MRS element, such as boron, into a patient. A known quantity of boron is distributed in a calibration sample. The subject tissue and calibration sample are simultaneously imaged by conventional techniques in a magnetic resonance imaging machine to create an image from a plurality of intensities corresponding to a multiplicity of concentrations of the boron at a respective multiplicity of locations in the subject tissue and at least one location in the calibration sample. The multiplicity of intensities obtained from the imaged tissue are then calibrated with respect to at least one intensity obtained from the calibration sample. Imaging may be repeated at intervals, and the accumulated data may be processed to determine the rate of change of the multiplicity of concentrations of boron in the subject tissue with respect to time.

As typically practiced, a magnetic-resonance-sensitive (MRS) element, notably boron (ⁿB), is compounded in a carrier, such as borocaptate sodium. A subject, typically a patient with a brain tumor, is positioned in the imaging field of a magnetic resonance imaging machine, along with a reference containing a known concentration of the MRS element. The magnetic resonance imaging machine is operated to produce a signal encoded by a reconstruction method that takes data immediately following the transverse RF pulse. The data is reconstructed to provide, for example, a subject image and a reference image. The subject image is calibrated from the reference image to derive a data matrix relating the concentration of the MRS element in the subject to locations within the subject. The data matrix may then be displayed as an image or downloaded to a computer for controlling radiation dosage during neutron therapy. A specific method for the production of images representing boron concentration relies on displayed MRI images of a boron-infused subject and a boric acid reference. An in vivo pharmaco-kinetic study of a specific patient's absorption of a specific boron compound provides actual boron concentration data in live tissue. Boron images of targeted tissues to be treated are precisely mapped with the method.

The method provides actual boron concentration data for calculation of a total dose of neutron radiation. The method provides not only concentration determinations but integration of the boron concentration information with treatment planning. A complete treatment planning method may thus be based on actual boron concentration data.

The method uses external references for determining boron concentration in sodium borocaptate, also called borocaptate sodium or BSH. The method can be practiced using in vitro samples of biological materials such as tissue or serum albumin, called phantoms, to study the chemical binding effects of the biological material and boron.

Absolute concentration determinations are provided by an external reference positioned within the homogeneous imaging volume of the RF coil of the MRI machine. Because the boron compound BSH in biological material has a relaxation time, T2, on the order of 600 secs, the imaging method relies upon three dimensional projection reconstruction (3DPR) to acquire the imaging data before the spin effects of the boron have substantially reduced their signal.

The dose of neutrons to be delivered to the tumor for cellular damage while sparing healthy tissue is determined from known characteristic energies entering the target area (head). Because the actual path of a neutron cannot be predicted with accuracy and is somewhat random, a model of the neutron distribution relies on Monte Carlo stochastic methods. Once the calculated neutron fluxes throughout the treatment volume have been determined and integrated into definable volumes, iso-dose contours are mapped for controlling placement and length of exposure to the neutron beam.

BKTF.F DF-SCRIPTION OF THE DRAWINGS In the drawings, which illustrate what is currently regarded as the best mode for carrying out the invention:

FIG. 1 is a partial cutaway isometric view showing a subject positioned within an MRI coil assembly;

FIG. 2 is a schematic cross-sectional axial view of a coil showing references positioned near a subject; FIG. 3 is a chart showing a comparison of signals as a function of concentrations of borocaptate sodium (BSH) in bovine serum albumin (BSA) and as a function of temperature; FIG. 4 is a chart showing a comparison of boron concentration derived by spectroscopy from the whole head with boron concentration in whole blood, deteπnined from ICP-AES;

FIG. 5 is a chart comparing boron signal intensity of different tissues of equal volumes after infusion of BSH, where curve 1 is from the tongue, 2 from mucosa, and 3 and 4 from masticator muscle; and

FIG. 6 is a chart illustrating the differences in boron concentration based on signal intensity from different tissues and from blood as a function of time begmning infusion initiation.

D TAπ.Fn DESCRIPTION OF THE ILLUSTRATED EMBODIMENTS FIG. 1 shows a vessel 20 arranged for infusion into a patient 21. In the illustrated instance, the vessel contains borocaptate sodium (BSH) 22 as a carrier compound for boron. Boron is the MRS element of choice for the illustrated embodiments. The BSH 22 is infused into the patient 21 through a conventional intravenous drip 23, necessarily infusing the MRS element and creating a concentration of boron in the bloodstream. The bloodstream then carries the MRS element (boron) to a subject tumor 24. The tumor 24 absorbs the BSH; molecules in the cells bind the BSH chemically. Maximum MRS element concentration is desirable for BNCT procedures.

It typically requires between about 10 minutes to about an hour for the concentration of a boron compound to approach its maximum concentration in a tissue following initiation of an infusion procedure. Infusion times of less than one minute to more than three hours are considered within the practical operating range. Preferably before, but possibly after, the initiation of infusion, the patient's head 25 containing the subject tumor 24 is placed within the coil assembly 27 containing the field coils of an MRI machine (not shown in full). The subject tumor 24 is preferably well within the assembly 27 where the field represented by flux lines 28 is uniform in direction, intensity and density. A pair of calibration references 38 are shown positioned adjacent the patient's head 25. Each reference is prepared by fLlling an appropriate container with a reference material contarining the MRS element, in this instance boron. Because BSH is selected as the carrier 22, boric acid is selected as the ideal reference material. Each reference 38 is elongate, having a length similar to that of the coil assembly 27. It is considered important for a reference to be at least as long as the tumor 24. The references 38 are illustrated in approximate alignment with the flux lines 28, but other orientations are operable. Boric acid is relatively insensitive to temperature within the context of this invention. Variations of plus or minus 4°C have minimal impact on signal accuracy. Adequate temperature control may ordinarily be provided by a temperature soak (long exposure) at room temperature.

As illustrated, multiple references 38 are positioned as close to the tumor 24 as possible. Thus, the references 38 are subjected to approximately the same fields and field variations as is the tumor 24. Moreover, each reference 38 is characterized by a uniform concentration of boron and is of sufficient length to react to inhomogeneities and non-uniformities in the field in harmony with similar reactions induced in the tumor 24. With the patient 21 and the references 38 in place, the magnetic resonance imaging (MRI) machine is operated to produce encoded signals. The MRI machine is operated conventionally in a sequence of steps by which a transverse RF field is pulsed at a right angle to the field 28 to produce a signal. Contrary to conventional practice, however, signal acquisition is commenced nearly immediately (within a few microseconds) after imposition of the pulse of the transverse RF field, while encoding field gradients are still being imposed. Thus, the data signals are recorded during the time when the signal to noise ratio would normally be expected to be too poor to provide meaningful data following conventional MRI practice.

The signals are preferably encoded using three dimensional projection reconstruction (3DPR). This method provides a signal characterized by an amplitude corresponding to a position in space. The signal is acquired at intervals in each spatial dimension within the imaging field. Each spatial location corresponds to a value of the magnitude of the signal amplitude over a small volume of space associated with that location in the imaging field. Using 3DPR, or other suitable method, the signals are decoded. The signal processing methods used during encoding assure that signal amplitude will correlate directly with concentration of the MRS element. Specifically, the encoded signal should vary substantially linearly with concentration of the MRS element so that it can be decoded by known methods of signal processing for MRI.

Using 3DPR for encoding and decoding, amplitudes of signals may be accumulated as a data matrix which is useful in a variety of ways. Either the signals or the matrix may be converted into an image in which any intensity (pixel) of the image corresponds to an amplitude of the signal, and the intensity and amplitude correspond to the same location (region) in the imaging field.

The overall image resulting from decoding the MRI signals may be displayed digitally (e.g. on a computer monitor or device using a digital grey scale) or as an analog image (photograph or transparency). Intensity is directly related to the light reflected or transmitted, respectively, from the displayed image.

The intensity of the image of the reference reflects the signal corresponding to the known concentration of boron in the reference and to variations in the transverse field at the location represented by the image. The intensity of the image of the tumor 24 (the subject image) at any desired number of locations in the subject image is compared with the intensity of the image of the reference 38 (the calibration image) at a location in the reference 38 as near as possible to the location of interest in the subject image. A ratio of intensity in the subject image to an intensity in the reference 38 is calculated, and may be computed for each location of interest in the subject image.

Image intensity at any location reflects the concentration of the MRS element (boron). The ratio of intensities (intensity ratio) of the image corresponding to a location in the subject image and corresponding to a location in the reference may be calculated for each location of interest in the subject image. The intensity ratio at a location reflects the ratio of the concentration of the MRS element in the subject compared to the concentration in the reference at the location in the imaging field.

Calculating a concentration ratio for each location of interest in the image of the tumor 24 (subject image), calibrates the subject image. Thus, concentrations may be calculated for all locations of interest in the subject as the product of image intensity in the subject image, divided by the image intensity in the reference 38 and multiplied by the known concentration of boron in the reference 38. An image may be constructed to reflect isoconcentration contours in the tumor 24. A map of the concentration of boron in the tumor 24 at numerous locations in three-dimensional space is preferably calculated and stored for reference by a computer that is configured to control neutron dosing during BNCT.

EXAMPLES Example I These experiments used the method for quantitating boron in vitro using phantoms of borocaptate sodium (Na₂B,₂H_nSH or BSH) bound to bovine serum albumin (BSA). The BSH-BSA binding models predict what is expected with the proteins in blood and tissue.

Temperature and concentration affect the image intensity as a result of the binding and the exchange between the free BSH and bound BSH. However, when the BSH concentration exceeds about 150 ppm, the relationship between concentration and intensity changes due to a decrease in available binding sites on the albumin molecule. To determine the boron concentration, a relationship between the boron in the BSH-BSA mixture and boric acid references within the "same" Bl RF field was developed. Boric acid has a chemical shift of about 30 ppm and its intensity does not change over the 20-37°C temperature range. With a boric acid reference placed within the same Bl RF field as the BSH-BSA phantom and with the temperature of the phantom known, then it was possible to determine the boron concentration. For in vivo systems at 37 °C, the relationship can be determined for a single temperature eliminating a degree of complexity. However, the binding factor for each tissue of interest may be different. For an in vivo or in vitro system, the unknown differences in RF field intensity (Bl) due to coil imperfections and from loading by the subject can be calibrated by an empirical method. In the image (FIG. 2), measurements were obtained over a region of interest (ROT) for each of the boric acid reference phantoms, I, and I₂. Then, for any pixel intensity, I₃, in the subject for which it is desired to obtain the boron concentration,

I, = C, ΫfoMl-er™"! ,(1)

1₂ = C₂ Ϋtøll-er™! ,(2)

1₃ = C₃

B(T) ,(3) where I; is the intensity for the ith measurement, is the boron concentration, Ψ(r) is the coil calibration factor at the location r, TR is the repetition time for the MR experiment, Tic, Tls are the relaxation times for the calibration solution and the subject, respectively, and B(T), which is determined empj-rically, takes into account the binding effects which are dependent on temperature. This system can be simplified to the following:

C₃ = k(I₃/I₂) ,(4)

Reference measurement 1 (corresponding to Ij) is not strictly needed. However, in practice it is difficult to keep the calibration vial 2 in place during an entire study procedure, so a quick calibration scan with both references in place is obtained, and then reference 2 is removed for the rest of the study. In this case (I₂/I,) from the calibration scan is used as well as the fact that C, =C₂.

Five phantoms were constructed of 50 ml plastic bottles (8.6 cm long with 3.8 cm diameter) and filled with BSH mixed with BSA. The BSA solution was prepared from Sigma^® A6793 and is 4% weight to volume. The phantoms were refrigerated for storage, and, before each experiment requiring MR scans, thermalized in a warm water bath to the desired temperature. Concentration of the boric acid reference was about 200 ppm. All experiments were performed with the spectrometer frequency centered on the BSH peak and with a TR of 6 ms. The resulting scanning time was eight minutes and 28 seconds. The boric acid has a Tl of about 8 ms, and the BSH-BSA has a Tl of about 2 ms. The boron concentration of BSH-BSA phantoms was between 75 and 250 ppm and was determined by Inductively Coupled Plasma Atomic Emissions Spectroscopy (ICP-AES). Temperature of the phantoms varied from 20°C to 36°C. The resolution of the images for a 24 cm field of view (FOV) was about

7.5mm by 7.5mm by 7.5 mm. The signal-to-noise ratio (SNR) for the phantom containing 250 ppm was about 17 (36°C), but with a decrease in temperature (21 °C), there is also a decrease in the signal level and thus in the SNR (~ 12 for 250 ppm and ~ 3 for 75 ppm). The effects of temperature and concentration on the image intensity are displayed in FIG. 3. When compared with the ICP-AES measurements, the imaging method resulted in concentrations that varied by as much as 10-20% from actual. An advantage of the in vivo system is the higher and constant temperature. However, the binding relation between BSH and the tissue component may have to be calculated for different tissues.

Example II In these experiments, 3DPR was used for determining the in vivo distribution and kinetics of BSH using a canine model. The following assumptions were made: For effective treatment, at least 30 ppm boron (¹⁰B) must exist in the tumor for a fluence of 2-8 xlO¹² neutrons.

For a method to be useful in determining distribution and concentration, it must be able to detect boron at this low level concentration with a resolution that would allow a clear distinction between tumor (- 1-3 cm) and normal tissue. Boron should not be found in normal brain tissue.

To be useful kinetically, the method here was required to acquire data within a "short" time frame (10 minute intervals; less than 2 hours elapsed time).

To conduct pharmaco-kinetic studies more quickly and noninvasively for new boron compounds developed for BNCT, the method used here is a suitably dynamic and noninvasive technique.

Although BNCT requires the ¹⁰B isotope, the imaging and spectroscopy techniques used ^UB, since ¹⁰B is eight times less sensitive for equal numbers of nuclei. This substitution was considered appropriate for studying different boron compounds because ⁿB and ¹⁰B in BSH behave the same, pharmaco-kinetically. However, even with an expected concentration of 100 ppm boron in tissue, ^MB imaging is about 1.6 million times lower in SNR than that achieved with proton imaging. The 3DPR technique was chosen to increase the SNR as much as possible. The Tl of ⁿB in BSH is short (— 1 msec) permitting a short TR and, subsequently, allowing averaging as needed.

The T2 of ^UB in BSH while bound to protein was expected to be on the order of 600 μse , intervening phase-encoding gradients produced a delay time (— 1-2 msec) which was too long before the recording of the free induction decay (FDD) signal, resulting in a high SNR, could be obtained. To make up for the lower SNR, many total acquisitions were necessary ( — 30,000). Instead of repeating lines in two-dimensional (2D) k-space by averaging, the same imaging time was more profitably spent acquiring a larger, three-dimensional (3D) volume of k-space data. The 3D projection reconstruction method was used to improve the SNR.

An anisotropic voxel (7.5 mm x 7.5 mm x 15.0 mm) was chosen to obtain a better SNR for the initial in vivo studies. With a TR of 6 msec and 1324 ray projections averaged 64 times, a 3D image series (scan) took approximately eight minutes. The 3D data were collected by a GE Signa™ 1.5 Tesla MRI system (Milwaukee, Wisconsin, U.S.A.) and then reconstructed on a Sun™ workstation (Sun Microsystems, Mountain View, California, U.S.A.). The reconstructed data were typically displayed in a 256 X 256 matrix containing the central 16-32 X 32 pixel image slices interpolated to 64 X 64 each. The resulting boron images were superimposed over corresponding proton images of higher resolution or displayed separately.

In vivo studies were performed using dogs with and without tumors. The large animal model was chosen to allow examination of larger tissues because of the resolution limits of the imaging technique. The BSH uptake and elimination of different tissue types were compared qualitatively by comparing the relative boron MRI signals obtained by using region of interest (ROI) measurements that encompass several voxels. Since the BSH is delivered to the brain via a concentration gradient from the blood, cephalic blood was obtained for analysis by inductive coupled plasma atomic emission spectroscopy (ICP-AES), and the boron concentration in blood was compared to the whole head intensity data .

Non-enriched BSH (80% ⁿB, 20% ^,0B) was obtained indirectly from Callery Chemical Co., Pittsburgh, Pennsylvania, U.S.A. and Boron Biologicals, Inc., Raleigh, North Carolina, U.S.A., after certification of purity by the Idaho National Engineering Laboratory (INEL). Solutions for IV administration at 100 mg "B/kg dog weight were prepared by weighing appropriate amounts, equivalent to 173-175 mg BSH/kg dog weight, into sterile containers and adding sterile physiological saline, 11 ml/kg dog weight. The solution was well mixed and transferred to sterile IV bags for infusion.

All protocols were approved by the Institutional Animal Use and Care Committee. Because of their larger head size, 20-30 kg, breed characteristic black Labrador Retrievers were obtained from various sources and entered into an approved conditioning program.

A reproducible tumor model was selected to obtain consistency for statistical analysis. Based upon reported success with a canine gliosarcoma in beagles, cells were obtained from Dr. John Hilton, Division of Oncology, The Johns Hopkins University in frozen injection-ready aliquots. Circulation and blood supply recruit¬ ment differences between the tumor model selected and spontaneously occurring tumors was compensated for by the reproducibility of the gliosarcoma tumor model and by the effectiveness of the tumor in destroying the blood-brain-barrier. The tumor cells were injected in the area of the brain that was desired for growth. After dogs under isoflurane anesthesia were draped in a sterile field, a 2.54 centimeter (one inch) incision was made, usually at a location above the left parietal lobe as determined with prior proton sagittal scout images. Electrocautery through the temporalis muscle was followed by scraping the periosteum and drilling a 0.635 centimeter (1/4 inch) burr hole with a hand-held drill. One to 1.2 million tumor cells in 40-50 μL volumes were injected 5-10 mm into the parietal cortex using a 100 μL Hamilton syringe. Muscle and scalp were closed using normal surgical procedures. The animals were then placed in recovery rooms where they were maintained between subsequent imaging procedures. Twelve animals were inoculated with single doses in the left parietal areas.

Three other animals received bilateral inoculations, one with a "sham" of physiological saline. Of these 15 cell administrations, 14 tumors were produced with only two tumor-bearing animals not surviving prior to boron imaging. Tumor growth periods were usually 9-12 days post inoculation, eventually producing classical gadolinium-enhancing nodules 2-3 cm in diameter. In most cases, evidence of abnormal tissue was noted 3-4 days post inoculation on T2-weighed proton images.

After the gliosarcoma cells were implanted, the animals were monitored using proton imaging every other day during the growth period. With the dogs sedated and under anesthesia an imaging series of sagittal Tl-weighted images for scouting purposes was followed by T2-weighted axial imaging with a 24 cm field of view (FOV). When the tumor was successfully implanted and the growth followed the expected pattern, an additional scan series was conducted using gadolinium DTPA (source), usually during the later stages of development (7-9 days) to measure the extent of the tumor growth. At about 10-12 days, the tumor typically grew to about 2-3 cm with edema extending some-2 cm into the brain tissue.

At this point the BSH pharmaco-kinetic studies were performed. The dog was once again immobilized with Telazol and then maintained on 1.5-2% isofiurane anesthesia for the remainder of the study. Proton sagittal and axial Tl- and T2- weighted images were obtained for anatomical referencing and tumor enhancement. BSH was infused by IV into the cephalic vein over 1 hour (dose rate: 3 mg/kg/min). Blood was sampled every 10 minutes during the infusion and every 20-30 minutes after the infusion during the elimination phase of BSH. The sampled blood was placed in vials with heparin and frozen until ICP-AES analysis.

Whole head spectroscopic data, averaged over 100 samples during about one minute, were obtained before each blood sampling. ^πB images with a FOV of 24 cm were collected every 20-30 minutes after the maximum uptake was determined by the whole head spectroscopy. "B" -euthanasia was performed at the end of ⁿB imaging followed by necropsy to collect representative tissues, mostly from the head, for chemical analysis of boron content by ICP-AES.

The boron RF coil (20.5 MHz) was a bird-cage configuration with 16 elements, 16 cm in diameter and 18 cm long, utilizing quadrature drive with inductive coupling. With true quadrature the expected increase in SNR was about 1.4 over conventional linear coils. The vendor preamplifier has a MOSFET front end with a noise figure (NF) of 0.5 dB when modified to operate at the "B frequency. Connecting coaxial cables between the coil and quadrature combiner/splitter and preamplifier are kept to a minimum to reduce signal loss due to inherent dielectric cable losses. However, even the short length of coaxial cable contributed additional noise on the order of about 0.27 dB. The quadrature splitter added another 0.5 dB to the NF giving a total NF of about 1.3 dB. The NF was calculated from the noise produced by a 50 ohm metal film resistor at the input for two different temperatures, 77°K and 300°K. An acrylic form was designed and built to hold the dog's head stationary either in the boron coil or the standard proton head coil provided by the MR manufacturer. This procedure allowed easy exchange of coils without the need to move the dog or the head which would have required realignment for each imaging protocol. The landmark was the same for each imaging protocol and was positioned over the location of the tumor growth as indicated from earlier scans.

Individual dogs displayed different kinetics. In one case the kidneys stopped functioning and the boron concentration remained the same for the duration of the boron imaging experiment, about 2 hours. But for the dogs with normal elimination, a quick redistribution and elimination resulted during the first hour and was followed by a subsequent slower elimination.

Differences between the blood and tissue kinetics for BSH are displayed in FIG. 4 which compares the boron in blood from ICP-AES with the whole head intensity data from a representative single dog experiment. Both ⁿB concentration in blood and ⁿB MR signal intensity were normalized to the maximum level observed. A strong correlation during uptake was observed, indicative of a boron concentration-driven gradient.

The maximum signal intensity in the ^πB images occurred about 5-10 minutes after the cessation of the IV for all the experiments. The curve depicting the spectroscopy data departs from the blood curve indicating a longer retention of boron in tissue/blood for the first hour of the elimination phase (redistribution phase). With the volume of blood in the brain being only 5-10%, this last observation shows that the MR signal came primarily from the tissue, and that a mechanism slowing the release of BSH from tissue existed during the redistribution phase. However, after the initial phase, the spectroscopy data began to resemble the blood curve, indicating an expected elimination of the BSH compound based upon concentration driven kinetics rather than any active transport.

A boron image (SNR « 30) of a normal dog with a 48 cm FOV and an isotropic resolution of 1.5 cm x 1.5 cm x 1.5 cm was superimposed over a filtered (Laplacian) proton image to show anatomical outlines. Using a time series of the resulting image, measurements (FIG. 5) using an ROI of 2 voxels were taken for several different tissues over time. Vascularized tissue (nasal mucosa, tissue 1 , and tongue, tissue 2) showed faster changes over time, as expected. The intensity of each measurement was normalized against the average intensity from a vial filled with 50 ppm boric acid which was placed on the dog's head over the eyes. Since these data are from a dog whose blood-brain-barrier was intact, negligible concentration of BSH appeared in the brain. To assess the sensitivity of the measurements, the tissues taken at necropsy were analyzed by ICP-AES for boron content. The end points of the graph were just before euthanasia. Tongue had a concentration of 27 ppm, and the muscles were 25 ppm ± 3 ppm. A T2-weighted proton image (slice thickness of 5 mm) of a dog with a gliosarcoma (nine days from inoculation) located in the left temporal lobe and a corresponding boron image taken about 53 minutes after the termination of the BSH infusion evidenced a greater concentration of boron in the left temporalis muscle above and to the left of the cranium with the site of the surgery. Boron was detected, as expected, in all of the tissue without a blood-brain-barrier, and, within the brain, only at the tumor sites.

Comparing a set of 16 axial boron images taken 7 minutes post infusion with another set of 16 taken 53 minutes later revealed pharmaco-kinetic data. Comparing intensity data (SNR « 5-9 for the second set) for two different sets of data showed that the boron in the surrounding tissue (muscle and surgical edema) decreased by 30% over 53 minutes whereas the boron decreased only by 14% in the tumor. Initial preferential retention of BSH appeared in tumor. However, in other dogs at 2-3 hours post infusion, the concentrations for all tissues including tumor seemed to equilibrate, again suggesting that the kinetics of BSH follows the behavior of a system that is concentration gradient driven. The normal brain showed negligible amounts of boron throughout all these studies.

A tumor more deep seated could be more easily distinguished from the muscle surrounding the brain. Closer to the surface of the brain more difficulty occurred in distinguishing between tumor-boron and muscle-boron. FIG. 6 shows a comparison between traumatized tissue at a surgery site and the tumor with associated edema. There is a definite lag of boron uptake in the tumor as compared to the tissue. Boron was eliminated faster in the tissue during a redistribution phase ( — one hour following infusion).

Example HI ⁿB MRI was performed on a 49 year old male preparing for BNCT in Japan. He had undergone a surgical resection and volumetric debulking of a large (7 cm) glioblastoma multiforme in the left frontal lobe. "B was infused as a part of borocaptate sodium (Na₂B₁₂H_uSH or BSH). In preparation for the ⁿB imaging, a sterile IV solution of BSH in physiological saline was prepared by dissolving six grams of BSH in 760 ml of sterile saline (50 mg ⁿB/kg in 11 ml/kg saline) and filtered through a 0.22 micron filter before transferring to a one liter IV bag. Two plastic vials of boric acid (200 ppm) with diameters of 1.5 cm and lengths of about 12 cm were attached with tape to the side of the patient's head. The left vial was placed longitudinally inferior to superior and the right vial was placed on top of the ear orthogonal to the left vial.

Prior to administration of BSH, a spin echo sequence was used to acquire Tl weighted, axial slices for anatomical reference. Boron imaging commenced five minutes after a 60 minute infusion of the BSH compound into an antecubital vein of the right arm. With a TR of 3.3 ms and with each projection averaged 128 times, each 3D scan lasted about 9 minutes. Four scans with a FOV of 24 cm were acquired over about 2 hours. Urine and blood were collected toward the end of the two hours for analysis.

The RF coil used for boron imaging was a quadrature bird-cage with 16 elements, 24 cm in diameter and 23 cm long. In order to obtain better RF homogeneity in the region near the top of the head, a copper end cap was installed. Axial boron images and corresponding Tl weighted proton images were taken. A boric acid reference vial was imaged in both the proton and boron images. BSH was found to accumulate at the border of the resection where the blood-brain barrier (BBB) had been compromised.

Because of the short T2 (0.64 msec) of ⁿB in BSH when bound to protein, 3DPR for boron was implemented. Using this method, 3D projection data was acquired within a few microseconds after a 90° RF pulse, giving a signal-to-noise (SNR) advantage over methods requiring the interposition of Cartesian k- space phase-encoding pulses before acquiring signal. With a resultant higher SNR, a resolution of 7.5 mm x 7.5 mm x 7.5 mm and scan times on the order of 4 minutes were achieved. Based upon the boron analysis of the blood by inductively coupled plasma atomic emissions spectroscopy and using the 3DPR technique, the concentration of ^πB was found to be about 100 ppm. BNCT required at least 30-35 ppm for successful treatment. Since there is no vascularization in the center of the resection, it was considered reasonable not to have boron accumulation there in the begύining. Indeed, the boron image confirmed this expectation. The higher accumulation of BSH also corresponded to the gado-enhanced images taken 24 hours post-surgery. The boron image taken immediately after termination of the IV had a SNR of 3-4. Succeeding images (one hour later) showed a rapid 30% decrease in boron and diffusion of boron into the center of the resection.

These preliminary results demonstrated ^UB MR images of a human brain, and showed differential uptake and washout in pathology.

While preferred methods of practicing the invention are described with specific reference to the drawings, it should be understood that the invention is not intended to be limited thereby. Other variations may be practiced without departing from the invention. The detailed embodiments are by way of illustration only, the invention being limited only by the claims.

Claims

CLAIMS What is claimed is:

1. A method for quantitation of an element taken up by a tissue of a subject, the method comprising: infusing a first material containing said element into said subject; positioning said subject within an imaging field region of a magnetic resonance imaging machine; positioning in said imaging field region a reference comprising a second material containing a known concentration of said element; operating said magnetic resonance imaging machine to produce a signal having a plurality of amplitudes, each of said plurality of amplitudes corresponding to a location in said field region and at least one such location being occupied by said reference; converting said plurality of amplitudes to an intensity matrix comprising a plurality of intensities corresponding to said plurality of amplitudes; and calibrating said plurality of intensities with respect to said known concentration of said element in said reference, thereby obtaining a concentration matrix reflective of the concentration of said element in said subject as a function of position within the subject.

2. The method of claim 1 wherein the subject comprises a sample of biological material.

3. The method of claim 2 wherein said biological material is serum albumin.

4. The method of claim 1 wherein the subject comprises a member of a body of a live being in the imaging field region.

5. The method of claim 4 wherein said element is infused by injection into said being.

6. The method of claim 1 wherein said element is boron.

7. The method of claim 1 wherein said element is infused over a time period having a duration of at least about 1 minute.

8. The method of claim 7 wherein said duration is selected from between about 10 minutes to about 3 hours.

9. The method of claim 1 further including recording repeatedly said plurality of amplitudes of said signal at a plurality of instances of time, thereby determύiing the rate of change of each said amplitude as a function of time.

10. The method of claim 1 further including mapping a concentration of said element in the subject using three-dimensional projection reconstruction.

11. The method of claim 1 wherein said intensity matrix is an intensity image.

12. The method of Claim 1 wherein said concentration matrix is a concentration image.

13. A method for quantifying the boron uptake in a subject tissue, the method comprising: u fusing a quantity of boron into said subject tissue; distributing a known quantity of boron in a calibration sample; imaging said subject tissue and said calibration sample simultaneously in a magnetic resonance imaging machine to create an image from a plurality of intensities corresponding to a multiplicity of concentrations of the boron at a respective multiplicity of locations in the subject tissue and at least one location in said calibration sample; and calibrating said multiplicity of intensities obtained from said tissue with respect to said at least one intensity obtained from said calibration sample.

14. The method of claim 13 wherein the boron is compounded in a material characterized by a spin relaxation time of from about 100 microseconds to about 100 milliseconds.

15. The method of claim 13 wherein the boron is contained in sodium borocaptate.

16. The method of claim 15 wherein the calibration sample comprises a vessel containing boric acid.

17. The method of claim 13 wherein said imaging is repeated at intervals.

18. The method of ctøim 17 further including calculating the rate of change of said multiplicity of concentrations of boron in the subject tissue with respect to time.

19. A method for quantitation of boron, the method comprising: infusing boron into a subject; positioning the subject in an imaging field of a magnetic resonance imaging machine; positioning a reference comprising boron distributed in a known concentration proximate the subject in the imaging field; operating the magnetic resonance imaging machine to produce a signal having a plurality of amplitudes, each amplitude of the plurality of amplitudes corresponding to a location in the imaging field; converting the plurality of amplitudes to an image having a plurality of intensities conesponding to the plurality of amplitudes; and calibrating the plurality of intensities with respect to the known concentration of boron in the reference thereby displaying an image reflective of a concentration of boron in the subject as a function of position within the subject.