1.

Introduction

Cerebral aneurysm rupture occurs in approximately 30,000 people per year in the United States, with most resulting in death or neurological debilitation. Drawbacks of current endovascular treatment using detachable embolic coils¹ to fill the aneurysm include the time-consuming delivery of multiple coils and the need for retreatment due to compaction of the coils and associated reexposure of the aneurysm wall.^{2, 3} Further, when the aneurysm neck is greater than $4\mathrm{mm}$ , complete occlusion using coils is achieved in only 15% of the cases.¹ In these instances, the wide neck allows the coils to migrate or unravel from the aneurysm into the parent artery, potentially increasing the risk for aneurysm regrowth and rupture.⁴

Recently, shape memory polymer (SMP) foam based on DiAPLEX Company, Ltd. (a subsidiary of Mitsubishi Heavy Industries) polyurethane has been investigated as a material to occlude aneurysms with promising embolic performance⁵ and biocompatibility.^{5, 6} SMPs can be formed into a specific primary shape, reformed into a stable secondary shape, and then controllably actuated to recover the primary shape. A review of SMP basics and representative polymers was given by Lendlein and Langer.⁷ For SMPs that are actuated thermally, raising the temperature of the polymer above the glass transition temperature $\left(T_g\right)$ results in a decrease in the elastic modulus from that of the glassy state $(\sim {10}^9\mathrm{Pa})$ to that of an elastomer $(\sim {10}^6\text{to}{10}^7\mathrm{Pa})$ . For SMP foams, this drop in modulus translates to an expansion force during thermal actuation ( $<1\mathrm{kPa}$ at recovered strains over 50%) that is much less^{8, 9} than the force required to rupture an aneurysm wall $\left(700\text{to}5000\mathrm{kPa}\right)$ . Upon cooling, the original modulus is nearly completely recovered and the primary form is stabilized. When the $T_g$ of the SMP is above body temperature $(37°\mathrm{C})$ , an external heating mechanism such as photothermal (laser heating) or electrical (resistive heating) is required.¹⁰

Researchers at Lawrence Livermore National Laboratory (LLNL) have developed an endovascular treatment technique to deliver thermally deployed SMP foams.¹¹ This paper presents fabrication of LLNL-synthesized SMP foam devices and laser deployment in a polydimethylsiloxane (PDMS) basilar-necked aneurysm model. Experimental analysis of the photothermally deployed foams in the model under bounding parent artery flow conditions (high and zero flow) is also presented.

2.

Materials and Methods

2.2.

Prototype Device Fabrication

A spheroid $(11\mathrm{mm}\times 10\mathrm{mm})$ was cut from the LLNL SMP foam using a scalpel and hollowed out by hand using a ball-tipped drill bit. The foam was positioned over a cylindrical light-diffusing fiber (made in-house from LLNL SMP) such that the open end of the spheroid faced proximally and collapsed using a crimping machine (Model W8FH, Interface Associates) at $93°\mathrm{C}$ (see Fig. 1 ). Upon reaching the collapsed diameter, the device was allowed to cool to room temperature and then released from the machine. The device maintained the collapsed form prior to actuation, as shown in Fig. 1b.

Fig. 1

(a) LLNL SMP hollow foam spheroid prior to collapse, $11.0\times 10.0\mathrm{mm}$ , and (b) post-collapse, $1.8\times 13.0\mathrm{mm}$ , on the light diffusing fiber. The open end of the spheroid is facing right. The light diffusing fiber is shown in the inset in (b) coupled to a red diode laser. (c) Image of the eight-blade cylindrical crimping machine with heated blades used to collapse the foam.

To fabricate the diffusing fiber, LLNL SMP $(T_g=81°\mathrm{C})$ was cast in a Teflon tube (inner $\text{diameter}=300\mu \mathrm{m}$ ) over a $100\text{-}\mu \mathrm{m}$ core cleaved optical fiber. The resulting SMP rod was then media blasted with $100\text{-}\mu \mathrm{m}$ sodium bicarbonate particles to create a diffusing surface. An ST connector was added to the proximal end of the optical fiber for coupling to the laser light source, an $810\text{-}\mathrm{nm}$ continuous-wave diode laser pigtailed into a $100\text{-}\mu \mathrm{m}$ core optical fiber (Model UM7800/100/20, Unique Mode). An overlying thin-walled stainless steel hypotube (inner $\text{diameter}=356\mu \mathrm{m}$ , outer $\text{diameter}=508\mu \mathrm{m}$ , Heraeus Vadnais, Inc.) was incorporated to add stiffness to the optical fiber for device delivery. Approximately 80 to 90% of the light was emitted radially, with the remaining light emerging from the distal end of the diffuser. The SMP formulation was specifically designed to be optically transparent so as not to absorb the laser wavelength (absorption $\text{coefficient}=0.01{\mathrm{cm}}^{-1}$ at $810\mathrm{nm}$ ).

2.3.

Benchtop Deployment of Foam

With the diffusing fiber inside the collapsed device, the proximal end of the optical fiber was coupled to the $810\text{-}\mathrm{nm}$ diode laser. The PDMS aneurysm model (Fig. 2 ) was filled with $21°\mathrm{C}$ water, and the collapsed device was delivered into the aneurysm. Although water at $37°\mathrm{C}$ (body temperature) would have provided a more accurate representation of physiological conditions and heat flow, the relatively low $T_g$ of the available foam $(45°\mathrm{C})$ required a lower water temperature to prevent spontaneous expansion prior to laser heating (the glass transition begins to occur $\sim 10°\mathrm{C}$ below the nominal $T_g$ ). Flow rates varied between 0 and $148\mathrm{cc}∕\min$ , where 0, 70, and $148\mathrm{cc}∕\min$ were typically used because they represent blocked, diastolic, and systolic flow in the basilar artery,¹³ respectively. Laser powers varied between 1 and $8.6\mathrm{W}$ . The foam expansions were video recorded, and the temperature was measured by a needle probe thermocouple (HYP1, Omega) protruding slightly into the aneurysm at the dome apex. While this single thermocouple may not represent the maximum temperature in the aneurysm, it provides a means of comparing the effects of the various laser powers and flow rates.

Fig. 2

(a) Clear PDMS (Sylgard 184, Dow Corning) generic basilar-necked aneurysm model made in house (dome dimensions: $11.3\mathrm{mm}$ apex to neck, $11.7\mathrm{mm}$ wide). The model consists of two PDMS halves pressed between polycarbonate plates with stainless steel alignment pins. The positive molds used to cast the PDMS were machined from polycarbonate stock using a computer-controlled rapid prototyping mill (Roland MDX-650). The thermocouple port permits measurement of the temperature at the apex of the aneurysm dome. The device is inserted into the model via a Touhy Borst valve. A peristaltic pump (Model 505Du, Watson-Marlow, Ltd.) provided flow through the system.

3.

Results

Figure 3 shows a combined timeline of laser power, flow rate, temperature, and images of the foam expansion. The flow in the aneurysm prevented sufficient heating of the outer region of the foam and thus prevented full expansion during $\sim 200\mathrm{s}$ of $148\mathrm{cc}∕\min$ flow and $8\mathrm{W}$ of laser power. Not until the flow was reduced did the foam fully expand. When the flow was reduced to $0\mathrm{cc}∕\min$ , the apex temperature quickly increased by over $20°\mathrm{C}$ . A subsequent experiment in which the flow was maintained at $0\mathrm{cc}∕\min$ and the laser power was set to $8.6\mathrm{W}$ resulted in full expansion of the foam within $60\mathrm{s}$ with a temperature rise of $\sim 30°\mathrm{C}$ (not shown in figures).

Fig. 3

Temporal history of aneurysm apex temperature, foam expansion, flow rate, and laser power. The bars above the plot display the laser power and flow rate time lines. The plot shows the corresponding apex temperature and the letters (A to H) on the plot correspond to the foam expansion images. Full expansion was prevented by convective cooling due to the high flow rate ( $148\mathrm{cc}∕\min$ ) and did not occur until the flow was turned off ( $0\mathrm{cc}∕\min$ ).

Rapid overheating associated with zero flow and the cooling impact of high flow were further demonstrated in a fully expanded foam undergoing laser heating (not shown in figures). At $148\mathrm{cc}∕\min$ , the steady-state apex temperature rise was limited to 1, 2, and $3°\mathrm{C}$ for laser powers of 2, 4, and $6\mathrm{W}$ , respectively. At $0\mathrm{cc}∕\min$ , the apex temperature increased by approximately 10, 20, and $30°\mathrm{C}$ within the first minute of laser heating for 2, 4, and $6\mathrm{W}$ , respectively.

To overcome the cooling effect of flow, the foam was doped with additional dye to increase laser heating; the hollow spheroid was coated with a solution of $\sim 4400\mathrm{ppm}$ dye in tetrahydrofuran followed by vacuum drying to remove the solvent. Figure 4 shows the results of foam deployment at a flow rate of $70\mathrm{cc}∕\min$ as the laser power was gradually increased to $8.6\mathrm{W}$ over $\sim 190\mathrm{s}$ . Full expansion was achieved in $\sim 180\mathrm{s}$ , and the apex temperature never exceeded $2°\mathrm{C}$ . A subsequent experiment using the same foam spheroid under the same flow condition $(70\mathrm{cc}∕\min )$ in which the laser power was increased to $8.6\mathrm{W}$ more quickly (over $\sim 100\mathrm{s}$ ) showed similar results (not shown in figures).

Fig. 4

Apex temperature and expansion of the foam at a flow rate of $70\mathrm{cc}∕\min$ . The darker color of the foam is due to the added dye. The laser power was gradually increased to $8.6\mathrm{W}$ over $\sim 190\mathrm{s}$ . The large arrow indicates the time of full expansion ( $\sim 180\mathrm{s}$ from laser on).

4.

Discussion and Conclusion

The foam device presented here will require several significant modifications before it is clinically viable. First, foam with a higher $T_g$ will need to be used to prevent spontaneous expansion in the body. Second, the foam device would need to be delivered through a microcatheter with a lumen diameter of $300\text{to}600\mu \mathrm{m}$ , requiring a reduction in the current collapsed diameter $\left(1800\mu \mathrm{m}\right)$ . Work on lower density foams with greater volume expansion is in progress. By further removing volume (e.g., holes, dimples, channels, etc.) without sacrificing embolic performance, foam devices capable of filling a $10\text{-}\mathrm{mm}$ aneurysm may potentially be compacted to the $\sim 500\mu \mathrm{m}$ diameter necessary for microcatheter delivery. Third, due to the impact of flow on foam expansion, the device may require some engineering of the flow via a baffle to reduce the flow near the aneurysm neck or a balloon to block flow¹⁴ in the parent artery during laser heating. At a temperature of $\sim 67°\mathrm{C}$ ( $\sim 30°\mathrm{C}$ above body temperature), thermal damage of arterial tissue occurs within a few minutes,¹⁵ a time scale similar to that of device deployment. Given the minutes time scale for thermal damage at $67°\mathrm{C}$ , our design goal is a maximum deployment time of $60\mathrm{s}$ and desired deployment time under $30\mathrm{s}$ . To avoid causing thermal damage to the already compromised aneurysm wall under zero or low flow conditions, the optimal combination of foam $T_g$ , laser power, dye concentration, and flow will need to be determined.

This preliminary study demonstrated that SMP foams can be laser deployed in an in vitro aneurysm model. The flow rate affected the deployment. Zero flow resulted in fast, full expansion with overheating at the aneurysm wall. Low (diastolic) flow resulted in slow, full expansion with minimal temperature increase at the aneurysm wall. High (systolic) flow resulted in incomplete expansion. Advances in the materials with the basic techniques presented here suggest the potential for a single device that can treat a $10\text{-}\mathrm{mm}$ aneurysm.