About TCRT

TCRT has been on continuous publication since 2002. It is covered currently by all the major data systems such as Medline, PubMed, Web of Science, Thomson's ISI and SCI and Scopus.

The 2012 Impact factor for TCRT is 1.943

TCRT Open Access
TCRT seeks original articles
Cancer Watch
Open Access
Magnetic Resonance

The Impact of Mid-Treatment MRI on Defining Boost Volumes in the Radiation Treatment of Glioblastoma Multiforme (p. 303-308)

Radiation therapy is a central modality in the treatment of glioblastoma multiforme (GBM). Integral to adequate radiation therapy delivery is the appropriate determination of tumor volume and extent at the time treatment is being delivered. As a matter of routine practice, radiation therapy treatment fields are designed based on tumor volumes evident on pre-operative or immediate post-operative MRIs; another MRI is generally not obtained for planning boost fields. In some instances the time interval from surgery to radiotherapy initiation is up to 5 weeks and the boost or ?cone-down phase? commences 4-5 weeks later. The contrast enhanced T1 MRI may not be a totally reliable indicator of active tumor, especially in regions where such blood-brain barrier breakdown has not occurred. Moreover, these volumes may change during the course of treatment. This may lead to a geographic miss when mid-treatment boost volumes are designed based on a pre-radiotherapy MRI. The goal of this study is to examine how a mid-treatment MRI impacts the delineation and definition of the boost volume in GBM patients in comparison to the pre-treatment MRI scan, particularly when the tumor-specific agent Motexafin-Gadolinium is used.

Key words: Motexafin gadolinium, Target volumes, Glioblastoma treatment planning.


Despite advances in the diagnosis and treatment of glioblastoma multiforme (GBM), the prognosis of patients with this disease remains poor: median survival times are approximately 9-12 months (1). Standard treatment for GBM involves surgical resection, and external beam radiation therapy, with or without systemic chemotherapy. Radiation therapy planning for GBM is typically based either on pre-operative MRI scans (usually for patients undergoing biopsy only) or immediate post-operative MRI scans (usually for patients undergoing tumor resection).
Although whole brain radiation therapy was used in the past, this practice has been abandoned in the United States. A common U.S. standard for defining radiation fields follows RTOG guidelines. A larger volume consisting of the T2 MRI abnormality plus a 2.0-cm margin is treated to 46Gy. This is followed by a 14Gy ?cone down boost? to 60Gy, which includes the T1 enhancing abnormality plus a 2.5-cm margin, with the rationale that microscopic disease could be treated to lower doses, but gross residual bulky disease requires a higher dose. The size of the margins is based on the following lines of reasoning: i) serial biopsies of patients undergoing craniotomy for malignant gliomas reveal tumor cells more than 3-cm from the CT contrast enhancing margin (2, 3, 4); and ii) approximately 80% of relapses occur within a 2-cm margin from the original tumor location (5, 6, 7). This strategy permits the delivery of ?graded? radiation doses without treating very large volumes of brain to the highest doses.

Integral to adequate radiation therapy is the appropriate determination of tumor volume and extent at the time treatment is being delivered. In some instances the time interval from surgery to radiotherapy initiation is up to 5 weeks (as in RTOG protocols) and the boost or ?cone-down phase? commences 4-5 weeks later. During this greater than 9 week window, considerable change in tumor volume could potentially occur. As a matter of routine practice, another MRI is generally not obtained for planning boost fields; the original MRI is most commonly used to identify the T1 abnormality for boost field design. In general, the gadolinium-enhancing lesion, as seen on T1 weighted MRI, reflects regions of blood-brain barrier breakdown (8). Contrast enhanced MRI may not be a totally reliable indicator of active tumor, especially in regions where such blood-brain barrier breakdown has not occurred. Recent methyl-[11C]-L-methionine (MET) and 3?-deoxy-3-[18F]fluoro-L-thymidine (FLT) PET studies provide evidence for this (9). These studies often show radiotracer uptake in regions where MRI contrast is not taken up. Moreover, MR spectroscopy (MRS) data suggest that standard T1 enhanced MRI may not adequately represent the volumes of metabolically active tumor (10).

Motexafin gadolinium (MGd) belongs to a class of compounds known as texaphyrins, which are expanded porphyrin macromolecules capable of forming complexes with large cations such as lanthanides. In tumor cells, MGd generates reactive oxygen species (ROS) by accepting electrons from cellular reducing metabolites and transferring them to oxygen. This results in disruption of cellular metabolism and predisposes the tumor cell to apoptosis when an additional stressor, such as radiation or chemotherapy, is applied (11, 12). Moreover, elegant tumor-specificity and intracellular localization of MGd has been demonstrated by DeStasio et al. and Woodburn et al. (11, 13). The precise mechanisms for this tumor specificity are not well identified. In other clinical trials, such as the one reported by Carde et al. (14), MGd has been shown not only to be tumor specific, but also long-lived with tumor retention periods of weeks to months. Data from UCLA by Ford et al., suggest that this retention and consequential ?enhancement? on T1 MRI is a function of the dose of MGd (15).

Currently, radiation therapy planning for GBM is performed on a contrast-enhanced pre-treatment MRI (pre-operative and/or post-operative). Our goal was to examine how a mid-treatment MRI impacts the delineation and definition of the boost volume in GBM patients in comparison to the pre-treatment MRI scan. We also wanted to investigate whether the tumor-specific, long-lived enhancement from MGd could potentially influence this definition of boost volumes. In this study we quantified how the treatment volume changed at the midway point during a course of radiation, compared to the contrast enhanced baseline (post-operative) MRI scan.

Materials and Methods

Following Institutional Review Board approval and individual informed consent (which included a planned analysis of the mid-treatment MRI scans) 25 newly diagnosed GBM patients were enrolled in protocol PCYC-0206 between March 30, 2001 and August 3, 2001. All patients underwent a post-operative contrast enhanced MRI scan which was used to define the radiation fields, specifically the initial field. This post-operative scan was performed as soon as possible after surgery. Ideally it was obtained within 24 hours after surgery, but could be obtained anytime within the 4 weeks prior to the first radiation treatment. As per protocol, the definition of the initial and boost fields was follows: i) the initial target volume included the T2 abnormality demonstrated on the postoperative MRI plus a 2.0-cm margin; and ii) the boost volume included the enhancing volume on the mid-treatment MRI scan plus a 2.0-cm margin. The initial field was treated to 46 Gy in 23 fractions. In the seven days prior to the 24th fraction a second (mid-treatment) MRI was obtained to define the boost volume. The boost volume was treated to a final dose of 60 Gy. MGd was given at a dose of 5mg/kg/dose, for a total of 22 administrations: a daily administration schedule (Monday-Friday) was used during weeks 1 and 2 of radiation therapy, while in weeks 3-6, MGd was given on a Monday, Wednesday, and Friday schedule.

For the specific purpose of this analysis of the impact of the mid-treatment MRI scan on boost definition, relative to the initial post-operative contrast enhanced MRI scan, our focus was on reliable definition of the boost volume; the boost volume is also referred to as the gross tumor volume (GTV). The GTV was independently contoured by three investigators. We defined GTV as the contrast-enhancing abnormality on the T1 MR sequence. This definition is consistent with that employed in most RTOG GBM protocols. The Pinnacle3 version 6.0i system was used for digitizing and comparing post-operative and mid-treatment MRI volumes (GTVpre and GTVmid respectively). The GTVmid volumes were defined from the contrast-enhanced mid-treatment MRI scans, using the T1 images. It was difficult to sort out the specific cause of enhancement on the post-contrast MRI images in this clinical situation since both standard Gadolinium containing contrast and MGd would contribute to the enhancing pattern. Although it was possible to identify contrast-enhancing regions that did not show MRI ?enhancement? with MGd, such areas were, in general, very small and not easily quantifiable. It was also difficult to identify areas that ?enhanced? solely as a consequence of MGd and not the MRI contrast agent, as the MGd effect is long lived and could not be subtracted out (17). Therefore, in all cases, the total enhancing volume on the post-contrast MRI scan was contoured to define the mid-treatment GTV.

After GTVpre was digitized into the system, GTVmid was appropriately scaled and overlapped on GTVpre by fusion of complementary anatomy. Three-dimensional volumetric expansions of 1.0 and 2.0-cm were carried out on GTVpre and designated CTV1 and CTV2 respectively. GTVmid was then superimposed on CTV1 and CTV2, and the degree of geographic miss was evaluated using pre-defined criteria (Figure 1A). A complete geographic miss was defined as any component of GTVmid extending beyond CTV2. If any component of GTVmid fell beyond CTV1, but not outside CTV2, the scenario was scored as a partial geographic miss. The situation was scored clinically acceptable if GTVmid was completely bounded by CTV 1 (Figure 1).

Figure 1A: GTVmid represented by the middle dashed line, CTV1 by the inner thin solid line, and CTV2 by the outer thin solid line, depict geographic miss criteria. (A) Demonstrates a ?clinically acceptable? scenario, GTVmid completely bounded by CTV1. (B) Demonstrates a ?partial geographic miss? whereby a component of GTVmid extends beyond CTV1, but not CTV2. (C) Demonstrates a ?complete geographic miss?, GTVmid extends beyond CTV2.

Figure 1B: 3D rendition of complete ?geographic miss.? This demonstrates the GTVmid extending beyond CTV2 in three dimensional space.

To verify the accuracy of our fusion algorithm we evaluated the precision of overlapping fixed anatomy of mid-treatment MRIs on post-operative MRI sets. Thus, normal anatomic structures (i.e., brain stem, cranial vault) were digitized, scaled, and overlapped in the same fashion as described above. Mid-treatment MRI anatomic structures were fused on the post-operative MRI anatomic structures. A 1.0-cm volumetric expansion was carried out on the digitized post-operative MRI anatomic structures. The data sets were then evaluated to determine the magnitude of error in the fusion process by examining the distance of the mid-treatment MRI fixed anatomy from the post-operative MRI expansion margin. The fusion program permitted the user to adjust the fusion until an accurate fit was achieved, defined as a fit-error of less than half the MRI slice thickness, which in most cases was on the order of 2-mm or less, with the extreme outlier being 5-mm. A description of the exact fusion methodology is beyond the scope of this paper and is published elsewhere (15).


Of the 25 patients enrolled in PCYC-0206, a multi-institutional trial, 24 post-operative as well as repeat mid-course enhanced and unenhanced hard-copy MRI sets were available for our review; data were not available in digital format. Nine sets of MRIs were excluded for the following reasons: i) five were excluded because of a large disparity in magnification and head angulation between post-operative and mid-treatment MRI images (not software correctable); and ii) four were poor quality scans (generally copies with originals not retrievable). Thus, 15 post-operative and mid-treatment MRIs were available for evaluation.

Of the MRI sets evaluated 12/15 (80%) were scored as having any degree of geographic miss. Of all geographic misses, 8/15 (53%) were scored as partial and 4/15 (27%) were scored as complete geographic misses. Three of the 15 (20%) MRI sets represented clinically acceptable situations in which GTVmid was within CTV1 (Table I). It is important to note that in this trial boost volumes were designed and patients were treated based on mid-treatment MRI volumes. Therefore, these geographic misses represent potential, rather than true misses.

Of the 15 MRI sets, all had mid-treatment contrast enhanced images and five had both contrast and non-contrast mid-treatment MRI scans available for our review. We further evaluated these five pre- and post-contrast mid-treatment MRI scans. ?Enhancing? tumor could be identified in all five non-contrast scans verifying the previous observation that MGd is MR detectable. In these five MRI sets, we fused the conventional contrast enhanced tumor volume from the post-operative MRIs with the ?enhancing? tumor volume on the non-contrast mid-treatment MRIs. This fusion was performed with the notion that these image sets would provide pure data for MGd accumulation without the confounding variable of MRI contrast enhancement (Figure 2). The fusion was then examined to see if the score changed with respect to the initial assigned score. The scores assigned by MGd alone changed in two of five cases. In one of these cases the score changed from partial geographic miss (as defined by standard MRI contrast) to clinically acceptable (as defined by MGd alone), suggesting that in this case, MRI contrast defined a larger GTVmid than MGd alone. In the second case the score changed from complete geographic miss (as defined by standard MRI contrast) to partial geographic miss (as defined by MGd alone). These observations suggest that in all five cases where non-contrast scans were available, ?tumor enhancement? secondary to MGd was visualized. In 3/5 cases, administration of conventional MRI contrast did not change the volume of MGd ?enhancement,? but in 2/5 cases a volumetric change secondary to conventional MR contrast, larger than the effect of MGd, was observed.

Figure 2: Illustrated here are a contrast enhanced post-operative MRI (A), a contrast enhanced mid-treatment MRI (B), as well as a non-contrast mid-treatment MRI (C) of a patient evaluated in this study. The tumor volumes on the mid-treatment MRIs are clearly different than on the post-operative MRI. Moreover, tumor volume is visible on the mid-treatment non-contrast MRI four weeks after MGd injection.

To test whether our observations were due to error in the fusion process, several anatomic structures were digitized, overlapped, and fused. The volume bounded by the digitized mid-treatment MRI anatomic structures all fell within 2 to 5-mm of the volume defined by post-operative MRI anatomy. Given that we used a minimum expansion margin of 1-cm on the GTV for treatment planning purposes, we defined half of this, or 5-mm as acceptable error in the fusion process. With the use of digital source images, fusion errors can be minimized further; however, we only had hard-copy MRI images available to us and the error ≤ 5-mm is a reflection of this.


The GTVmid, defined as the enhancing abnormality on mid-treatment MRIs, changed 80% of the time when compared to post-operative MRI GTVs. Therefore, boost radiation therapy fields designed using post-operative MRI scans alone could potentially lead to scenarios where significant geographic miss occurs. This phenomenon of geographic miss could be the result of change in the morphology of the tumor, which would be observed if the tumor demonstrated growth during treatment. Alternatively, geographic miss could be a partial effect of MGd. Like the porphyrins, MGd is selectively taken up by tumor cells and not surrounding tissues (11, 13, 14, 15. 16). This tumor specificity would potentially allow a greater and more accurate extent of tumor to be visualized on MRI scans after MGd administration; resulting in an alteration of the tumor volume on the mid-treatment scans. Of course, the true impact of MGd alone cannot be ?subtracted? out of a post-contrast MRI scan.

The geographic misses observed could also be the result of several non-tumor associated factors resulting in MRI signal changes. Post-operative changes occur, most often, in the second week after surgery. These changes result in linear enhancement of varying intensity and thickness around the resection cavity and usually persist for two months (17). In approximately 10% of patients there is gyriform enhancement of adjacent brain parenchyma, which may be related to ischemia (17). This type of post-operative enhancement is more commonly observed after surgery involving the temporal lobes. In addition, several acute and sub-acute reactions after radiation therapy have been described clinically. However, neither group of reactions has been imaged extensively (18). Acute reactions are a result of vasogenic edema of early onset and may appear during the course of radiation therapy or shortly thereafter. These manifest as a transient encephalopathy and are not associated with specific MRI or CT findings (19). Sub-acute reactions (early delayed radiation complications) occur a few weeks to 3 months after radiation therapy. Findings on MRI include transient enhancement involving the white-matter, basal ganglia, and cerebral peduncles. This condition is usually transient and responsive to steroids.

In the post-operative context the resection bed can change due to inflammatory processes, scarring, hemorrhage, and the effects of radiation, which by altering blood-brain-barrier permeability, may cause changes in MRI enhancement patterns (8, 17, 19). It can be challenging to differentiate these changes from tumor progression. Often, enhancement as a result of tumor progression is irregular, nodular, or mass like (17). The GTVmid enhancement characteristics on the MRI sets exhibiting a geographic miss were in this series, for the most part, non-linear, irregular, nodular, and near or adjacent to the resection cavity. There was no evidence of gyriform enhancement of adjacent brain parenchyma, or enhancement of basal ganglia and cerebral peduncles; white matter changes were evident, but usually in proximity to the resection cavity. These findings suggest that the changes associated with geographic miss were likely reflective of tumor.

The observations made in this study do not conclusively identify the source of the change in GTV, but do point to the potential value of a mid-treatment MRI scan in designing a boost fields. A thorough definition of the value of such mid-treatment imaging can of course, only come from further larger studies to characterize and quantify the magnitude of this observation.

  1. Leibel, S. A., Scott, C. B., Loeffler, J. S. Contemporary Approaches to the Treatment of Malignant Gliomas with Radiation Therapy. Seminars in Oncology 21, 198-219 (1994).
  2. Kelly, P. J., Daumas-Duport, C., Scheithauer, B. W., et al. Stereotactic Histologic Correlations of Computed Tomography and Magnetic Resonance Imaging-defined Abnormalities in Patients with Glial Neoplasms. Mayo Clinic Proceedings 62, 450-459 (1987).
  3. Masukado, Y. M. C., Kernohan, J. W. The Growth of Glioblastoma Multiforme (Astrocytoma Grades 3 and 4) in Neurosurgical Practice. J. Neurosurg. 18, 636-644 (1961).
  4. Burger, P. C., Heinz, E. R., Shibata, T. et al. Topographic Anatomy and CT Correlations in the Untreated Glioblastoma Multiforme. Journal of Neurosurgery 68, 698-704 (1988).
  5. Hochberg, F. H., Pruitt, A. Assumptions in the Radiotherapy of Glioblastoma. Neurology.
  6. Oppitz, U., Maessen, D., Zunterer, H. et al. 3D-Recurrence-Patterns of Glioblastomas After CT-planned Postoperative Irradiation. Radiotherapy & Oncology 53, 53-57 (1999).
  7. Wallner, K. E., Galicich, J. H., Krol, G. et al. Patterns of Failure Following Treatment for Glioblastoma Multiforme and Anaplastic Astrocytoma. International Journal of Radiation Oncology, Biology, Physics 16, 1405-1409 (1989).
  8. Holodny, A. I., Nusbaum, A. O., Festa, S. et al. Correlation Between the Degree of Contrast Enhancement and the Volume of Peritumoral Edema in Meningiomas and Malignant Gliomas. Neuroradiology 41, 820-825 (1999).
  9. Jacobs, A. H., Winkler, A., Dittmar, C. et al. Molecular and Functional Imaging Technology for the Development of Efficient Treatment Strategies for Gliomas. Technology in Cancer Research & Treatment 1, 187-204 (2002).
  10. Pirzkall, A., McKnight, T. R., Graves, E. E. et al. MR-spectroscopy Guided Target Delineation for High-grade Gliomas. International Journal of Radiation Oncology, Biology, Physics 50, 915-928 (2001).
  11. Miller, R. A., Woodburn, K., Fan, Q. et al. In Vivo Animal Studies with Gadolinium (III) Texaphyrin as a Radiation Enhancer. International Journal of Radiation Oncology, Biology, Physics. 45, 981-989 (1999).
  12. Young, S. W., Qing, F., Harriman, A. et al. Gadolinium(III) Texaphyrin: A Tumor Selective Radiation Sensitizer that is Detectable by MRI. [Erratum appears in Proc Natl Acad Sci USA 96, 2569 (1999)]. Proceedings of the National Academy of Sciences of the United States of America 93, 6610-6615 (1996).
  13. De Stasio, G., Casalbore, P., Pallini, R. et al. Gadolinium in Human Glioblastoma Cells for Gadolinium Neutron Capture Therapy. Cancer Research 61, 4272-4277 (2001).
  14. Carde, P., Timmerman, R., Mehta, M. P. et al. Multicenter Phase Ib/II Trial of the Radiation Enhancer Motexafin Gadolinium in Patients with Brain Metastases. Journal of Clinical Oncology 19, 2074-2083 (2001).
  15. Ford, J. M. E. J., Alger, J. R., Earle, J., Miles, D., Phan, S. A Phase I Dose Escalation Trial of Motexafin Gadolinium (Gadolinium Texaphyrin, XcytrinR) as a Radiation Sensitizer in Patients with Newly Diagnosed Glioblastoma Multiforme [Abstract]. Int. J. Radiat. Oncol. Biol. Phys. 51, 205 (2001).
  16. Rosenthal, D. I., Nurenberg, P., Becerra, C. R. et al. A Phase I Single-dose Trial of Gadolinium Texaphyrin (Gd-Tex), A Tumor Selective Radiation Sensitizer Detectable by Magnetic Resonance Imaging. Clinical Cancer Research 5, 739-745 (1999).
  17. Albert, F. K., Forsting, M., Sartor, K. et al. Early Postoperative Magnetic Resonance Imaging After Resection of Malignant Glioma: Objective Evaluation of Residual Tumor and its Influence on Regrowth and Prognosis. [See Comment]. Neurosurgery 34, 45-60; discussion 60-41 (1994).
  18. Rabin, B. M., Meyer, J. R., Berlin, J. W. et al. Radiation-induced Changes in the Central Nervous System and Head and Neck. Radiographics 16, 1055-1072 (1996).
  19. Rowley, H. A. D. W. Iatrogenic White Matter Diseases. Neuroimaging Clinics of North America 3, 379-404 (1993).

TCRT June 2004

category image
Volume 3
No. 3 (p 229-308)
June 2004
ISSN 1533-0338

Rafael Manon, M.D.1*
Susanta Hui, Ph.D.1
Prakash Chinnaiyan, M.D.1
John Suh, M.D.2
Eric Chang, M.D.3
Robert Timmerman, M.D.4
See Phan, M.D.5
Rupak Das, Ph.D.1
Minesh Mehta, M.D.1

1University of Wisconsin
Department of Radiation Oncology
K4/B100, 600 Highland Ave
Madison, WI 53792 USA
2Cleveland Clinic Foundation
Dept. Radiation Oncology Desk T28
9500 Euclid Ave
Cleveland, OH 44195 USA
3MD Arderson Cancer Center
1515 Holcombe Blvd., Box 97
Houston, TX 77030 USA
4Indiana University Medical Center
535 Barnhill Dr. RT 041
Indianapolis, IN 46202 USA
5Pharmacyclics Inc.
995 East Arques Ave
Sunnyvale, CA 94085-4521 USA

Open Access Article
The authors, the publisher, and the right holders grant the right to use, reproduce, and disseminate the work in digital form to all users.

Click here to download PDF

Related Articles