REVIE\N

Cerebrovascular
• •neurolmaglng:

""hich tests and
""hen?

Michael
Hoffmann

Durban Cerebrovascular Group,
Department of Surgery,

University of Natal, Durban

Cerebrovascular
neuroimaging
requires careful
clinical and
rad iological
correlation for
interpretation.
With a
profusion of
tests avai lable,
what is the best
investigation
and when?

Determination of brain pathology and
pathophysiology was for a long term largely
deductive with no objective evidence. Early
neurologists had to wait for autopsy infor-
mation for a definitive diagnosis. Both the
impervious, intact skull and the dearth of
clinically useful brain treatment strategies
contributed to the "dark ages of the brain".
The situation today is vastly differentin that
we are overwhelmed with an ever increas-
ing array of new investigative tools. These
include colour duplex Doppler flow
imaging (CDFI), transcranial Doppler
(TCD), magnetic resonance imaging (MRI),
magnetic resonance angiographic imaging
(MRA), single photon emission computed
tomographic imaging (SPECf), spiral com-
puterised tomography (helical or spiral Cf)
and positron emission tomography (pET).
Coincidentwith such a wide choice ofbrain
imaging methods, are the imminent acute
stroke therapies (antithrombotics,
thrombolytics, neuroprotective and gene
based therapies) that are antidpated within
the next few months to years. Despite the
complexity of the brain, outward manifes-
tations of impairment remain limited to
long tract signs,visual deficits and cognitive
deficits, underscoring the need for
paraclinical evidence of malfunction.

Colourduplexflowimaging (CDF!) has
become a routine test for extra cranial arte-
rial assessment with excellent sensitivity and
specificity as compared to angiography.'
Plaque characteristics are most accurately

6 SA JOURNAL OF RADIOLOGY. September 1996

assessed with CD FI and area stenosis as
measured by Doppler as opposed to diam-
eter stenosis as measured by angiography
is a more accurate correlate of the true ana-
tomical degree of stenosis.2 A recent and at
present still clinical investigation extension
of CDF! is power Doppler imaging (POD.
This new modality generates intravascular
colour signals from the amplitude of the
echo signal, which in tum depends on the
intensity of the sampled red blood cells.It
overcomes the two major limitations of
CDFI which are the assessment of high
grade stenosis and the intrastenotic diam-
eter and area reduction in heavily calcified
plaques, which are commonly encoun-
tered. In addition the method is angle in-
dependent and free from aliasing artifacts
which is a limitation of CD PI. The POI
contrasts the lumen with colour similar to
angiography. In this respect it essentially
bridges the gap between angiography and
CDF! whose major limitations were inabil-
ity to image the outer vessel boundary and
residual vessellumen."

Transcranial Doppler (TCD) has estab-
lished applications for the reliable diagno-
sisof intracranial vessel stenosis,vasospasm,
Willisian collateralisation and
arteriovenous malformations. Inaddition,
it is useful for the detection of vasomotor
testing and monitoring and brain death.' A
more recent and highly clinically relevant
application is in the detection of
intracranial emboli, whether of cardiac,
large artery or exogenous source. The di-
agnosis of patent foramen ovale, an impor-
tant predisposing factor in a significant pro-
portion of young stroke patients of "un de-
termined aetiology", is most easily and
cheaply done by contrast ("bubble'] TCD.6
The differentiation of symptomatic from
asymptomatic carotid stenoses can also be
fadlitated by determining the degree of ar-
tery to artery embolisation byTCD by per-
forming HITS (high intensity transient sig-
nals) counts. 7 Both duplex and powerTCD

t:-:o~pa':"':'g-e7:---......



REVIE\I\I

frompage6

are currently in the clinical investigational
stage with promise for considerable widen-
ing of their applications

Magnetic resonance imaging has af-
forded the excellent anatomical detail of the
brain especially in areas which do not show
up weil on computerised tomographic
(CT) scanning such as the brainstem, cere-
bellum and spinal cord. Routine MR
imaging by spin echo sequences, the best
known beingT1, T2 and proton density
weighted images.Tl images enable high tis-
sue to tissue differentiation (anatomy), T2
images are most useful for intratissue char-
acteristics (pathology) and proton density
weighted images aid differentiation of fo-
cal tissue changes adjacent to free fluid such
as the subarachnoid space and ventricles.
This has resulted in unheralded accuracy
of diagnoses of posterior circulation
infarction especially of the brainstem, cere-
bellum and diencephalon. Thrombolysis
with tissue plasminogen activator is now
indicated for a subset of acute stroke pa-
tients'The accurate, early detection of this
treatable subgroup of patients is depend-
ent on the appropriate investigative tech-
nique.As a consequence, the requirement
for ultra early detection of infarctive and
haemorrhagic components is of vital impor-
tance. Two new investigative techniques,
MR diffusion (MRD) and MR perfusion
(MRP) allow tissue changes of cytotoxic
oedema to be seen within 1 hour of onset. 9
Ultrafast echoplanar imaging may reduce
this to about 30 minutes and allowsimaging
of ischaemie tissue that has been experi-
mentally reversed in the animal model by
drugs such as NMDA antagonists.P!' Al-
though not routine in clinicalpractice, these
methods hold promise, MR angiography
(MRA) has already reached routine clini-
cal practice with the two principle tech-
niques; namely, time of flight (TOF) and
phase contrast (PC) each having their par-
tienlar attributes" Most usefully combined
with transcranial Doppler and duplex

Doppler to grade stenoses, a two and three
dimensional cerebrovascular tree can be
imaged entirely noninvasively. MRA adds
only about 10 minutes to the conventional
MRl scan of the brain and it is easy to see
that this may rapidly become a routine part
of cerebrovascular imaging. Finally, mag-
netic resonance spectroscopy (MRS) al-
though not yet in clinical use, may detect
the earliest possible changes of ischaemia
at cellular level by measuring increases of
lactate and a fallin high energy phosphates. 13

Computerised tomographic brain
scanning has improved with respect to defi-
nition and speed of examination with each
new generation of scanners. Because of its
inherent drawbacks itis relegated to the task
of characterisation of sudden brain syn-
dromes into either mass lesions or stroke,
and in the case of stroke, to differentiate
haemorrhage from bland infarction, Ho~-
ever due to its popularity and relatively
lower cost it will probably retain its posi-
tion as the most widely used brain scan-
ning method. The novel application ofheli-
cal CT scanning of blood vessels has be-
come so useful clinically that even dissec-
tion of carotid arteries, until recently the
domain of conventional angiography,can be
reliably diagnosed by this method."

SPECT has been shown to be a sensi-
tive indicator of cerebral perfusion and the
most sensitive brain scan for the demon-
stration of ischaemia and infarction in the
acute phase. Three different tracers are
available for clinical use; HMPAO, Iodine
123 and Tc ethyl cysteinate dimer. The
radionucleotide is injected intravenously
and accumulates in different areas of the
brain proportionate to the rate of delivery
of nutrientslblood flow to that volume of
brain tissue. Technology has allowed the
combination of gamma cameras and com-
puterised tomography to create three di-
mensional construction of images. The
major advantage ofSPECT is its functional
imaging capadtywhich complements the

7 SA JOURNAL 'OF RADIOLOGY. September 1996

standard anatomical imaging using mag-
netic resonance and the older computer-
ised tomography scans. SPECT is also af-
fordable and considerably cheaper than
positron emission tomography (pET). The
latter is globally limited by cost and the ne-
cessity for proximity to radionucleotide
processing (a cyclotron). In fact, the cost of
SPECT is similar to the conventional CT
scan without contrast. The major advantage
of PET over SPECT isimproved resolution
and its capability for the measurement ofre-
gional cerebral metabohsm. Biochemical in-
formation from receptor activity measure-
ment such as muscarinic, benzodiazepine,
serotonergic and dopaminergic receptor sys-
tems are possible but are not in routine clini-
cal use and remain experimental atthe time
ofwriting IS

The quest for pathophysiological
subtyping in acute stroke has never been
more urgent. The long hst of animal-effec-
tive, human-failure stroke drugs has been
blamed squarely on the unreahstically long
time window of24 - 48 hours or more for
patient recruitment. Now that we appreci-
ate that "time is brain" and strive for sub 6
hour recruitment period for acute stroke
intervention, this may nevertheless be un-
necessarily rigid and exclude potentially
salvageable patients with slightly longer
time frames, 16 Such subtyping can only be
done with the strategic choice of a combi-
nation of functional and structural
neuroimaging. By means oITCD andMRA
the site of arterial occlusion can be defined
in order to determine the desirability of
thrombolytic therapy and to exclude pa-
tients with distal arterial or branch occlu-
sion or those without occlusion. Tissue vi-
ability can be determined to some degree
by early SPEer, diffusion weighted MRl
and extent of hypo perfusion can be de-
duced by Cl.The exact choice, timing and
interpretation of these modalities remains
within the domain of the art and science
ofmedidne.

lopage8



REVIEVV

!rampage 7
All the tests have the capability of me as-

uring either anatomical or functional de-
rangement of the brain or indeed both and
many have the added capability of depict-
ing pathophysiological events in three axes
(x, y and z axes) and four (time) dimen-
sional domains. In combination with the
specialities of cardiovascular medicine and
vascular surgery it can readily be appreci-
ated that the heart and the entire
cerebrovascular tree can be evaluated
noninvasively.This constitutes not only the
heart, but also the aortic arch and its major
.branches, the cervicocephalic vessels (ver-
tebral arteries Vl- V3, common, extemal
and internal carotid), the intracranial carotid
(siphon region and V 4 vertebral sections)
and the circle of Willis and its major
branches. Conventional angiography today
seems destined largely for the exclusion of
infrequent vasculitic disorders and clarifica-
tion of discrete anatomical features of aneu-
rysms and arteriovenous malformations.

The current approach to
cerebrovascular diagnostics demands an
immediate appraisal of the stroke mecha-
nisms as all early and secondary preventa-
tive treatment is entirely dependent on
determination of the pathophysiology. In
the acute stroke situation, the initial three
questions the clinician has to answer before
initiating treatment are - is it a stroke (and
not a neoplasm, seizure or migraine), if it is
a stroke, is it bland infarction or haemor-
rhage; and ifblandinfarction what subtype
(because these have different treatments).
Broad categories include 1) small vessel
atherosclerotic disease (lacunarj.Z) large
vessel atherosclerotic disease (carotid, ver-
tebral or intracranial stenoses), 3)
cardioembolism (disrhythmias, valvular,
dyskinetic segments, paradoxical embo-
hsm),4) other (dissection, prothrombotic,
vascuhtic) and 5) undetermined groups.
Briefly, treatment options include:
antiaggregant therapy and risk factor con-
trol for small vessel disease, carotid

endarterectomy and at times aortic branch
vessel bypass for large vessel disease. Most
cardiac diseases are eminently treatable:
antithrombotics for dissection, anticoagu-
lants for many prothrombotic disorders and
immunosuppressive agents for many of the
vasculitides constitute effective albeit not
always curative treatment today.

Using combinations of the
neurodiagnostic armamentarium to advan-
tage is the key to accurate diagnosis and
hence cost efficacy.The latter isparticularly
relevant given current (and very likely fu-
ture) financial restraints and managed
health care. In the clinical context it may
be readily appreciated how MRA may
clearly depict a stenosis in a large brain sup-
plying vessel which may be graded by
transcranial Doppler. Likewise, spiral CT
may complement duplex Doppler of a
brain supplying cervicocephalic vessel. In
both instances invasive intra-arterial
angiography is obviated. Transcranial Dop-
pler and duplex Doppler not only rapidly
and cost effectively diagnose large artery
disease but may also diagnose high inten-
sity transient signals (HITS), diagnostic of
emboli in most instances and whether the
embologenic site is cardiac,major vessel (ca-
rotid bifurcation) or intracranial stenosis.
SPECT andTCD have been usefully com-
bined for prognostication and stroke
subtype classification,to guide therapy and
to deduce a cerebral perfusionindex." Both
TCD and SPECT may be used to estimate
cerebrovascular reserve using CO2,
Acetazolamide or paper bag rebreathing as
vasodilatory agents.' Such subtyping of cer-
ebral ischaemie symptomatology allows
refinement of treatment options such as
earlier and more appropriate carotid
endarterectomy (CEA). SPECT iscurrently
the best early stroke imaging test and pro-
vides objective evidence for diaschisis type
neurological deficits, seen with a variety of
stroke patients both cortical and
subcortical.' 5

S SA JOURNAL OF RADIOLOGY- September 1996

As one of many possible illustrative ex-
amples, it is easy enough to appreciate that
a clinical diagnosis of posterior circulation
ischaemia or minor infarction (so called
vertebrobasilar insufficiency) is rarely made
with confidence. Such patients are notori-
ously difficult to diagnose and usually have
had numerous speciality consultations and
a long hst of costly nondiagnostic investiga-
tions, mainly because they present with a
plethora of symptoms that may fluctuate
with time. Even isolated tinnitus and ver-
tigo may be presenting features and mas-
querade as otologic problems, but more
commonly a nondescript "wooziness" is
reported, much to the despair of the clini-
cian. With a clinical suspicion though, in
the absence of other disease, MRA of the
vertebral and basilar arteries, usefully com-
bined with transcranial Doppler might de-
tect the responsible vertebral artery origin
disease or basilar stenosis. Antiaggregant
treatment may be then prescribed with
conviction or changed to Tielepidine or
Warfarin as required. Both doctor and pa-
tient will benefit from more precise diag-
nosis of the neurological symptomatology.

Biotechnology has rewarded us with
remarkably clear pictures of the brain struc-
ture and function that impact on manage-
ment. The most challenging issue facing the
clinician and radiologist today isthe knowl-
edge of discriminatory use of these
modalities. Clearly, no exact step by step
flow diagram for stroke investigation can be
presented as every stroke patient is differ-
ent. Investigation must be tailored and in-
dividualised. Stroke should be recognised
as a multidisciplinary syndrome and so too
isthe investigation. The way forward seems
certain to be a close multidisciplinary hai-
son for each individual patient. The effec-
tive, orderly use of tests can save time, speed
up diagnosis and save the patient from un-
dergoing unnecessary, expensive and at
times potentially harmful tests. Despite the
ever increasing fractionation of medical sub

to page 15