BJR
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
British Journal of Radiology (2006) 79, S16-S26
© 2006 British Institute of Radiology
doi: 10.1259/bjr/84072695
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Payne, G S
Right arrow Articles by Leach, M O
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Payne, G S
Right arrow Articles by Leach, M O

Full paper

Applications of magnetic resonance spectroscopy in radiotherapy treatment planning

G S Payne, DPhil and M O Leach, PhD, FInstP, FMedSci

Cancer Research UK Clinical Magnetic Resonance Research Group, Institute of Cancer Research and Royal Marsden NHS Trust, Downs Road, Sutton, Surrey SM2 5PT, UK


   Abstract
Top
 Abstract
Introduction
 Discussion
 References
 
Following advances in conformal radiotherapy, a key problem now facing radiation oncologists is target definition. While MRI and CT provide images of excellent spatial resolution, they do not always provide sufficient contrast to identify tumour extent or to identify regions of high cellular activity that might be targeted with boost doses. Magnetic resonance spectroscopy (MRS) is an alternative approach that holds great promise for aiding target definition for radiotherapy treatment planning, and for evaluation of response and recurrence. MRS is able to detect signals from low molecular weight metabolites such as choline and creatine that are present at concentrations of a few mM in tissue. Spectra may be acquired from single voxels, or from a 2D or 3D array of voxels using spectroscopic imaging. The current state of the art achieves a spatial resolution of 6–10 mm in a scan time of about 10–15 min. Co-registered MR images are acquired in the same examination. The method is currently under evaluation, in particular in brain (where MRS has been shown to differentiate between many tumour types and grades) and in prostate (where cancer may be distinguished from normal tissue and benign prostatic hypertrophy). The contrast achieved with MRS, based on tissue biochemistry, therefore provides a promising alternative for identifying tumour extent and regions of high metabolic activity. It is anticipated that MRS will become an essential tool for treatment planning where other modalities lack the necessary contrast.


   Introduction
Top
 Abstract
 Introduction
Discussion
 References
 
Magnetic resonance spectroscopy (MRS) is a non-invasive technique for measuring biochemicals in tissue. It uses the same general principles and equipment as its widely used partner, MRI. However, while MRI builds images using signals from 1H nuclei in tissue water (and sometimes lipid), present at concentrations of approximately 35 M, MRS is used to measure signals from magnetic nuclei (usually 1H, but 31P has also been extensively studied) of tissue metabolites such as choline, creatine and lactate that are present at much lower concentrations (typically of the order of a few mM). Example spectra from normal brain and from brain tumour are shown in Figure 1Go, to illustrate the type of data that are obtained. Further details of specific metabolites detected in brain and prostate are listed in Table 1Go and discussed in the corresponding sections towards the end of the article. More detailed introductions to the use of in vivo MRS may be found in the literature [1, 2].


Figure 1
View larger version (9K):
[in this window]
[in a new window]
  		Figure 1. Example of1H MRS spectra from (a) normal brain and (b) brain stem tumour. The major peaks are from choline (Cho), Creatine (Cr), myo-Inositol (mI), N-acetyl aspartate (NAA) and lipids (lip). Repetition time (TR) = 1.5 s; echo time (TE) = 30 ms. Each spectrum shows acquired data (grey line), the fit (bold line) and fitted baseline. Note the increase in choline (Cho) and reduction in NAA in the tumour compared with normal brain.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Brief details of some metabolites seen in1H spectra of tissues

 
The frequency of MRS signals depends only on the gyromagnetic ratio ({gamma}) of the nucleus (which is a constant for a particular nuclear species, e.g. {gamma}(1H) = 2.675x108 rad/T/s; {gamma}(31P) = 1.083x108 rad/T/s), and on the local magnetic field, Blocal, experienced by the nucleus (Equation (1)).


Formula 001

Blocal is determined primarily by the applied magnetic field (e.g. 1.5 T for a standard clinical MR system). However, it is modified slightly by shielding effects of the electron cloud associated with covalent bonds to neighbouring atoms (usually of a few parts per million (ppm)), by magnetic field inhomogeneity, and by the presence of local nuclear or electronic magnetic dipoles. The horizontal "frequency" axis is usually expressed in parts-per-million of the main precession frequency (approximately 63 MHz at 1.5 T) so that spectra acquired at different magnetic field strengths may be compared. By convention (and for historical reasons), frequency increases from right to left.

Following excitation of magnetic nuclei using a radiofrequency (RF) pulse, signal is acquired as the spins relax back towards their equilibrium state. The time taken for this relaxation to occur is described by two time constants, T1 (for longitudinal relaxation) and T2 (for transverse relaxation). T1 is typically of the order of 1s, and T2 (never longer than T1, and usually much shorter) may be about 100 ms. These relaxation time constants determine how quickly the measurement may be repeated to improve the signal-to-noise ratio. Incomplete relaxation can be used to weight signals according to their characteristic T1 or T2, which under some circumstances (and particularly for MRI) yields better contrast than concentration alone.

One key advantage of MRS relative to radio-labelled studies is that the same type of magnetic nucleus (e.g. 1H) produces signals at different frequencies, depending on the bonding to neighbouring nuclei. Thus, different nuclei of the same species within a single molecule produce distinct peaks within a spectrum (unless they are "magnetically equivalent"), which usually permits different compounds of biochemical interest to be distinguished. In addition, one can study compounds naturally present in tissue without the requirement for specific labelling. The major disadvantage of MRS is that the signals from metabolites are relatively small, so that compared with MRI much larger "pixels" or "voxels" are required to obtain an adequate signal-to-noise ratio. Methods for data acquisition and some related issues are described in the section on "Technical Issues" below.

In contrast to CT and MRI, which provide morphological information about tissue, MRS gives biochemical information (Figure 1Go). Since the biochemistry of tumours is substantially different from that of normal tissue, MRS has the potential to aid identification of tumours when there is insufficient contrast in the morphological image. In particular, it can be used for differential diagnosis, both between tumours and benign pathology (e.g. prostate cancer and benign prostatic hypertrophy [3]), and between different tumour types (e.g. different brain tumours [4]). Since MR spectra are acquired using MRI scanners, they are automatically co-registered with MR images, which may then themselves be co-registered with planning CT scans in the normal way if required (see also the paper by Kessler in this issue).

In the context of radiotherapy, MRS has several potential applications. These include:

  1. Identifying tumour extent and metabolically active regions to aid targeting of radiotherapy. This includes distinguishing tumour from normal tissue, benign lesions, or radiation necrosis.
  2. Evaluating response to treatment
  3. Identifying recurrence

With the advent of new conformal techniques for radiotherapy, target definition has become one of the main issues in further improving the effectiveness of radiotherapy treatment. In particular, it is necessary to delineate accurately the extent of tumour tissue, and one may also wish to identify regions of high clonogenic cell density, which could be targeted with boost doses of radiotherapy using either intensity-modulated external beam radiotherapy or brachytherapy. Conventional imaging methods are often limited in this respect. For example, in the prostate, MRI has better soft tissue contrast than CT [5], but still only has a positive predictive value of 50% [6]. In addition, owing to the different sources of contrast in the two cases, the boundaries shown are often different [7, 8]. Since MRS measures tissue biochemistry, it has the potential to identify abnormalities and the nature of the abnormality more selectively than morphological imaging can achieve.

Current status of MRS in target planning for radiotherapy
1H MRS packages are currently available for most commercial clinical MR scanners, usually including both single-voxel localization techniques and MR spectroscopic imaging (MRSI; for more details see subsection on localization strategies below). However, owing to the extra expense, and the time required to perform MRS examinations, many hospitals choose not to have this option. The packages are increasingly easy to use, but previous experience is important to achieve good results, and to understand the reasons for technical failures. Assuming diagnostic MR images have already been acquired, the extra time required for 1H MRS is typically a few minutes for single-voxel techniques, and 10–15 min for MRSI. Some software for spectral processing is usually provided, but for quantitative work many users choose to export their data to specialist spectral processing software. MRS of nuclei other than 1H is less used owing to the requirement for additional hardware and (usually) longer scan times.

While MRS has great potential to aid target identification in radiotherapy treatment planning, it is not currently used to any significant extent for this purpose. This is partly because it is only recently becoming available in a form that is relatively easy to use. In order to alter current treatment protocols to utilize MRS, the following steps need to be achieved:

To improve spatial resolution (currently about 6–10 mm) would be helpful but probably not essential, as practical radiotherapy techniques will include a margin of this order. As usual, there is a trade-off between signal amplitude and spatial resolution, so improving spatial resolution requires an improved methodology to increase the sensitivity of signal detection.

Some of these issues are already being addressed, in particular in studies of the brain and prostate (see below).

Technical issues
Magnetic field strength
Most clinical MR studies are performed at a magnetic field strength of 1.5 Tesla and most published MRS work, including most of that referred to in this review, has been performed at this field strength. In research environments, 3 T scanners are becoming increasingly used, as the signal-to-noise ratio and spectral resolution both increase (approximately linearly) with field strength. The doubling of signal-to-noise ratio at 3 T can be traded for a 50% reduction in voxel volume (i.e. a reduction in voxel size of (0.5)1/3 {approx} 0.8 in each spatial dimension) compared with 1.5 T scanners, but otherwise the functional information is very similar. Systems with much higher magnetic fields for human studies at 7 T and 8 T have been installed in a handful of research sites. However, these are very expensive, present a large number of technical challenges to operate (in particular the correspondingly higher gradient amplitudes and radiofrequency power deposition can exceed safety limits unless duty cycles are reduced) and generally (except in the brain) do not produce good diagnostic images owing to dielectric resonance effects [9]. Problems with image distortion (see below) will also increase at the higher fields.

Shimming
The sharpness of spectral lines depends on the homogeneity of the magnetic field. This is because a spread in values of the local magnetic field leads to a spread of frequencies (Equation (1)). The effect of placing a sample into a homogeneous field usually disturbs the field homogeneity. MR systems are equipped with a series of "shim coils" which create small additional magnetic fields that can be adjusted to help counteract this effect. This "shimming" process is usually performed at the beginning of each MRS study, and takes a few minutes. On current MR systems the process is normally automated.

Magnetic nuclei studied
Because clinical MR scanners are designed to detect signal from 1H nuclei, only additional software is required to acquire 1H MR spectra. In addition 1H is the most sensitive magnetic nucleus, and hydrogen is present in nearly all biologically relevant compounds. A list of metabolites seen in 1H spectra of several tissues is given in Table 1Go. More details for brain metabolites can be found in the work of Govindaraju, Young and Maudsley [10]. If nuclei within a molecule are "equivalent" then they all contribute to the same peak. Peaks appear as multiplets (doublets, triplets etc.) when the nucleus of interest is covalently bound directly or through other nuclei to another (but non-equivalent) magnetic nucleus (1H, 31P etc.). This splitting (known as J-coupling) decreases as the number of bonds between the nuclei increases.

In order to study compounds at millimolar concentrations, it is necessary to suppress the large signals from tissue water, and also sometimes from lipids. Suppression techniques are available to accomplish this.

For comparison of spectra it is necessary to have a reference frequency. 1H MRS studies in solution often include tetramethylsilane (TMS) for this purpose, with a frequency (or "chemical shift") set to 0 ppm. TMS is used because it is stable and produces a sharp single peak. For consistency, the positions of spectral peaks in vivo are still expressed relative to TMS, even though it is not present in tissue; other peaks that are present, and which have shifts that are relatively independent of environmental conditions such as pH and metal ion concentration etc., are used as an internal secondary reference.

Other nuclei that may produce useful MR spectra are 31P, 13C and 19F. Phosphorylation is an important biochemical process, and several compounds of key importance therefore contain phosphorus. Several metabolites seen in 31P MR spectra are affected by signalling pathways up-regulated in cancer [11, 12] and therefore 31P MRS could play an important role in target identification and evaluation of therapy, especially of new therapeutic agents. One study [13] in soft tissue sarcoma treated with thermoradiotherapy has shown strong correlations between pre-treatment MRS data (in particular the ratios of phosphodiesters and of phosphomonoesters to inorganic phosphate) and pathological complete response. However, since 31P has only 6% of the sensitivity of 1H at a given field strength, larger voxels are required to obtain an adequate signal-to-noise ratio. In addition, extra hardware (RF amplifiers, filters, RF coils) is required that is tuned to the 31P resonance frequency.

Currently, little work has been performed using 13C nuclei (owing to only 1% natural abundance and a lower sensitivity). While no endogenous metabolites contain MR-visible fluorine, several studies have demonstrated the value of 19F MRS in following drug metabolism [14].

Localization strategies
MRS localization techniques fall into two groups. Single-voxel methods include those known as PRESS [15], STEAM [16] and ISIS [17], in which data are acquired from a single voxel positioned using MR images. The alternative is MRSI, also known as chemical shift imaging, (CSI) [18], in which a matrix of spectra are acquired either over a plane (2D-MRSI) or a volume (3D-MRSI). The choice of localization method affects the efficiency of signal detection from within the target region and the contribution of unwanted signals from outside the volume (contamination). The effectiveness of the shimming method in optimizing the magnetic field homogeneity in the selected region, the effectiveness of suppressing the very large signals from water (and sometimes from lipid present in the selected volume) and the ability to suppress signals from outside the volume of interest can also affect spectral quality. In the brain, cerebral spinal fluid flowing into the selected region can also adversely affect shimming and water suppression. Sequence timing parameters determine the T1 and T2 weighting of the sequence, and also affect modulation of coupled spins. Common echo times (TE) are 270 ms (minimizing contributions from lipids and macromolecules), 135 ms (inverting the lactate doublet) and 20 ms (maximizing signal from macromolecules and lipids). A summary of the merits of single voxel MRS relative to MRSI is given in Table 2Go. For most applications in radiotherapy treatment planning use of MRSI will be preferred, with single voxel methods restricted to cases where it is required to characterize a single well-defined lesion, when time is short, or when adjacent structures make it hard to shim over larger regions.


View this table:
[in this window]
[in a new window]
  		Table 2. Comparison of single voxel techniques with spectroscopic imaging

 
Registration and motion issues
The volumes selected for acquisition of the MR spectra are positioned using MR images acquired in the same examination. The MRS and MRI data are therefore inherently co-registered. However, although MR images have better soft tissue contrast compared with CT images, they have two major disadvantages for radiotherapy treatment planning – they suffer distortions and they do not intrinsically provide electron density information. These issues have been extensively discussed elsewhere [19]. There are two approaches to solve this problem. One approach under development is to correct the distortions [20, 21] and to make use of the MR images to estimate the electron density information [22]. The alternative is to co-register the MR images with CT images – usually using either external fiducial markers that can be seen by both imaging modalities, or using methods such as "mutual information" [23] based on the features within the images (see also the paper by Kessler in this issue). It is also necessary to ensure that the MR scanner is equipped with a flat-topped couch (similar to that used for treatment), and with suitable laser positioning beams.

Motion during the scan is a potential problem. Being rigid, immobilization of the head is fairly straightforward, although there is still some residual pulsatile motion from the circulation of cerebral spinal fluid. Other parts of the body are more challenging. For studies of the prostate it is common to use agents like Buscopan to reduce peristalsis. Data acquisition can be synchronised to the heart or breathing cycle using cardiac or respiratory triggering, while "navigator echoes", an MR technique that identifies a column of tissue and monitors movement along this column [24], may also be used. It is not possible to eliminate motion entirely and, as with conventional CT techniques, this must be remembered when prescribing the margins for irradiation. As the application of MRI and MRS for radiotherapy planning develops, it is likely that use of immobilization devices will be evaluated in more detail.

Validation
Before radiation oncologists have the confidence to alter treatment plans based on MRS data, they need evidence that the abnormality detected using MRS represents tumour. The "gold standard" for comparison is histopathology (where available). In some cases, such as prostate (see below), studies have been performed where the prostate has been removed very shortly following the in vivo MRS examination, so that MR spectra can be directly compared with histopathology of the whole gland. This is the best possible situation for validation. In other cases biopsy samples are available, although it is often difficult to be certain of their precise location. Results of some of these studies are described below. Interpretation of the in vivo spectra can be aided by high-resolution ex vivo MRS of intact tissue samples acquired using magic angle spinning (MAS) at high magnetic fields [25]. Such spectra have much higher spectral resolution and sensitivity than spectra acquired in vivo, and permit detection and quantification of metabolites that in the in vivo spectrum are either at too low a concentration or are in regions of spectral over-lap. Other imaging methodologies (MRI, PET etc.) give supporting evidence, but may themselves lack the necessary validation.

Example 1: MRS and brain cancer
General issues regarding diagnosis and treatment
There is a strong incentive for non-invasive assessment of brain cancer, particularly in children, and especially where sequential evaluations are required. In managing patients, the first priority is diagnosis, currently usually confirmed by histology, although this suffers from known limitations of sampling error and tumour heterogeneity. Guidance of biopsy is therefore also important. Assessment of tumour grade is important in defining treatment – this is usually limited to one time point, either at biopsy or during tumour debulking. Assessment of response following treatment is of increasing value, so that the treatment can be modified according to the response observed.

MRI appearance and location alone have limited power to differentiate brain tumours. In particular, early stage disease and infiltrative disease may have little impact on the blood–brain barrier, limiting the sensitivity of contrast-enhanced MRI and leading to the requirement for an alternative method to identify tumour boundaries. Identification of residual disease from necrosis can also be difficult with MRI. The metabolic profile provided by MRS offers considerable potential to distinguish different tumour types and grade.

Most MRS measurements of tumours have employed the 1H nucleus, owing to its greater sensitivity. 1H MRS of brain tumours has developed rapidly due to the less exacting technical requirements of spectroscopy in the brain compared with measurements elsewhere in the body. In particular, it is easier to control motion and most parts of the brain are fairly homogeneous, making it easier to achieve a good shim. Care should be taken with MRS when using contrast agents as this can broaden the choline peak, without affecting peak area. However, if shimming takes place immediately prior to MRS acquisition, this effect can be minimized or eliminated [26].

Features of 1H brain tumour spectra
Spectra from brain tissues (Figure 2Go) include contributions from total creatine (Cr) (3.94 ppm and 3.01 ppm), total choline (Cho) (3.22 ppm), N-acetyl aspartate (NAA) (2.01 ppm), myo-inositol (mI) and glycine (3.55 ppm), lactate (Lac) (1.35 ppm), alanine (Ala) (1.47 ppm), contributions from a range of lipid resonances (lip) (0.9 ppm, 1.3 ppm), broad resonances due to macromolecules (2.05–2.8 ppm, 5.4 ppm) and poorly resolved amino acids such as glutamate and glutamine (Glx) (approximately 2.3 ppm). In tumours other resonances may be present, or may contribute to peaks, but their contribution can be hard to confirm unless tissue specimens are available.


Figure 2
View larger version (18K):
[in this window]
[in a new window]
  		Figure 2. Mean and standard deviation(vertical lines) of normalized STEAM (echo time (TE) = 30 ms) spectra acquired from a number of subjects: Normal white matter (n = 6); meningioma (n = 8); astrocytoma grade II (n = 5); anaplastic astrocytoma (n = 7); glioblastoma (n = 13). Published with permission [28].

 
Application in differential diagnosis and assessing the grade of brain tumours
While MRI is an essential part of the evaluation of cranial neoplasms, indicating location, morphology, boundary definition and physiological characteristics, there remains considerable overlap between the appearance of primary tumours, metastases and radiation necrosis [27–33]. However, as illustrated in Figure 2Go, MR spectra do display differences between tumour types. For example, compared with normal brain, MR spectra of astrocytomas show elevated Cho, reduced Cre and significantly reduced NAA. A summary of the characteristics found in some common brain tumours is given in Table 3Go. More details can be found in Howe et al and Devos et al [28, 34]. While visual inspection of spectra can be helpful, there is considerable interest in automated evaluation of multiple spectral lines [29, 34, 35]. In particular this permits comparison of new spectra with those from tumours of known histology, provided the spectra are acquired under similar conditions (magnetic field, sequence timing parameters etc.). This approach is showing promising indications of improved discrimination of tumour type and grade. Since there are many tumour types, some of which are relatively rare, effective databases require many samples, thus necessitating multi-centre cooperation.


View this table:
[in this window]
[in a new window]
  		Table 3. Summary of1H MRS characteristics of different brain tumours

 
Therapeutic guidance, assessment of response and recent developments
Recently there has been increasing interest in using functional imaging techniques, together with metabolic imaging using MRS [36] to aid target definition in radiotherapy, for example to identify areas of infiltration not evident with MRI. Figure 3Go illustrates an example where the boundaries of the lesion detected using MRS (based on levels of Cho and NAA) are quite different from those suggested by contrast-enhanced MRI [37]. In a further study [38] of patients with high grade glioma who were scanned following surgery (but prior to radiotherapy), MRS indicated areas of residual abnormality that did not enhance with contrast at that time, but of which a proportion did enhance subsequently at follow-up. In this study, MRS volumes were also greater than the T2 weighted area of abnormality. A study of survival following gamma knife surgery [39] showed that patients with poor overlap (<50%) between treatment volume and volume of metabolic abnormality had reduced survival compared with those having good overlap.


Figure 3
View larger version (79K):
[in this window]
[in a new window]
  		Figure 3. Comparison of lesion extent measured usingT2 weighted MRI (red), T1 weighted MRI following injection of contrast agent (green), and MRS to measure high Cho/NAA (orange) for a patient with a grade 4 glioma. Published with permission [43].

 
Elevated Lip/Cre and Lac/Cre in the peri-tumoural region of high grade glioma may help to identify patients at risk of recurrence [40]. It is suggested that areas of relatively high Cho/NAA may indicate high cellular activity, and hence radio-sensitivity, and Lac may indicate hypoxic areas with reduced radio-sensitivity [37, 41]. The technique can also be helpful in identifying areas missed by radiation fields, and in separating recurrence from radiation necrosis. Reduction in Cho [42] and in lipid and lactate [43] can also reflect response to chemotherapy and radiotherapy.

MRS has been compared with MRI and with 201TlCl-SPECT (a radio-nuclide technique that is widely used to assess brain tumours) in monitoring response to treatment in metastatic brain tumours. Reductions in Cho and Lac and an increase in Lip (believed to represent necrosis) in responding tumours were detected at an earlier time (1 week to 1 month) than contrast-enhanced MRI or 201TlCl [44]. A short term change in metabolites reflecting both pathology and normal brain tissues has been reported within 1 day of radiotherapy [45] although the significance of this is not clear, as a short term metabolic response to immediate damage delivered by radiotherapy may not translate into a therapeutic gain.

Example 2: MRS and prostate
Introduction
In the prostate, T2 weighted MRI has much superior contrast to CT, transrectal ultrasound and digital rectal examinations. However, the sensitivity and positive predictive value are still only of the order of 83% and 50%, respectively [6]. Several investigators are currently evaluating the use of 1H MRSI in the prostate to aid target definition.

The main spectral peaks observed in normal prostate (Figure 4Go, right) are those of the choline-containing compounds (Cho) at 3.2 ppm, creatine and phosphocreatine (3.02 ppm) and a large peak from citrate (2.6 ppm). This latter peak is tightly coupled, and the appearance depends on the echo time used to acquire the data [47]. As in other tissues, choline compounds are associated primarily with membrane synthesis, while creatine is involved in energy metabolism. Citrate is a product of normal epithelial cell metabolism in the prostate, where high levels of zinc inhibit the enzyme aconitase and hence prevent the oxidation of citrate in the Krebs cycle that occurs in other cells. It is regulated by testosterone and prolactin [48]. Sometimes an additional peak may be detected between the choline and creatine peaks that arises from spermine, a polyamine characteristic of healthy prostate and which is also under androgen control.


Figure 4
View larger version (29K):
[in this window]
[in a new window]
  		Figure 4. Centre: T2 weighted transverse image of prostate with tumour in right mid-gland, overlain with grid of voxels from which spectra were acquired. Left: Spectrum acquired from region of tumour, illustrating dramatically elevated choline and absence of citrate and polyamines. Right: Spectrum from healthy peripheral zone tissue with high citrate and presence of polyamines. (Reproduced with permission from J Kurhanewicz et al [46]).

 
In prostate cancer (Figure 4Go, left) choline is elevated and the normal production of citrate is reduced. In contrast, benign prostatic hyperplasia, an enlargement of the prostate commonly found in older men, is characterized by high levels of citrate. Hence the choline/citrate ratio is a fairly reliable measure of the presence of cancer. Lipids are also sometimes seen in cancerous tissues, although the significance of these is yet to be established.

1H MRSI data may be acquired from the prostate using an external phased-array coil. However, the best signal-to-noise ratio is achieved using an endorectal coil. This is usually well tolerated by patients. The main disadvantage of the endorectal approach is slight deformation of the prostate which needs to be allowed for in using the images for radiotherapy treatment planning. One study suggests that rigid endorectal coils are less of a problem in this respect than the more usual balloon coils [49]. Buscopan is often used to reduce involuntary motion. While many studies [50, 51] have used slice-localized 2D-MRSI, full 3D-MRSI is much preferred to obtain data from the whole of the prostate. Currently, the state-of-the-art at 1.5 T is to achieve voxels with a 3D isotropic resolution of 6.25 mm in an acquisition time of 17 min [52]. Studies have shown that owing to haemorrhage there is some degradation of in vivo spectra in the 8 weeks following transrectal biopsy (18.5% of peripheral zone voxels have been reported as degraded within 8 weeks of biopsy, and 7% for those examined more than 8 weeks after biopsy [53]).

Validation of MRS in prostate for target definition
A strong correlation has been found between negative MRSI and negative biopsy findings, and between positive MRSI and positive biopsy findings [54]. However, there is only a weak correlation between the concentration of prostate specific antigen (PSA, the current "gold standard") and either biopsy or MRSI findings [54]. Step-section pathological examination of radical prostatectomy specimens demonstrated that MRI combined with MRSI yielded a significant improvement in cancer localization to a prostate sextant (left or right; base, mid-gland or apex) compared with MRI alone [55]. Several studies have shown that adding MRSI to an MRI examination increases the accuracy of diagnosis [55–57]. One particular area of high current interest is in discriminating the many patients who present with elevated PSA, but who have pathologically indolent cancer from those with aggressive disease; preliminary studies suggest MRSI has a useful role to play here as well [52]. High-resolution studies of tissue samples using MAS support these findings, with linear correlations measured between metabolite levels characteristic of normal epithelium or of prostate cancer, and the proportions of the corresponding cells as measured by computer-aided image analysis of prostate pathology slides [58].

Recurrence
Use of MRSI together with MRI has been shown to improve substantially the identification of tumour recurrence following external beam radiotherapy (the area under the ROC curve, a measure of the effectiveness of a test, increased from 0.5 to 0.81 [59]). The presence of 3 or more suspicious voxels in a hemiprostate showed a sensitivity and specificity of 89% and 82%, respectively, for the diagnosis of local recurrence.

Planning
MRSI has been used in combination with MRI to define regions for dose escalation within the prostate [60–62], permitting a dose of > 90 Gy to the high-risk region while treating the remainder of the prostate to about 70 Gy.


   Discussion
Top
 Abstract
 Introduction
 Discussion
References
 
MRI already plays a major role in identifying the extent and position of tumours, aiding delineation of target volumes. Increasingly, there is interest in defining the functional extent of tumours to better select target volumes, or areas that might receive a boost dose. MRI already provides the capability to image several characteristics of tissues which reflect the physiology and microenvironment, such as perfusion and diffusion. MRS provides the ability to observe aspects of tissue metabolism, which is a more direct reflection of tumour activity and of therapeutic response. The ability to obtain this information at the same examination as the anatomical and functional MRI images is a major advantage compared with other techniques. While the number of studies so far is limited, there is strong evidence that MRS may have a valuable role to play in radiotherapy treatment planning.

The metabolic information provided by MRS provides signals associated principally with tumour (elevated Cho, Lac, Lip) and with normal tissue (NAA, Cre, Cit), where Lip is likely to be associated with necrosis, allowing identification of different tissue components, and providing measures that may reflect tumour grade, normal tissue response, and areas of mixed tumour and normal tissue. The mismatch between metabolic information, function and anatomical abnormality is not surprising. In brain tumours, contrast enhancement depends on breakdown of the blood–brain barrier, which is not apparent in many low grade gliomas, and may not be present in early areas of infiltrative disease. Thus metabolism may be a better guide to early disease and disease extent compared with contrast enhancement. High signal on T2 weighted images reflects increased water content or local oedema, which may reflect inflammatory processes. Again, metabolic change could occur sooner, depending on the process controlling increased water content.

The main limitation with MRS is that voxel sizes of typically 8 mm3 to 10 mm3 are required to achieve an adequate signal-to-noise ratio, with a large part of this volume needing to be occupied by abnormal tissue for a change in signal to be detected. Thus, small infiltrating lesions are unlikely to be detectable. It also needs to be remembered that the edges of target volumes defined from spectroscopic images for treatment planning, although smoothed and interpolated for presentation, are in fact limited by this same spatial resolution and will also include some blurring from point spread function effects. While some improvement in sensitivity and spectral specificity is expected with higher field scanners and improved sensitivity coils, this is likely to yield only a small improvement in spatial resolution. On the other hand, it is increasingly clear that while in principle both MRI and CT have much better spatial resolution, MRS, has the potential to improve identification of the gross tumour volume and hence improve treatment using radiotherapy.

GSP is funded on a research grant from Cancer Research UK ([CUK] grant number C1060/A808).


   Acknowledgments
 
We wish to thank Dr Phil Evans and Dr Andrew Jackson (Institute of Cancer Research and Royal Marsden Hospital NHS Foundation Trust) for useful discussions.

Received for publication January 3, 2006. Revision received April 7, 2006. Accepted for publication May 30, 2006.


   References
Top
 Abstract
 Introduction
 Discussion
 References
 

  1. Gadian DG. Nuclear magnetic resonance and its applications to living systems. Oxford University Press, 1982
  2. de Graaf RA. In vivo NMR spectroscopy - principles and techniques. John Wiley & Sons Ltd, 1998
  3. Cornel EB, Smits GA, Oosterhof GO, Karthaus HF, Deburyne FM, Schalken JA, et al. Characterization of human prostate cancer, benign prostatic hyperplasia and normal prostate by in vitro 1H and 31P magnetic resonance spectroscopy. J Urol 1993;150:2019–24.[Medline]
  4. Preul MC, Caramanos Z, Collins DL, Villemure JG, Leblanc R, Olivier A, et al. Accurate, noninvasive diagnosis of human brain tumors by using proton magnetic resonance spectroscopy. Nat Med 1996;2:323–5.[CrossRef][Medline]
  5. Khoo VS, Padhani AR, Tanner SF, Finnigan DJ, Leach MO, Dearnaley DP. Comparison of MRI with CT for the radiotherapy planning of prostate cancer: a feasibility study. Br J Radiol 1999;72:590–7.[Abstract]
  6. Beyersdorff D, Taupitz M, Winkelmann B, Fischer T, Lenk S, Loening SA, et al. Patients with a history of elevated prostate-specific antigen levels and negative transrectal US-guided quadrant or sextant biopsy results: value of MR imaging. Radiology 2002;224:701–6.[Abstract/Free Full Text]
  7. Rasch C, Barillot I, Remeijer P, Touw A, van Herk M, Lebesque JV. Definition of the prostate in CT and MRI: a multi-observer study. Int J Radiat Oncol Biol Phys 1999;43:57–66.[CrossRef][Medline]
  8. Roach M III, Faillace-Akazawa P, Malfatti C, Holland J, Hricak H. Prostate volumes defined by magnetic resonance imaging and computerized tomographic scans for three-dimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 1996;35:1011–8.[CrossRef][Medline]
  9. Norris DG. High field human imaging. J Magn Reson Imaging 2003;18:519–29.[CrossRef][Medline]
  10. Govindaraju V, Young K, Maudsley AA. Proton NMR chemical shifts and coupling constants for brain metabolites. NMR Biomed 2000;13:129–53.[CrossRef][Medline]
  11. Chung YL, Troy H, Banerji U, Jackson LE, Walton MI, Stubbs M, et al. Magnetic resonance spectroscopic pharmacodynamic markers of the heat shock protein 90 inhibitor 17-allylamino,17-demethoxygeldanamycin (17AAG) in human colon cancer models. J Natl Cancer Inst 2003;95:1624–33.[Abstract/Free Full Text]
  12. Ronen SM, Jackson LE, Beloueche M, Leach MO. Magnetic resonance detects changes in phosphocholine associated with Ras activation and inhibition in NIH 3T3 cells. Br J Cancer 2001;84:691–6.[CrossRef][Medline]
  13. Dewhirst MW, Poulson JM, Yu D, Sanders L, Lora-Michiels M, Vujaskovic Z, et al. Relation between pO2, 31P magnetic resonance spectroscopy parameters and treatment outcome in patients with high-grade soft tissue sarcomas treated with thermoradiotherapy. Int J Radiat Oncol Biol Phys 2005;61:480–91.[CrossRef][Medline]
  14. Findlay MP, Leach MO, Cunningham D, Collins DJ, Payne GS, Glaholm J, et al. The non-invasive monitoring of low dose, infusional 5-fluorouracil and its modulation by interferon-alpha using in vivo 19F magnetic resonance spectroscopy in patients with colorectal cancer: a pilot study. Ann Oncol 1993;4:597–602.[Abstract/Free Full Text]
  15. Bottomley PA. Spatial localization in NMR spectroscopy in vivo. Ann N Y Acad Sci 1987;508:333–48.[Medline]
  16. Frahm J, Merboldt K, Hanicke W. Localized proton spectroscopy using stimulated echoes. J Magn Reson 1987;72:502–8.
  17. Ordidge RJ, Connelly A, Lohman AB. Image-selected in vivo spectroscopy (ISIS). A new technique for spatially selective NMR spectroscopy. J Magn Reson 1986;66:283–94.
  18. Brown TR, Kincaid BM, Ugurbil K. NMR chemical shift imaging in three dimensions. Proc Natl Acad Sci USA 1982;79:3523–6.[Abstract/Free Full Text]
  19. Chen Z, Ma CM, Paskalev K, Li J, Yang J, Richardson T, et al. Investigation of MR image distortion for radiotherapy treatment planning of prostate cancer. Phys Med Biol 2006;51:1393–403.[CrossRef][Medline]
  20. Reinsberg SA, Doran SJ, Charles-Edwards EM, Leach MO. A complete distortion correction for MR images: II. Rectification of static-field inhomogeneities by similarity-based profile mapping. Phys Med Biol 2005;50:2651–61.[CrossRef][Medline]
  21. Doran SJ, Charles-Edwards L, Reinsberg SA, Leach MO. A complete distortion correction for MR images: I. Gradient warp correction. Phys Med Biol 2005;50:1343–61.[CrossRef][Medline]
  22. Lee YK, Bollet M, Charles-Edwards G, Flower MA, Leach MO, McNair H, et al. Radiotherapy treatment planning of prostate cancer using magnetic resonance imaging alone. Radiother Oncol 2003;66:203–16.[CrossRef][Medline]
  23. Veninga T, Huisman H, van der Maazen RW, Huizenga H. Clinical validation of the normalized mutual information method for registration of CT and MR images in radiotherapy of brain tumors. J Appl Clin Med Phys 2004;5:66–79.[CrossRef][Medline]
  24. Ehman RL, Felmlee JP. Adaptive technique for high-definition MR imaging of moving structures. Radiology 1989;173:255–63.[Abstract/Free Full Text]
  25. Cheng LL, Lean CL, Bogdanova A, Wright SC Jr, Ackerman JL, Brady TJ, et al. Enhanced resolution of proton NMR spectra of malignant lymph nodes using magic-angle spinning. Magn Reson Med 1996;36:653–8.[Medline]
  26. Murphy PS, Dzik-Jurasz AS, Leach MO, Rowland IJ. The effect of Gd-DTPA on T(1)-weighted choline signal in human brain tumours. Magn Reson Imaging 2002;20:127–30.[CrossRef][Medline]
  27. Smith JK, Castillo M, Kwock L. MR spectroscopy of brain tumors. Magn Reson Imaging Clin N Am 2003;11:415–29, v–vi
  28. Howe FA, Barton SJ, Cudlip SA, Stubbs M, Saunders DE, Murphy M, et al. Metabolic profiles of human brain tumors using quantitative in vivo 1H magnetic resonance spectroscopy. Magn Reson Med 2003;49:223–32.[CrossRef][Medline]
  29. Tate AR, Majos C, Moreno A, Howe FA, Griffiths JR, Arus C. Automated classification of short echo time in in vivo 1H brain tumor spectra: a multicenter study. Magn Reson Med 2003;49:29–36.[CrossRef][Medline]
  30. Majos C, Alonso J, Aguilera C, Serrallonga M, Perez-Martin J, Acebes JJ, et al. Proton magnetic resonance spectroscopy ((1)H MRS) of human brain tumours: assessment of differences between tumour types and its applicability in brain tumour categorization. Eur Radiol 2003;13:582–91.[Medline]
  31. Leclerc X, Huisman TA, Sorensen AG. The potential of proton magnetic resonance spectroscopy ((1)H-MRS) in the diagnosis and management of patients with brain tumors. Curr Opin Oncol 2002;14:292–8.[CrossRef][Medline]
  32. Maldjian JA, Patel RS. Cerebral neoplasms in adults. Semin Roentgenol 1999;34:102–22.[Medline]
  33. Lee PL, Gonzalez RG. Magnetic resonance spectroscopy of brain tumors. Curr Opin Oncol 2000;12:199–204.[CrossRef][Medline]
  34. Devos A, Lukas L, Suykens JA, Vanhamme L, Tate AR, Howe FA, et al. Classification of brain tumours using short echo time 1H MR spectra. J Magn Reson 2004;170:164–75.[CrossRef][Medline]
  35. Julia-Sape M, Acosta D, Mier M, Arus C, Watson D. A multi-centre, web-accessible and quality control-checked database of in vivo MR spectra of brain tumour patients. MAGMA 2006;19:22–33.[CrossRef][Medline]
  36. Jacobs AH, Kracht LW, Gossmann A, Ruger MA, Thomas AV, Thiel A, et al. Imaging in neurooncology. NeuroRx 2005;2:333–47.[Abstract/Free Full Text]
  37. Nelson SJ, Graves E, Pirzkall A, Li X, Chan AA, Vigneron DB, et al. In vivo molecular imaging for planning radiation therapy of gliomas: an application of 1H MRSI. J Magn Reson Imaging 2002;16:464–76.[CrossRef][Medline]
  38. Pirzkall A, Li X, Oh J, Chang S, Berger MS, Larson DA, et al. 3D MRSI for resected high-grade gliomas before RT: tumor extent according to metabolic activity in relation to MRI. Int J Radiat Oncol Biol Phys 2004;59:126–37.[CrossRef][Medline]
  39. Chan AA, Lau A, Pirzkall A, Chang SM, Verhey LJ, Larson D, et al. Proton magnetic resonance spectroscopy imaging in the evaluation of patients undergoing gamma knife surgery for Grade IV glioma. J Neurosurg 2004;101:467–75.[Medline]
  40. Walecki J, Tarasow E, Kubas B, Czemicki Z, Lewko J, Podgorski J, et al. Hydrogen-1 MR spectroscopy of the peritumoral zone in patients with cerebral glioma: assessment of the value of the method. Acad Radiol 2003;10:145–53.[CrossRef][Medline]
  41. Nelson SJ. Multivoxel magnetic resonance spectroscopy of brain tumors. Mol Cancer Ther 2003;2:497–507.[Abstract/Free Full Text]
  42. Murphy PS, Viviers L, Abson C, Rowland IJ, Brada M, Leach MO, et al. Monitoring temozolomide treatment of low-grade glioma with proton magnetic resonance spectroscopy. Br J Cancer 2004;90:781–6.[CrossRef][Medline]
  43. Laprie A, Pirzkall A, Haas-Kogan DA, Cha S, Banerjee A, Le TP, et al. Longitudinal multivoxel MR spectroscopy study of pediatric diffuse brainstem gliomas treated with radiotherapy. Int J Radiat Oncol Biol Phys 2005;62:20–31.[CrossRef][Medline]
  44. Kimura T, Sako K, Tanaka K, Gotoh T, Yoshida H, Aburano T, et al. Evaluation of the response of metastatic brain tumors to stereotactic radiosurgery by proton magnetic resonance spectroscopy, 201TlCl single-photon emission computerized tomography, and gadolinium-enhanced magnetic resonance imaging. J Neurosurg 2004;100:835–41.[Medline]
  45. Chernov MF, Hayashi M, Izawa M, Abe K, Usukura M, Ono Y, et al. Early metabolic changes in metastatic brain tumors after Gamma Knife radiosurgery: 1H-MRS study. Brain Tumor Pathol 2004;21:63–7.[CrossRef][Medline]
  46. Kurhanewicz J, Swanson MG, Nelson SJ, Vigneron DB. Combined magnetic resonance imaging and spectroscopic imaging approach to molecular imaging of prostate cancer. J Magn Reson Imaging 2002;16:451–63.[CrossRef][Medline]
  47. Cunningham CH, Vigneron DB, Marjanska M, Chen AP, Xu D, Hurd RE, et al. Sequence design for magnetic resonance spectroscopic imaging of prostate cancer at 3 T. Magn Reson Med 2005;53:1033–9.[CrossRef][Medline]
  48. Costello LC, Franklin RB. Concepts of citrate production and secretion by prostate. 1. Metabolic relationships. Prostate 1991;18:25–46.[Medline]
  49. Kim Y, Noworolski S, Pouliot J, Hsu IC, Kurhanewicz J. Analysis of prostate deformation due to different MRI/MRS endorectal coils for image fusion and brachytherapy treatment planning. Med Phys 2004;31:1728
  50. Kaji Y, Wada A, Imaoka I, Matsuo M, Terachi T, Kobashi Y, et al. Proton two-dimensional chemical shift imaging for evaluation of prostate cancer: external surface coil vs. endorectal surface coil. J Magn Reson Imaging 2002;16:697–706.[CrossRef][Medline]
  51. van Dorsten FA, van der GM, Engelbrecht MR, van Leenders GJ, Verhofstad A, Rijpkema M, et al. Combined quantitative dynamic contrast-enhanced MR imaging and (1)H MR spectroscopic imaging of human prostate cancer. J Magn Reson Imaging 2004;20:279–87.[CrossRef][Medline]
  52. Shukla-Dave A, Hricak H, Pucar D, Kattan MW, Kuroiwa K, Chen H, et al. Indolent prostate cancer-prediction by magnetic resonance imaging and spectroscopy. Proceedings of the 13th Scientific Meeting of the International Society for Magnetic Resonance in Medicine. 2005. 262:60
  53. Qayyum A, Coakley FV, Lu Y, Olpin JD, Wu L, Yeh BM, et al. Organ-confined prostate cancer: effect of prior transrectal biopsy on endorectal MRI and MR spectroscopic imaging. AJR Am J Roentgenol 2004;183:1079–83.[Abstract/Free Full Text]
  54. Pickett B, Kurhanewicz J, Coakley F, Shinohara K, Fein B, Roach M III. Use of MRI and spectroscopy in evaluation of external beam radiotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 2004;60:1047–55.[CrossRef][Medline]
  55. Scheidler J, Hricak H, Vigneron DB, Yu KK, Sokolov DL, Huang LR, et al. Prostate cancer: localization with three-dimensional proton MR spectroscopic imaging--clinicopathologic study. Radiology 1999;213:473–80.[Abstract/Free Full Text]
  56. Hasumi M, Suzuki K, Taketomi A, Matsui H, Yamamoto T, Ito K, et al. The combination of multi-voxel MR spectroscopy with MR imaging improve the diagnostic accuracy for localization of prostate cancer. Anticancer Res 2003;23:4223–7.[Medline]
  57. Scheenen T, Weiland E, Futterer J, van Hecke P, Bachert P, Villeirs G, et al. Preliminary results of IMAPS: an international multi-centre assessment of prostate MR spectroscopy. Proceedings of the 13th Scientific Meeting of the International Society for Magnetic Resonance in Medicine. 2005. 260:59
  58. Burns MA, He W, Wu CL, Cheng LL. Quantitative pathology in tissue MR spectroscopy based human prostate metabolomics. Technol Cancer Res Treat 2004;3:591–8.[Medline]
  59. Coakley FV, Teh HS, Qayyum A, Swanson MG, Lu Y, Roach M III, et al. Endorectal MR imaging and MR spectroscopic imaging for locally recurrent prostate cancer after external beam radiation therapy: preliminary experience. Radiology 2004;233:441–8.[Abstract/Free Full Text]
  60. Pickett B, Vigneault E, Kurhanewicz J, Verhey L, Roach M. Static field intensity modulation to treat a dominant intra-prostatic lesion to 90 Gy compared to seven field 3-dimensional radiotherapy. Int J Radiat Oncol Biol Phys 1999;44:921–9.[CrossRef][Medline]
  61. Pouliot J, Kim Y, Lessard E, Hsu IC, Vigneron DB, Kurhanewicz J. Inverse planning for HDR prostate brachytherapy used to boost dominant intraprostatic lesions defined by magnetic resonance spectroscopy imaging. Int J Radiat Oncol Biol Phys 2004;59:1196–207.[CrossRef][Medline]
  62. Xia P, Pickett B, Vigneault E, Verhey LJ, Roach M III. Forward or inversely planned segmental multileaf collimator IMRT and sequential tomotherapy to treat multiple dominant intraprostatic lesions of prostate cancer to 90 Gy. Int J Radiat Oncol Biol Phys 2001;51:244–54.[Medline]
  63. Murphy PS, Rowland IJ, Viviers L, Brada M, Leach MO, Dzik-Jurasz AS. Could assessment of glioma methylene lipid resonance by in vivo (1)H-MRS be of clinical value? Br J Radiol 2003;76:459–63.[Abstract/Free Full Text]
  64. Chiang IC, Kuo YT, Lu CY, Yeung KW, Lin WC, Sheu FO, et al. Distinction between high-grade gliomas and solitary metastases using peritumoral 3-T magnetic resonance spectroscopy, diffusion, and perfusion imagings. Neuroradiology 2004;46:619–27.[Medline]
  65. Yeung DK, Chan Y, Leung S, Poon PM, Pang C. Detection of an intense resonance at 2.4 ppm in 1H MR spectra of patients with severe late-delayed, radiation-induced brain injuries. Magn Reson Med 2001;45:994–1000.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 JNMHome page
C. Plathow and W. A. Weber
Tumor Cell Metabolism Imaging
J. Nucl. Med., June 1, 2008; 49(Suppl_2): 43S - 63S.
[Abstract] [Full Text] [PDF]

Home page
 JCOHome page
J. L. Speight and M. Roach III
Advances in the Treatment of Localized Prostate Cancer: The Role of Anatomic and Functional Imaging in Men Managed With Radiotherapy
J. Clin. Oncol., March 10, 2007; 25(8): 987 - 995.
[Abstract] [Full Text] [PDF]

Home page
 Br. J. Radiol.Home page
BJR review of the year - 2006
Br. J. Radiol., March 1, 2007; 80(951): 147 - 151.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Payne, G S
Right arrow Articles by Leach, M O
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Payne, G S
Right arrow Articles by Leach, M O

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
BJR DMFR IMAGING  ALL BIR JOURNALS