BJR
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
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 Sannazzari, G L
Right arrow Articles by Guarneri, A
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sannazzari, G L
Right arrow Articles by Guarneri, A
British Journal of Radiology 75 (2002),603-607 © 2002 The British Institute of Radiology

Full Paper

CT–MRI image fusion for delineation of volumes in three-dimensional conformal radiation therapy in the treatment of localized prostate cancer

G L Sannazzari, MD 1 R Ragona, PhD 1 M G Ruo Redda, MD 1 F R Giglioli, PhD 2 G Isolato, MD 1 and A Guarneri, MD 1

1 Department of Radiation Oncology and Diagnostic Imaging, University of Turin, Turin and 2 Medical Physics Division, S. Giovanni Battista Hospital, Via Genova 3, 10126 Turin, Italy

Correspondence: Dr G L Sannazzari, Department of Radiation Oncology and Diagnostic Imaging, S. Giovanni Battista Hospital, Via Genova 3, 10126 Turin, Italy


   Abstract
Top
 Abstract
Introduction
 Material and methods
 Results
 Conclusion
 References
 
The objective of this study was to assess the utility of CT–MRI image fusion software and compare both prostate volume and localization with CT and MRI studies. We evaluated the differences in clinical volumes in patients undergoing three-dimensional conformal radiation therapy for localized prostate cancer. After several tests performed to ensure the quality of image fusion software, eight patients suffering from prostate adenocarcinoma were submitted to CT and MRI studies in the treatment position within an immobilization device before the start of radiotherapy. The clinical target volume (CTV) (prostate plus seminal vesicles) was delineated on CT and MRI studies and image fusion was obtained from the superimposition of anatomical fiducial markers. A comparison of dose–volume histograms relative to CTV, rectum, bladder and femoral heads was performed for both studies. Image fusion showed a mean overestimation of CTV of 34% with CT compared with MRI. Along the anterior–posterior and superior–inferior direction, CTV was a mean 5 mm larger with CT study compared with MRI. The dose–volume histograms resulting from CT and MRI comparison showed that it is possible to spare a mean 10% of rectal volume and approximately 5% of bladder and femoral heads, respectively. This study confirmed an overestimation of CTV with CT images compared with MRI. Because this finding only allows a minimal sparing of organs at risk, considering the organ motion during each radiotherapy session and the excellent outcomes of prostate cancer treatment with CT based target identification, we are still reluctant to reduce the CTV to that identified by MRI.


   Introduction
Top
 Abstract
 Introduction
Material and methods
 Results
 Conclusion
 References
 
Prostate cancer is the most common malignancy and second leading cause of cancer death in men. High-precision three-dimensional conformal radiation therapy (3D-RT) using sophisticated computer-aided treatment planning, along with surgery, has been accepted as standard therapy for localized prostate cancer [1]. 3D-RT allows rapid decrease in dose delivered to surrounding normal tissues since it is a function of distance from the targeted tumour. It often permits effective exclusion of surrounding normal tissues from the volume exposed to high radiation dose levels of up to 80 Gy. This, in turn, allows significant increases in tumour dose to levels beyond those feasible with conventional two-dimensional radiotherapy, with a concomitant decrease in the risk of normal tissue complications and improved local tumour control [2].

Since patients may be treated radically either by surgery or radiation, with equivalent disease-free and overall survival rates, the choice of treatment modality is often based on a comparison of treatment related morbidity [3, 4].

Progress in medical imaging, as well as computer hardware and software developments, has led to improvements in the delineation and modeling of tumour volumes for radiation therapy, depending on the clinical stage of the disease (prostate gland and a variable portion of the seminal vesicles). These volumes, defined by 5 mm thick axial CT slices [1], are the basis upon which conformal treatment plans using three-dimensional (3D) treatment planning systems are designed. However, while CT easily recognizes the upper prostate and seminal vesicles, the lower extent of the gland is difficult to distinguish from adjacent normal structures, i.e. urogenital diaphgram, because of small CT numbers [5]. Since MRI yields more contrast than CT when differentiating the prostate gland from the periprostatic soft tissues, as well as allowing more precise delineation of normal critical structures and more accurate definition of treatment volumes. Diagnostic information available from MRI can be incorporated into that of CT [6, 7]. This integration of complementary information from two or more separate imaging studies into a single consistent imaging study is called "image fusion". This technique can better enable the radiation oncologist and medical physicist to design and execute a successful course of therapy and more closely follow the progress of the patient after therapy.

However, imaging data from MRI introduces some geometric distorsion because all magnetic fields possess inhomogeneities of the main field and non-linearities of magnetic field gradients and eddy current effects. In general, system distortion is particularly important for larger fields of view (FOVs) as this distorsion tends to increase with increasing distance from the centre of the magnet [8]. Also MRI does not provide the necessary geometric accuracy and physical information required in CT based 3D treatment planning systems, such as electron density of body tissues. Nor can MR image complex bone/air heterogeneity. This information is essential for patient dose calculation and for designing compensators and modulators to shape the beam profile. Therefore, the unique information provided by MRI studies must be registered to, and then integrated with, the treatment planning CT data set.

A variety of quantitative methods have been developed to determine transformation parameters, including point matching, line or curve matching, surface matching and volume matching.

This paper evaluates the possible contribution of MRI images in accurately defining clinical target volume (CTV) (prostate and seminal vesicles), by giving the necessary information for computerized delineation leading to image fusion.

This study is based on the evaluation of the results achieved with CT–MRI image fusion in a group of patients suffering from localized prostate carcinoma, and is an effort in analyzing and rationalizing the use of this methodology for better treatment planning in radiation oncology.


   Material and methods
Top
 Abstract
 Introduction
 Material and methods
Results
 Conclusion
 References
 
This study was conducted on eight patients affected with localized prostate adenocarcinoma who underwent radical 3D-RT at the Radiation Oncology Division, University of Turin from January 1999.

For all patients, diagnosis was based on prostate specific antigen (PSA) levels and biopsy. Patient age ranged from 61 years to 76 years (median 71 years), and palpation T stage ranged from T1a to T2b. The median pre-treatment PSA was 12.6 ng ml-1 (range 7.6–14.8 ng ml-1). Gleason Score ranged from 3 to 9 (median 6).

To ensure that patients could be positioned in a reproducible fashion for planning and treatment, individualized thermoplastic casts were produced for each patient. Each patient underwent CT and MRI in the planned treatment position within the immobilization device from level L5–S1 to 10 mm caudal to the ischial tuberosities. CT and MR images were 3 mm thick and had a 3 mm interval, these conditions being necessary for correct image fusion. Pelvic MRI was taken with a 1 T scanner, using phase array multicoil. In order to minimize MRI geometric distortion, patients were positioned to set the CTV at the magnet centre.

Image fusion was obtained through transversal T2 weighted images (TR/TE= 4600 ms/100 ms). The contours of CTV and critical organs and structures, such as rectum, bladder and femoral heads, were later delineated by the same radiation oncologist on the acquired CT and MR images (Figure 1Go). Moreover, in 3D treatment planning it is necessary to take into account uncertainties in tumour delineation, organ motion and day to day patient positioning. To compensate for these uncertainties, additional safety margins have usually been added to the CTV to shape the planning target volume (PTV), extending into surrounding normal tissue to decrease the risk of marginal tumour miss. Therefore, PTV included the entire prostate and seminal vesicles plus 1.0 cm margins, except at the prostate–rectum interface where a 0.6 cm margin was used to decrease the risk of rectal toxicity.



View larger version (135K):
[in this window]
[in a new window]
  		Figure 1. CT–MRI image fusion: delineation of the clinical target volume (CTV) and critical organs. Red line, CT CTV; yellow line, MRI CTV; blue line, rectum (CT); green line, bladder (CT).

 
On the acquired CT and MR images, a cluster of anatomical fiducial markers was selected, varying for each patients from 9 to 15, localized in connection to the pubic symphysis and the acetabulum margins. The image fusion software can give the total deviation to the superimposition between CT and MRI fiducial markers. Several tests were carried out to achieve a total standard deviation to the superimposition of fiducial markers lower than 1.5 mm; in all patients, optimal number and disposition of anatomical landmarks were found. Image fusion and CTV measuring were then carried out. The study aimed to evaluate the accuracy of MRI in defining CTV compared with that of CT. For this purpose, the CTVs delineated on images from both methods were compared through the survey of dimensions and differences observed after point superimposition and image fusion.


   Results
Top
 Abstract
 Introduction
 Material and methods
 Results
Conclusion
 References
 
From data analysis it was noted that along the anterior–posterior direction of CTV, mean values only show minimal non-significant differences at all analyzed levels (base, centre, apex). The linear regression line shows that differences are more evident only in the larger sections where CT values 5 mm larger than MRI values were registered (Figure 2Go).



View larger version (17K):
[in this window]
[in a new window]
  		Figure 2. The relationship of CT–MRI defined prostate dimension (mm) in the anterior–posterior direction.

 
Significant mean differences (p<0.05) are registered along the lateral–lateral direction at the base and central body of the CTV. The linear regression line shows that 80 mm MRI images correspond to CT images overestimated by 10 mm (Figure 3Go).



View larger version (16K):
[in this window]
[in a new window]
  		Figure 3. The relationship of CT–MRI defined prostate dimension (mm) in the lateral–lateral direction.

 
Image fusion also allowed the analysis of organ position, showing a mean forward shifting of the CTV in MRI of approximately 5 mm (±9.25 mm) compared with CT. The associated standard deviation yields highly varying results, with little statistical significance [9]. Detailed analysis resulted in a higher discrepancy at the base and central body of the CTV showing, respectively, a 6 mm and 8 mm forward displacement in MRI.

Measures taken on the superior–inferior direction show a 5 mm mean overestimation of CT compared with MRI. The linear regression line shows similar values for small sections of the prostate, whereas a noticeable difference is registered for larger sections of the gland (7–8 mm discrepancy at dimensions of 60 mm) (Figure 4Go).



View larger version (18K):
[in this window]
[in a new window]
  		Figure 4. The relationship of CT–MRI defined prostate dimension (mm) in the superior–inferior direction.

 
Analyzing prostate volume (Figure 5Go), a 34% mean overestimation in CTV volume by CT compared with MRI is registered.



View larger version (16K):
[in this window]
[in a new window]
  		Figure 5. The relationship of CI–MRI defined prostate volume (cm3).

 
A personalized treatment plan for a patient with mean prostate anatomical characteristics was designed. The first step was the conformation of six irradiation fields to CT-calculated CTV. The same operation was then repeated for MRI-calculated CTV. Dose distribution was calculated for both situations and the ensuing dose–volume histograms were compared.

The dose–volume histogram resulting from CT–MRI comparison shows that it is possible to spare approximately 10% of rectal volume and approximately 5% of bladder and femoral head volumes, by delineating CTV on MRI acquired images. All in all, comparison between the two methods only shows small differences, and it is important to evaluate its clinical yield, taking into account the closeness between the prostate, rectum and bladder.


   Conclusion
Top
 Abstract
 Introduction
 Material and methods
 Results
 Conclusion
References
 
Routine application of CT–MRI pelvic image fusion would require that images be taken at the same time, in order to reduce inaccuracy owing to prostate volume shifting caused by the different degree of filling in the patient's rectum and bladder. Regarding the variation of PTV position, Melian et al [10] proved that such shifting is mainly in a superior–inferior direction in connection with the base of the prostate, whereas the apex area is the firmer part of the gland. On the grounds of this observation, none of the patients underwent retrograde urethrography.

Our study showed that the volume measured through CT, including prostate and seminal vesicles, was 34% greater than the same volume acquired through MRI. Kagawa et al [11], in a similar study on 22 patients, proved a mean increase in CT acquired prostate volume with or without seminal vesicles of 27% compared with MRI. Studies conducted by Rasch et al [12] and by Roach et al [13], where the volume only included the prostate, proved an increase of 43% and 32%, respectively, on CT acquired measurements compared with MRI. In our study, mean absolute values are higher than reported by other authors (Table 1Go). This may be due to our delineation of volume, which always included the prostate capsule and seminal vesicles. Some authors [14–16] evaluated MRI prostate volume before radical prostatectomy compared with surgical specimens. Results from these studies indicate that MRI accuracy consistently underestimated prostate volume.


View this table:
[in this window]
[in a new window]
  		Table 1. Comparision of prostate volume studies on image fusion

 
From data analysis, a significant 5 mm CTV reduction in the prostate superior–inferior dimension was registered, comparing favourably with Roach et al (4.5 mm) [13]. Considering the anterior–posterior and lateral–lateral dimensions, the differences were 1.9 mm and 4.4 mm, respectively.

A mean forward shift of 5 mm in MRI-acquired CTV images was noted, probably due to the different degree of rectal and bladder filling.

The comparison of dose–volume histograms in a patient presenting morphological and volumetric characteristics similar to mean values of all analyzed patients, showed that it is possible to spare approximately 10% of the rectal volume included in the radiation field and approximately 5% of bladder and femoral head volumes.

This study has identified both where and why the CT imaged prostate is larger than the MR imaged prostate. These findings may help radiation oncologists in target identification and regarding normal structures adjacent to the prostate, when the latter is imaged by CT for treatment planning. An accurate definition of target volume is the base of overall treatment procedure. More precise targeting may result not only in improved disease free survival and post-treatment biopsy but also in decreased morbidity by reducing radiation induced damage to normal tissues.

Anatomical organ motion and patient positioning (set-up) are an especially large barrier to conformal radiation therapy. However, considering the excellent outcomes of prostate cancer treatment with CT-based target identification and low treatment related toxicity, we are still reluctant to reduce CTV to that identified by MRI until we can control organ motion during each radiotherapy session. Both accurate targeting and adequate control of organ motion are crucial to optimize the CTV.

Received for publication August 3, 2001. Revision received November 19, 2001. Accepted for publication January 8, 2002.


   References
Top
 Abstract
 Introduction
 Material and methods
 Results
 Conclusion
 References
 

  1. Leibel SA, Zelefsky MJ, Kutcher GJ, et al. The biological basis and clinical application of three-dimensional conformal external beam radiation therapy in carcinoma of the prostate. Semin Oncol 1994;21:580–97.[Medline]
  2. Leibel SA, Kutcher GJ, Mohan R, Harrison LB, Armstrong JG, Zelefsky MJ, et al. Three dimensional conformal radiation therapy at the memorial Sloan-Kettering Cancer Center. Semin Radiat Oncol 1999;2:274–89.
  3. Hanks GE. External beam radiation therapy for clinically localized prostate cancer: patterns of care studies in the united States. NCI Monogr 1988;7:75–84.
  4. National Institutes of Health Consensus Development Panel. Consensus statement: the management of clinically localized prostate cancer. NCI Monogr 1988;7:3–6.
  5. Sandler HM, Lee LB, McLaughlin PW, et al. Localizaton of the prostatic apex for radiation therapy using implanted markers. Int J Radiat Oncol Biol Phys 1993;27:915–9.[Medline]
  6. Schnall MD, Pollack HM. Magnetic resonance imaging of the prostate gland. Urol Radiol 1990;12:109–14.[Medline]
  7. Khoo VS, Padhani AR, Tanner SF, Finnigan DJ, Leach MO, Dearnaley DP. Comparison of MRI with CT for radiotherapy planning of prostate cancer: a feasibility study. Br J Radiol 1999;72:590–7.[Abstract]
  8. Khoo VS, Dearnaley DP, Finnigan DJ, Padhani A, et al. Magnetic resonance imaging (MRI): considerations and applications in radiotherapy treatment planning. Radiother Oncol 1997;42:1–15.[Medline]
  9. Zelefsky MJ, Crean D, Mageras GS, Lyass D, Happersett L, Ling CC, et al. Quantification and predictors of prostate position variability in 50 patients evaluated with multiple CT scans during conformal radiotherapy. Radiother Oncol 1999;50:225–34.[Medline]
  10. Melian E, Mageras GS, Fuks Z, Leibel SA, Niehaus A, Lorant H, et al. Variation in prostate position quantitation and implications for three-dimensional conformal treatment planning. Int J Radiat Oncol Biol Phys 1997;38:73–81.[Medline]
  11. Kagawa K, Lee WR, Schultheiss TE, Hunt MA, Shaer AH, Hanks GE, et al. Initial clinical assessment of CT-MRI image fusion software in localization of the prostate for 3D conformal radiation therapy. Int J Radiat Oncol Biol Phys 1997;37:35–42.
  12. Rasch C, Barillot I, Reneijer P, Touw A, Van Herk M, Lebesque JV. Definition of the prostate in CT and MRI: a multi-observation study. Int J Radiat Oncol Biol Phys 1999;43:57–66.[Medline]
  13. Roach M 3rd, 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.[Medline]
  14. Quint LE, Van Erp JS, Bland PH, Mandell SH, et al. Carcinoma of the prostate: MR images obtained with body coils do not accurately reflect tumor volume. AJR 1991;156:511–6.[Abstract/Free Full Text]
  15. Sommer FG, Nghiem HV, Herfkens R, et al. Determining the volume of prostatic carcinoma: value of MR imaging with an external-array coil. AJR 1993;161:81–6.[Abstract/Free Full Text]
  16. Jager GJ, Ruijter ET, van de Kaa CA, et al. Local staging of prostate cancer with endorectal MR imaging: correlation wtih histopathology. AJR 1996;166:845–52.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Br. J. Radiol.Home page
A S N Jackson, S A Reinsberg, S A Sohaib, E M Charles-Edwards, S A Mangar, C P South, M O Leach, and D P Dearnaley
Distortion-corrected T2 weighted MRI: a novel approach to prostate radiotherapy planning
Br. J. Radiol., November 1, 2007; 80(959): 926 - 933.
[Abstract] [Full Text] [PDF]

Home page
 Br. J. Radiol.Home page
V S Khoo and D L Joon
New developments in MRI for target volume delineation in radiotherapy
Br. J. Radiol., September 1, 2006; 79(Special_Issue_1): S2 - S15.
[Abstract] [Full Text] [PDF]

Home page
 Br. J. Radiol.Home page
N Charnley, A Morgan, E Thomas, S Wilson, S Bacon, D Wilson, and D Bottomley
The use of CT-MR image registration to define target volumes in pelvic radiotherapy in the presence of bilateral hip replacements
Br. J. Radiol., July 1, 2005; 78(931): 634 - 636.
[Abstract] [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 Sannazzari, G L
Right arrow Articles by Guarneri, A
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Sannazzari, G L
Right arrow Articles by Guarneri, A

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