Child's Nervous System

https://doi.org/10.1007/s00381-020-04746-9

FOCUS SESSION

Selective dorsal rhizotomy: functional anatomy of the conus-cauda

and essentials of intraoperative neurophysiology
Claudia Pasquali 1 & Vedran Deletis 2,3 & Francesco Sala 1

Received: 6 June 2020 / Accepted: 11 June 2020

# Springer-Verlag GmbH Germany, part of Springer Nature 2020

Abstract
Introduction Spasticity is the result of an exaggeration of the monosynaptic muscle stretch reflex due to lesions affecting the
central nervous system, in particular an upper motor neuron lesion. Selective dorsal rhizotomy (SDR) is a surgical technique
developed to treat spastic diplegia, one of the common forms of cerebral palsy, resulting from the lack of supraspinal inhibitory
controls. The aim of SDR is to identify and cut a critical amount of the sensory rootlets, in particular those contributing the most to
spasticity, in order to relieve the patient from lower limb spasticity while preserving motor strength and sphincter control. Various
surgical techniques to perform SDR have been proposed over time. Similarly, intraoperative neurophysiology (ION)—first
introduced by Fasano and colleagues in 1976—is a safe and effective tool to guide the surgeon in the procedure of SDR, but
different ION strategies are used by different authors, and the value of ION itself has been questioned.
Methods The purpose of this paper is to review the anatomo-physiological background of SDR, the historical development of the
surgical technique, and the essential principles of ION.
Results While some surgeons privilege a single-level approach and others a multi-level approach, nowadays, there are still
neither agreement nor guidelines on the percentage of roots to be cut. Rather, a tailored approach based on both the preoperative
functional status as well as intraoperative ION findings seems reasonable. ION is considered not essential to decide the percent-
age of roots to cut, but it assists to distinguish between ventral and dorsal roots, and to preserve sphincterial function, whenever
S2 rootlets are included in SDR.
Conclusions To optimize the balance between reduction of spasticity and preservation of motor strength while minimizing the
neurological damage remains the main goal of SDR.

Keywords Selective dorsal rhizotomy . Intraoperative neurophysiology . Neurophysiological mapping . Cauda equina .
Anatomy . Spasticity . Cerebral palsy

Introduction to the developing brain that cause life-long impairments of

movement and posture.
Selective dorsal rhizotomy (SDR) is a surgical technique de- Despite progress in neonatal care, the incidence of cerebral
veloped to treat spastic diplegia, which is one of the common palsy is increasing, affecting approximately 2–4 children per
forms of cerebral palsy. Cerebral palsy (CP) is a heteroge- 1000 live births, being the most common cause of physical
neous group of neuromotor disorders resulting from an insult disability in childhood [1]. The neurological insult is not
evolutive, but the clinical manifestation of cerebral palsy is
subject to change as children and their developing nervous
* Francesco Sala system mature with a maximum expression between 7 and
[email protected]; [email protected] 10 years of age.
Cerebral palsy was described for the first time as a clinical
1
Section of Neurosurgery, Department of Neurosciences, entity in 1861 by William John Little, a British orthopedic
Biomedicine and Movement Sciences, University Hospital,
Verona, Italy
surgeon, who wrote a monograph in which he proposed an
2
association between perinatal asphyxia and poor neurological
Department of Neurosurgery, University Hospital Dubrava,
Zagreb, Croatia
outcomes later in life [21]. Three decades later, Sigmund
3
Freud and Sir William Osler, observing that children with
Albert Einstein College of Medicine, New York, NY, USA
 Childs Nerv Syst

cerebral palsy had multiple medical comorbidities such as At present, there are also some evidence supporting the
intellectual disability, epilepsy, and visual disturbances, pro- theory that multiple genetic factors contribute to the cause of
posed that cerebral palsy could begin during fetal brain devel- cerebral palsy. In particular, there is the evidence that muta-
opment [10, 31]. tions in multiple genes result in Mendelian disorders that pres-
It is now known that the etiology of cerebral palsy could be ent with cerebral palsy-like features [24]. It is also known that
related to prenatal, perinatal, or postnatal events [24, 32] the prevalence of congenital anomalies in individuals with
(Table 1). cerebral palsy is significantly higher (11–32%) than in the
The onset period of these events is also critical to determine general population (2–3%). Many studies have recently
the kind of cerebral lesion, as summarized in Table 2. underlined that the cerebral palsy rate in monozygotic twins
Supernumerary gestations, infection, ante-partum vag- and in dizygotic twin pairs is statistically different [27].
inal bleeding, and maternal exposure to alcohol are Depending on the dominant type of tone or movement
common causes of prenatal brain injuries, which repre- abnormality, the Executive Committee for the Definition and
sent 70–80% of CP causes. Low-birth-weight infants Classification of Cerebral Palsy (2004) define the disorder as
have an augmented risk of perinatal brain injuries due spastic, dyskinetic (dystonia or choreoathetosis), ataxic, or
to fetal anoxia, bradycardia, hyperbilirubinemia, hemor- mixed groups [4]. The spastic type, either in the form of di-
rhage, and infection. The developing brain is subject to plegia or asymmetric quadriparesis, is the most common, ac-
injuries until about 3 years of age and traumatic brain counting for approximately 60–80% of cases. The dyskinetic
injury, tumors of the CNS, and near-drowning, as well group has a prevalence ranging from 3 to 17%, and then, the
as meningo-encephalitis, could be responsible for post- ataxic group represents about 10% of children [23, 38].
natal onset of cerebral palsy [32, 51]. The mechanism of injury and the timing can define the type
Cerebral hypoperfusion tends to result in periventricular of disorder: in children born at or before 30 weeks of gesta-
leukomalacia which damages the structural integrity of the tion, cerebral palsy is typically characterized by spastic diple-
corticospinal and other descending tracts, and in the develop- gia. Children born later usually exhibit rigidity, dystonia, or a
ment of germinal matrix hemorrhages, which often will lead mixed type of hypertonia [1, 30].
to asymmetric deficits. So, although the diagnosis of CP is Nowadays, children with cerebral palsy can benefit, ac-
clinical, MRI can contribute to define the diagnosis in equiv- cording to the type of hypertonia, of several treatment options.
ocal cases [14, 33]. None of these treatments can definitely repair the damaged

Table 1 Etiology of cerebral palsy (adapted from Panteliadis CP, Cerebral palsy – a multidisciplinary approach (3rd ed.), 2018, Springer International
Publishing (8))

Prenatal Perinatal Postnatal

Maternal Vascular, maternal epilepsy and mental Labor complications (infections)

retardation, cervical or third-trimester bleeding,
maternal trauma and infection during
pregnancy, antiphospholipid autoantibodies,
fertilization problems
Infection Congenital infections of the neonate (herpes, Bacterial and viral neonatal Bacterial and viral CNS infections
toxoplasma, rubella, CMV, syphilis) infections
Metabolic Iodine deficiency, maternal thyroid disfunction Hyperbilirubinemia (kernicterus), Hypoglycemia
hypoglycemia, vitamin K
deficiency
Prematurity Prematurity, intracranial
associated hemorrhage, periventricular
leukomalacia
Birth associated Placental complications,
hypoxia-anoxia, trauma during
labor, meconium aspiration
syndrome
Teratogenic Maternal drug, alcohol abuse High-dose steroids in preterm
drugs infants
Genetic Migrational disorders, chromosomal Thrombophilia, metabolic Thrombophilia, cytokine polymorphisms,
abnormalities, family history of CP or epilepsy diseases metabolic and neurodegenerative
diseases
Other Intrauterine stroke, intrauterine hypoxia, placenta Cerebral infarction, venous sinus Near drowning, suffocation, cardiac arrest,
previa, multiple pregnancies, low birth weight, thrombosis CNS injuries (head trauma), acquired
intrauterine death of a co-twin thrombophilia, infarction, child abuse
Childs Nerv Syst

Table 2 Types of pre-, peri-, and

postnatal hypoxic-ischemic brain Lesion Onset period
injury in children and their date of
origin Periventricular Perinatal in preterm births, declining prevalence with increasing
leukoencephalopathy gestational age
Germinal/ventricular hemorrhage Perinatal, predominantly in preterm neonates
Por-/hydranencephaly 5th month of gestation
Pontosubicular necrosis 30th gestational week to 2nd postnatal month
Border zone infarction/ ulegyria Mature neonate
Territorial infarction Mature neonate to infancy
Marbled state Perinatal period until the age of 6–9 months
Multicystic encephalopathy Early infancy
Global hemispheric necrosis Postnatally up to an age of 18 months

brain areas that control muscle coordination and movement, IV, V, VI, and IX, the smaller ones project into laminae I, II,
but can improve the quality of daily life and functional per- V, and IX.
formance of these children. The “essence of proprioception” lies in lamina VI and in
Well-known treatments include physical and occupational minor part in lamina IX: the I-A fiber project directly into
therapy, oral myorelaxant, botulinum toxin, intrathecal baclo- these layers and the interneurons give rise to the ascending
fen therapy, orthopedic surgery, and SDR. pathways and the activation of alpha-motoneuron and the
Over the decades, the SDR technique has changed and has myotatic reflex.
been perfectionated also thanks to the advent of intraoperative The epicritic tactile and conscious proprioception stimuli
neurophysiology, but the main target remains the dorsal feed- are carried by A-beta fibers located dorsomedially in the dor-
back system of the sensory rootlets from T12 to S2. These sal root entry zone (DREZ) and they give rise to the lemniscus
roots are grouped in the cauda equina at the distal end of the pathway which reaches the brainstem.
spinal cord. The nociceptive A-delta and C fibers project to the ventro-
The aim of SDR is to specifically identify and cut a critical lateral region of the DREZ and immediately they decussate
amount of the sensory rootlets, in particular those contributing and ascend to the brainstem in the Lissauer tract.
the most to spasticity, in order to reduce spasticity while pre- The laminae I and II release inhibitory amino acid, all the
serving motor strength and sphincter control. others have mainly excitatory effects.
In this paper, we aim to review the anatomo-physiological In the intermediate layers of laminae (VII and VIII) there
background of SDR and the essential principles of intraoper- are the cell bodies of spinal interneurons: they receive periph-
ative neurophysiology, when performing the procedure. eral information and central information from the descending
pathways and then modulate muscular response sending stim-
uli to the motoneurons of limbs through the dorso- and ven-
trolateral tract. In particular, the dorsolateral interneurons pro-
Functional anatomical background ject to the flexor muscles and the ventrolateral ones to the
extensor muscles.
The gray matter of the spinal cord can be functionally classi- In the laminae IV, VI, and IX, the neuronal axons of the
fied in different ways: into four columns, or into six nuclei, or Clarke nuclei give rise to the spinocerebellar tracts. The fibers
into ten laminae [23, 30, 38] (Fig. 1 and Table 3): ascend in the lateral columns divided into fasciculus of
Gowers, ventral, which carry proprioceptive sensation from
Spinal cord neurophysiology contralateral limbs and into fasciculus of Flechsig, dorsal, car-
rying information from the ipsilateral trunk.
The alpha-motoneurons are located in the ventral horn of the The spinoreticular tract originates from the deep layers of
spinal cord and they innervate different muscle groups accord- laminae VII, VIII, and X and carries ipsi- and contralateral
ing to a somatotopic organization (Fig. 2). From medial to fibers. Its connections with the reticular formation have a role
lateral columns, we find alpha-motoneurons for axial muscles in motor control and not only in nociception [16, 38].
and then upper and lower limb muscles with the motoneuron
for the extensors muscles located more ventrally than flexor Supraspinal control
ones.
Except for lamina VIII, all the laminae receive primary The descending pathways have a role in the generation and
afferents directly: the major fibers project to the laminae III, modulation of segmental spinal cord motor output.
 Childs Nerv Syst

Fig. 1 Laminae (left) and nuclei

(right) organization

Numbers of corticospinal fibers project into alpha- are from 11.1 to 7.7 rootlet for each segment from L1
motoneurons of the ventral horn, some of them reach the lam- to S2. The lumbosacral roots originate from or terminate
inae IV to X and spread over the intermediate zone and the in the conus medullaris which is the cone-shaped termi-
dorsal horn. nal end of the spinal cord whose tip is usually located
Also, the rubrospinal and the reticulospinal pathways orig- on average at L1-L2 level. The conus medullaris gives
inating from the gigantocellular nucleus of the pons and retic- rise also to the lumbar sympathetic, sacral somatic, and
ular formation descend into the dorsolateral column and pro- sacral parasympathetic nerves. It continues with the
ject contralaterally to lamina I and the lateral part of laminae epiconus (L4 to S1 segments) which consists of S2 to
V, VI, VII, and IX. S5 as well as the coccygeal segments.
An important role is played by the reticular formation Axons of 3–5 lumbar, 5 sacral, and the single coccygeal
whose neuronal bodies are located into the brainstem tegmen- nerve give rise to the cauda equina: a bundle of spinal nerves
tum. Its function is modulatory and premotor and it performs and spinal nerve rootlets which contains also the filum
is through the ascending pathways directed to the cortex (as- terminale. This latter is a delicate strand of fibro-elastic tissue
cending reticular activating system (ARAS)) and the descend- about 20 cm in length which takes its origin during the em-
ing ones directed to the spinal cord (reticulospinal tracts). bryogenesis from cellular degeneration and it has a stabilizing
This is the major impaired circuit in cerebral palsy. This role connecting the conus to the coccyx through the coccygeal
fact is underlined by the study on H-reflex, a reflex evoked by ligament.
percutaneous nerve stimulation, which shows a reduced laten- In the cauda equina, motor ventral nerve root and sensory
cy period in patients affected by cerebral palsy, due to a re- dorsal root come together in the lateral portion of the dural sac,
duced supraspinal inhibitory control [8, 22, 49]. and they finally leave the dural sac through the dural root
sleeve containing an extension of the subarachnoid space. At
the end of the root sleeve, the roots separate into a number of
Functional anatomy of conus-cauda region fascicles: the sensory components blend in the dorsal root
ganglion, while the motor fibers merge into the mixed spinal
Muscles and dermatomes of the lower extremities are nerve distal to the dorsal root ganglion as the nerve is formed.
innervated by lumbosacral segments of the spinal cord. Its ventral ramus becomes part of the lumbosacral plexus
Spinal nerves arise from the union of dorsal and ventral while the dorsal ramus innervates local spinal structures.
roots of a spinal cord segment. There is a pair of gan- It is important to remember that ventral roots from Tl to L2
glia, named dorsal root ganglia, for each segment of the and S2–S4 contain a group of thinly myelinated axons which
spinal cord. Each dorsal root is composed by a series of are part of the sympathetic and parasympathetic systems [12,
rootles, and through them, the afferent (sensory) fibers 17, 53].
reach the spinal cord. Likewise, the ventral root origi- Numeration of lumbar nerve roots depends on the vertebra
nates as a series of rootlets, and through them, the ef- under whose pedicle the roots enter the intervertebral foramen.
ferent (motor) fibers leave the spinal cord. There is a Table 4 summarizes for each nerve root the innervated
certain interindividual variability but, on average, there muscle/s and the related function.
Childs Nerv Syst

Table 3 Spinal cord organization

Columns 1. The dorsal horn (posterior horn) which neurons receive somatosensory information from
organization the body, and then transmit these information to the brain through the ascending
pathways.
2. The ventral horn (anterior horn) containing alpha-motoneurons which, abandoning the
spinal cord, innervate skeletal muscle.
3. The intermediate column and the lateral horn whose neurons innervate visceral and
pelvic organs.
Nuclei 1. Marginal zone: at the tip of the dorsal horn, transmits pain and temperature sensation to
organization the brain.
2. Substantia gelatinosa: located at the top of the dorsal horn, transmits pain, temperature,
and light touch sensation to the brain.
3. Nucleus proprius: located in the superior-third of the dorsal horn, transmits mechanical
and temperature sensations to the brain.
4. Dorsal nucleus of Clarke: the most dorso-medial nuclei located only from C8 to L3
segments, relays unconscious proprioceptive informations to the brain.
5. Inferomediolateral nucleus: situated in the intermediate column and lateral horn, it sends
sensory information from viscera to the brain, and also transmits autonomic signals from
the brain to the visceral organs.
6. Lateral motor neurons and medial motor neurons: located in the ventral horn. Its motor
neurons innervate visceral and skeletal muscles.
Rexed laminae 1. Lamina I: corresponding to the marginal zone, its cells receive noxious and thermal
stimuli and send informations to the brain via contralateral spinothalamic tract.
2. Lamina II: corresponding to the substantia gelatinosa, has a role in the modulation of
sensory painful signals. It sends information to laminae III and IV.
3. Lamina III: corresponding to the nucleus proprius: relays proprioceptive and tactile
stimuli. It sends information to laminae IV, V, and VI.
4. Lamina IV: partially corresponding to the nucleus proprius, has a role in registration of
non-painful stimuli. It is in connection with lamina II.
5. Lamina V: sends sensory, including nociceptive and potentially painful stimuli,
informations to the brain through the contralateral and spinothalamic tracts and receives
descending information from the brain via the corticospinal and rubrospinal tracts.
6. Lamina VI: receives sensory information from muscle spindles and sends information to
the brain via ipsilateral spinocerebellar pathways. It plays a role in spinal reflexes with its
interneurons.
7. Lamina VII: partially corresponding to the nucleus of Clarke, large and heterogenous,
receives information from Lamina II to VI, and from viscera, sends motor information to
the viscera and has a role in the regulation of the autonomic system.
8. Lamina VIII: modulates motor outputs to the skeletal muscles. It is the only lamina which
does not receive primary afferents directly, because of its location in the
intercommissural area.
9. Lamina IX: sends motor outputs to the skeletal muscles.
10. Lamina X: surrounds the central canal, its neurons decussate from a side of the spinal
cord to the other.

Bases of spasticity one given by Lance describing spasticity as “a motor disorder

characterized by a velocity-dependent increase in tonic stretch
Spasticity is the result of an exaggeration of the monosynaptic reflexes (“muscle tone”) with exaggerated tendon jerks,
muscle stretch reflex due to lesions affecting the central ner- resulting from hyperexcitability of the tonic stretch reflex, as a
vous system, in particular an upper motor neuron lesion. component of the upper motor neuron syndrome” [49].
Alternatively, spasticity can be secondary to spinal cord injury Spasticity can be considered as a sensorimotor phenome-
resulting in intraneuronal dysfunction and motoneuronal non, associated with reflexive movement responses to sensory
hyperactivity. inputs: clinically it manifests as an increased resistance offered
Spasticity has been considered as an expression of nervous by muscles to passive stretching and can be associated with
system dysfunction for a long time and it has been described in increased tendon reflexes, clonus, flexor and extensor spasms,
different ways. Maybe, the most appropriate definition is the and clasp-knife phenomenon.
 Childs Nerv Syst

Fig. 2 Somatotopic organization of ventral horn motor neurons: a medial extensor muscles (E) are located more ventrally, those for flexor muscles
group of fibers innervates the axial musculature (neck and trunk) and a more dorsally (F). Laterally there are motoneurons for distal limb muscles
lateral group innervates the extremities musculature. In this latter group, (D) while medially are located those for proximal limb muscles (P). CS:
the motoneurons are also somatotopiccally organized: motoneurons for corticospinal tract

There are two main mechanisms of spasticity: spinal and observe an overactivation of antigravity muscles (in the upper
supraspinal. This last one was demonstrated by Sherrington, limbs prevail shoulder abduction and elbow and wrist flexion,
who won the Nobel prize for his research on the neuromuscu- while hip abduction, knee extension, and plantar flexion are
lar system in 1932 [37]. prominent in the lower limbs).
In the spinal cord lesion mechanism there is a disinhibition Interestingly, the interruption of the dorsal tracts in patients
of the local multisynaptic pathways and so the impulses orig- with cerebral palsy does not inhibit the antigravity character of
inating from the muscle spindle are transmitted in a centripetal spasticity [49]: many other tracts are involved in muscular
way and spread uninhibited both in cranial and caudal direc- control, in particular the corticospinal, reticulospinal and
tions via propriospinal pathways, causing massive motor re- vestibulospinal tract.
sponses from several muscle groups (Fig. 3). Periventricular malacia is observed in over 70% of patients
In the supraspinal mechanism—namely a cerebral lesion— with bilateral diplegia. It compromises supratentorial influ-
a rapid motor response follows a muscular stretch because the ence to the spinal neuronal pool sending abnormal inputs
cerebral lesion disrupts the balance of supraspinal inhibitory through the vestibular and reticular nuclei and their tracts
and excitatory inputs directed to the spinal cord, leading to a causing a loss of inhibition to the spinal reflex arcs and an
state of disinhibition of the stretch reflex. Clinically, we can increase in muscular tone.

Table 4 Dermatomeric
organization Nerve roots Functions Muscles

L1–L2–L3 (*) Hip flexion Ileo-psoas, abductors

L3–L4 Knee extension Quadriceps femoris
L4–L5 Ankle dorsiflexion, Hip extension Medial hamstrings, tibialis anterioris
L5–S1 Plantarflexion of the ankle, knee Lateral hamstrings, posterior tibialis, extensor
flexion, great toe extension digitorum longus, extensor hallucis longus
L5–S1–S2 Grand toe flexion Peroneals, extensor digitorum
S1–S2 Plantar flexion Gastrocnemius, soleus, first dorsal interosseous
S2 (**) Hallucis flexion Abductor hallucis
S2–S3–S4(**) Bladder, lower bowel, and anal
control

(*) L1 and L2 contain also sympathetic innervation of some pelvic and abdominal organs
(**) Sacral nerve roots provide also parasympathetic innervation of pelvic and abdominal organs
Childs Nerv Syst

Fig. 3 Supraspinal control. Corticospinal tract, corticoreticular tract, systems (lateral reticulospinal tract and lateral corticospinal tract) result
bulbopontine tegmentum, and vestibular nucleus are facilitating in activation of the segmental interneuronal network. (2) Incomplete
structures. The medial reticulospinal tract and the vestibulospinal tract spinal cord lesions affecting the corticospinal fibers and the lateral
are excitatory systems. The lateral reticulospinal tract and the lateral reticulospinal tract result equally in activation of the segmental
corticospinal tract are inhibitory systems. The level of lesion interneuronal network. (3) Complete spinal lesions affect both
determinates the clinical pattern: (1) Lesions affecting the corticospinal excitatory and inhibitory systems
tract and the corticoreticular tract which activate the main inhibitory

Due to the topographic arrangement of the periventricular dorsal root fibers [15, 35, 41]. This type of approach was
white matter, small lesions adjacent to the trigonal regions of followed by the persistence of spasticity in many patients
the ventricles primarily affect the lower limbs, while more but also a great weakness. So, later, he introduced for the first
extensive lesions can also cause upper limb spasticity [3]. time in this kind of surgery intraoperative electromyography
with the purpose to help identify the abnormal rootlets. In that
period, great importance was given to the clinical aspects, so
Historical development of the SDR technique there was a great dialog between the neurophysiologist who
stimulated each root -using a 1 msec stimulus with an ampli-
There are multiple options in order to deal with the spasticity. tude of 1 V, and the physical therapist who detected which
Selective dorsal rhizotomy is a surgical procedure which con- groups of muscles responded to the stimuli, as well as the
sists in selective sectioning of the lumbosacral afferent nerve grade of their contraction, in order to precisely select the path-
rootlets and, when associated with an extensive program of ological root: with this technique, asymmetrical SDR was not
physical and occupational therapy, it is considered the corner- uncommon.
stone of the options for children with spastic diplegia and, In 1976, Fasano postulated that stimulation induce abnor-
possibly, with a rather good level of independence. mal muscular response due to a spreading of muscle contrac-
This surgical procedure has changed and has been tion to adjacent myotomes of either ipsi- or contralateral mus-
perfectioned during the few past decades, but even today is cle groups due to a lack of suprasegmental inhibition on the
not yet fully standardized (Table 5). alpha-gamma neuronal loop. During surgery, he utilized neu-
The first dorsal rhizotomy to treat spasticity in humans was rophysiological guidance in order to identify when a repetitive
performed by Foerster in the early 1900s: he sectioned all the stimulation at 30 to 50 Hz caused abnormal patterns of muscle
nerve root from T12 to S2 sparing only L4 because of its role response due to the lack of normal inhibitory circuits. These
in knee extension control which is mainly responsible for patterns include a sustained and synchronous activation of
standing posture and gait [9]. This technique was burdened muscles, and/or a spread of the muscle response to the contra-
by multiple complications such as sensory loss, propriocep- lateral lower limb or even to upper limbs or trunk. The path-
tion alteration, and bladder dysfunctions [6]. ological rootlets were cut [7].
In order to reduce post-operative morbidity, in the 1960s, It is with W. Peacock—who worked at Red Cross War
Gros et al. revisited the technique and he sectioned 80% of the Memorial Children’s Hospital, in Cape Town, South
 Childs Nerv Syst

Table 5 Development of different SDR techniques and the role of intraoperative neurophysiology

Site of SDR % of roots cut Role of intraoperative Parameters of stimulation

neurophysiology

Forrest, early Approach at T12 to S2 Sparing 100% of L4

1900 conus level
Gros, 1960 Approach at 80% of posterior rootlets Identify the abnormal EMG and Stimulation of each fascicle—
conus level rootlets using a 1-ms stimulus with an amplitude
of 1 V
Fasano, 1976 Approach at Abnormal patterns of Repetitive stimulation at 30 to 50 Hz
conus level muscle response
Peacock, 1982 Approach at Preservation of S2
cauda
equina level
Park, 1993 Approach at 60–70% of posterior rootlets Abnormal patterns of
conus level muscle response,
pudendal mapping
Steinbok, 1997 20–70% of posterior rootlets
Morota, 2007 15–85%, conservation of 50% of L4 Abnormal patterns of Repetitive stimulation at 10 Hz
No more than 3 successive roots muscle response
totally cut
Sindou, 2014 Keyhole approach

Africa—that we arrive at the modern technique for SDR. He The first step consists in the topographic identification of the
shifted the site of the rhizotomy from the conus medullaris ventral and dorsal roots and confirmation of their lumbosacral
region to the cauda equina. The surgical access was created level. A low threshold response to a single stimulus is tested.
with a L2–L5 laminectomy and this made it easier to identify Normally dorsal roots require an intensity three times higher than
the rootlets in the attempt to avoid bladder dysfunction, pos- ventral roots in order to reach a muscular response [50].
sibly the worst complication of SDR. Moreover, Peacock pre- Once every dorsal root from L2 to S2 has been identified,
ferred not to section S2 fibers because they might innervate and the threshold determined by 1-Hz stimulation, a stimulus,
the anal and urinary sphincters, even if there was a concomi- of 1-ms duration. After that, stimulation is delivered with a
tant abnormal activation of the triceps surae muscle [19, 36]. bipolar electrode at a frequency of 50 Hz and intensity ranging
Sindou, who in the 1970s observed that lesioning the DREZ between 0.1 and 1 mA. The stimulus is sustained for 1 s, and
in order to control pain reduced also spasticity, in 2014, the intensity is increased until a response appears. The re-
revisited Peacock’s approach and developed a keyhole inter- sponses are monitored by both palpation of the appropriate
laminar dorsal rhizotomy technique which reduces spinal liga- limb/muscle myotomal group and observation of the EMG
mentous injuries and further instability problems [12, 35, 40]. response. The level of excitability is quantified with the 5-
grade SDR Reflex Grading Criteria (Philips, 1989) [7, 34,
41, 50] (Table 6).
A spastic root is recognized on the basis of:
Role of intraoperative neurophysiology
during SDR 1. A threshold response to a single stimulus
2. A sustained response to a train at a frequency of 50 Hz for
The aim of selective posterior rhizotomy is to relieve the pa- 1s
tient from lower limb spasticity, improving motor function. In 3. The diffusion of the response to other muscle groups ipsi-
order to minimize the neurological damage optimizing the and contralateral not directly involved with the stimulated
balance between reduction of spasticity and preservation of rootlet’s segmental innervation
strength, many surgeons rely on intraoperative electrophysio-
logical testing [26]. The pathological dorsal roots are then carefully sepa-
It was introduced for the first time by Fasano and col- rated into 3–10 rootlets. Each rootlet is stimulated again,
leagues in 1976 with the aim to recognize the abnormal roots with a train of stimuli after having determined the thresh-
and then study them in order to section only the non- old level, and the ones producing pathological responses
functioning rootlets [30]. A similar approach was followed are cut. More specifically, the rootlets that produce a
by other authors afterwards [42–45]. response of grades 3 and 4 are cut and the rootlets with
Childs Nerv Syst

Table 6 Classification of EMG

responses Grade EMG response

0 Unsustained or single discharge to a train of stimuli

1 Sustained discharges from muscles innervated through the segment stimulated in the
ipsilateral lower extremity
2 Sustained discharges from muscles innervated through the stimulated and immediately adjacent
segments
3 Sustained discharges from segmentally innervated muscles and muscles innervated
through segments distant from the one stimulated
4 Sustained discharges from contralateral muscles w/ or w/o sustained discharges from the
ipsilateral muscles

a normal response (grade 0) are spared. The rootlets with than those with only mild spasticity and relatively intact
a grade 1 or 2 response are generally left intact, but if no motor function [20, 28, 29].
higher grade of elicited responses is detected, some of Nowadays, the percentage of sectioned radicle varies from
these rootlets are cut too. In order to avoid sensory loss, 40 to 75% and can be adapted by integrating preoperative and
at least one radicle has to be spared [6, 45]. Since it has intraoperative findings.
been shown that a train stimulation with a frequency of L2 and L3, which control flexion, adduction, and internal
50 Hz and a duration of 1 s can evoke a low threshold rotation of hip and thigh, as well as knee extension, are nor-
and sustained response of dorsal roots and rootlets also mally cut until two-thirds. L4 is very important for quadricep
in non-spastic patients, the most center gives more im- innervation and rarely is cut more than one-third except when
portance to the diffusion of the response to segments not the quadriceps is highly spastic and there is associated patella
related to the one stimulated [5, 43, 45]. ascension. In this case, more extended sections are considered.
Recently, the H-reflex testing has been introduced: the L5 plays a role in the antigravity function and hip stability so
stimulation of a dorsal root normally evokes a motor response, rarely is cut more than one third. The grade of S1 and S2 cut
via H-reflex, in the corresponding muscle bypassing the direct depends on toe posture: if it coexists, equinus or varus foot S1
activation of the spindle mechanism. Increasing stimulation is cut for three- to four-fifths, while S2 is cut by one-half if
intensity, in roots with a normal response, this H-reflex tends claw toes are present [39, 40].
to disappear because of the refractory period of spinal inter- It is known that S2 plays an important role in bladder con-
neurons [48]. On the contrary, if one tests H-reflex by eliciting trol: S2 lesion is responsible up to 24% of post-operative uri-
it through the stimulation of peripheral nerves and dorsal roots nary disturbance. It has been demonstrated that testing the
in patients with upper motor neuron lesion (such as the one roots to determine the number of pudendal afferents, after
with cerebral palsy), there is a more rapid and complete re- electrical stimulation of dorsal clitoral or penile nerves, can
covery period of spinal interneurons with less inhibition than reduce post-operative urinary dysfunction down to 4%
in roots with a normal response. The ratio between the second (Fig. 4) [2]. Furthermore, it has been shown that including
and the first H-reflex (H2/H1), is considered pathological S2 in the rhizotomy reduced up to 71% the number of patients
when it is more than 50% at any frequency [22]. with residual plantar flexor spasticity [18].
The criteria for selecting the pathological roots and the In recent years, with the diffusion of mini-invasive single-
percentage of cutting are controversial because of the lack of level approach, the importance of intraoperative neurophysi-
guidelines; therefore, every Center follows its own combina- ology has grown because of its role in permitting safe identi-
tion of neurophysiological criteria and clinical criteria such as fication of dorsal roots [13, 52].
age, spasticity degree and its distribution, preoperative motor At the same time, it has been questioned if intraoperative
function (Gross Motor Function Measure (GMFM) score), neurophysiological testing is needed to decide which part of
and also the surgeon’s expertise [25, 47]. the dorsal roots to cut [11, 40, 43]. At least one meta-analysis
In order to reduce post-operative morbidity, it is important of three randomized controlled trials [25] as well as a very
to correlate the intraoperative finding with a preoperative neu- recent, large, clinical series [28] have suggested that the pre-
rological evaluation tailored for SDR patients. operative functional level may be related with the extent of
In 1999, Lazareff showed that functional improvement resection needed. Ten years ago, a provocative study by
in spastic children could be reached also without exten- Steinbok et al. showed that at 1-year follow-up, there is no
sive selective deafferentation. Later, it has been hypoth- clinical difference in complications and outcome measures
esized that children with severe spasticity and limited between patients operated with and without electrophysiolog-
motor function would require more extensive cutting ical guidance. The only significant difference was a shorter
 Childs Nerv Syst

Fig. 4 Dorsal root action potentials recorded from a 4-year-old boy after stimulation of the penile nerve showing that all afferent activity is carried on by
the right S2 root only. Two averages of 100 responses from each root are superimposed

duration of the operative time without intraoperative neuro- on the “H-reflex recovery curve,” which tests the excitability
physiology, but the percentage of sectioned roots was similar of local circuitry within the spinal cord. Under normal condi-
[46]. The study by Steinbok et al. did not address, specifically, tions excitability of local circuitry is regulated by sensory and
whether or not the use of IN may have resulted in a similar proprioceptive input coming to the spinal cord by dorsal roots.
outcome by sectioning a smaller percentage of the dorsal This input is regulated locally within the gray matter of the
roots. So, it may well be that the use of IN increases surgeons’ spinal cord by interneurons and excitatory/inhibitory synapses
confidence in cutting the roots. In addition, although and local feedback mechanisms (e.g., Renshaw cells). The
Steinbok’s study has been often cited as an argument against final control of excitability is regulated by descending input
the value of IN in SDR, in fact, the main message of that study from supraspinal pathways, the most important of them is the
was that IN is not essential to the point of denying this surgical corticospinal tract, but also by other descending tracts
procedure to patients in areas of the world where IN is (reticulospinal, vestibulospinal, cerebellospinal, etc).
unavailable. In children with cerebral palsy, this mechanism of
This being said, IN still has a value to distinguish ventral supraspinal control is affected in a different grade by their
from dorsal roots and, with the use of pudendal mapping, may anatomical lesion, resulting in spasticity, increased muscle
assist to include S2 rootlets in the cutting, while preserving tone, exaggerated tendon reflexes, pathologic reflexes, and
sphincterial function. Overall, the importance of intraopera- other signs of upper motor neuron lesion. The corticospinal
tive neurophysiological testing is reported in many articles tract, per se, might not necessarily be affected very much, but
[12, 26, 28, 42] and a recent meta-analysis [25] of long-term a combination of lesions of descending pathways is a critical
outcomes which agree in saying that intraoperative neuro- factor resulting in spasticity. In order to better regulate the
physiology is an important and effective tool that permits bet- affected suprasegmental control, one of the neurosurgical tar-
ter preservation of roots involved in antigravity and pelvic gets is to decrease sensory inputs to the spinal cord by selec-
stability, namely L4 and L5, and a patient-tailored procedure tive lesioning of the dorsal roots, resulting in decreased
allowing “asymmetrical” rhizotomies for a better clinical re- spasticity.
sult in most patients.
Compliance with ethical standards

Conclusions Conflict of interest The authors declare that they have no conflict of
interest.

During 100 years of applying the concept of SDR, many neu-

rophysiologic techniques have been developed but principally
all of them are based on the strategy to decrease sensory input
to the spinal cord in order to decrease spasticity and increased
References
better suprasegmental control over voluntary movement, gait, 1. Abbott R (2007) Editorial on ‘The management of childhood hy-
and balance. pertonia’. Childs Nerv Syst 23(9):937–941
Most of the neurophysiologic methods used to test roots 2. Abbott R (1996) Sensory rhizotomy for the treatment of childhood
with abnormal responses after electric stimulation are based spasticity. J Child Neurol 11(Suppl. 1):36–42
Childs Nerv Syst

3. Aquilina K, Graham D, Wimalasundera N (2015) Selective dorsal 25. McLaughlin J, Bjornson K, Temkin N, Steinbok P, Wright V,
rhizotomy: an old treatment re-emerging. Arch Dis Child 100(8): Reiner A, Roberts T, Drake J, O'Donnell M, Rosenbaum P,
798–802 Barber J, Ferrel A (2002) Selective dorsal rhizotomy: meta-
4. Bax M, Goldstein M, Rosenbaum P, Levinton A, Paneth N (2005) analysis of three randomized controlled trials. Dev Med Child
Proposed definition and classification of cerebral palsy. Dev Med Neurol 44(3):17–25
Child Neurol 47(April):571–576 26. Mittal S, Farmer JP, Poulin C, Silver K (2001) Reliability of intra-
5. Cohen AR, Webster HC (1991) How selective is selective posterior operative electrophysiological monitoring in selective posterior rhi-
rhizotomy? Surg Neurol 35(4):267–272 zotomy. J Neurosurg 95(1):67–75
6. Enslin JMN, Langerak NG, Fieggen AG (2019) The evolution of 27. Moreno De Luca A, Ledbetter DH, Martin CL (2012) Genetic
selective dorsal rhizotomy for the management of spasticity. insights into the causes and classification of the cerebral palsies.
Neurotherapeutics 16(1):3–8 Lancet Neurol 11:283–292
7. Fasano VA, Barolat-Romana G, Ivaldi A, Sguazzi A (1976) La 28. Morota N (2019) Clinically practical formula for preoperatively
radicotomie postérieure fonctionnelle dans le traitement de la estimating the cutting rate of the spinal nerve root in a functional
spasticité cérébrale Premieres observations sur la stimulation posterior rhizotomy. Childs Nerv Syst 35(4):665–672
électrique peropératoire des racines postérieures, et leur utilisation 29. Morota N (2007) Functional posterior rhizotomy: the Tokyo expe-
dans le choix des racines à sectionner. Neurochirurgie 22(1):23–24 rience. Childs Nerv Syst 23(9):1007–1014
8. Filloux FM (1996) Neuropathophysiology of movement disorders 30. Netter F (2014) Atlas of human anatomy. 6th ed, Edra
in cerebral palsy. J Child Neurol 11(Suppl. 1):S5–S12 31. Osler SW (1889) The cerebral palsies of children: a clinical study
9. Foerster O (1913) On the indication and results of the excision of for the infirmary for nervous diseases, classics of neurology and
posterior roots in men. Surg Gynecol Obstet 16:463–475 neurosurgery Library, Philadelphia
10. Freud S (1897) Die infantile Cerebrallähmung. In: Specielle 32. Panteliadis CP (2018) Cerebral palsy - a multidisciplinary ap-
Pathologie und Therapie Bd IX, Teil III, Holder ed (Wien), pp. proach, 3rd edn. Springer International Publishing, pp 35–47
1–327 33. Papadelis C, Kaye H, Shore B, Snyder B, Grant PE, Rotenberg A
11. Fukuhara T, Nakatsu D, Namba Y, Yamadori I (2011) Histological (2019) Maturation of corticospinal tracts in children with hemiple-
evidence of intraoperative monitoring efficacy in selective dorsal gic cerebral palsy assessed by diffusion tensor imaging and trans-
rhizotomy. Childs Nerv Syst 27(9):1453–1458 cranial magnetic stimulation. Front Hum Neurosci 13(July):1–9
34. Park TS, Dobbs MB, Cho J (2018) Evidence supporting selective
12. Georgoulis G, Brînzeu A, Sindou M (2018) Dorsal rhizotomy for
dorsal rhizotomy for treatment of spastic cerebral palsy. Cureus
children with spastic diplegia of cerebral palsy origin: usefulness of
10(10):e3466
intraoperative monitoring. J Neurosurg Pediatr 22(1):89–101
35. Park TS, Johnston JM (2006) Surgical techniques of selective dor-
13. Graham D, Aquilina K, Cawker S, Paget S, Wimalasundera N
sal rhizotomy for spastic cerebral palsy.Technical note. Neurosurg
(2016) Single-level selective dorsal rhizotomy for spastic cerebral
Focus 21(2):1–6
palsy. J Spine Surg 2(3):195–201
36. Peacock WJ, Aren LJ, Berman B (1987) Cerebral palsy spasticity.
14. Graham D, Aquilina K, Mankad K, Wimalasundera N (2018)
Selective posterior rhizotomy. Pediatr Neurosci 13:61–66
Selective dorsal rhizotomy: current state of practice and the role
37. Sherrington CS (1898) Decerebrate rigidity and reflex coordination
of imaging. Quant Imaging Med Surg 8(2):209–218
of movements. J Physiol 22:319–327
15. Gros C, Ouaknine G, Vlahovitch B, Frerebeau P (1967) La
38. Sindou M, Georgoulis G, Mertens P (2014) Neurosurgery for
radicotomie sélective postérieure dans le traitement neuro-
spasticity- a practical guide for treating children and adults, 1st
chirurgical de l’hypertonie pyramidale. Neurochirurgie 13:505–
edn. Springer-Verlag, Wien
518
39. Sindou M (1995) Microsurgical DREZotomy (MDT) for pain,
16. Hultborn H (2006) Spinal reflexes, mechanisms and concepts : spasticity, and hyperactive bladder: a 20-year experience. Acta
from Eccles to Lundberg and beyond. Prog Neurobiol 78:215–232 Neurochir 137:1–5
17. Kothbauer KF, Deletis V (2010) Intraoperative neurophysiology of 40. Sindou M (2019) Surgery in the dorsal root entry zone for spastic-
the conus medullaris and cauda equina. Childs Nerv Syst 26(2): ity. In: Lozano AM, Gildenberg PL, Tasker RR (eds) Textbook of
247–253 stereotactic and functional neurosurgery. Springer, Berlin, pp 1–26
18. Lang FF, Deletis V, Cohen HW, Velasquez L, Abbott R (1994) 41. Smyth MD, Peacock WJ (2000) The surgical treatment of spastic-
Inclusion of the S2 dorsal rootlets in functional posterior rhizotomy ity. Muscle Nerve 23:153–163
for spasticity in children with cerebral palsy. Neurosurgery 34(5): 42. Steinbok P, Gustavsson B, Kestle JRW, Reiner A, Cochrane DD
847–853 (1995) Relationship of intraoperative electrophysiological criteria
19. Langerak NG, Lamberts RP, Fieggen AG, Peter JC, Peacock WJ, to outcome after selective functional posterior rhizotomy. J
Vaughan CL (2007) Selective dorsal rhizotomy: long-term experi- Neurosurg 83(1):18–26
ence from Cape Town. Childs Nerv Syst 23(9):1003–1006 43. Steinbok P, Keyes R, Langill L, Cochrane DD (1994) The validity
20. Lazareff JA, Garcia-Mendez MA, De Rosa R, Olmstead C (1999) of electrophysiological criteria used in selective functional posterior
Limited (L4-S1, L5-S1) selective dorsal rhizotomy for reducing rhizotomy for treatment of spastic cerebral palsy. J Neurosurg
spasticity in cerebral palsy. Acta Neurochir 141(7):743–752 81(3):354–361
21. Little WJ (1966) On the influence of abnormal parturition, difficult 44. Steinbok P, Langill L, Cochrane DD, Keyes R (1992) Observations
labours, premature birth, and asphyxia neonatorum, on the mental on electrical stimulation of lumbosacral nerve roots in children with
and physical condition of the child, especially in relation to defor- and without lower limb spasticity. Childs Nerv Syst 8:376–382
mities. Clin Orthop Relat Res 46:7–22 45. Steinbok P, Reiner A, Beauchamp RD, Cochrane DD, Keyes R
22. Logigian EL (1994) H reflex studies in cerebral palsy patient un- (1992) Selective functional posterior rhizotomy for treatment of
dergoing partial dorsal rhizotomy. Muscle Nerve 17(5):539–549 spastic cerebral palsy in children. Pediatr Neurosurg 18(1):34–42
23. MacKinnon CD (2018) Sensorimotor anatomy of gait, balance, and 46. Steinbok P, Tidemann AJ, Miller S, Mortenson P, Bowen-Roberts
falls. Handb Clin Neurol 159:3–26 T (2009) Electrophysiologically guided versus non-
24. MacLennan AH, Thompson SC, Gecz J (2015) Cerebral palsy: electrophysiologically guided selective dorsal rhizotomy for spastic
causes, pathways, and the role of genetic variants. Am J Obstet cerebral palsy: a comparison of outcomes. Childs Nerv Syst 25(9):
Gynecol 213(6):779–788 1091–1096
 Childs Nerv Syst

47. Steinbok P (2007) Selective dorsal rhizotomy for spastic cerebral 52. Xiao B, Constatntini S, Browd SR, Zhan Q, Jiang W, Mei R (2019)
palsy: a review. Childs Nerv Syst 23(9):981–990 The role of intra-operative neuroelectrophysiological monitoring in
48. Storrs BB, Nishida T (1988) Use of ‘H’ reflex recovery curve in single-level approach selective dorsal rhizotomy. Childs Nerv Syst:
selective posterior rhizothomy. Pediatr Neurosci 14:120–123 1–9. https://doi.org/10.1007/s00381-019-04408-5
49. Trompetto C, Marinelli L, Mori L, Pelosin E, Currà A, Molfetta L, 53. Zhou M, Wang W, Huang H, Zhu G, Chen Y, Zhou C (2010)
Abbruzzese G (2014) Pathophysiology of spasticity: implications Microsurgical anatomy of lumbosacral nerve rootlets for highly
for neurorehabilitation. Biomed Res Int 2014:1–8. https://doi.org/ selective rhizotomy in chronic spinal cord injury. Anat Rec 293:
10.1155/2014/354906 2123–2128
50. Turner RP (2009) Neurophysiologic intraoperative monitoring dur-
ing selective dorsal rhizotomy. J Clin Neurophysiol 26(2):82–84
Publisher’s note Springer Nature remains neutral with regard to jurisdic-
51. Wang KK, Munger ME, Chen BP, Novacheck TF (2018) Selective
tional claims in published maps and institutional affiliations.
dorsal rhizotomy in ambulant children with cerebral palsy. J Child
Orthop 12(5):413–427