Vous êtes sur la page 1sur 5

Trigeminal Neuralgia Guidelines

John R. Adler, MD
Stanford University

The following description outlines the CyberKnife rhizotomy method used for trigeminal neuralgia (TN)
at Stanford. I welcome comment.


The initial challenge is to meticulously define the retrogasserian sensory root of CN V in the prepontine
cistern. Imaging the nerve was originally done with CT contrast enhanced cisternography alone.
However, after detailed comparison studies, and now several years of experience, thin section MRI
scanning using a thin slice FIESTA (heavily T-2 weighted) sequence has largely replaced
cisternography at Stanford for identifying the trigeminal nerve target. Multiplan image fusion is utilized
to develop a composite data set for final target definition.

In the occasional patient in whom MR imaging is not possible, most commonly because of a pacemaker,
CT cisternography is still performed. In such patients the procedure begins with a lumbar puncture and
the instillation of 5-8 cc of iodinated contrast (IsoView 300M) into the lumbar subarachnoid space. The
patient is kept in a prone trendelenberg position for 10-20 minutes allowing contrast to traverse the
foramen magnum. Ideally, but not necessarily, this can be monitored under fluoroscopy. Subsequent
to the dye equilibrating throughout the intracranial cisterns, the patient is expeditiously imaged with thin
slice high resolution CT through the entire head as is typically done for any other routine intracranial
CyberKnife case.

Fig 1 Depicts a typical FIESTA-weighted MR (left) and a CT cisternogram (right) used for identify the
right trigeminal nerve prior to CyberKnife rhizotomy
Target Delineation

Defining the optimal target is generally the biggest technical challenge confronting the novice who
embarks upon the initial treatment of patients with trigeminal neuralgia. Even with ideal MR imaging the
anatomy can be far from ideal in some patients. However, with experience it should be possible to
delineate the prepontine trigeminal nerve (i.e. retrogasserian sensory root) in all patients. A few
important tricks for this include:

1) Identify the opposite trigeminal nerve with the understanding that anatomy tends to be
2) It is almost always possible to find the trigeminal eminence where the dorsal root merges with
the lateral pons.
3) It is generally straightforward to find Meckel’s cave in every patient by virtue of its characteristic
notch on CT, and on MRI, the splaying of the three trigeminal divisions.
4) Once the nerve has been identified at the brainstem and within Meckel’s cave, it is a fairly safe
bet to conclude that the trigeminal sensory root will travel from one to the other.
5) Don’t just check the target volume on axial slices; it is important to reference the reconstructed
coronal and sagittal images throughout the entire process of target definition.
6) Generally speaking the target volume will be encompassed by 2 to 3, two mm thick MR slices

The most common mistakes include:

1) Mistaking adjacent blood vessels for CN V itself; these are often the offending vessel
responsible for producing the disorder in the first place. Note that vessels can be traced from
slice to slice, and will enhance on CT, which is a good reason to use contrast for CT scanning.
2) In previously operated (MVD) patients, the delineation of the intracisternal nerve segment can
be quite challenging; in such cases the teflon pledget distorts the normal anatomy and the mass
itself often obscure the nerve. Such cases require extra diligence but following the above rules
will generally keep one from going too far astray.

It is critical to realize that the process of delineating the trigeminal nerve is never a simple “paint by
numbers” process. In fact, many novices can badly misplace the target resulting in either an ineffectual
or dangerous procedure. However, credible target definition by a knowledgeable user is almost always
possible utilizing contemporary imaging.

The Target

There is no universally accepted portion of the trigeminal nerve that is best lesioned when
radiosurgically operating on trigeminal neuralgia. Anecdotally, lesions closest to the brainstem result in
more numbness, while those closer to the semilunar ganglia in Meckel’s cave are associated with a
higher incidence of unpleasant dysesthesia. At Stanford we choose to lesion a nerve segment that
spares the proximal 2 to 3 mm adjacent to the brainstem-as measured along the length of the nerve.
The lesion should extend over a 6 to 6.5 mm segment of the sensory root within the prepontine cistern
as measured on the reconstructed sagittal imagery; generally speaking this is the maximal length of the
target. We have been reasonably generous in defining the lateral margins of the nerve and rather
routinely use a width of ~3mm even if the nerve is slightly thinner. We reason that it is safe to be
generous in this dimension, and by doing so, one may be able to compensate for some inaccuracy in
treatment delivery. The dimensions of the final target volume are 6.5 x 3 x 3 mm and the volume of this
often times staircase region (on reconstructed sagittal imaging) varies between 0.035 and 0.040 cc.
Treatment Planning

The specifics of treatment planning have evolved significantly over the years, and may still evolve in the
future, but currently are outlined as follows:

Node Set: head 650_trigeminal node set

Collimator: 5 mm (remember this is defined at a nominal SAD of 80 cm)
Targeted segment: ~6.5 mm of prepontine 5th nerve on the SYMPTOMATIC SIDE: the targeted
segment of nerve begins 2-3 mm from the DREZ.
Critical structures: The brain stem (pons) beginning 1 cm above and ending 1 cm below the level of the
trigeminal nerve is delineated. Brainstem dose is kept to <4000 cGy but the constraint index is only set
at strict (generally not more than 0.3 cm3 receiving a dose of 12 Gy or higher; i.e V12 of 0.3 cm3.
Furthermore, the trigeminal/semi-lunar ganglia should be defined as a volume of anatomy within which
dose should be minimized (generally 38-40 Gy maximum point dose). This requirement need not be
strictly followed since only full dose to the ganglia appears problematic. The ipsilateral cochlea and
7th/8th nerve complex should also be contoured and limited to approximately 5-8 Gy and 10-15 Gy
Dose: Performed with a high resolution dose calculation matrix
Prescribed dose: 6000 cGy to ~80th percentile, covering 95% of the delineated target
Dmax: 73-75 cGy

Target(s) Total Prescription Dose Number of Sessions

left trigeminal nerve 60 Gy to 80% ISL 1
right trigeminal nerve

Critical Structure Dose volume constraints Maximum dose constraint

brainstem ≤ 0.3 cm3 ≥ 12 Gy 40 Gy
left cochlea ≤ cm3 ≥ Gy 10 Gy
right cochlea
ipsilateral ≤ cm3 ≥ Gy 40 Gy
Gausserian ganglion
7th and 8th nerve complex ≤ cm3 ≥ Gy 15 Gy

The above treatment scheme solves in 30 minutes or less and utilizes 140-160 beams with max
MU/beam set between 100 and 140 MU.

A range of dose is specified in part because of continued uncertainty regarding optimal dose and to
allow users to respond to variable clinical conditions; it remains possible that increased dose might
correlate with faster pain relief. In addition, the planning system is rarely able to arrive at a solution
exactly matching too narrowly specified values. Where necessary we compromise on percentage of the
target covered by the prescribed dose.
Figure 2 shows an example of a satisfactory treatment plan.

While we permit the 50% isodose line to touch the lateral aspect of the brainstem, it is an absolute
requirement that this isodose line does not intrude/overlap the brainstem. Attention must also be paid
to the dose reaching the undersurface of the adjacent temporal lobe.
NOTE:….It is absolutely critical to pay attention to the lower isodose lines (e.g. 20th percentile)
prior to accepting a plan as adequate!! The treatment planning system can generate a plan that is
completely satisfactory at the higher dose levels covering and immediately adjacent to the treated nerve,
but still allow an unacceptable dose within a too large volume of brainstem; the risk in such cases is a
serious brainstem injury. Generally speaking we have found any plan that limits <0.3 cm3 of
brainstem to >12 Gy will be safe.


The clinical benefits of the above radiosurgical technique will occasionally become noticeable in the first
couple days. With current dose schemes, the time to pain relief ranges from a few weeks up to 3
months after treatment. We generally wait until one month after SRS to see patients back in follow-up
clinic, and during this outpatient visit, anti-epileptics and other pain relief medication are gradually
tapered. In some patients, TN pain remission will require even more time; we have seen brain relief in
one patient take as long as eight months. At Stanford we wait least 6 months before declaring the
procedure a failure. Approximately 10% of patient will not get any meaningful benefit from an initial
effort at CK rhizotomy. We have generally elected to retreat them with a CyberKnife rhizotomy; at the
time of retreatment cases the dose is typically lowered 10 Gy, using a marginal prescription of 50 Gy.


It should be emphasized that, although all radiosurgery requires diligence to be done safely and
effectively, trigeminal rhizotomy is especially exacting. There is little room for error. However, when
done correctly, there can be few clinical outcomes that are more rewarding for both patient and