ARTICLE IN PRESS

Advances in Radiation Oncology (2025) 000, 101778

Scientific Article

Clinical Outcomes of Fractionated Radiation

Therapy for Spinal Intramedullary Arteriovenous
Malformations: A Single Institutional
Retrospective Review
Ruja Vichitvejpaisal, MD,a Kentaro Nishioka, MD, PhD,a,*
Toshiya Osanai, MD, PhD,b Motoyuki Iwasaki, MD, PhD,b
Takashi Mori, MD, PhD,a Yusuke Uchinami, MD, PhD,a
Tomohiko Miyazaki, MD,a Fuki Koizumi, MD,a Takayuki Hashimoto, MD, PhD,a
Miki Fujimura, MD, PhD,b and Hidefumi Aoyama, MD, PhDa
a
Department of Radiation Oncology, Hokkaido University Hospital, Hokkaido, Japan; and bDepartment of Neurosurgery,
Hokkaido University Hospital, Hokkaido, Japan

Received 26 December 2024; accepted 24 March 2025

Purpose: Intramedullary spinal arteriovenous malformations (im-sAVMs) have a bleeding risk of up to 4% to 10% annually. Given the
rarity of im-sAVMs and the lack of established evidence regarding the efficacy of radiation therapy, we retrospectively reviewed the
treatment outcomes of fractionated radiation therapy (fRT) administered for im-sAVMs at our institution.
Methods and Materials: We examined patients of im-sAVMs treated with fRT at Hokkaido University Hospital between January
2003 and December 2023. The incidences of post-fRT bleeding from the AVMs, nidus occlusion/reduction, and late radiation-induced
complications were assessed.
Results: Of the 25 patients, we analyzed 21 patients, excluding 4 patients with a follow-up period of less than 180 days. Prior to
radiation therapy, 10 patients underwent surgical occlusion of feeding vessels because of coexisting arteriovenous fistulas, 2 and 3
patients underwent surgical occlusion and endovascular embolization for arteriovenous malformation, respectively. The radiation dose
was 20 Gy in 4 fractions for 19 patients and 24 Gy in 4 fractions for 2 patients. The follow-up period ranged from 12 to 236 months
(median, 56.4 months). Complete nidus occlusion was observed in 2 patients (9.5%), whereas 8 patients (38.1%) showed partial
reduction with residual nidus, and 10 patients (47.6%) showed no reduction. One patient showed slight nidus enlargement 12 months
after fRT, and another patient developed an aneurysm formation in the feeding artery at the margin of the irradiation field 57 months
after fRT. No bleeding and late radiation-induced complications were observed during the follow-up.
Conclusions: Fractionated radiation therapy at doses of 20 to 24 Gy in 4 fractions for im-sAVMs was shown to be safely administered
and may be effective in reducing the risk of bleeding.
© 2025 The Author(s). Published by Elsevier Inc. on behalf of American Society for Radiation Oncology. This is an open access
article under the CC BY-NC-ND license ([Link]

Sources of support: This work had no specific funding.

Research data are stored in an institutional repository and will be shared upon request to the corresponding author.
*Corresponding author: Kentaro Nishioka, MD, PhD; Email: [Link]@[Link]
[Link]
2452-1094/© 2025 The Author(s). Published by Elsevier Inc. on behalf of American Society for Radiation Oncology. This is an open access article under
the CC BY-NC-ND license ([Link]

Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.
 ARTICLE IN PRESS
2 R. Vichitvejpaisal et al Advances in Radiation Oncology: Month 2025

Introduction C) Perimedullary AVF: The arteriovenous shunt is

located on the surface of the spinal cord. If the shunt
Spinal arteriovenous malformations (sAVMs) are rare is positioned at the cervical level, subarachnoid or
vascular malformation with an unknown cause and natu- intramedullary hemorrhage often occurs. In contrast,
ral histories. The symptoms include bleeding (4% per year symptoms from increased venous pressure are more
and can increase to 10% per year after first bleeding1), common when the shunt is located at the thoracic
compressive myelopathy, venous hypertension, and steal level.
phenomenon. This type of vascular abnormality could D) Intramedullary sAVMs (im-sAVMs): Primarily found
cause an inferior prognosis if left untreated. The patients in younger patients, this type of arteriovenous malfor-
might suffer from acute, subacute, and chronic spinal mation involves the shunt or nidus being located
cord dysfunction, leading to severe permanent neurologic within the spinal cord, which can cause intramedul-
deterioration such as myelopathy and radiculopathy.2,3 lary hemorrhage.
Generally, there are 4 types of sAVMs, classified based
on the location of the arteriovenous shunt,4,5 as described The mainstay of treatment for sAVMs types A to C is
below (Fig. 1): microsurgery and/or endovascular embolization. How-
ever, these treatment options are not suitable for every
patient, particularly those with im-sAVMs, because of the
A) Dural arteriovenous fistula (AVF): Most common in risk of neurologic deficits, such as spinal infarction.
middle-aged male patients, this type of shunt is Bostr€om et al6 reported a 20% rate of neurologic worsen-
located on the dura surface near the intervertebral ing following microsurgery, with 5% of cases resulting in
foramen and refluxes backward into the vein on the permanent neurologic deterioration. Similarly, Velat et al7
surface of the spinal cord. Symptoms usually arise reported a surgical morbidity rate of 5% after microsur-
from increased venous pressure, leading to progres- gery.
sive flaccid paraplegia. As a result, fractionated radiation therapy (fRT) is con-
B) Epidural AVF: In this type, the arteriovenous shunt is sidered an excellent noninvasive alternative for treating
located outside the dura but does not reflux inward as im-sAVMs when microsurgery or endovascular emboliza-
in the dural AVF type. Instead, the shunt flows tion is not feasible. The radiation therapy technique has
directly into the epidural vein. This type is usually advanced over the past few decades, enabling radiation
asymptomatic but can cause symptoms because of oncologists to treat patients more accurately and effec-
nerve root compression. tively. At Hokkaido University Hospital, we have treated

Figure 1 Classification of spinal arteriovenous malformations based on the location of the arteriovenous shunt (A) dural arte-
riovenous fistula (AVF); (B) epidural AVF; (C) perimedullary AVF; and (D) intramedullary arteriovenous malformation.

Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.
 ARTICLE IN PRESS
Advances in Radiation Oncology: Month 2025 Outcome of fRT for spinal intramedullary AVMs 3

patients with im-sAVMs for over a decade. This retro- arteriovenous malformations (AVMs) with microsurgery or
spective review was conducted at our institution to evalu- endovascular embolization before radiation therapy (10
ate whether fRT is beneficial in treating im-sAVMs with microsurgery for AVF, 2 with microsurgery and 3 with
without causing long-term, permanent radiation toxicity. endovascular embolization for AVMs). Only 4.8% (1 of 21)
underwent microsurgery for the residual coexisting AVF
after fRT, while no patients received treatment for AVM.
Methods and Materials The details of the patients’ characteristics are shown in
Table 1 and supplementary material (Table E1).
This single-center, retrospective study was approved by All patients underwent angiography and MRI before
the institutional review board. The materials and methods radiation therapy to locate the AVM nidus. CT simulation
are as follows. was done for radiation therapy planning. Among the cases
that underwent surgical treatment prior to fRT, fiducial
markers were placed in the vertebrae near the nidus in 9
Patient selection
patients and were used for patient setup. Patient immobi-
lization techniques are shown in Table 1. In the first 19
The paper and electronic medical records of patients cases, we used the 3-dimensional conformal radiation
with im-sAVMs treated with fRT at Hokkaido University therapy (3DCRT) technique, whereas the intensity modu-
Hospital between January 2003 and December 2023 were lated radiation therapy (IMRT) technique was used for
reviewed. Patients with a follow-up period of less than planning in the last 2 cases. For 3DCRT planning, the
180 days were excluded from the study. clinical target volume included the entire spinal canal,
covering all levels of the nidus region. In the last 2
patients where IMRT was used, the clinical target volume
Evaluation and treatment included only the AVM nidus, which was delineated by
an experienced neuro-oncologist using MRI combined
All patients were evaluated using angiography and with angiography to accurately localize the nidus (Fig. 2).
magnetic resonance imaging (MRI) before radiation ther- The nidus volume was calculated in 2 patients treated
apy. CT simulation was performed for treatment plan- with the IMRT technique, showing volumes of 1.7 cc and
ning. Linac-based fRT was used, with doses ranging from 2.3 cc, respectively. However, it was not possible to access
20 to 24 Gy delivered in 4 fractions to the nidus area. the treatment planning system for the patients treated
Treatment response was evaluated using either MRI or with 3DCRT, and therefore, volume data for the nidus
angiography. could not be collected. The prescribed dose was 20 Gy in
4 fractions for 19 patients and 24 Gy in 4 fractions for 2
patients. In cases treated with 3DCRT, the dose was pre-
Study endpoints scribed to the isocenter set at the center of the nidus. For
cases treated with IMRT, the prescribed dose was set to
The endpoints of this study were the incidence of ensure that over 95% of the gross tumor volume (GTV, ie,
bleeding after fRT, the obliteration rate of the nidus, and nidus) was covered by the prescribed dose and that the
late radiation-induced toxicity. Basic patient characteris- minimum dose to the planning target volume (PTV, ie,
tics, treatment planning procedures, and all study end- GTV + 1 mm) reached at least 90% of the prescribed
points were analyzed using descriptive statistics, dose.
presented as means, medians, and percentages. During the follow-up period, clinical and radiographic
evaluations (either angiography or MRI) were performed
for all patients at each follow-up visit. At a median fol-
Results low-up time of 56.4 months (range, 12-236), the AVM
nidus was obliterated in 9.5% of patients (2 of 21),
Among 25 patients with im-sAVMs who underwent reduced in size in 38.1% (8 of 21), remained stable in
fRT between January 2003 and December 2023, 4 patients 47.6% (10 of 21), and increased in size in 4.8% (1 of 21).
were excluded because of follow-up times of less than The median time from the first day of radiation therapy
180 days, leaving 21 patients enrolled in the study. The to a reduction in size was 14 months (range, 4.8-37.5). Of
median age at the time of radiation therapy was 27 years the 2 patients whose AVM was completely obliterated,
(range, 14-55); 10 were male and 11 were female biologi- one occurred at 39.7 months after radiation therapy and
cally. Of the 21 patients, the distribution of sAVMs the other at 52 months. All details regarding the nidus
included 9 at the cervical level, 10 at the thoracic level, response after radiation therapy, including additional
and 2 at the lumbar level. Hemorrhage was present at ini- treatments, are shown in the swimmer plot (Fig. 3). No
tial presentation in 42.8% of patients, and 71.4% (15 of patients experienced hemorrhagic symptoms or proven
21) had undergone prior treatment of coexisting AVF or hemorrhage by radiographic evaluation after fRT until

Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.
 ARTICLE IN PRESS
4 R. Vichitvejpaisal et al Advances in Radiation Oncology: Month 2025

Table 1 Patient characteristics (n = 21) the last follow-up. No radiation therapy-induced late
adverse event was observed. However, one patient devel-
Characteristic Value Ratio (%)
oped an aneurysm in the feeding artery 57 months after
Ages Median (y) - radiation therapy. Because the aneurysm occurred outside
27 (14-55) the radiation field, it is unlikely to have been caused by
Follow-up time (mo) - the radiation (Fig. 4).
Median 56.4 (12-236) -
<36 5 23.8 Case study 1
37-72 9 42.9
A 26-year-old male presented with paralysis of both
73-108 4 19.0
lower limbs and later developed weakness in the upper
>108 3 14.3 limbs as well as the neurogenic bladder (patient No. 20,
Biological gender Fig. 2). MRI revealed flow voids on T2-weighted images
Male 10 47.6 at the C6 to C7 level with intramedullary hemorrhage
extending from C5 to C7. Digital subtraction angiography
Female 11 52.4
(DSA) showed im-sAVMs with a nidus measuring
Distribution approximately 2.8 £ 1.4 cm at the C6 to C7 level. The
Cervical 9 42.9 feeder vessel originated from the left T4 intercostal artery,
Thoracic 10 47.6 flowing via the anterior spinal artery (ASA). Following
conservative treatment, the patient’s weakness improved,
Lumbar 2 9.5
and he was referred to Hokkaido University Hospital for
Prior treatment of coexisting AVF definitive treatment. Because of the high surgical and
Microsurgery 10 47.6 embolization risks associated with ASA involvement, we
Embolization 0 0 decided to use radiation therapy as the primary treatment
approach.
None 11 52.4
The patient underwent fRT at a dose of 24 Gy in 4
Prior treatment of AVM fractions over consecutive days using the IMRT tech-
Microsurgery 2 9.5 nique. The GTV included the nidus, which was delineated
Embolization 3 14.3 by combining CT angiography with CT simulation
images and MRI. The PTV consisted of the GTV plus a
None 16 76.2
1 mm margin to account for setup error. The prescribed
Treatment after fRT dose was designed to cover at least 95% of the GTV,
Microsurgery for residual AVF 1 4.8 ensuring that the entire PTV was within the 90% isodose
Embolization 0 0 line. At 4.8 months posttreatment, both MRI and DSA
showed a dramatic reduction in nidus size. At the 39.7-
None 20 95.2
month follow-up, the MRI revealed complete nidus reso-
Immobilization lution, which was confirmed by DSA. His symptoms con-
Shell/gold marker 3 14.3 tinued to improve over time, with only mild left cervical
Shell 6 28.6 radiculopathy at the C7 level, and he was able to resume
daily life as usual.
Cushion 2 9.5
Gold marker 6 28.6
None 4 19.0
Case study 2
Radiation therapy technique
3DCRT 19 90.5 A 31-year-old female presented with weakness in the
IMRT 2 9.5 left lower limb (patient No.5, Fig. 4). She had previously
Radiation therapy dose undergone 3 endovascular embolizations for a dural AVF
at another hospital. After 12 years of follow-up, her weak-
20 Gy/4 F 19 90.5
ness worsened, and she developed a neurogenic bladder,
24 Gy/4 F 2 9.5 requiring self-catheterization. She was subsequently
Abbreviations: 3DCRT = 3-dimensional conformal radiation ther- referred to Hokkaido University Hospital. Angiography
apy; AVF = arteriovenous fistula; AVM = arteriovenous malforma- revealed im-sAVMs at the T3 to T4 level, with the feeder
tion; fRT = fractionated radiation therapy; IMRT = intensity vessel originating from the right thyrocervical trunk and
modulated radiation therapy.
flowing via the ASA. Following surgical treatment for a

Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.
 ARTICLE IN PRESS
Advances in Radiation Oncology: Month 2025 Outcome of fRT for spinal intramedullary AVMs 5

perimedullary AVF, she received fRT at a dose of 20 Gy in

4 fractions over consecutive days using the 3DCRT tech-
nique. At 20.9 months of follow-up, the MRI showed that
the nidus had reduced in size, with no bleeding detected.
However, an MRI scan at 57 months post-fRT showed a
new aneurysm located at the T2 level. A review of the
prior fRT plan confirmed that this aneurysm was located
above the irradiated area. Although the cause of the new
aneurysm cannot be fully determined, it is likely unrelated
to the radiation therapy.

Discussion

As a result of the rarity of im-sAVMs, there is no

established standard treatment regarding optimal modali-
ties, their sequencing, or the definitive role of radiation
therapy. Because of noninvasiveness, radiation therapy is
theoretically an appropriate option in contrast with sur-
gery or embolization, which carries a risk of severe, per-
manent treatment-related adverse events, such as spinal
cord infarction. Additionally, details regarding radiation
therapy, including the technique, dose, and fractionation,
remain subjects of debate.
Several articles have been published on the outcome of
fRT as a part of im-sAVM treatment for more than a
decade (Table 2).8-16 Most of these studies are small retro-
spective series or case reports. The largest study was
reported by Kalani et al,13 which included 37 patients
with im-sAVMs treated with fRT using the CyberKnife
system. This study found a complete obliteration rate of
19% for the nidus, with no cases of rebleeding after fRT.
However, the prescribed doses varied in this study, with a
biologically effective dose (a/b = 2) (BED2) range of 70 to
120 Gy, and one patient developed radiation-induced
myelopathy (1 of 37, 2.7%), making it difficult to adopt
this protocol as a reference. Therefore, we conducted a
retrospective review of our patients treated with our fRT
protocol. Although there was a previously published

vessel originated from the left T4 intercostal artery, flowing via

the anterior spinal artery. The patient underwent fractionated
radiation therapy (fRT) at a dose of 24 Gy in 4 fractions using
the intensity modulated radiation therapy technique. (B)
Shows the gross tumor volume (GTV) (ie, nidus, in red) and
isodose distribution. The 100% isodose line (navy) covered at
least 95% of the GTV, ensuring that the entire planning target
Figure 2 (Patient No. 20) A 26-year-old male presented with volume (GTV + 1mm) was included within the 90% isodose
paralysis of both lower limbs and later developed weakness in line (green). (C) (Left) Shows MRI and (right) DSA at 4.8
the upper limbs as well as the neurogenic bladder. (A) (Left) months after fRT. The nidus (white arrow) was dramatically
Magnetic resonance imaging (MRI) showed flow voids on T2- decreased in size. (D) (Left) shows MRI at 39.7 months after
weighted images at the C6 to C7 level (white arrow) with intra- fRT where all flow voids had disappeared. (right) The DSA
medullary hemorrhage extending from C5 to C7. (right) Digi- confirmed the disappearance of the nidus compared with
tal subtraction angiography (DSA) revealed intramedullary before fRT. Markedly high T2 signal was noted at the C6 to C7
spinal arteriovenous malformation with a nidus measuring level (yellow arrow), indicating fatty marrow changes from
approximately 2.8 £ 1.4 cm at the C6 to C7 level. The feeder radiation therapy.

Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.
 ARTICLE IN PRESS
6 R. Vichitvejpaisal et al Advances in Radiation Oncology: Month 2025

Figure 3 Swimmer plot showing nidus response and additional treatments after radiation therapy in relation to the duration of
follow-up.

Figure 4 (Patient No. 5) A 31-year-old female presented with weakness in the left lower limb, and angiography revealed intra-
medullary spinal arteriovenous malformation at the T3 to T4 level. After undergoing surgical treatment for a perimedullary arte-
riovenous fistula, she received fractionated radiation therapy (fRT) at a dose of 20 Gy in 4 fractions. Flow voids had decreased in
size by 20.9 months after fRT, with no bleeding detected. However, magnetic resonance imaging at 57 months post-fRT (left)
showed a new aneurysm at the T2 level (white arrow). A review of the previous fRT planning (right) confirmed that the aneu-
rysm was located above the irradiated area.

Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.
 Advances in Radiation Oncology: Month 2025
Table 2 Summary of studies reporting details and outcomes of RT for treating im-sAVMs, including a comparison with our study
Median Sex Hemorrhage Treatment Dose Nidus (%)
Studies No. age (y) F: M before RT Type before RT (Gy/F) (BED2) Target volume Technique F/U RM Rebleed
CR PR SD PD
Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier

Wolkov et al, 19888 2 12, 75 2:0 NR NR 2 surgery 45-50/25-28 7, 12.5 cm Conventional 5,12 y NR 0% 0%
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.

(1963-1981) (85.5-125)
Hida et al, 20039 10 25 6:4 100% Intramedullary 3 embolization 32-40/20 NR 3DCRT Median 49 mo 0 50 50 0 0% 0%
(1982-1999) (15-50) 2 Surgery (57.6-80) (26-124)
32/16, 32.4/18
(61.56,64)
20/4,30/8
(70,86.5)
Sinclair et al, 200610 15 NR NR 53% Intramedullary 7 embolization 25/5 (87.5) Mean 2.36 cc CyberKnife Mean 27.9 mo 1/5* 3/5* 1/5* 0 0% 0%
(1997-2005) 1 surgery 21/3-4 stainless fiducial (3-59)
(76.7-94.5) average isodose 83%
20/2 (120)
Gekka et al, 201411

ARTICLE IN PRESS
1 37 0:1 100% Juvenile type None 20/4 (70) NR NR 5y 100 0% 0%
(2007)
Potharaju et al, 201412 3 26 2:1 100% Intramedullary 2 embolization 21/3 (94.5) Mean 4.05 cc CyberKnife Mean 66.6 0 33.3 0 0% 0%
(2010-2011) (25-57) Vaclok, X-sight 35.7 mo
spine tracking (24-39)
Isodose
80%-85%
Kalani et al, 201613 37 30 19:18 50% Intramedullary 16 embolization 20/3-4 Mean 2.3 cc CyberKnife Mean 39.7 mo 19 46 35 0 3% 0%
(1996-2014) (9-56) 7 surgery (70-86.7) use fiducial (2-121)

Outcome of fRT for spinal intramedullary AVMs

5 SBRT 21/3 (94.5)
20/2 (120)
Rashad et al, 201714 5 39 1:4 40% Intramedullary 2 embolization 18/3 (72) NR CyberKnife Mean 37.2 mo 0 20 80 0 0% 0%
(2010-2014) (31-70) (4 compact 1 juvenile) (16-62)
Mori et al, 201815 5 32 3:2 60% Intramedullary 2 embolization 20/4 (70) Mean 2.62 cc Novalis Median 40 0 60 0 20% 0%
(2006-2014) (23-54) 5.5 y (temporary)
(3-8.5)
Suzuki et al, 202316 10 17.5 3:7 90% Intramedullary 1 embolization 14-24/2-7 Mean 0.72 cc CyberKnife Mean 47.2 mo 0 10 90 0 0% 0%
(2011-2022) (7-55) 1 surgery (37.5-94.5) isodose 78.2% (6-125)
21 27 11:10 42.8% Intramedullary 3 embolization 20/4 (70) 1,7, 2.3 cc 3DCRT 19 Median 10 38 48 4 0% 0%
Present study (2003-2023) (14-55) 12 surgery 24/4 (96) (IMRT patients) IMRT 2 56.4 mo
(12-236)

Abbreviations: NR = not reported; 3DCRT = 3-dimensional conformal radiation therapy; BED2 = biological effective dose (a/b = 2); CR = complete response; IMRT = intensity modulated radiation therapy;
im-sAVMs = intramedullary spinal arteriovenous malformation; PD = progressive disease; PR = partial response; RM = radiation-induced myelopathy; RT = radiation therapy; SD = stable disease.
*Only 5 patients were able to undergo imaging evaluation in this study.

7
 ARTICLE IN PRESS
8 R. Vichitvejpaisal et al Advances in Radiation Oncology: Month 2025

result,9 the initial study included mostly conventional experience treating patients with im-sAVMs suggests that
fractionation radiation therapy (70% used 1.8-2 Gy/d). immobilization with a shell or cushion, combined with
With additional patients, a focus solely on fRT, continu- bony landmarks for image guidance, is sufficient. This
ous advancement in radiation therapy techniques, and approach has enabled us to deliver effective and safe treat-
extended follow-up, we now present the outcomes as one ment without the need for costly equipment. This method
of the largest series on fRT in im-sAVMs. may be beneficial for nontertiary care centers where spe-
Our study demonstrated that fRT with a dose of 20 to cialized treatment systems are not available.
24 Gy in 4 fractions was both effective and safe, with no
cases of rebleeding after treatment and no radiation- Limitation
related myelopathy observed, even after more than
10 years of follow-up. Although the complete obliteration
Our study has a few limitations. Because radiation
rate in our study appears lower compared to other studies,
therapy techniques and treatment systems have evolved
this may be because of the relatively low BED2 of 70 to 96
over time, only 2 patients in this study were treated with
Gy. Nevertheless, the rebleeding rate was zero.
the IMRT technique. IMRT provides a highly conformal
In contrast, there is evidence that partial obliteration of
dose distribution, making it well-suited for fRT planning.
the intracranial arteriovenous malformation (bAVMs)
Additionally, as im-sAVMs typically occur in younger
does not reduce the annual hemorrhage rate.17 As a result,
patients and the condition is benign, concerns about sec-
the goal in treating bAVMs is complete nidus obliteration,
ondary malignancy or tumor development should be
which requires high-dose radiation therapy for definitive
more than usual. IMRT has the advantage of reducing
treatment. Flickinger et al18 analyzed the dose-response
radiation exposure to normal tissues by decreasing the
relationship between radiosurgery dose and nidus obliter-
irradiated target volume. Several studies have reported
ation rates, finding that radiosurgery doses of 20 to 24 Gy
that using radiosurgery with advanced treatment systems
achieved obliteration rates of 65% to 86%. However, such
does not increase the incidence of secondary malignancy
high doses cannot be used in cases of im-sAVMs because
or malignant transformation in benign central nervous
of spinal cord dose tolerance limitations.19-21 According
system tumors compared to the general population and
to previous studies, including ours (Table 2), most of the
shows a lower incidence than conventional or 3-field radi-
series used radiation doses with a BED2 of less than 100
ation therapy.24-26 However, with only 2 patients treated
Gy, and no patients in these studies experienced
with advanced IMRT in our study, we were unable to
rebleeding. This suggests that treatment of im-sAVMs
report certain data, such as GTV nidus volume, or to eval-
may not require high doses to achieve complete oblitera-
uate the clinical advantages of IMRT over older 3DCRT
tion, thereby reducing the risk of radiation-induced mye-
planning.
lopathy.
Another limitation is that each patient received other
The difference in treatment response between im-
treatments before and/or after radiation therapy (57.1% and
sAVMs and bAVMs remains unclear because of the
14.3% of the patients received microsurgery and emboliza-
limited understanding of the causes and natural his-
tion before fRT, respectively, whereas 4.8% underwent
tory of im-sAVMs. One possible explanation lies in
microsurgery for the residual coexisting AVF after fRT), pre-
the possible differences in blood flow and pressure
venting us from isolating the effect of radiation therapy
between intracerebral and intramedullary vessels.
alone on rebleeding prevention or accurately assessing its
Blood flow within the spinal cord may be lower than
impact on the annual bleeding rate compared to others.
in the brain,22 allowing radiosurgery-induced flow
Although a prospective or randomized controlled study
reduction to be sufficient to prevent bleeding. In con-
could address these limitations, conducting such a study is
trast, the higher blood pressure in intracerebral vessels
challenging because of the rarity of this disease. As of now,
may need complete obliteration of the nidus, as merely
all available literature, including our study, consists of retro-
decreasing blood flow might not be enough. However,
spective reviews and represents the current best evidence.
this theory lacks direct evidence. Furthermore, most
studies on spinal cord blood flow have been conducted
in animal models because of the technical challenges
of measuring or sampling venous drainage from the Conclusions
small spinal cord venous system.23 Future advance-
ments in technology for measuring human spinal cord Fractionated radiation therapy at doses of 20 to 24 Gy in 4
blood flow, combined with accumulated treatment out- fractions for im-sAVMs has been shown to be safe and may
comes from more patients, may help clarify this issue. effectively reduce the risk of bleeding, with no severe perma-
Another important aspect to discuss is the fRT tech- nent radiation-induced myelopathy observed even after long-
nique. Although many institutions use radiosurgery sys- term follow-up. We conclude that fRT is a reasonable option,
tems equipped with image guided radiation therapy either as an adjunct to other modalities or as a primary treat-
modalities to correct intrafraction movement, our ment choice when other treatments are not feasible.

Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.
 ARTICLE IN PRESS
Advances in Radiation Oncology: Month 2025 Outcome of fRT for spinal intramedullary AVMs 9

Disclosures 10. Sinclair J, Chang SD, Gibbs IC, Adler Jr JR. Multisession CyberKnife
radiosurgery for intramedullary spinal cord arteriovenous malfor-
mations. Neurosurgery. 2006;58:1081-1089.
The authors declare that they have no known compet- 11. Gekka M, Seki T, Hida K, Osanai T, Houkin K. Surgical manage-
ing financial interests or personal relationships that could ment of combined intramedullary arteriovenous malformation and
have appeared to influence the work reported in this perimedullary arteriovenous fistula within the hybrid operating
paper. room after five years of performing focus fractionated radiotherapy:
Case report. Neurol Med Chir (Tokyo). 2014;54:936-940.
12. Potharaju M, John R, Venkataraman M, Gopalakrishna K, Subrama-
nian B. Stereotactic radiosurgery results in three cases of intrame-
Acknowledgments dullary spinal cord arteriovenous malformations. Spine J.
2014;14:2582-2588.
13. Kalani MA, Choudhri O, Gibbs IC, et al. Stereotactic radiosurgery
Ruja Vichitvejpaisal was responsible for statistical anal- for intramedullary spinal arteriovenous malformations. J Clin Neu-
ysis. rosci. 2016;29:162-167.
14. Rashad S, Endo T, Ogawa Y, et al. Stereotactic radiosurgery as a fea-
sible treatment for intramedullary spinal arteriovenous malforma-
tions: A single-center observation. Neurosurg Rev. 2017;40:259-266.
Supplementary materials 15. Mori Y, Hashizume C, Tsugawa T, Kato S, Shibamoto Y. Stereotac-
tic radiotherapy for intramedullary spinal arteriovenous malforma-
tions. Cureus. 2018;10:e2908.
Supplementary material associated with this article can 16. Suzuki T, Kagawa K, Sato K, Nomura R, Irie K, Ichi S. CyberKnife
be found in the online version at doi:10.1016/[Link].2025. radiosurgery for spinal intramedullary arteriovenous malformations:
101778. A single-center experience. World Neurosurg. 2023;175:e230-e237.
17. Lv X, Wu Z, Li Y, Yang X, Jiang C. Hemorrhage risk after partial
endovascular NBCA and ONYX embolization for brain arteriove-
nous malformation. Neurol Res. 2012;34:552-556.
References 18. Flickinger JC, Kondziolka D, Maitz AH, Lunsford LD. An analysis of
the dose-response for arteriovenous malformation radiosurgery and
1. Gross BA, Du R. Spinal glomus (type II) arteriovenous malforma- other factors affecting obliteration. Radiother Oncol. 2002;63:347-
tions: A pooled analysis of hemorrhage risk and results of interven- 354.
tion. Neurosurgery. 2013;72:25-32. 19. Sahgal A, Chang JH, Ma L, et al. Spinal cord dose tolerance to ste-
2. Ozpinar A, Weiner GM, Ducruet AF. Epidemiology, clinical presen- reotactic body radiation therapy. Int J Radiat Oncol Biol Phys.
tation, diagnostic evaluation, and prognosis of spinal arteriovenous 2021;110:124-136.
malformations. Handb Clin Neurol. 2017;143:145-152. 20. Timmerman R. A story of hypofractionation and the table on the
3. Flores BC, Klinger DR, White JA, Batjer HH. Spinal vascular malfor- wall. Int J Radiat Oncol Biol Phys. 2022;112:4-21.
mations: Treatment strategies and outcome. Neurosurg Rev. 21. Milano MT, Usuki KY, Walter KA, Clark D, Schell MC. Stereotactic
2017;40:15-28. radiosurgery and hypofractionated stereotactic radiotherapy: Nor-
4. Miyasaka K, Asano T, Ushikoshi S, Hida K, Koyanagi I. Vascular mal tissue dose constraints of the central nervous system. Cancer
anatomy of the spinal cord and classification of spinal arteriovenous Treat Rev. 2011;37:567-578.
malformations. Interv Neuroradiol. 2000;6(1_suppl):195-198. 22. Hickey R, Albin MS, Bunegin L, Gelineau J. Autoregulation of spinal
5. Hida K, Asano T, Aoyama T, Houkin K. Multidisciplinary approach cord blood flow: Is the cord a microcosm of the brain? Stroke.
for spinal AVMs(<Special Issue>Recent progress in the diagnosis 1986;17:1183-1189.
and treatment of cerebral and spinal vascular malformation). Jpn J 23. Joshi S, Ornstein E, Young WL. Cerebral and spinal cord blood. In:
Neurosurg. 2011;20:20-28. Cottrell JE, Young WL, eds. Cottrell and Young’s Neuroanesthesia.
6. Bostr€om A, Krings T, Hans FJ, Schramm J, Thron AK, Gilsbach JM. 5th ed Mosby; 2010:17-59.
Spinal glomus-type arteriovenous malformations: Microsurgical 24. Wolf A, Naylor K, Tam M, et al. Risk of radiation-associated intra-
treatment in 20 cases. J Neurosurg Spine. 2009;10:423-429. cranial malignancy after stereotactic radiosurgery: A retrospective,
7. Velat GJ, Chang SW, Abla AA, Albuquerque FC, McDougall CG, multicentre, cohort study. Lancet Oncol. 2019;20:159-164. Published
Spetzler RF. Microsurgical management of glomus spinal arteriove- correction appears in Lancet Oncol. 2019;20:e72.
nous malformations: Pial resection technique: Clinical article. J Neu- 25. Sherry AD, Bingham B, Kim E, et al. Secondary malignancy fol-
rosurg Spine. 2012;16:523-531. lowing stereotactic radiosurgery for benign neurologic disease: A
8. Wolkov HB, Bagshaw M. Conventional radiation therapy in the cohort study and review of the literature. J Radiosurg SBRT.
management of arteriovenous malformations of the central nervous 2020;6:287-294.
system. Int J Radiat Oncol Biol Phys. 1988;15:1461-1464. 26. Minniti G, Traish D, Ashley S, Gonsalves A, Brada M. Risk of sec-
9. Hida K, Shirato H, Isu T, et al. Focal fractionated radiotherapy for ond brain tumor after conservative surgery and radiotherapy for
intramedullary spinal arteriovenous malformations: 10-year experi- pituitary adenoma: Update after an additional 10 years. J Clin Endo-
ence. J Neurosurg. 2003;99(1 Suppl):34-38. crinol Metab. 2005;90:800-804.

Downloaded for Atanu Panda (atanupanda1995@[Link]) at Chittaranjan National Cancer Institute from [Link] by Elsevier
on July 15, 2026. For personal use only. No other uses without permission. Copyright ©2026. Elsevier Inc. All rights reserved.