Re: angiographic evidence of aneurysm neck healing following endovascular treatment with bioactive coils.

Full text

Page 1

Reporting Terminology for Lumbar Disk
Herniations: Axial Segmentation of the
Preneural Foraminal Portion of the
Lumbar Nerve Roots
Through the past decades, many attempts have been made
to standardize lumbar disk herniation–related terminology (1,
2). Two different, and not necessarily parallel, sets of charac-
teristics of a herniated disk (HD) are taken into account in the
reporting of cross-sectional images of the lumbar spine. First, is
the absolute attributes of the HD (ie, morphology, location,
volume, etc), and second is the relationship of the HD to its
neighboring neural structures. The absolute attributes of a HD
are only one set of factors that determine the pattern and
degree of compromise of the neighboring neural structures.
The morphology and cross-sectional areas of the involved lum-
bar vertebral and neural foraminae, along with the degree of
the epidural fat, are the other determining factors.
Although standardized nomenclature has been proposed for
the absolute attributes of a HD, descriptive terminology is used
for the elaboration of the relationship of a HD to its neighbor-
ing neural structures. Terms such as “abuts,” “impinges,” “flat-
tens,” “indents,” “compresses,” “displaces,” and “distorts” are
useful in the description of the effects of a HD on the adjacent
thecal sac (TS) and intracanalicular dural nerve root sleeves
(DNRSs). The effects of a HD on specific intracanalicular
nerve roots may at times also be ascertained by using high-
resolution axial T2-weighted sequences of the lumbar spine. In
particular, compromise (or maximal degree of compromise) of
a lumbar nerve root (LNR) at a level just proximal to its
respective neural foraminal exit could be pinpointed to a spe-
cific segment of the nerve.
In this context, a 3-tiered descriptive division of the pre-
neural foraminal portion of a LNR, into the lateral thecal
sac, junctional, and dural root sleeve segments (Figs 1 and 2)
may be of value. This partitioning is particularly suitable at
FIG 1.
portion of the S1 nerve root. In this patient, individualization of all of the 3 segments is possible within a single zone.
A, Cranial-most image demonstrates the lateral thecal sac segment of the right S1 nerve root (arrow) in the right central zone.
B, Image shows the junctional zone segment of the right S1 nerve root (long arrow) in the right central zone. This segment, which is
situated in the proximal portion of the dural root sleeve, is separated from the contents of the neighboring thecal sac by 2 adjacent layers
of dura mater (arrowhead). The thickness of a single layer of dura mater is demarcated (short arrow) in the posterior portion of the thecal
sac for comparison.
C, More caudally, image demonstrates the dural root sleeve of the right S1 nerve root (arrow) in the right central zone. Individualization
of the contained dural root sleeve segment of the S1 nerve root is not possible in this patient.
Contiguous axial T2-weighted (3800/97.8) MR images through the L5–S1 level show the 3 segments of the preneural foraminal
FIG 2.
(3700/106.7) MR images through the
L5–S1 level show the varying appear-
ances (also refer to Fig 1B) of the pro-
posed “junctional segment” of a lumbar
nerve root.
A, Image demonstrates the junctional
segment of the left S1 nerve root (long
arrow) within the “pinched” portion (ar-
rowheads) of the left ventrolateral angle
of the thecal sac. The unrestricted con-
tour of the right ventrolateral aspect of
the thecal sac harbors the lateral thecal
sac segment of the ipsilateral S1 nerve
root (short arrow).
B, Image at a slightly more caudal level
shows the proximal-most portion of the
left S1 dural nerve root sleeve (arrow-
heads), the medial wall of which is in close apposition to the ventrolateral wall of the neighboring thecal sac. This portion of the dural
root sleeve is also regarded as housing the junctional segment of the S1 nerve root (arrow).
Contiguous axial T2-weighted
Letters
2430

Page 2

the lower lumbar levels that demonstrate a distinguishable
lateral recess and parallels the compartmentalization of the
preneural foraminal portion of dural coverings of the LNRs.
The intracanalicular portions of the LNRs demonstrate
an intradural topography (the LNRs pierce the dura at the
level of their respective neural foraminae) (3), either within
the TS (Fig 1A) or in a DNRS (Fig 1C). These sleeves are
pinched off from the ventrolateral angles of the TS. This
TS-DNRS transition may be evident as “waisting” (Fig 2A)
of the ventrolateral angles of the TS. More distally, the
passage of the DNRS within the epidural fat (Fig 1C) is
reminiscent of the intraorbital course of the optic nerve
within the orbital fat.
These proposed segments do not necessarily comply with
the localization system proposed by Wiltse et al (4) and later
endorsed by the 2001 multidisciplinary task force (1). A
particular “zone” may harbor different segments of a LNR
(Fig 1), whereas a specific “level” may accommodate differ-
ent segments of the opposing LNRs (Fig 3).
Awareness of the varying appearances of a LNR within
the vertebral foramen may be of value in localizing the
level—or maximal level—of neural structural compromise
and correlating the imaging findings with the patient’s symp-
toms. This is especially valuable in herniated disks that
demonstrate a volume-neural structural compromise discor-
dance (eg, a small-volume posteriorly HD may cause signif-
icant impingement of a neighboring ventrally located pre-
neural foraminal LNR) (Fig 4), and may have a potential
role in patient management.
Kimia Khalatbari, Mazyar Azar
and Farid Kazemi Gazic
Departments of Radiology and Neurosurgery
Rassoul Akram University Hospital
Tehran, Iran
References
1. Fardon DF, Millette PC. Nomenclature and classification of
lumbar disc pathology: recommendations of the combined task
forces of the North American Spine Society, American Society of
Spine Radiology, and American Society of Neuroradiology.
http://www. asnr.org/spine_nomenclature/ (accessed March 20,
2005)
2. Milette PC. The proper terminology for reporting lumbar interver-
tebral disk disorders. AJNR Am J Neuroradiol 1997;18:1859–1866
3. Clemente CD. Gray’s anatomy. Philadelphia: Lea & Febiger; 1985:
1191–1192
4. Wiltse LL, Berger PE, McCulloch JA. A system for reporting the
size and location of lesions
22:1534–1537
ofthespine.
Spine1997;
FIG 3.
suprapedicular level of L5–S1 demonstrates the nonparallel seg-
mentation of the S1 nerve roots. On the right, the lateral thecal
sac segment (arrow), and on the left, the junctional segment
(arrowhead), of the S1 nerve roots are visualized.
Axial T2-weighted (4000/98) MR image at the diskal-
FIG 4.
ualization of the dural root sleeve segments of the S1 nerve roots within their respective dural sleeves is not possible in this particular
patient).
A, Cranial-most image demonstrates the right (long arrow) and left (short arrow) S1 dural nerve root sleeves within the subarticular
zones.
B, Image at a slightly more caudal level shows a mild central-subarticular zone disk protrusion that causes a relatively severe
compression of the right S1 dural nerve root sleeve (long arrow). The left-sided intact complex is demarcated (short arrow) for
comparison.
C, The S1 dural nerve root sleeves (long and short arrows) assume a symmetric appearance at a slightly more caudal level.
Contiguous axial T2-weighted (3800/97.8) MR images through the L5–S1 level with the S1 dural nerve root sleeves (individ-
AJNR: 26, October 2005 Letters2431

Page 3

Reply
The multidisciplinary task force on lumbar disk nomen-
clature to which Dr. Khalatbari refers tried to devise a
practical and simple classification of disk herniations, with
the fewest categories, so that substantial interobserver
agreement could be achieved (1). This is the main reason
why the proposed system does not require observers to grade
compromise of nerve roots by displaced disk material, an
unreliable exercise in frequent situations (eg, suboptimal
technical quality images, spinal stenosis, postoperative
changes). In this system, the impact of a disk herniation on
the thecal sac and individual nerve roots is suggested by
assessing volume of displaced disk material with respect to
available space. Thus, herniations are graded as mild, mod-
erate, or severe, depending on extension of displaced disk
material in the proximal, middle, or distal third of the
available spinal lumen, in a specific anatomical zone, and at
a specific anatomical level. In the example provided by Dr.
Khalatbari (Fig 4), the right subarticular zone is completely
obliterated by disk material at the suprapedicular level, and
the term mild is inappropriate to classify such a herniation
according to the system proposed by the multidisciplinary
task force: this lesion qualifies as a severe herniation. This is
an interesting case because it illustrates an important con-
cept; that is, in the subarticular zone, displacement of a
relatively small amount of disk material can cause severe
nerve root compression. However, I can well understand the
reluctance to use the term severe to describe such a localized
herniation and, in retrospect, a numerical grading system
(grades 1, 2, 3) would have been a better choice to describe
the volume of herniated disk material. Radiologists who
report on lumbar spine imaging studies according to the
classification and definitions proposed by the task force are
of course welcome to provide referring physicians all addi-
tional anatomical details that they feel may be clinically
relevant, such as those discussed by Dr. Khalatbari. A grad-
ing system for nerve root compromise (0, normal; 1, contact;
2, deviation; 3, compression) has been recently proposed (2)
and might be considered as a complement.
Pierre C. Milette
Montreal, Quebec, Canada
References
1. Milette PC. Reporting lumbar disk abnormalities: at last, consen-
sus! AJNR Am J Neuroradiol 2001;22:429–430
2. Pfirrmann CWA, Dora C, Schmid MR, Zanetti M, Hodler J, Boos
N. MR image–based grading of lumbar nerve root compromise due
to disk herniation: reliability study with surgical correlation. Ra-
diology 2004;230:583–588
Location, Location, Location: Angiography
Discerns Early MR Imaging Vessel Signs Due
to Proximal Arterial Occlusion and Distal
Collateral Flow
The location or topography of hypoperfusion is likely a
critical determinant of outcome in acute ischemic stroke. The
spatial distribution of a perfusion defect is directly influenced
by the site of vascular occlusion and corresponding collateral
flow. Such vascular localization is readily apparent on angiog-
raphy and may also be evident on MR imaging as early vessel
signs (EVS) or intravascular signal intensity derangements,
including fluid-attenuated inversion recovery (FLAIR) vascu-
lar hyperintensity (FVH), and gradient-echo susceptibility
(GRE SVS). Although similar in appearance, proximal and
distalEVSmay represent
substrates.
Schellinger et al (1) provide an important analysis on the
differentpathophysiologic
diagnostic and prognostic value of such relatively subtle EVS in
hyperacute stroke cases treated with intravenous thrombolysis.
They explicitly note that their study is unique, because “no
clinical study to date has taken thrombus composition as being
reflected by GRE (and FLAIR) images and its role for rtPA
[recombinant tissue plasminogen activator] response into ac-
count” (p. 623). Similar to prior reports of EVS, the authors
combine proximal and distal EVS as a single entity. Proximal
and distal EVS are both associated with perfusion abnormali-
ties induced by proximal arterial occlusion or thrombosis, yet
distal EVS may not be due to the MR imaging signal intensity
characteristics of thrombus.
Distal EVS may be associated with retrograde collateral
flow distal to thromboembolic occlusion of a proximal artery
(Fig 1). Distal EVS are typically manifest in middle cerebral
artery occlusion as serpiginous structures in the sylvian fis-
sure extending distally through the convoluted architecture
of the cerebral sulci. It remains highly unlikely and unproven
that such intravascular signal intensity derangements are
due to thrombus extending along the entire course of the
middle cerebral artery. Contemporaneous angiography may
reveal patency of these vascular segments filled by slow,
retrograde collateral perfusion via leptomenigeal anastomoses
(Fig 1). FVH likely represents exceedingly slow collateral flow
induced by retrograde perfusion of an arterial tree with dimin-
ished intravascular pressure. Distal EVS on gradient echo, or
hypointensity in distal segments, may be associated with intravas-
cular deoxygenation of collateral blood due to precapillary oxygen
loss (2). The appearance of such distal EVS on gradient echo is
dissimilar to GRE SVS described in the report by Schellinger et al
(Fig 1A). Intravascular gradient-echo hypointensity in distal seg-
ments rarely demonstrates the prominent, and frequently coexis-
tent, blooming artifact apparent as GRE SVS due to proximal
thrombosis.
The authors also explore the enticing hypothesis that clot
burden evident on MR imaging as manifest by EVS may be
associated with thrombolytic efficacy. In contrast to endo-
vascular reperfusion techniques using mechanical clot dis-
ruption, intravenous thrombolysis is dependent on the ex-
tent of clot surface that is exposed to the lytic agent at only
the proximal and distal ends of the clot. Clot location, not
necessarily clot burden, may influence thrombolytic expo-
sure to the distal clot surface via collaterals. Conversely,
proximal diversion of blood flow to collateral routes may
also diminish forward pressure on the clot, hindering recan-
alization of thromboembolic occlusion. Clot location and
corresponding collateral routes may therefore be critical
variables in assessing not just tissue perfusion, but also
thrombolytic efficacy in acute stroke.
EVS on MR imaging may provide valuable information
regarding collaterals, not just thrombosis. Conventional perfu-
sion MR imaging and MR angiography may be used for detec-
tion of proximal arterial occlusion, yet selective techniques
such as regional perfusion MR imaging may more reliably
delineate collateral flow (3). Until such noninvasive techniques
are refined, angiography will remain the gold standard for
defining the location of proximal occlusion and corresponding
collaterals.
The location and not the mere presence of EVS on MR
imaging may disclose distinct vascular correlates. In acute isch-
emic stroke, it is imperative to consider the location of proxi-
mal flow cessation, the location of thrombus, and the location
of compensatory collateral flow.
David S. Liebeskind
UCLA Stroke Center
Department of Neurology
University of California
Los Angeles, CA
2432 LettersAJNR: 26, October 2005

Page 4

References
1. Schellinger PD, Chalela JA, Kang DW, et al. Diagnostic and prog-
nostic value of early MR imaging vessel signs in hyperacute stroke
patients imaged <3 hours and treated with recombinant tissue
plasminogen activator. AJNR Am J Neuroradiol 2005;26:618–624
2. Liebeskind DS, Ances BM, Weigele JB, Hurst RW. Intravascular
deoxygenation of leptomeningeal collaterals detected with gradi-
ent-echo MRI. Stroke 2004;35:266
3. Hendrikse J, van der Grond J, Lu H, et al. Flow territory mapping
of the cerebral arteries with regional perfusion MRI. Stroke
2004;35:882–887
Reply
We very much appreciated the letter of Dr. Liebeskind as
well as the opportunity to reply to his comments. Dr. Liebe-
skind raises an important point with regard to location, throm-
bus extent, and early vessel signs (EVSs). His major concern is
that not only the presence or location of EVSs, which indicate
a thrombus, disclose a distinct vascular pathophysiologic cor-
relate. EVS also represent areas with slow blood flow and
therefore may indicate the location and extent of compensatory
distal collateral flow (1).
First, the statement that the hyperintense fluid-attenuated
inversion recovery sign (FLAIR HVS) and the gradient-echo
susceptibility vessel sign (GRE SVS) may represent thrombus,
as well as slow and/or retrograde flow (FLAIR HVS) of deox-
ygenated blood (GRE SVS), is entirely accurate and acknowl-
edged in our manuscript in the second paragraph of the dis-
cussion (2, p. 622).
As stated in the introduction, our study had 3 objectives, the
first and foremost being diagnostic accuracy of early MR im-
aging vessel signs. For this, the mere presence as compared
with MR angiography (MRA)/perfusion-weighted imaging
(PWI) and the concordance of location of the occlusion with
the most proximal end of the EVS were analyzed. If present
(sensitivity of 65.9%), the proximal end of thrombus/occlusion
on FLAIR was correlated significantly (r ? 0.66; P ? .001) with
MRA/PWI. For this diagnostic objective it is not relevant
whether the further distal extent of the vessel sign is thrombus
or slow flow. We do acknowledge, however, the fact that GRE
besides its low sensitivity (34.1%) was not at all correlated with
MRA/PWI and is often seen more distally. Whether this rep-
resents the distal deoxygenated part of the thrombus or collat-
eral flow of deoxygenated blood is not clear. In fact, this could
imply that the sensitivity of the hypointense GRE sign for the
clot is substantially ?34.1%, because the hyperacute, still-
FIG 1.
and angiography in acute left middle ce-
rebral artery stroke. EVSs due to proxi-
mal thrombosis and distal collateral flow
are evident on gradient-echo (A) and
FLAIR (B). Angiography demonstrates
proximal arterial occlusion (C) and slow,
retrograde collateral flow (D).
Contemporaneous MR imaging
AJNR: 26, October 2005Letters 2433

Page 5

oxygenated clot is not seen within 3 hours. Therefore, if nec-
essary, the FLAIR HVS rather than the GRE SVS should be
used diagnostically.
Second, we aimed to assess the prognostic value of early
MR imaging vessel signs for clinical outcome, recanaliza-
tion, and intracranial hemorrhage. Here we did not take
collateral status into account. We agree with Dr. Liebeskind
that a sufficient collateral status likely is a predictor for a
good outcome, if recanalization and reperfusion can be
rapidly achieved. We believe, however, that there are also
patients with thrombus extending from the proximal middle
cerebral artery into distal MCA branches beyond the trifur-
cation. This has been seen on CT scans where the hyperat-
tenuated MCA sign extends into the MCA branches in the
Sylvian fissure (dot sign). We concur that, in acute ischemic
stroke, it is imperative to consider the location of proximal
flow cessation, the location of thrombus, and the location of
compensatory collateral flow. We doubt, however, that early
MR imaging vessel signs provide this information reliably
and arterial spin-labeling techniques for the assessment of
local perfusion are far from being routinely applicable in a
timely fashion (5 sections, 20 minutes in healthy patients)
(3).
Finally, we hypothesized that the MR imaging signal in-
tensity characteristics of early MR imaging vessel signs may
reflect the structure of the intraluminal thrombus in patients
with a vessel occlusion and can predict response to recom-
binant tissue plasminogen activator. Here, the presence or
absence of good collaterals could be a stronger predictive
parameter—and therefore a confounding variable—we take
into account. This issue would have been more important if
we had detected that clot composition as supposedly re-
flected by MR imaging EVS is a positive predictor of ther-
apeutic response. In that case a type 1 error might have
occurred.
Peter D. Schellinger
National Institute of Neurologic
Disorders and Stroke
National Institutes of Health
Bethesda, MD
and
Steven Warach
Department of Neurology
University of Heidelberg
Heidelberg, Germany
References
1. Liebeskind DS. Collateral circulation. Stroke 2003;34:2279–2284
2. Schellinger PD, Chalela JA, Kang DW, et al. Diagnostic and prog-
nostic value of early MR imaging vessel signs in hyperacute stroke
patients imaged <3 hours and treated with recombinant tissue
plasminogen activator. AJNR Am J Neuroradiol 2005;26:618–624
3. Hendrikse J, van der Grond J, Lu H, et al. Flow territory mapping
of the cerebral arteries with regional perfusion MRI. Stroke
2004;35:882–887
Are Cervical Nerve Root Blocks “Safe and
Effective”?
Wagner (1) reported on the use of CT fluoroscopy to guide
cervical nerve root blocks (CNRB). He detailed the technique
and described the initial experience at his institution. He con-
cluded that CT fluoroscopy is a safe and effective alternative
during CNRB.
We want to emphasize the serious complications associated
with CNRB. Five well-documented cases of catastrophic neu-
rologic complications following CNRB have been reported
recently in the literature (2–6). Other publications have al-
luded to several other unreported cases. Described injuries
include brain stem and midbrain hemorrhage, massive cerebel-
lar infarct, occipital cerebral edema, and 2 cases of anterior
spinal artery syndrome. Four of the reported patients died as a
result of the injuries, and 1 had permanent neurologic
sequelae.
The exact mechanism of central nervous system injury fol-
lowing CNRB is not known. In the case reports citing compli-
cations following CNRB, a standard technique of radiologic
guidance with radiocontrast material and either CT or fluoros-
copy was used. A series of mechanisms have been proposed to
explain the different neurologic injuries associated with this
procedure. One explanation is that CNRB can compromise a
radicular artery and jeopardize blood supply to the cervical
spinal cord. Possible mechanisms include occlusion by partic-
ulate corticosteroids or spasm of the blood vessel. Direct injury
of the vertebral artery, embolic microvascular occlusion by
corticosteroids, and direct neurotoxicity of radiocontrast agents
placed into vertebral intracranial circulation have been pro-
posed to explain injury to more proximal central nervous sys-
tem structures.
Cervical radicular pain is a common problem that is esti-
mated to occur in 1 of 1000 people per year. Recently, CNRB
has gained increased acceptance in the treatment of cervical
radicular pain. However, use of CNRB for radicular pain has
been advocated on the basis of only observational studies.
Although good results have been claimed, this treatment has
not been subjected to rigorous controlled studies. Further-
more, CNRB has not been proven to change the natural history
of cervical radiculopathy or any other disorder associated with
chronic spinal pain.
Given the increased recognition of serious complications
and the scarce evidence supporting the effectiveness of CNRB,
it is imperative for clinicians to examine the safety and efficacy
of this procedure and evaluate potentially safer forms of treat-
ment. An objective assessment of the risks and benefits of
CNRB needs to be performed before continued performance
of this procedure can be justified. Although many practitioners
consider the risk of spinal cord injury and stroke extremely
small, the exact incidence and mechanism of injury during
CNRB are not known. There is an urgent need to determine
the incidence of severe complications, the circumstances asso-
ciated with such complications, and assess the safest manner to
perform CNRB. Only then will we be able to state with confi-
dence that this is a safe and effective procedure.
Juan Santiago-Palma
Orthopedic Associates of
Kankakee Spine and Pain Center
Bradley, IL
Ricardo Vallejo
The Millennium Pain Center
Bloomington, IL
Craig Kornick
Riverside Spine and Pain Physicians
Jacksonville, FL
Steven Barna
Massachusetts General Hospital Pain Center
Harvard Medical School
Boston, MA
References
1. Wagner AL. CT fluoroscopic-guided cervical nerve root blocks.
AJNR Am J Neuroradiol 2005;26:43–44
2. Tiso RL, Cutler T, Catania JA, et al. Adverse central nervous
system sequelae after selective transforaminal block: the role of
corticosteroids. Spine J 2004;4:468–474
2434 LettersAJNR: 26, October 2005