Introduction
Stent-assisted endovascular coil embolization of an intracranial aneurysm is an established
interventional procedure, allowing treatment of complex and challenging broad-necked
aneurysms, by preventing coil protrusion or dislocation of coils into the parent artery
[1 ]
[2 ]. Thrombosis in the aneurysm and an inflammatory reaction in the surrounding tissue
occlude and stabilize the aneurysm, followed by formation of endothelial layer at
the aneurysm neck [3 ].
Cho et al. investigated the physical properties of four commercial self-expanding
intracranial stents, showing that the Neuroform EZ stent was generally in the midrange,
but achieves above-average to excellent results regarding radial force and surface
roughness [4 ]. The open cell Nitinol stent with four radiopaque markers at the proximal and distal
end can access aneurysms in tortuous vascular passages [5 ].
Due to the implanted foreign material, inhibition of platelet aggregation is usually
performed as a dual therapy during the first year after implantation, followed by
lifelong administration of acetylsalicylic acid (ASA). However, there is still insufficient
long-term experience with regard to implanted intracranial stents [6 ].
Bioresorbable stents (BRS) made of metals or polymers, particularly magnesium alloys
and poly L-lactic acid (PLLA) offer a new therapeutic approach in the treatment of
intracranial aneurysms as well as stenoses and are part of current studies and reviews
[7 ]. In contrast to conventional metallic stents, BRS dissolve after a certain time
and are metabolized by the body [8 ]. Platelet inhibition might be avoidable after the BRS dissolves, thus decreasing
the risk of hemorrhagic complications [9 ]
[10 ]. Finally, nonmetal BRS have fewer artifacts on computed tomography (CT) and magnetic
resonance imaging (MRI), thus facilitating follow-up imaging. However, these BRS are
generally of low radiopacity needing radiopaque markers for X-ray visibility [11 ].
The aim of the current study was to design, manufacture, and examine a neurovascular
bioresorbable microstent (NBRS) prototype based on a commercially available product
suitable to treat broad-based intracranial aneurysms.
Within this proof-of-concept study, an NBRS prototype was investigated regarding morphology
and fundamental mechanical properties. Furthermore, simulated use tests including
implantation of the device into a 3D-printed patient-specific phantom model under
physiological-orientated conditions and an in vitro coiling procedure to assess the stent’s ability to support a coil package were performed.
Materials and Methods
Design of a neurovascular bioresorbable microstent
In the current study, an NBRS design oriented on the Neuroform EZ stent (Stryker Corp.,
USA) was developed (see [Fig. 1 ]). This commercial device has a nominal outer diameter of DO
= 3.5 mm and length of LS
= 20 mm. This stent is approved to treat wide neck, intracranial, saccular aneurysms
combined with the use of embolic coils and is recommended for a parent vessel size
of 3.0 mm < DV
≤ 3.5 mm.
Fig. 1 Design of bioresorbable self-expanding microstent: stent length (LS
), stent circumference (CS
) or stent outer diameter (DO
), cell-to-cell length (LCTC
), cell-to-cell height (HCTC
), connector length (LC
), connector height (HC
), peak-to-peak cell spacing (LCS
), cell opening radius (RCO
), and strut width (WS
).
Since PLLA shows inferior mechanical properties compared with Nitinol, the NBRS wall
thickness TW
and strut width WS
were increased to 100 µm to improve the mechanical properties of the NBRS prototype
– particularly to enhance the bending stiffness, the radial stiffness, and the resistance
to local deformation. An outer stent diameter of DO
= 4.0 mm in the fully expanded state was chosen, aiming for an indicated use in the
parent vessels of DV
= 3.0 mm. To improve the overall stiffness of the NBRS structure, the connector elements
between meander structures were increased in length as well as height to about LC
= HC
= 360 μm and cell spacing was reduced by about 55 µm. As a result, the described
design changes led to a minimum crimping diameter of 1 mm. The stent design was implemented
using Creo Parametric 6.0 software (Parametric Technology Corp., USA).
Manufacturing of semi-finished products
Tubes as semi-finished products for the NBRS were manufactured based on a PLLA (Resomer
L210 S, Evonik Health Care GmbH, Germany) in chloroform (Carl Roth GmbH + Co. KG,
Germany) solution (3 g of PLLA dissolved in 100 ml chloroform) using a semiautomatic
dipping process as described previously [12 ]. In short, stainless steel mandrels with a diameter of 4.0 mm and a length of 55 mm
were used for repeated dipping into the polymer solution. This process was performed
using a KSV NIMA Dip Coater (Biolin Scientific Holding AB, Sweden) at (22 ± 2) °C
and 25 % humidity. After each repetition, the mandrels were dried for 10 minutes and
turned over by 180°. After twelve cycles the desired wall thickness of 100 µm was
achieved.
The tubing outer diameter was measured by means of a biaxial laser scanner (ODAC 32
XY, Zumbach Electronic AG, Switzerland) in 0.5 mm increments along the longitudinal
axis. The tube wall thickness TW
along the longitudinal axis was determined using the arithmetic mean of the outer
tubing diameter DT
and the outer mandrel diameter DM
according to equation 1. Furthermore, the average wall thickness variation TVAR
was calculated from the mean of the maximum DTmax
and minimum outer tube diameter DTmin
as well as the mean tube wall thickness according to equation 2.
The mandrels were removed and PLLA tubes were dried for 24 h at 40 °C. In order to
minimize residual solvent content, the PLLA tubes were washed in methanol and deionized
water using a platform-shaking device (Unimax 1010, Heidolph Instruments GmbH & Co.
KG, Germany) at 100 rpm and 20 °C, for two days each. Finally, the PLLA tubes were
dried for two days in a vacuum at 40 °C.
For NBRS fs-laser manufacturing, PLLA tubes with a minimum length of 21 mm and a maximum
average wall thickness variation of TVAR
= 20 % were selected.
Femtosecond-laser manufacturing of microstents
NBRS manufacturing was conducted using an fs-laser system (Monaco 1035-80-60, Coherent
Inc., USA) embedded into a 4-axis CNC system (Star Cut Tube Monaco, Coherent Munich
GmbH & Co. KG, Germany). The NBRS design was transferred into a numerical control
(NC) file using CAGILA (CAM-Service GmbH, Germany). PLLA tubes were mounted on a 4 mm
ceramic mandrel and clamped in the laser system jaw chuck. Fs-laser processing was
conducted at a pulse rate of 10 kHz, with a pulse width of 276 fs and power of 0.3 W
at a cutting speed of 1 mm s-1 using Argon at 0.5 bar as the process gas.
Geometrical and morphological examination of microstents
Geometrical and morphological analyses were performed utilizing three NBRS prototypes
and one Neuroform EZ. Macroscopic images of the NBRS as well as a Neuroform EZ stent
were taken using a digital camera (EOS 70 D with Zoom Lens EF-S 18–55 mm 1:3.5–5.6
II, EF 55–200 mm 1:4.5–5.6 II USM, Canon, Japan) and reversal adapter (EOS-Retro,
Novoflex, Germany).
Stent length measurements were performed with a micrometer gauge (No. 164-163, Mitutoyo,
Japan). Further geometrical dimensions ([Fig. 1 ]) were investigated by optical microscopy (SZX16 with UC30 camera, Olympus, Japan).
For measuring the stent length at an intended vessel diameter, the stents were deployed
into a transparent rigid tube with an inner diameter of 3.0 mm. Surface morphology
and cutting edges were analyzed by means of scanning electron microscopy (SEM; Quattro
S, Thermo Fisher Scientific, USA) in environmental scanning mode at 0.5 mbar and 15 kV.
The vessel coverage ratio VCR was determined according to equation 3 comparing the mass of the stent with the mass
of an ideal hollow cylinder with an appropriate wall thickness and density [13 ].
The mass was measured with a precision balance (Comparator XP6U, Mettler Toledo, Switzerland)
and the density was estimated based on values from the literature (1.24 g/cm³ for
PLLA; 6.40 g/cm³ for Nitinol) [14 ]
[15 ].
Radial force and radial stiffness
Investigation of the radial force behavior was performed with a segmented head testing
machine (TTR2 with J-Crimp station, Blockwise Engineering LCC, USA) at 37 ± 2 °C and
a velocity of 0.05 mm s‑1 . Radial forces were measured from a fully expanded state down to a minimum diameter
of 1.0 mm and evaluated at the intended vessel diameter of 3.0 mm during compression
(radial resistive force, RRF) as well as during expansion (chronic outward force,
COF). Additionally, the radial stiffness, represented by the initial increase of the
radial force curve, was assessed. After radial force testing, the stent outer diameter
DO,C
was measured by optical microscopy at five positions along the longitudinal axis
and compared to the nominal diameter DO
in a fully expanded state prior to radial force testing ([Fig. 2 ]).
Fig. 2 Procedure for analysis of crimping and expansion behavior of stents: 1) measurement
of initial stent outer diameter at five positions along the longitudinal axis, 2)
diameter reduction to a specific crimping diameter, 3) release of stent as well as
measurement of stent outer diameter after crimping.
Kink behavior
The kink behavior was investigated by bending the stents around cylindrical mandrels
with radii ranging from 32.5 mm to 7.5 mm in 2.5 mm steps and from 7.5 mm to 2.5 mm
in 1.25 mm steps [16 ]. A 0.014” guide wire (Cruiser F, BIOTRONIK, Switzerland) was inserted into the stent
and bent 180° around a mandrel to contact the stent and mandrel over the entire length.
The bent state was documented using a digital camera (EOS 70 D, Zoom Lens EF-S 18–55 mm,
Olympus, Japan). The analysis was conducted under ambient conditions at 21 ± 2 °C
in air.
Resistance to local deformation
During stent-assisted coiling, the stent is locally loaded in the radial direction.
For measurement of radial resistance to local deformation, the expanded stent was
deformed by a prismatic plunger with a 1 mm radius. The plunger was connected to a
universal testing machine (Zwick BT1-FR2.5TN.D14, ZwickRoell GmbH & Co. KG, Germany)
using a 20 N load cell (XForce HP, ZwickRoell GmbH & Co. KG, Germany).
The force distance curves were measured at 37 ± 2 °C in air using a cross head speed
of 5 mm min‑1 until a maximum local reduction of stent diameter of 25 % was achieved.
Simulated use and stent release in vitro
For analysis of NBRS applicability, a simulated use test setup combined with a patient-specific
neurovascular vessel model was used ([Fig. 3 ]), as described previously [17 ]. The test setup based on A. Kemmling’s flow model consists of a water bath including
the patient-specific vessel model connected to an 80 cm long arterial access vessel
and a 6F introducer sheath (Terumo Radifocus Introducer II, Japan) [18 ]. A pulsatile pump (FlowTek 125, United Biologics Inc., USA) was connected to circulate
water at 37 ± 2 °C, applying a flow rate of 1.4 l min‑1 at a pulse rate of 60 min‑1 .
Fig. 3 Schematic representation of the simulated use test setup: water bath (A ) including the patient-specific vessel model of the left intradural vertebral artery
(B ) connected to an 80 cm long arterial access vessel (C ) and a 6F introducer sheath (D ); Medium circulation via pulsatile pump (E ); Stent implantation was conducted using an angiography catheter (F ) and a guide wire (H ) as pusher wire; Continuous saline flushing (I ) was implemented using a Tuohy-Borst adapter (G ). The highly complex curves along the path to the aneurysm are illustrated in Detail
B and are quantified by angles and radii.
The vessel model was manufactured using a Form 2 3D printer and Clear Resin (Formlabs
Inc., USA). The complex model curvature is illustrated in [Fig. 3 ] and was quantified by angles and radii. NBRS implantation was conducted using a
straight 4F angiography catheter with an inner diameter of 0.038”/1.03 mm (Terumo
Radifocus Optitorque, Terumo, Japan) with continuous saline flush through the side
port of a Tuohy-Borst adapter (Rotating Hemostatic Valve 0.115”/2.92 mm, Abbot, USA).
The catheter tip was positioned distal to the aneurysm neck, using a fitting hydrophilic
guidewire (Radifocus Guide wire M, J-tip, Terumo, Japan). After crimping, the NBRS
was transferred manually into a straight 4F diagnostic catheter with an inner diameter
of 0.038”/1.02 mm (Merit Impress, Merit Medical Systems, USA), serving as the loading
device. The loading device tip was inserted into the Tuohy-Borst adapter and secured
in the distal catheter hub. The NBRS was transferred into the distal catheter using
a fitting straight hydrophilic guidewire (Radifocus Guide wire M, Terumo, Japan) as
the pusher wire, and continuously advanced to the distal catheter tip. NBRS expansion
was performed by keeping the pusher wire in position and pulling back the catheter
and placing it over the aneurysm neck.
In vitro coiling procedure
In vitro stent-assisted coiling was performed using a transparent technical vessel model with
a parent vessel diameter of 3 mm and a broad-based aneurysm of 4 mm. The expanded
Neuroform EZ stent and an NBRS prototype were subsequently inserted into the two-piece
acrylic glass aneurysm model which was secured using four plastic screws. The coiling
procedure was performed using a guide wire (Transend 0.010", Boston Scientific, USA),
a microcatheter (Excelsior SL-10, Stryker, USA) with a steam-shaped tip (60°) and
a coil (Target XL 360 Soft 3 mm × 9 cm, Stryker, USA) under combined fluoroscopic
(Philips Azurion ClarityIQ, Philips Healthcare, Netherlands) and video guidance (EOS
R5, Canon, Japan). Documentation and analysis of the vessel lumen size reduction were
carried out by single shot radiographs and macrographs.
Results
Manufacturing, geometrical and morphological examination of stents
PLLA tubes with inner and outer diameters of DM
= 4.00 ± 0.00 mm and DT
= 4.20 ± 0.01 mm and a resulting wall thickness of TW
= 100 ± 5 µm were manufactured (n = 20). A representative NBRS prototype and the Neuroform EZ stent are shown in [Fig. 4 ] as macrographs and SEM images in the expanded state. Both stents show sharp cutting
edges and a cylindrical shape without fractures or irregularities. The surface of
the electropolished Neuroform EZ stent appeared clearly smoother compared to the NBRS.
Fig. 4 Macrographs and SEM images of the Neuroform EZ stent (a, c ) as well as a representative NBRS prototype (b, d ); The SEM images show a stent ending (left), a detailed image of the connector (middle),
and a detailed image of the end segment (right).
The results of the geometrical examination of stents are summarized in [Table 1 ]. Stent lengths of the Neuroform EZ stent and the NBRS were comparable in the crimped
state as well as in the case of the indicated use diameter of 3.0 mm resulting in
a similar low length change for both stent types (1 % for Neuroform EZ and 0.7 % for
NBRS). The vessel coverage ratio was 43.53 % for the Neuroform EZ stent and 69.19 %
for the NBRS due to the larger strut width.
Table 1
Measured dimensions of NBRS prototypes (n = 3) and the Neuroform EZ stent (n = 1) in fully expanded state as well as measured lengths depending on stent diameter;
mean value ± standard deviation of n = 5 single measurements per dimension.
Dimension
Neuroform EZ
NBRS
Strut thickness TS
[µm]
70.58
± 0.7
94
± 5.44
Strut width WS
[µm]
51
± 2
109
± 12
Cell-to-cell length LCTC
[µm]
4791
± 29
4656
± 35
Cell-to-cell height HCTC
[µm]
1278
± 9
1200
± 44
Connector length LC
[µm]
293
± 3
358
± 8
Connector height HC
[µm]
222
± 7
357
± 9
Cell opening radius RCO
[µm]
68
± 1
50
± 2
Peak-to-peak cell spacing LCS
[µm]
158
± 22
103
± 24
Outer stent diameter DO
(fully expanded) [mm]
4.02
± 0.07
3.89
± 0.07
Outer stent diameter DO,C
(after crimping) [mm]
3.95
± 0.07
3.33
± 0.04
Stent length LS
(fully expanded) [mm]*
20.36
19.79
± 0.13
Stent length at DO
= 3 mm (vessel) [mm]*
20.65
20.51
± 0.04
Stent length at DO
= 1.2 mm (crimped) [mm]*
20.86
20.66
± 0.06
Stent length change (crimped to vessel) [mm]*
0.21
0.14
± 0.04
Stent length change (crimped to vessel) [%]*
1.01
0.70
± 0.21
*Stent length for Neuroform EZ measured without radiopaque markers
Radial force and radial stiffness
Exemplary radial force curves are shown in [Fig. 5 ]. Radial force increases with decreasing stent diameter for both stent types. For
small diameters below 1.2 mm, the NBRS shows an increasing radial force which is attributed
to the self-contacting stent struts and does not represent a realistic radial force
of the NBRS. RRF and COF at the indicated vessel diameter of 3.0 mm were 1.13 ± 0.22
N and 1.05 ± 0.17 N for the Neuroform EZ stent as well as 0.95 ± 0.09 N and 0.61 ± 0.03 N
for the NBRS. The Neuroform EZ stent showed a higher radial stiffness (5.23 ± 0.24 kPa/mm)
compared with the NBRS (3.99 kPa/mm) (see also [Table 2 ]). After radial force testing (equivalent to crimping down to DC
= 1.0 mm and subsequent release), the mean stent diameters DO,C
were 3.95 ± 0.07 mm (n = 1) for the Neuroform EZ stent and 3.33 ± 0.04 mm (n = 1) for the NBRS.
Fig. 5 Representative radial force curve progression for the Neuroform EZ stent and NBRS
during diameter reduction and expansion.
Table 2
Measured stent properties of NBRS prototypes (n = 3) and the Neuroform EZ stent (n = 1).
Measurements
Neuroform EZ (n = 1)
NBRS (n = 3)
Radial stiffness [kPa/mm]*
5.23
± 0.24
3.99
± 0.51
Radial resistive force [N] (at DO
= 3 mm)*
1.13
± 0.22
0.95
± 0.09
Chronic outward force [N] (at DO
= 3 mm)*
1.05
± 0.17
0.61
± 0.03
Force at local deformation (25 %) [mN]*
22.46
± 5.82
7.78
± 1.54
Stent mass [mg]**
12.71
5.20
± 0.03
Vessel coverage ratio [%]***
43.53
69.16
*Mean value ± standard deviation of n = 8 single measurements per stent
**Mean value ± standard deviation of n = 3 single measurements per stent
***Based on mean values for NBRS prototypes (n = 8)
Kink behavior
[Fig. 6 ] shows exemplary images of the bent NBRS and Neuroform EZ stent at different bending
radii. Both stent structures withstand the deformations and follow the respective
radius. The entire stent structure is in contact with the mandrel and no protruding
struts were observed on their inner and outer radii. Self-contact of both stent structures
was observed for radii smaller than 12.5 mm. No stent kinked at the minimum radius
of 2.5 mm.
Fig. 6 Representative macrographs documenting the kink behavior of the examined stents for
different radii R (a ) and the behavior of these stents during local compression CLOC
(b ) for the Neuroform EZ stent (top) and NBRS (bottom).
Resistance to local deformation
The analyzed Neuroform EZ stent (n = 1) and NBRS (n = 3) show a comparable deformation behavior under local deformation of CLOC
= 25 % based on the initial diameter. Respective forces increased to 24 ± 5 mN (Neuroform
EZ) and 8 ± 2 mN (NBRS) (n = 3 single measurements per stent).
Simulated use and microstent release in vitro
Consecutively, two NBRSs were loaded into the catheter, were advanced through a patient-specific
vessel model, and were deployed into the proximity of the aneurysm neck ([Fig. 7 ]). Microstent delivery could be performed without kinking, buckling, or stent relocation.
The stents expanded uniformly and adapted well to the rigid vessel wall.
Fig. 7 Crimping tool and partially loaded NBRS in 4F 0.038“ Merit Impress diagnostic angiography
catheter (a ). Implanted NBRS prototype in 3D-printed patient-specific vessel model of a vertebral
artery with an aneurysm (b ).
In vitro coiling procedure
[Fig. 8 ] shows the documented results of the coiling procedure for both stents. The microcatheter
was inserted through the stent struts into the aneurysm model. The coil was released
stepwise into the aneurysm sack. Reaching an appropriate coil package density was
verified by X-ray. The Neuroform EZ stent as well as the NBRS show only slight deformations
within the coiling region, indicating for the ability to retain the coil package.
Fig. 8 Representative macrographs documenting the coiling procedure in a transparent technical
vessel model for the examined stents (a ) and radiographs of the completed procedure of both stents (b ).
Discussion
The present study describes the design adaptation, manufacturing, and investigation
of a NBRS prototype in comparison to the commercially available Neuroform EZ stent.
To compensate for the inferior mechanical properties of PLLA, the NBRS design was
adapted with regard to wall thickness and larger connector elements. We investigated
the mechanical behavior including radial force, radial stiffness, and resistance against
local deformation as well as kinking. In addition, in vitro performance tests including a release procedure within a patient-specific vessel
model and an in vitro coiling procedure were performed.
Femtosecond laser parameter set led to true-to-shape contouring of the semi-finished
products with a small heat-affected zone and material re-deposition, thus achieving
comparably good results in morphological and geometrical examinations to the Neuroform
EZ stent [19 ]. In general, adequately equal quality of the produced prototypes was determined
(e. g., free length between two cells 4656 ± 35 µm for NBRS and 4791 ± 29 µm for Neuroform
EZ).
Good crimping capacity of the NBRS prototypes resulted in a minimum outer diameter
of 1.0 mm, allowing the use of small catheter profiles.
NBRS with a device diameter of 4.0 mm could potentially be used in target vessels
with a diameter of less than or equal to 3.0 mm, as the NBRS showed a potential for
self-expansion to 3.3 mm.
The radial force of the stent plays a key role regarding fixation inside the target
vessel and for supporting the coil package. Cho et al. analyzed the radial force of
four commercially available self-expanding stents for neuroradiological applications
and found radial force values ranging from 0.47 N to 2.59 N (e. g., 1.07 N to 1.51
N for Neuroform EZ 4.0/20 mm) when reducing its initial diameter to half [4 ]. Measured radial forces at an indicated use diameter of 3.0 mm for the Neuroform
EZ (1.05–1.12 N) were in good agreement to the results of Cho et al. The radial force
of the NBRS (0.61–0.95 N) was inferior but in the lower range compared to the results
of Cho et al. [4 ]. Radial stiffness as well as the reaction force after local deformation was lower
for the NBRS compared with the Neuroform EZ. However, within the in vitro coiling procedure both the Neuroform EZ and NBRS could successfully retain a coil
package. Further tests are needed to prove that radial forces are sufficient for a
successful in vivo application.
Stent flexibility is important to access the commonly tortuous cerebral target vessels.
In addition, stent apposition was observed as a key factor for neointimal coverage
[20 ]. Also, incomplete apposition is highly prevalent in patients with very late stent
thrombosis (> 30 days) after drug-eluting stent implantation, which can lead to abrupt
vascular closure and subsequently to adverse events [21 ]. Kink behavior results are promising as NBRS adaption to small radii was found,
which is generally favorable for use in intracranial vessels.
Patient-specific medical therapy procedures in the digital and analog space are currently
a main topic in medical research. Reconstructed 3D models of patient anatomies have
great potential for example to assist the education and training of medical professionals
as well as to improve implant and device development [17 ]. Current 3 D printing materials for the model prototypes represent patient-related
vessel physiology only to a limited extent. However, it was possible to perform implantation
procedures realistically and reproducibly.
To our knowledge, the current article is the first study using a self-expanding bioresorbable
polymeric microstent for stent-assisted coiling. Only one report about poly-tyrosine-derived
polycarbonate as bulk material in a balloon-expandable stent was found [22 ]. PLLA degrades completely in one to three years within the human body and can be
generally considered as non-toxic and non-inflammatory [23 ]. However, at an advanced stage of scaffold resorption, very late scaffold thrombosis
may occur [24 ]. Possible implications concerning endothelial damage and potential in-stent restenosis
at a late stage were reported, implying that PLLA degradation products have unfavorable
side effects on endothelial function [25 ].
The current study has several limitations: The small number of test samples does not
allow for statistical evidence of the test results. Furthermore, only in vitro tests were performed to characterize the acute performance properties of the NBRS. However,
an in vivo test will be needed to guarantee the safety and efficacy of the NBRS.
Future studies should concentrate on optimization of the NBRS stent design, especially
the reduction of the strut thickness while maintaining radial force properties and
resistance to local deformations. Stent designs should be verified with the help of
Finite Element Analysis. To guarantee X-ray visibility, it will be crucial to develop
specific X-ray markers to be added to the NBRS stent design.
Future studies should also cover the investigation of fatigue behavior considering
relevant load cases such as radial and torsional loading, as well as trackability
testing [26 ]
[27 ].
