Int. J. Radiation Oncology Biol. Phys., Vol. 64, No. 3, pp.

892– 897, 2006

Copyright © 2006 Elsevier Inc.
Printed in the USA. All rights reserved
0360-3016/06/$–see front matter

doi:10.1016/j.ijrobp.2005.05.067

CLINICAL INVESTIGATION Brain

INTENSITY-MODULATED RADIOTHERAPY IN HIGH-GRADE GLIOMAS:

CLINICAL AND DOSIMETRIC RESULTS

ASHWATHA NARAYANA, M.D.,* JOSH YAMADA, M.D.,* SEAN BERRY, M.S.,† PRITI SHAH, B.S.,*
MARGIE HUNT, M.S.,† PHILIP H. GUTIN, M.D.,‡§ AND STEVEN A. LEIBEL, M.D.储
Departments of *Radiation Oncology, †Medical Physics, and ‡Surgery, Memorial Sloan-Kettering Cancer Center;
§
Department of Neurological Surgery, Weill Medical School and Cornell Medical Center, New York, NY;
and 储Stanford University Cancer Center, Stanford, CA

Purpose: To report preliminary clinical and dosimetric data from intensity-modulated radiotherapy (IMRT) for
malignant gliomas.
Methods and Materials: Fifty-eight consecutive high-grade gliomas were treated between January 2001 and
December 2003 with dynamic multileaf collimator IMRT, planned with the inverse approach. A dose of 59.4 – 60
Gy at 1.8 –2.0 Gy per fraction was delivered. A total of three to five noncoplanar beams were used to cover at least
95% of the target volume with the prescription isodose line. Glioblastoma accounted for 70% of the cases, and
anaplastic oligodendroglioma histology (pure or mixed) was seen in 15% of the cases. Surgery consisted of biopsy
only in 26% of the patients, and 80% received adjuvant chemotherapy.
Results: With a median follow-up of 24 months, 85% of the patients have relapsed. The median progression-free
survival time for anaplastic astrocytoma and glioblastoma histology was 5.6 and 2.5 months, respectively. The
overall survival time for anaplastic glioma and glioblastoma was 36 and 9 months, respectively. Ninety-six
percent of the recurrences were local. No Grade IV/V late neurologic toxicities were noted. A comparative
dosimetric analysis revealed that regardless of tumor location, IMRT did not significantly improve target
coverage compared with three-dimensional planning. However, IMRT resulted in a decreased maximum dose to
the spinal cord, optic nerves, and eye by 16%, 7%, and 15%, respectively, owing to its improved dose
conformality. The mean brainstem dose also decreased by 7%. Intensity-modulated radiotherapy delivered with
a limited number of beams did not result in an increased dose to the normal brain.
Conclusions: It is unlikely that IMRT will improve local control in high-grade gliomas without further dose
escalation compared with conventional radiotherapy. However, it might result in decreased late toxicities
associated with radiotherapy. © 2006 Elsevier Inc.

Intensity-modulated radiotherapy, Glioma, Radiotherapy.

INTRODUCTION The goal of treatment planning is to spare the critical

The combination of surgery, radiotherapy, and chemotherapy
represents the standard approach to the treatment of malignant
gliomas. However, survival remains poor. The median survival
time with conventional therapy for patients with glioblastoma
multiforme is 10 –12 months, with a 3-year survival rate of
6 – 8% (1). The median survival time for patients with anaplas-
tic astrocytoma is 36 months, and the 3-year survival rate is
approximately 50%. The pattern of recurrence in high-grade
gliomas is almost always local (2). The inability to deliver
adequate radiation to the target volume often has been cited as
the cause for local failure (3).
Radiation treatment planning for gliomas can be challenging.
structures while delivering the complete prescription dose to
the target volume. The tumor often lies in close proximity or
surrounds several radiosensitive normal tissues, including
the eyes, optic nerves, chiasm, brainstem, and spinal cord.
The tolerance of these tissues is less than the desired pre-
scription dose, and the associated morbidity can be severe.
Any effort to decrease the treatment toxicity might result in
compromised target coverage with conventional therapy,
thus increasing the risk of local recurrence.
Intensity-modulated radiotherapy (IMRT) is a more ad-
vanced form of three-dimensional conformal radiotherapy
(3D-CRT) that relies on advanced accelerator and multileaf

Reprint requests to: Ashwatha Narayana, M.D., Department of Presented at the 40th Annual Meeting of the American Society
Radiation Oncology, Memorial Sloan-Kettering Cancer Center, of Clinical Oncology, New Orleans, Louisiana, June 5– 8, 2004.
Room SM16, 1275 York Avenue, New York, NY 10021. Tel: (212) Received Jan 13, 2005, and in revised form April 15, 2005.
639-6773; Fax: (212) 639-2417; E-mail: narayana@mskcc.org Accepted for publication May 6, 2005.

892
 IMRT in high-grade gliomas ● A. NARAYANA et al. 893

collimator technology to deliver nonuniform beam intensi- immobilized with an Aquaplast face mask (WFR/Aquaplast Cor-
ties. This approach has been demonstrated to improve tumor poration, Wyckoff, NJ). Patients then underwent CT simulation
coverage while decreasing the dose to critical structures in with intravenous contrast, and 3-mm slice images were obtained
the head and neck and other sites (4, 5). At present there is from the vertex to approximately the level of C2. Postresection
magnetic resonance images with 3-mm axial fluid attenuation
very little clinical or dosimetric data regarding the dose
inversion recovery (FLAIR) sequence images obtained within 1
distribution and outcomes from IMRT in the treatment of week before simulation were registered with the CT images, with
malignant gliomas (6). This retrospective analysis presents the image registration software available on the CT simulator
the preliminary data on the potential benefits of IMRT in (AcQSim, Philips Medical Systems, Bothell, WA). The clinical
high-grade gliomas. target volume was defined as the gross disease observed in the
FLAIR sequence MRI series plus a 1.5-cm margin (7). The plan-
ning target volume (PTV) was determined by adding an additional
METHODS AND MATERIALS 0.5-cm margin to the clinical target volume to account for treat-
Fifty-eight consecutive patients with high-grade gliomas received ment uncertainties. The normal tissues delineated included the
radiotherapy with IMRT between January 2001 and December 2003 spinal cord, brainstem, optical structures, and “normal brain,”
at Memorial Sloan-Kettering Cancer Center (MSKCC). Institutional which were defined as all brain tissue outside of the PTV. The dose
review board approval was obtained for retrospective review of the limits that defined an acceptable plan included a maximum point
data. Patient characteristics are shown in Table 1. Glioblastoma dose of 4500 cGy to the spinal cord, 5400 cGy to the optic nerves
accounted for 70% of the cases. Anaplastic astrocytoma and ana- and the chiasm, 4000 cGy to the retina, and 5500 cGy to the
plastic oligodendroglioma histology (pure or mixed) constituted brainstem. An attempt was made to keep the lens dose as low as
15% each of the cases. Surgery was limited to biopsy alone in 26% possible. Acceptable target coverage was defined by a D95 (dose
of patients. Eighty percent of patients received adjuvant chemo- received by 95% of the PTV) of at least 95% of the prescription.
therapy. In all but 2 cases, radiotherapy was started within 1 month As in 3D-CRT, multiple noncoplanar beams were commonly used.
of surgical resection. An attempt was made to limit the number of treatment fields to
The technique of IMRT treatment planning for brain tumors has three to five, in consideration of both treatment complexity and the
been previously described by us (7). Simulation was performed volume of normal brain tissue irradiated to low dose levels. All
generally 2 to 3 weeks after surgical resection. All patients were planning was performed with the MSKCC in-house-developed
treatment-planning system, which relies on the inverse planning
method developed at our institution by Spirou and Chui (8). This
algorithm relies on a gradient minimization technique and a qua-
Table 1. Patient characteristics
dratic objective function. It supports the application of dose con-
Gender (n) straints for targets and both dose and dose–volume constraints for
Male 31 normal tissues. Treatment was delivered with dynamic multileaf
Female 27 collimation. The intensity profiles generated during optimization
Age (y) were converted into leaf motion with the method described by
Median 54 Spirou and Chui (8). All the patients were treated with 6-MV
Range 24–80 X-rays from Varian accelerators (Varian Medical Systems, Palo
Location (n) Alto, CA) equipped with multileaf collimators with either a 1-cm
Temporal 22 or 0.5-cm leaf width. The prescription dose for all patients was
Frontal 20
either 59.4 or 60 Gy, given in 1.8-Gy or 2.0-Gy fractions. The
Parietal 11
Occipital 2 average time from simulation to the start of radiotherapy was 7
Posterior fossa 3 days (range, 5–10 days).
Histology (n) A comparative dosimetric analysis was performed for 20 pa-
Glioblastoma 41 tients treated with IMRT. Conventional 3D-CRT plans were gen-
Anaplastic astrocytoma 9 erated retrospectively with the identical beam arrangement used
Anaplastic 4 for the IMRT, wedges, and static multileaf collimator apertures.
Oligodendroglioma Doses to the critical structures, PTV, and normal brain tissue were
Anaplastic mixed glioma 4 evaluated for both plans through the evaluation of dose–volume
Karnofsky performance status histograms and dose distributions. In addition to maximum normal
Median 80
tissue doses and the PTV D95, the maximum and mean PTV doses
Range 60–90
Radiation dose (Gy) and V18 Gy and V24 Gy (percentage of volume receiving at least
Median 59.4 18 or 24 Gy) for the normal brain were recorded. V18 Gy and V24
Range 59.4–60 Gy were chosen on the basis of neurotoxicity data available in
Surgical resection (n) pediatric populations receiving whole-brain radiotherapy.
Total/near total 27 The baseline evaluations included a complete medical history,
Partial 16 physical examination, determination of performance status, hema-
Biopsy 15 tology and clinical chemistry assessments, and gadolinium-en-
Chemotherapy (n) hanced MRI of the brain. Gadolinium-enhanced MRI was per-
Adjuvant/concomitant 46 formed 2– 4 weeks after the completion of radiotherapy and then
None 12
every 3 months during the first year and every 3 to 4 months
894 I. J. Radiation Oncology ● Biology ● Physics Volume 64, Number 3, 2006

Table 2. Progression-free survival

Grade Total 1-y DFS (%) 2-y DFS (%)

GBM (Gr IV) 41 0.0 0.0

All Gr III 17 21.8 10.8
AA 9 27.7 0.0
AO/mixed 8 18.8 18.8

Abbreviations: DFS ⫽ disease-free survival; AA ⫽ anaplastic

astrocytoma; AO ⫽ anaplastic oligodendroglioma; GBM ⫽
glioblastoma.

subsequently. Toxicity was graded according to the Radiation

Therapy Oncology Group grading system and the National Cancer
Institute Common Toxicity Criteria (version 3.0). Overall survival
was calculated from the time of diagnosis until death or last
follow-up, according to the Kaplan-Meier method.
RESULTS
With a median follow-up of 24 months (range, 12– 48
months), 49 of the 58 patients (85%) have relapsed. Forty-
seven (96%) of the recurrences were local. Of the remaining
2 patients, 1 recurred in a different lobe of the brain outside
the radiated region, whereas the other had a leptomeningeal
relapse.
Local control
Overall survival (OS) and progression-free survival
(PFS) data are shown in Tables 2 and 3 and Figures 1 and
2, respectively. The median PFS for Grade III and IV
tumors was 5.6 and 2.5 months, respectively. The median
OS for Grade III and IV tumors was 36 and 9 months,
respectively. The subset of patients with anaplastic oligo-
dendroglioma histology (pure or mixed) had a similar PFS
but a better OS compared with patients with anaplastic
astrocytomas.
Toxicity
Twenty-one patients experienced acute Grade I/II neuro-
toxicities. Four Grade III acute neurotoxicities were seen,
including seizures, altered consciousness, and worsening
confusion. Two Grade IV neurotoxicities occurred because
of brainstem bleed and intractable seizures. Six late Grade
I/II neurotoxicities were noted, all related to white matter
Fig. 1. Progression-free survival in patients with high-grade gli-
omas.
changes. Four Grade III late toxicities related to speech or
cognitive problems were seen. There were no Grade IV/V
late neurotoxicities. The rate of freedom from late neuro-
toxicity at 24 months was 85% (Fig 3).
Dosimetry
A comparison of the IMRT treatment plans with the more
conventional 3D-CRT plans was performed in 20 patients
representative of the overall patient population in terms of
tumor location, size, grade, volume, and the prescription
dose. It showed that regardless of tumor location, IMRT
does not lead to clinically significant differences in PTV
coverage. Of all the PTV parameters evaluated, a statisti-
cally significant difference was observed only for PTV
mean dose, with the mean dose from 3D-CRT being, on
average, approximately 1% higher than that with IMRT. No
differences were observed in PTV maximum dose, mean
dose, or D95 coverage (Table 4, Figs. 4 and 5). However,
the use of IMRT resulted in a significantly lower mean
brainstem dose (7%) and lower maximum doses to the
spinal cord, optic nerve, and eye by 16%, 7%, and 15%,
respectively (Table 4, Fig 6). The use of IMRT did not
result in an increase of the total dose received by the normal

Table 3. Overall survival

Grade Total 1-y OS (%) 2-y OS (%)

GBM (Gr IV) 41 30.0 0.0

All Gr III 17 86.3 61.6
AA 9 88.8 29.6
AO/mixed 8 87.5 87.5

Abbreviation: OS ⫽ overall survival. Other abbreviations as in

Table 2. Fig. 2. Overall survival in patients with high-grade gliomas.
 IMRT in high-grade gliomas ● A. NARAYANA et al. 895

Fig. 4. Comparison of intensity-modulated radiotherapy (IMRT)

and three-dimensional (3D) dosimetry in a patient with frontal
Fig. 3. Delayed Grade III neurologic event-free survival in high-
glioma.
grade gliomas.

thermore, IMRT has been shown to improve the dose dis-

brain (Fig 6). On the contrary, the volume of normal brain
irradiated was significantly less with IMRT. The volume
receiving at least 18 Gy and 24 Gy decreased by 7% and
8%, respectively.
DISCUSSION
There has been a dramatic improvement in radiotherapy
techniques over the last 2 decades. Improvements in dose
distribution and local control have been observed with 3D-
CRT as compared with conventional two-dimensional treat-
ment planning for prostate cancer and subsequently for
other tumor locations (9, 10). It has also been shown that the
morbidity of therapy decreased with the use of 3D-CRT
compared with conventional treatment planning (11). Fur-
tribution to the target volume, as well as to reduce the dose
to normal tissues when compared with 3D-CRT techniques
(4). The outcomes of clinical studies of IMRT for dose
escalation and critical organ sparing in prostate and head-
and-neck cancer sites are particularly encouraging (12, 13).
Our hypothesis was that a similar benefit might be achieved
in patients with high-grade gliomas.
Our results indicated that neither PFS nor OS improved
with the use of IMRT in high-grade gliomas. Why did the
improvement in local control noted at other sites not trans-
late to the brain? The reasons are several. Our PTV was
defined as a 1.5–2-cm margin around the FLAIR sequence–
defined tumor, which is the conventional volume definition
(7). Several studies have shown that gliomas can extend
more than 2 cm from the contrast-enhancing lesion (14). In

Table 4. Dosimetric data comparing 3D and IMRT plans

Structure 3D IMRT IMRT/3D p

PTV
Max dose 108.9 ⫾ 2.5 109.1 ⫾ 4.4 1.002 0.86
Min dose 67.3 ⫾ 24.1 65.8 ⫾ 23.2 0.98 0.33
Mean dose 103.6 ⫾ 2.1 102.3 ⫾ 1.7 0.99 0.04
Brainstem max dose 94.1 ⫾ 17.2 92.3 ⫾ 18.6 0.98 0.19
Brainstem mean dose 60.9 ⫾ 25.2 56.4 ⫾ 23.8 0.93 0.001
Spinal cord max dose 29.1 ⫾ 16.2 24.4 ⫾ 13.2 0.84 0.01
Optical structures
Chiasm max dose 81.4 ⫾ 24.8 78.9 ⫾ 24.0 0.97 0.10
Optic nerve max dose 53.9 ⫾ 37.5 49.9 ⫾ 36.5 0.93 0.02
Eye max dose 28.4 ⫾ 31.8 24.0 ⫾ 29.1 0.85 0.04
Lens mean dose 2.8 ⫾ 2.9 3.3 ⫾ 4.1 1.18 0.04
Normal brain*
Mean dose 41.9 ⫾ 12.2 38.9 ⫾ 10.6 0.93 0.0004
% volume receiving ⱖ18 Gy 56.2 ⫾ 18.0 52.0 ⫾ 17.4 0.93 0.01
% volume receiving ⱖ24 Gy 43.3 ⫾ 17.5 39.7 ⫾ 16.8 0.92 0.01

Abbreviations: 3D ⫽ three-dimensional conformal radiotherapy; IMRT ⫽ intensity-modulated

radiotherapy; PTV ⫽ planning target volume; Max ⫽ maximum; Min ⫽ minimum.
Values are percentages of total prescription dose, unless otherwise noted.
* “Normal brain” includes all the brain tissue, excluding the PTV.
896 I. J. Radiation Oncology ● Biology ● Physics
Fig. 5. Dose–volume histogram of planning target volume (PTV)
with three-dimensional (3D) and intensity-modulated radiotherapy
(IMRT) treatment.
addition, the radiation dose used in this study was only
59.4 – 60 Gy, which, although certainly adequate by con-
ventional standards, might be insufficient for long-term
control in a relatively radioresistant tumor like glioma. As in
conventional techniques, IMRT was used in this study to
deliver a uniform dose throughout the target volume. It is
clear that the density of tumor throughout the volume could
vary, and failure to give a higher dose to areas with highest
tumor density might have resulted in tumor recurrence. In
addition, the use of conventional fractionation of 1.8 –2.0
Gy per fraction might not be the optimal fractionation
regimen for these aggressive tumors.
Our dosimetric analysis confirmed that IMRT did not
lead to significant improvements in target coverage when
Fig. 6. Dose–volume histogram of critical structures and normal brain tissue with three-dimensional (3D) and
intensity-modulated radiotherapy (IMRT) treatment volume (%).
Volume 64, Number 3, 2006
compared with conventional 3D-CRT planning. For many
gliomas, target coverage and dose uniformity are excellent
with standard 3D techniques, owing to the nearly spherical
or cylindrical shape of the lesion. Therefore, it was not
surprising that significant further improvement was not ob-
served with IMRT. Target coverage and uniformity im-
provements with IMRT have been primarily reported in
sites such as the head and neck or prostate (4, 12, 13), where
the target is concave, surrounding normal tissues with dose
limits much less than that of the tumor. Gliomas can be
highly irregular but typically exhibit few concavities. When
concavities do exist, such as when the tumor surrounds the
chiasm, the required dose gradient between tumor and nor-
mal tissue is often less than that observed in other sites. As
a result, very good target coverage is often achieved with
3D planning. However, as we escalate the prescription dose
for these tumors, even if only to areas of suspected high
tumor density, the benefit of IMRT might increase because
steeper dose gradients and more concave dose distributions
will be necessary. Indeed, our group and others have already
shown the clear benefit of IMRT in concave brain lesions
(15, 16).
The use of IMRT did result in a moderate decrease of
dose to the critical structures in the brain. This might be
responsible for the decreased acute and late neurotoxicity
we observed compared with historical results. Contrary to
the conventional thinking (17), the use of IMRT with three
to five beams did not result in increased dose to the normal
brain. In fact, the volume of the brain receiving 18 and 24
 IMRT in high-grade gliomas
Gy decreased with the use of IMRT, owing to optimal beam
modulation. This has obvious implications for reducing
neurocognitive deficits in long-term survivors after radio-
therapy. Decreasing the toxicity would also enable possible
dose escalation, although the potential results of this ap-
proach are uncertain (18).
In summary, it is unlikely that the use of IMRT with
conventional doses and volume definition will improve
local control in high-grade gliomas. However, it can
REFERENCES
1. Medical Research Council Brain Tumor Working Party. Ran-
domized trial of procarbazine, lomustine, and vincristine in the
adjuvant treatment of high grade astrocytoma: A medical
research council trial. J Clin Oncol 2001;19:509 –518.
2. Wallner KE. Radiation therapy treatment planning for malig-
nant astrocytomas. Semin Radiat Oncol 1991;1:17–25.
3. Hochberg FH, Pruitt A. Assumptions in the radiotherapy of
glioblastomas. 9 Neurology 1980;30:907–911.
4. Hunt MA, Zelefsky MJ, Wolden S, et al. Treatment planning
and delivery of intensity-modulated radiation therapy for pri-
mary nasopharynx cancer. Int J Radiat Oncol Biol Phys 2001;
49:623– 632.
5. Huang E, Teh BS, Strother DR, et al. Intensity-modulated
radiation therapy for pediatric medulloblastoma: Early report
on the reduction of ototoxicity. Int J Radiat Oncol Biol Phys
2002;52:599 – 605.
6. Shaw EG, Stieber V, Tatter S, et al. A phase I dose escalating
study of intensity modulated radiation therapy (IMRT) for the
treatment of glioblastoma multiforme (GBM) [abstract]. Int J
Radiat Oncol Biol Phys 2002;54:206.
7. Narayana A, Leibel SA. Primary and metastatic brain tumors
in adults. In: Leibel S, Philips TL, editors. Textbook of radi-
ation oncology. 2nd ed. Philadelphia: Saunders; 2004. 471–
495.
8. Spirou SV, Chui CS. A gradient inverse planning algorithm
with dose-volume constraints. Med Phys 1998;25:321–333.
9. Zelefsky MJ, Marion C, Fuks Z, et al. Improved biochemical
disease-free survival of men younger than 60 years with
prostate cancer treated with high dose conformal external
beam radiotherapy. J Urol 2003;170:1828 –1832.
10. Leibel SA, Kutcher GJ, Harrison LB, et al. Improved dose
● A. NARAYANA et al. 897
result in decreased acute and late toxicities associated
with radiotherapy. There is a need to better define the
target volume and then to perform selective dose inten-
sification according to the tumor density to optimize the
radiation technique. A prospective phase I/II dose-esca-
lation clinical trial with the use of multivoxel magnetic
resonance spectroscopy to define the target volume and
then use IMRT dose painting on the basis of tumor cell
density is planned.
distributions for 3D conformal boost treatments in carcinoma
of the nasopharynx. Int J Radiat Oncol Biol Phys 1991;20:
823– 833.
11. Skwarchuk MW, Jackson A, Zelefsky MJ, et al. Late rectal
toxicity after conformal radiotherapy of prostate cancer (I):
Multivariate analysis and dose-response. Int J Radiat Oncol
Biol Phys 2000;47:103–113.
12. Zelefsky MJ, Fuks Z, Hunt M, et al. High-dose intensity
modulated radiation therapy for prostate cancer: Early toxicity
and biochemical outcome in 772 patients. Int J Radiat Oncol
Biol Phys 2002;53:1111–1116.
13. Lee N, Xia P, Fischbein NJ, et al. Intensity-modulated radia-
tion therapy for head-and-neck cancer: The UCSF experience
focusing on target volume delineation. Int J Radiat Oncol Biol
Phys 2003;57:49 – 60.
14. Chamberlain MC, Murovic JA, Levin VA. Absence of con-
trast enhancement on CT brain scans of patients with supra-
tentorial malignant gliomas. Neurology. 1988;38:1371–1374.
15. Narayana A, Yenice KM, Chang J, et al. Intensity-modulated
stereotactic radiotherapy for the treatment of skull base
meningioma—a novel approach. Skull Base 2004;14:18–19.
16. Baumert BH, Norton IA, Davis JB. Intensity-modulated ste-
reotactic radiotherapy vs. stereotactic conformal radiotherapy
for the treatment of meningioma located predominantly in the
skull base. Int J Radiat Oncol Biol Phys 2003;57:580 –592.
17. Glatstein E. Intensity-modulated radiation therapy: The in-
verse, the converse, and the perverse. Semin Radiat Oncol
2002;12:272–281.
18. Chan JL, Lee SW, Fraass BA, et al. Survival and failure
pattern of high-grade gliomas after three-dimensional confor-
mal radiotherapy. J Clin Oncol 2002;20:1635–1642.