Keywords
temporal lobe - epilepsy - multiple hippocampal transection - memory - ultrasound
guidance
Introduction
Temporal lobe resections are the most widely performed epilepsy surgeries in adults.
These procedures include standard anteromesial temporal lobe resection[1] and several modifications of amygdalohippocampectomy.[2] Both good postoperative seizure outcomes and low risk of surgical complications
have been reported.[3] Nevertheless, the temporal lobe epilepsy surgery is underutilized worldwide.[4] One of the reasons for this fact are attending physicians' or patients' concerns
about postoperative cognitive function decline, especially after operations on language-dominant
side in patients with morphogically normal hippocampus and intact memory. Thus, there
is a strong mandate for an ongoing search for minimally invasive nondestructive procedures.
Hypothetically, there are two ways to improve functional outcomes after temporal lobe
epilepsy surgery: (1) minimize collateral damage in the access route (e.g., laser
thermal ablation)[5] and (2) minimize destruction to the target and restrict the intervention only to
structures important for seizure generation, early propagation, and synchronization.[6]
One possible way to combine both the points is to utilize multiple hippocampal transection
(MHT). The principle of this operation is based on observations that both interictal
and ictal epileptic discharges in hippocampus spread and synchronize by means of anterior-to-posterior
propagation via the longitudinal hippocampal fibers.[7] Thus, transection of these fibers, if performed perpendicularly to the long axis
of hippocampus, has the potential to stop seizures while keeping intact the transverse
circuits. MHTs have been performed in small groups of nonlesional temporal lobe epilepsy
patients[8] and patients with amygdala enlargement[9] with encouraging results.
In all of the published series, the main acute procedural end point was the abolition
of spikes on perioperative intracranial electocorticography (iECoG). Once these changes
were observed, the surgery was terminated; hence, the real extent of the transections
was not studied in detail.[10] Whether abolition of iECoG epileptic discharges is a valid procedural end point
(especially since these iECoG changes may only be transitory) is debatable.[11]
[12]
[13]
In contrast, we hypothesize that the completeness of intrahippocampal longitudinal
fiber transection is the crucial factor for seizure control in mesial temporal lobe
epilepsy treated with MHT. With this premise in mind, we conducted an anatomic study
of a cadaveric brains, in which we aimed to achieve complete transections of the hippocampal
body perpendicular to the long axis except fimbria. On one side, we performed hippocampal
transection based solely on anatomical landmarks, while on the contralateral side
we employed a perioperative ultrasound examination of a navigational adjunct. The
basic null hypothesis was that there was no difference in the extent of transection
on both sides.
In addition, we used this anatomical study to design and develop specific shapes and
sizes of several transectors to facilitate precise and minimally invasive hippocampal
transections.
Material and Methods
In this anatomical comparative study, we have performed MHT in five formalin-fixed
human cadaveric brains supplied by our institute's pathology department. The study
took place at a tertiary referral hospital with an epilepsy surgery program during
a 2-year period (2020–2021).
During the cadaveric surgery, the middle temporal gyrus was resected, and via a large
corridor through white matter on the bottom of the gyrus, the temporal horn of the
lateral ventricle was approached on both sides of the brain. The head and body of
the hippocampus were well exposed ([Fig. 1]).
Fig. 1 Exposure of the head and body of the hippocampus via the middle temporal gyrus.
After this step, the parameters of the hippocampus were measured and registered by
ultrasound examination (Craniotomy probe 8862, 4–10MHz, Ultrasound BK Medical Flex
Focus 800, BK Medical A/S, Herlev, Denmark) ([Fig. 2]). The transection procedure followed. The first two transections of the hippocampus
were executed at the level of the hippocampus head followed by three or four transections
of the hippocampus body. The spacing between each transection was 5 mm. The aim of
each transection was to completely transect the hippocampus perpendicularly to its
long axis, hence interrupting all of the longitudinal hippocampal connections except
fimbria that could be easily visualized during the surgery and injury of it could
be avoided. On one side of the brain, the transections were performed based on visible
anatomical landmarks and according to the pretransection utrasound examination (control
group). On the other side of the same specimen, the transections were governed by
ultrasound guidance during each transection so the transector could be visualized
while inserted in the hippocampus (study group) ([Fig. 3]). The side that was transected under ultrasound supervision was assigned randomly
for each cadaver to enable blinded reading by radiologist.
Fig. 2 Ultrasound examination of the hippocampus (images oriented similarly to coronal magnetic
resonance imaging scans).
Fig. 3 Ultrasound examination of hippocampus without (a) and with (b) transector (arrow) placed in the hippocampus.
The transectors that were designed by the first author (J. S.) include bayonet loop
curettes ([Fig. 4a]). The loop has different diameter and can be chosen according to the size of hippocampus
cross-section ([Fig. 4b]). A 6-mm-diameter loop is used for the head and proximal part of hippocampus, and
a 4-mm loop for the subiculum. The dorsal part and tail of the hippocampus is transected
by 4-mm transectors that have extra bend designed to facilitate access. The edge of
the loop is slightly blunted to enable the surgeon to feel the resistance of the pia
mater and the arachnoid layer, thus avoiding injury to deeply located ambient cistern
structures. The cadaveric surgery was performed solely by the first author (J. S.).
Fig. 4 Transectors. (a) Straight transector overview. (b) Different types of transectors' tips in detail: 4-mm straight loop for subiculum
transection, 6-mm straight loop for transection of the head and proximal body of hippocampus,
and 4-mm bended loop for transection of the distal hippocampal body and tail.
Postoperative MRI Scans
After surgery, we scanned the cadaveric brain with magnetic resonance imaging (MRI)
and measured and calculated the percent transected cross-sectional area of each transection.
MRI was performed using a 3T Siemens Skyra (software VE11C, Siemens, Erlangen, Germany)
scanner equipped with 20-channel head and neck coil (to accommodate the plastic container
with the cadaver) using 3D magnetization-prepared rapid gradient echo T1-weighted
sequence (slice thickness: 1.1 mm; inplane resolution: 0.9 × 0.9 mm; TR: 1,540 ms;
TE: 2.48 ms; TI: 900 ms; FA, 8 degrees) and T2-weighed sequences in the sagittal plane,
oblique transverse plane aligned with the long axes of both hippocampi, and oblique
coronal plane perpendicular to this plane. All three T2-weighted sequences shared
common parameters—slice thickness: 2 mm; TE: 82 ms; TI: 5,160 ms; NEX 2: 2 concatenations
with zero interslice gap; 871 phase-encoding steps; ETL: 15; inplane resolution 361 × 361
μm interpolated to 181 × 181 μm (matrix 1,024 × 1,024 pixels).
When imaged in the coronal plane, the shape of the cross-section of the body of the
hippocampus (perpendicular to the long axis of the hippocampus) can be approximated
as an ellipse ([Fig. 5]).
Fig. 5 The surface of the planned and realized transections can be estimated and calculated
by approximating this region to an ellipse (white ellipse: cross-section of hippocampus, yellow line: diameter a; green line: diameter b).
Therefore, the surface of this cross-section can be calculated using two diameters
directly measured on the MRI, using the following expression: S = a/2 × b/2 × π. Here, a is the diameter of the hippocampus body in axial plane, while b is its diameter in sagittal plane. The diameters and the surface of the accomplished
transection can be measured and calculated in a similar fashion ([Fig. 6]); the ratio of these two surface estimates (planned and realized transection) can
be expressed as percent transected area.
Fig. 6 Transection visible on the cadaveric brain magnetic resonance imaging in the (a) axial plane and (b) sagital plane.
Statistics
Statistical analysis was performed in the SPSS statistical program from IBM, version
19. For descriptive statistics, medians and interquartile means were established.
Nonparametric Mann–Whitney U test of the distribution of differences (percent of transection extent) was used
for comparison of the study and control group.
Ethics
All data collection, storage, and processing were done in compliance with the Helsinki
Declaration. The study was approved by the Ethical Committee of the Na Homolce Hospital.
It was supported by Ministry of Health of the Czech Republic (grant MH CZ and DRO
[NHH] IG 171501 and IG 193001).
Results
Each cadaver had five to six transections of the hippocampus on each side, and each
transection was measured and analyzed separately. Altogether 27 transections were
measured in the control group and 27 transections in the study group.
MRI measurement of transected diameters was conducted by a side-blinded radiologist
(J. K.) and the comparison of the two groups was statistically evaluated.
The average cross-section of the body of the hippocampus was 53 mm2 in the control group versus 47 mm2 in the study group. The median percentage of transection was 58% in the control group
and 73% in the study group ([Table 1]). A Mann–Whitney test indicated statistically significant difference of distribution
in favor of the study group: U (N
control = 27, N
study = 27) = 153, Z = 3,641, p (exact two-sided) = 0.0014. When evaluating each cadaver individually and comparing
the control group side with the study group, we reached similar results albeit without
reaching statistical significance. However, there was clear increase in the transected
planes in both groups, possibly reflecting the learning curve of the surgeon. Detail
results of transected ranges of the hippocampus for both groups are shown in [Fig. 7].
Fig. 7 The percentage of transected cross-section of the hippocampus of each transection
with and without ultrasound guidance.
Table 1
Descriptive statistics of both groups
|
Group
|
N (transections)
|
Median
(percent transected area %)
|
Interquartile range
|
|
Control group
|
27
|
58.3
|
51.3–65.9
|
|
Study group: ultrasound guided
|
27
|
72.9
|
63.9–83.0
|
Note: U (N
control = 27, N
study = 27) = 153, Z = 3.641, p (exact two-sided) = 0.0014.
Discussion
MHT is an epilepsy surgery performed in patients at high risk of postoperative memory
decline (e.g., nonlesional mesial temporal lobe epilepsy). Some authors used MHT successfully
as an adjunct to lesionectomy in patients with epilepsy due to hippocampal lesions[14] or in tandem with multiple subpial transections.[15]
The principal goal of this procedure is the disruption of the longitudinal pathway
of the hippocampus, which plays a role in the synchronization and propagation of epileptic
discharges, while sparing transverse lamellae, responsible for memory function.
This study was made on cadaveric brains. Its only aim is to prove that the extent
of longitudinal hippocampal fibers transections is more complete when using ultrasound
navigation. Thus, it has no ambition to asses how MHT will work in real surgery and
if it will yield better cognitive results. Moreover, the study is based on yet unproven
hypothesis that completness of longitudinal fibers transections has the potential
to stop seizures while keeping intact the transverse circuits that are responsible
for memory function. Furthermore, it is not clear whether this procedure lowers the
risk of epileptogenic network reorganization after the operation and late seizure
reccurence. To our knowledge, the extent of transections has never been studied in
published clinical series.
The goal was to reach full anatomic hippocampal body transection via three or four
transections, perpendicular to its long axis. We employed an ultrasound probe in the
area of the body of the hippocampus and compared these navigated transections to transections
based only on anatomical landmarks, i.e., without the benefit of visualizing the tip
of the transector at the depth of the hippocampus. The interval spacing between each
transection was kept to ∼5 mm. We hypothesized that ultrasound visualization of the
transector and anatomical landmarks of the hippocampus during the operation would
have the potential of improving the precision of this operation. Even if executed,
we did not include the transection of the head and subiculum in the analysis of our
cadaver study as the surface of the these cross-sections is not possible to approximate
with easy mathematical formula (as for the body of the hippocampus as ellipse); hence,
the calculation would not meet adequate scientific level.
We found that ultrasound navigation led to a significantly larger extent of transections
in the study group. Subjectively, the ultrasound-assisted transections in the study
group allowed for much higher operator confidence than the anatomy-guided transections
in the control group.
Several operative challenges had to be overcome in this study. Due to the stiffness
of the formalin-fixed cadaveric tissue, our approach to the temporal horn via white
matter on the bottom of superior temporal sulcus had to be wide. In a living specimen,
the white matter dissection can be limited to less than 1 cm to open the temporal
ventricle as needed and can be provided via well known approaches (transsylvian, transtemporal,
transulcal). We also recognized that the white matter dissection should be limited
in its dorsal extent so as to avoid damage to Meyer's loop. Furthermore, during this
experiment, we encountered serious difficulties in achieving transverse transection
of hippocampus in its dorsal parts (dorsal body and tail of hippocampus). Thus, we
created a specifically bent transector, enabling the surgeon to reach the target in
a living patient via a minimal opening in the temporal horn above the head of the
hippocampus.
The most commonly used form of the MHT techniques relies on preoperative chronic intracranial
video electroencephalographic monitoring to prove hippocampal seizure onset zone and
intraoperative electrocorticography (iECoG) to demonstrate hippocampal spike abolition
during the operation.[16]
However, it is debatable whether abolishment of ECoG spikes provides an adequate end
point for these transections. Some authors considered iECoG-based guidance to be of
limited value in temporal lobe epilepsy surgery.[12]
[13] Another group found that a high proportion (approximately one-third) of patients
may not even have any spikes on their preoperative iECoG.[17]
In addition to MHT methodology, the question of the utility of adjunctive lesions
also remains open. For example, some MHT case series also reported transections or
resections of the parahippocampal gyrus. However, such approach remains controversial,
given that the main afferent cortical connections of the polysynaptic intrahippocampal
pathway project via parahippocampal gyrus and that the direct intrahippocampal pathway
proceeds via the entorhinal and perirhinal cortices. In addition, the direct intrahippocampal
pathway efferents communicate to the cortex via the entorhinal cortex, which can be
destroyed by parahippocampal transection or resection.[18] Hence, it is our opinion that entorhinal and perirhinal cortex should be excluded
from the surgery for the sake of preserving memory function.
MHT is a technique that resembles multiple subpial transections. In this technique,
62 to 71% of patients had greater than 95% seizure reduction.[19] During longer follow-up, however, only 16% were seizure-free.[20]
[21] The logical explantion of this fact is that the destruction of the routes of seizure
propagation was incomplete and seizure recurrence is the consequence of neural network
reorganization. What cannot be avoided in neocortex can be solved in hippocampus by
technically feasible nearly complete transections perpendicular to the long axis of
the hippocampus.
Based on the known organization of memory circuits,[18] we also hypothesized that the truly perpendicular transection to the long axis of
the hippocampus is the critical condition required to avoid postoperative memory decline.
However, we are aware that this model is oversimplifiyed and some memory functions
are supported by connections that do not run perpendicular to long axis of the hippocampus.
That said, we recognize that some MHT authors present postoperative MRI scans in which
the transection of the dorsal part and tail of the hippocampus is actually parallel
to the long axis, yet the postoperative memory results are excellent.[22] Further studies are needed to reconcile these reported findings with the theory
of memory circuits.
Despite several inherent limitations in our study, we think that our findings may
lead to an improvement in the MHT operative technique and we plan to use it, build
on this foundational experience, by designing a clinical study of MHT in the treatment
of nonlesional temporal lobe epilepsy. One of the main drawbacks of the present work
stems from the different physical attributes of formalin-fixed human cadaveric brains
compared with a living brain; this imposes problems such as the need of larger neocortical
resection and different tissue resistance to the transector penetration. Another limitation
is the use of different navigation system that will be routinely used (MRI-based navigation
with implemented planned trajectories of transections) were the brain shift should
be corrected with the 3D ultrasound navigation.
Conclusion
This anatomical study on cadaveric brains proves the possibility to transect at least
73% of cross-section of the hippocampus perpendicular to the long axis with the aid
of real-time ultrasound visualization. Modified instruments for this procedure were
designed.
