Magnetic Resonance in Radiotherapy
Magnetic Resonance in Radiotherapy Magnetic Resonance in Radiotherapy Supriyanto Ardjo Pawiro Supriyanto Ardjo Pawiro
Magnetic Nuclei studied Magnetic Nuclei studied I ntroduction I ntroduction
1
1 MRS MRS H is the most sensitive magnetic nucleus and H is the most sensitive magnetic nucleus and
• •
hydrogen is present in nearly all biologically relevant
hydrogen is present in nearly all biologically relevant compound. compound.
- Magnetic resonance spectroscopy is a non invasive technique for Magnetic resonance spectroscopy is a non invasive technique for
For comparison of spectra it is necessary to have a
• •
For comparison of spectra it is necessary to have a measuring biochemical in tissue. measuring biochemical in tissue.
reference frequency reference frequency
- The advantage of MRS produce signals at different frequencies,
- The advantage of MRS produce signals at different frequencies,
For consistency, the position of spectral peak in vivo
• •
For consistency, the position of spectral peak in vivo depending on the bonding to neighboring nuclei.
depending on the bonding to neighboring nuclei.
expressed relative to expressed relative to tetrametylsilane tetrametylsilane (TMS) even (TMS) even though it is not present in tissue.
though it is not present in tissue.
Other nuclei that may procedure useful MR spectra are Other nuclei that may procedure useful MR spectra are
• •
31
31
13
13
19
C, and F.
19 P,
P,
C, and F. Localization strategy and Validation Localization strategy and Validation
- MRS localization techniques fall into two groups. Single
- MRS localization techniques fall into two groups. Single
voxel voxel methods include those known as PRESS/ STEAM/ I SI S methods include those known as PRESS/ STEAM/ I SI S
in which data are required from a single in which data are required from a single voxel voxel positioned positioned using MR images. using MR images.
The alternative is MRSI as known as chemical shift The alternative is MRSI as known as chemical shift
imaging (CSI ) in which a matrix of spectra are acquired imaging (CSI ) in which a matrix of spectra are acquired either over a plane (2D either over a plane (2D - - - MRSI ) or a volume (3D MRSI ) or a volume (3D MRSI ). MRSI ). -
- Registration MR CT images and motion issues
Registration MR - CT images and motion issues
Before radiation oncologist have the confidence to alter Before radiation oncologist have the confidence to alter
treatment plans based on MRS data, they need evidence
treatment plans based on MRS data, they need evidence
that the abnormality detected using MRS represents that the abnormality detected using MRS represents tumor.
tumor.
The gold standard for comparison is histopathology The gold standard for comparison is histopathology
Case 1: MRS and Brain cancer Case 1: MRS and Brain cancer
- There is a strong incentive for non There is a strong incentive for non invasive assessment invasive assessment -
- of brain cancer, particularly for children. of brain cancer, particularly for children.
MRI appearance and localization alone have limited MRI appearance and localization alone have limited
power to differentiate brain tumor, particularly early
power to differentiate brain tumor, particularly early
stage disease and infiltrative disease stage disease and infiltrative disease
1
H MRS of brain tumors has developed rapidly due to the
less technical requirements of spectroscopy in the brain less technical requirements of spectroscopy in the brain compared with measurements elsewhere in the body. compared with measurements elsewhere in the body.
Recently there has been increasing interest in using
Recently there has been increasing interest in using functional imaging techniques, together with metabolic functional imaging techniques, together with metabolic
imaging using MRS to aid target definition in
imaging using MRS to aid target definition in radiotherapy. radiotherapy.
- Area of relatively high Cho/ NAA may indicate high
- Area of relatively high Cho/ NAA may indicate high
- cellular activity, and hence radio cellular activity, and hence radio sensitivity, and Lac may sensitivity, and Lac may indicate hypoxic area with reduces radiosensitivity .
indicate hypoxic area with reduces radiosensitivity .
The technique can also be helpful in identifying area The technique can also be helpful in identifying area
missed by radiation fields, and in separating recurrence
missed by radiation fields, and in separating recurrence from radiation necrosis. from radiation necrosis.
Reduction in Cho and Lipid and lactate can also reflect
Reduction in Cho and Lipid and lactate can also reflect response to chemotherapy and radiotherapy. response to chemotherapy and radiotherapy. Therapeutic guidance, assessment of response and recent developm Therapeutic guidance, assessment of response and recent developm ent ent
•
•
MRS has been compared with MRI and SPECT in monitoring MRS has been compared with MRI and SPECT in monitoring response to treatment in metastasis brain tumor. response to treatment in metastasis brain tumor.
•
Reduction in Cho and Lac and increasing Lip (believe to•
Reduction in Cho and Lac and increasing Lip (believe to
represent necrosis) in corresponding tumor were detected represent necrosis) in corresponding tumor were detected at earlier time (1 week to 1 months) than contrast
at earlier time (1 week to 1 months) than contrast enhanced MRI or SPECT. enhanced MRI or SPECT.
Case2 : MRS and Prostate Validation Case2 : MRS and Prostate
- A strong correlation has been found between negative
MRSI and negative biopsy findings, and between positive MRSI and positive biopsy findings
•
There is only a weak correlation between the concentration of prostate specific antigen (PSA, the current ‘‘gold standard’’) and either biopsy or MRSI findings• Several studies have shown that adding MRSI to an MRI examination increases the accuracy of diagnosis
- Use of MRSI together with MRI has been shown to
- MRSI has been used in combination with MRI to define improve substantially the identification of tumor recurrence regions for dose escalation within the prostate, following external beam radiotherapy (the area under the permitting a dose of > 90 Gy to the high-risk region
ROC curve, a measure of the effectiveness of a test, while treating the remainder of the prostate to about 70 increased from 0.5 to 0.81
Gy
- 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
Gamma Knife Radiosurgery planning Gamma Knife Radiosurgery planning and delivery and delivery
- At University California San Francisco, the clinically
- At University California San Francisco, the clinically
utilized radiological unit is the gamma knife from utilized radiological unit is the gamma knife from Elekta Elekta instrument instrument
The morning of treatment, a stereotactic frame is The morning of treatment, a stereotactic frame is attached to patient attached to patient ’ ’ s skull to provide a means of s skull to provide a means of immobilization as well as a coordinate reference system
immobilization as well as a coordinate reference system
The patient then undergoes an MR examination which is The patient then undergoes an MR examination which is
subsequently transferred to a workstation for treatment subsequently transferred to a workstation for treatment planning using planning using GammaPlan GammaPlan software software . .
Reconstruction of MRSI Data Reconstruction of MRSI Data Acquisition of MRSI Data Acquisition of MRSI Data
Reconstruction and analysis of 3D MRSI data is achieved Reconstruction and analysis of 3D MRSI data is achieved
• •
Three dimensional MRSI acquisition employing a point Three dimensional MRSI acquisition employing a point
by transferring the data from the scanner to an offline by transferring the data from the scanner to an offline resolved spectroscopy (PRESS) technique. resolved spectroscopy (PRESS) technique.
UltraSPARC UltraSPARC solaris solaris workstation. workstation.
A phase encoding matrix of 8x8x8, 16x8x8, or 12x12x8 A phase encoding matrix of 8x8x8, 16x8x8, or 12x12x8
The peak parameters (height, area, and width) for
•
•
The peak parameters (height, area, and width) for chosen based on the dimensions of the volume to be chosen based on the dimensions of the volume to be
choline choline , , creatine creatine (Cr), N (Cr), N - acetyl aspartate - acetyl aspartate (NAA), and (NAA), and excited. excited. lactate/ lipid (LL) are then calculated for each voxel
lactate/ lipid (LL) are then calculated for each voxel All MRSI acquisitions use a TR of 1000 ms, TE of 144 All MRSI acquisitions use a TR of 1000 ms, TE of 144
Statistical routine has been developed that computes a Statistical routine has been developed that computes a
• •
- – ms, and are acquired in 17 ms, and are acquired in 17 – 19 minutes 19 minutes measure of abnormality for each spectral voxel in
measure of abnormality for each spectral voxel in
dataset based on the extent to which its dataset based on the extent to which its choline choline and and NAA levels.
NAA levels.
•
•
The measurements is referred to as the The measurements is referred to as the choline choline / NAA / NAA
residual, and has been shown to reproducibly quantity residual, and has been shown to reproducibly quantity the degree of spectral abnormality in the degree of spectral abnormality in voxel voxel relative to relative to normal normal voxel voxel . .
•
• The authors have review more than 1500 combined The authors have review more than 1500 combined•
• 100 patients with recurrent primary brain tumors, 36%• •
•
• This phenomena useful for both expanding high dose This phenomena useful for both expanding high dose- This technique was been shown to reproducibility This technique was been shown to reproducibility
• •
• •
•
• I ts is important that the target only have diameter less I ts is important that the target only have diameter less• •
- The result of this analysis demonstrated a tendency for The result of this analysis demonstrated a tendency for
• •
radiosurgey
radiosurgey
were compared to
were compared to
the radiation target volume and to patients clinical the radiation target volume and to patients clinical outcome.
outcome.
poorer responses, in terms of survival time, time to poorer responses, in terms of survival time, time to further treatment, and change in contrast enhancing
further treatment, and change in contrast enhancing
volume, among patients with spectroscopic abnormalities volume, among patients with spectroscopic abnormalities that extended beyond the target volume at the time of
that extended beyond the target volume at the time of
treatment treatment
The gamma knife The gamma knife radiosurgery radiosurgery is focal ablative therapy, is focal ablative therapy, modification of the prescribed radiation target based on modification of the prescribed radiation target based on MRSI data has proceeded cautiously
MRSI data has proceeded cautiously
or equal to 4 cm can be effectively treated with
or equal to 4 cm can be effectively treated with
radiosurgery radiosurgery
The constructed plan is then viewed superimposed on The constructed plan is then viewed superimposed on the the choline choline / NAA residual contour labeled MRI image / NAA residual contour labeled MRI image within
within
Gamaplan
Gamaplan
I n situation where the extent of abnormal spectroscopic I n situation where the extent of abnormal spectroscopic voxels
voxels
is larger than the maximum permissible volume
is larger than the maximum permissible volume
for for radiosurgery radiosurgery , the strategy suggest target the use of , the strategy suggest target the use of concurrent chemotherapy
undergoing gamma knife
The pre The pre - - treatment metabolic lesion of 36 treatment metabolic lesion of 36 glioma glioma patient patient undergoing gamma knife
I mplications of Metabolic target delineation I mplications of Metabolic target delineation for gamma knife for gamma knife radiosurgey radiosurgey
with mid or low grade lesions had spectra with elevated
Correlation of MRSI with Treatment planning Correlation of MRSI with Treatment planning MRI MRI
Registration of the MRI component of Registration of the MRI component of MRSi MRSi examination examination to MRI dataset being used for GK treatment planning to MRI dataset being used for GK treatment planning
The The choline choline / NAA residual map determined from the / NAA residual map determined from the registered registered MRSi MRSi data is then data is then resampled resampled to the resolution to the resolution of the registered MRI , and represented as contour map of the registered MRI , and represented as contour map at levels of 2.0, 3.0, and 4.0 at levels of 2.0, 3.0, and 4.0
The contour are represented within the registered MRI The contour are represented within the registered MRI pixel data and the final image is sent to TPS. pixel data and the final image is sent to TPS.
The same coordinate frame as planning image dataset The same coordinate frame as planning image dataset and spectral contour can be used to guide the and spectral contour can be used to guide the delineation of radiation target.
delineation of radiation target.
identify the region of spectroscopic abnormality in identify the region of spectroscopic abnormality in glioma glioma patients within 3 mm. patients within 3 mm.
Differences between anatomic and Differences between anatomic and metabolic lesions in brain tumors metabolic lesions in brain tumors
MRI / MRI / MRSi MRSi examinations of patens with brain tumors examinations of patens with brain tumors
100 patients with recurrent primary brain tumors, 36%
of patients with high grade lesion and 50% of patient of patients with high grade lesion and 50% of patient with mid or low grade lesions had spectra with elevated
choline choline and decreased NAA outside the region of and decreased NAA outside the region of contrast enhancement
glioma
contrast enhancement
Only 50% of the T2 Only 50% of the T2 hyperintense hyperintense region had region had voxel voxel with with a
a
choline
choline
/ NAA residual greater that or equal 2.0
/ NAA residual greater that or equal 2.0
treatment volume beyond the contrast enhancing region treatment volume beyond the contrast enhancing region and limiting the dose delivered and limiting the dose delivered
I f T2 I f T2 hyperintense hyperintense regions with regions with choline choline / NAA residual / NAA residual value greater than 2.0, 3.0, and 4.0 in population of
value greater than 2.0, 3.0, and 4.0 in population of
patient with newly diagnosed grade I I I and grade I V patient with newly diagnosed grade I I I and grade I V glioma
concurrent chemotherapy
Subject: Subject:
•
Successful 2D CSI was performed in 28 of 29 patients.
29 patient ( 16 male, 13 female), 4
54 year
54 year
- 29 patient ( 16 male, 13 female), 4
•
The following parameters were used for all 2D CSI examinations: a point-resolved spectroscopy sequence (PRESS); TR/ TE, 1,500/ 144;- Tumor stages are I I Tumor stages are I I - -
• Field of view, 16 cm; matrix, 16 × 16; slice thickness,
All patient got fractionated radiation therapy ( 54 All patient got fractionated radiation therapy ( 54 – –
70
70 Gy Gy
)
)
I V
I V
Methods: Methods:
10–20 mm; acquisition, 1 average; scanning time, 4 min 20 sec
- The spectra were analyzed for the signal intensity of NAA, choline, and creatine and for the presence of lactates and lipids.
- Ratios were manually calculated for NAA/ Cr, Cho/ Cr, and Cho/ NAA
- Metabolite ratios among the recurrent tumor population and radiation injury population were compared.
- The highest Cho/ Cr, Cho/ NAA, and NAA/ Cr ratios in one voxel were
used for comparison
- Two locations in the brain of 26 of 28 patients were measured, one
in the contrast-enhancing lesion and one in the normal-appearing white matter.
- The generalized estimating equation technique was used to fit a
linear regression model for each of the three ratios to take into account the within-subject correlation
- Robust SEs were used to compute Wald test p values. The level for
p value of less than 0.05 statistical significance was set to a
- The reference is histopathology
•
•
Tumor had significantly higher Cho/ Cr than the radiation injury and Tumor had significantly higher Cho/ Cr than the radiation injury and had significantly higher Cho/ Cr ratios than the normal-appearing white matter
- The lesions in the tumor group also had significantly higher
Cho/ NAA ratios than those in the radiation injury group and lesions in this group again had significantly higher Cho/ NAA ratios than the normal-appearing white matter
- Lesions in the tumor group had significantly lower NAA/ Cr ratios
than those in the radiation injury group, which had insignificantly lower NAA/ Cr ratios than the normal-appearing white matter
Discussion Discussion Have presented identifying tumor extent and Have presented identifying tumor extent and
metabolically active regions to aid targeting of metabolically active regions to aid targeting of radiotherapy, evaluating response to treatment, radiotherapy, evaluating response to treatment, and identifying recurrence and radiation injury and identifying recurrence and radiation injury
3 The limitation voxel size is typically 8 mm
- 3
The limitation voxel size is typically 8 mm to 10
- to 10
3 mm to achieve SNR mm to achieve SNR
3
- None of the lesions classified as radiation injury reached the value
Need improvement of sensitivity and spectral of 1.8 in either Cho/ Cr or Cho/ NAA ratios. Two lesions classified as Need improvement of sensitivity and spectral • • radiation injury had Cho/ NAA ratios of 1.71 and 1.78, respectively. specificity to achieve the good resolution specificity to achieve the good resolution
- Two patients with tumor who had Cho/ Cr ratios lower than 1.8 had
For radiotherapy planning, MR doesn ’ t provide
- For radiotherapy planning, MR doesn ’ t provide
Cho/ NAA ratios higher than 1.8 (3.07 and 2.00). One patient with tumor who had a Cho/ NAA ratio less than 1.8 had a Cho/ Cr ratio electron density electron density greater than 1.8 (1.96)
- Only one patient with tumor had both Cho/ Cr and Cho/ NAA ratios lower than 1.8.
Assessment of geometrical accuracy of Assessment of geometrical accuracy of magnetic resonance images for radiation magnetic resonance images for radiation therapy of lung cancer therapy of lung cancer
Journal of Applied Clinical Medical Physics 4 (4), 2003 Journal of Applied Clinical Medical Physics 4 (4), 2003
N. Koch, H.H. Liu, L.E. Olsson, N. Koch, H.H. Liu, L.E. Olsson, and E.F. and E.F. Jakson Jakson
I ntroduction I ntroduction Methods: Theoretical simulation Methods: Theoretical simulation
I nterest in the assessment of respiration induced lung I nterest in the assessment of respiration induced lung
•
•
Difference in magnetic susceptibility between two materials Difference in magnetic susceptibility between two materials
tumor motion has increased with advent of gated therapy
tumor motion has increased with advent of gated therapy
creates microscopic pertubation in magnetic fields
creates microscopic pertubation in magnetic fields
- Complementary to x Complementary to x ray computed tomography, MR ray computed tomography, MR
•
•
I n the human body, interfaces of air ( x ~ 0.36 I n the human body, interfaces of air ( x ~ 0.36 ppm ppm ) and ) and
imaging offer unique capabilities for tracking the imaging offer unique capabilities for tracking the tomor tomor tissue ( X~ - 11 to 7.0 - ppm ) give rise to the largest natural
- tissue ( X~ - 11 to 7.0 ppm ) give rise to the largest natural with flexible imaging plane with flexible imaging plane magnetic susceptibility field magnetic susceptibility field pertubation pertubation . .
The distortion is also dependent on the shape of the The distortion is also dependent on the shape of the
• •
They exploited some features of MR imaging using two They exploited some features of MR imaging using two
interface or object and the orientation of its axis interface or object and the orientation of its axis pulse sequences: a cine imaging sequence designed for
pulse sequences: a cine imaging sequence designed for
Circular cross sections of vials with this axis oriented Circular cross sections of vials with this axis oriented
• •
tracking lung tumor during respiration. tracking lung tumor during respiration.
perpendicular and parallel with direction of main magnetic perpendicular and parallel with direction of main magnetic Localized fluctuation in the magnetic field originating from Localized fluctuation in the magnetic field originating from field.
- field.
- large susceptibility difference can cause distortion near air large susceptibility difference can cause distortion near air tissue interfaces tissue interfaces
Methods: Phantom experiments Methods: Phantom experiments
•
•
Phantom was designed and built in house to approximate Phantom was designed and built in house to approximate
the geometry of the upper thorax, including two air cavities the geometry of the upper thorax, including two air cavities that served as simulated lung that served as simulated lung
3
3 The phantom size was 35 x 41 x 17 cm The phantom size was 35 x 41 x 17 cm with cavity of the with cavity of the
• •
3
3
phantom measured 28 x 15 x 15 cm and contained the
phantom measured 28 x 15 x 15 cm and contained the inserts. inserts.
The inserts were designed to hold vial in place so vial The inserts were designed to hold vial in place so vial
• •
- cross sections could be imaged in the sagital and cor
- cross sections could be imaged in the sagital and coronal planes. planes.
Analysis
The inner diameter of the vials was about 1.5 cm and the The inner diameter of the vials was about 1.5 cm and the distance between two adjacent vial was approximately of distance between two adjacent vial was approximately of
3.8 5.6 cm depending on the arrangement. 5.6 cm depending on the arrangement. –
- – 3.8
The filling phantom and vials, except the air cavities was The filling phantom and vials, except the air cavities was
8.3 L of a solution consisting of deinonized water, 2.4
- gram/ L gram/ L NaCl NaCl , 18 , 18 mL mL Gd Gd DTPA doping agent. The solution DTPA doping agent. The solution has a measured T1/ T2 of 204/ 112 ms
has a measured T1/ T2 of 204/ 112 ms
MR images were acquired of phantom on a 1.5 T whole MR images were acquired of phantom on a 1.5 T whole
body signal echo speed MR scanner and an fGRE sequence
body signal echo speed MR scanner and an fGRE sequence
was used to get images with parameters : BW 83.3 kHz, was used to get images with parameters : BW 83.3 kHz,
o o
FOV= 44 cm, TE/ TR = 2/ 4 ms, flip angle 60 , NEX = 3.0,
FOV= 44 cm, TE/ TR = 2/ 4 ms, flip angle 60 , NEX = 3.0, matrix = 512 x 512 and slice thickness 1.0 mm. matrix = 512 x 512 and slice thickness 1.0 mm.
Second sequence fast spin echo (FSE) was used to get Second sequence fast spin echo (FSE) was used to get
images with parameters : BW 62.5 kHz, FOV= 44 cm, images with parameters : BW 62.5 kHz, FOV= 44 cm, TE/ TR = 9.4/ 600 ms, echo train length= 3, NEX = 3.0, and TE/ TR = 9.4/ 600 ms, echo train length= 3, NEX = 3.0, and matrix = 512 x 512 matrix = 512 x 512
Result Result
Sagital of 1 st sequence Coronal To measure the accuracy, the result compare standard from CT images
I n registering the MR and CT images, a common coordinate frame w I n registering the MR and CT images, a common coordinate frame w as as established with an original selected near the center of image. established with an original selected near the center of image. The rationale of such an approach is that the image distortion is minimum near the center of the magnetic field.
- Though the differences in the vial positions between the MR and CT
images could be subject to the specific location of the reference point, we expect that the effect of choosing the reference point would be negligible as long as it is near the center of the magnet and FOV.
- I n addition, the position differences measured from the fGRE images were all significantly less than 2 mm.
Conclusion Conclusion
- The vial positions in the sagittal and coronal fGRE images had displacements no greater than 2.0 mm in either the FE or PE direction compared with those from the CT images. >Discrepancies exceeding the acceptable limit of 2.0 mm were found along the periphery in the axial images using the FSE sequen
- The spatial accuracy for the sagital and coronal images
were found to be acceptable for subsequent patient imaging, whereas additional assessment for the axial image has to be made to improve their spatial accuracy for radiation therapy applications