ASAIO Journal 2002
Application of Stereolithography for Scaffold Fabrication for
Tissue Engineered Heart Valves
RALF SODIAN,* MATTHIAS LOEBE,* ANDREAS HEIN, DAVID P. MARTIN, SIMON P. HOERSTRUP, EVGENIJ V. POTAPOV,*
HARALD HAUSMANN,* TIM LUETH, AND ROLAND HETZER*
A crucial factor in tissue engineering of heart valves is the
functional and physiologic scaffold design. In our current
experiment, we describe a new fabrication technique for
heart valve scaffolds, derived from x-ray computed tomogra-
phy data linked to the rapid prototyping technique of stereo-
lithography. To recreate the complex anatomic structure of a
human pulmonary and aortic homograft, we have used ste-
reolithographic models derived from x-ray computed tomog-
raphy and specific software (CP, Aachen, Germany). These
stereolithographic models were used to generate biocompat-
ible and biodegradable heart valve scaffolds by a thermal
processing technique. The scaffold forming polymer was a
thermoplastic elastomer, a poly-4-hydroxybutyrate (P4HB)
and a polyhydroxyoctanoate (PHOH) (Tepha, Inc., Cam-
bridge, MA). We fabricated one human aortic root scaffold
and one pulmonary heart valve scaffold. Analysis of the heart
valve included functional testing in a pulsatile bioreactor
under subphysiological and supraphysiological flow and pres-
sure conditions. Using stereolithography, we were able to
fabricate plastic models with accurate anatomy of a human
valvular homograft. Moreover, we fabricated heart valve
scaffolds with a physiologic valve design, which included the
sinus of Valsalva, and that resembled our reconstructed aortic
root and pulmonary valve. One advantage of P4HB and
PHOH was the ability to mold a complete trileaflet heart
valve scaffold from a stereolithographic model without the
need for suturing. The heart valves were tested in a pulsatile
bioreactor, and it was noted that the leaflets opened and
closed synchronously under subphysiological and supraphysi-
ological flow conditions. Our preliminary results suggest that
the reproduction of complex anatomic structures by rapid
prototyping techniques may be useful to fabricate custom
made polymeric scaffolds for the tissue engineering of heart
valves. ASAIO Journal 2002; 48:1216.
T he principle of tissue engineering represents a new and
promising concept to create functional, viable tissue from
From the *Department of Thoracic and Cardiovascular Surgery,
Laboratory for Tissue Engineering, German Heart Institute Berlin, Ber-
lin, Germany; Tepha, Inc., Cambridge, Massaachusetts; Surgical
Robotics Lab, Department of Maxillofacial Surgery, Charit, Humboldt
University Berlin, Campus Virchow-Klinikum, Berlin, Germany; and
Clinic for Cardiovascular Surgery, University Hospital, Zurich,
Switzerland.
Submitted for consideration July 2000; accepted for publication in
revised form May 2001.
Reprint requests: Ralf Sodian, Department of Thoracic and Cardio-
vascular Surgery, Laboratory for Tissue Engineering, German Heart
Institute Berlin, Augustenburger Platz 1, 13353 Berlin, Germany.
12
autologous cells for surgical application. In our laboratory, we
focus on tissue engineering of heart valves for cardiac surgery.
Our intent is to use biodegradable polymer matrices as a
scaffold for tissue development until the cells produce their
own matrix. Autologous cells are seeded onto the polymer
scaffold, forming viable tissue while the polymer degrades.
This in vitro created construct could potentially then be im-
planted, integrated into the native tissue, and use biologic
mechanisms for repair, remodeling, and growth.1 Although the
feasibility of in vitro and in vivo formation of valvular tissue
has been demonstrated in previous animal experiments, the
fabrication of an optimum heart valve scaffold with satisfactory
hemodynamic function continues to be a significant
problem.13
To overcome this problem, we developed a new technique
to fabricate a custom made heart valve scaffold for tissue
engineering. The excellent hemodynamic characteristics of
allografts, such as pulmonary and aortic homografts, are
widely acknowledged.4,5 In our experiment, we used stereo-
lithographic models derived from x-ray computed tomography
to recreate the complex anatomic structure of a pulmonary and
aortic homograft. The polymeric materials we used for scaffold
fabrication were biodegradable, biocompatible, and suitable
for appropriate tissue formation in vitro and in vivo.6 8
Materials and Methods
Scaffold Materials
The following porous scaffolds were evaluated for scaffold
fabrication for tissue engineering of heart valves that used
rapid prototyping techniques: (1) Poly-3-hydroxyoctanoate-co-
3-hydroxyhexanoate (PHOH; molecular weight [MW]
100,000 by gas phase chromatography [GPC]) is a biopolyes-
ter that was produced through a proprietary fermentation pro-
cess (Tepha, Inc., Cambridge, MA). PHOH is a semicrystalline,
thermoplastic elastomer with a melting point of approximately
51C and a glass transition temperature of 35C. A salt
leaching technique was used to produce a porous film (poros-
ity, 60 65%; pore size, 180 200 m; thickness, 200 500
m).9,10
Poly-4-hydroxybutyrate (P4HB; MW 700,000 by GPC) is a
biopolyester, produced through a proprietary fermentation
process (Tepha, Inc.). P4HB is a semicrystalline, thermoplastic
elastomer with a melting point of approximately 60C and a
glass transition temperature of 51C. To fabricate a porous
scaffold for tissue engineering of heart valves, P4HB was dis-
solved in anhydrous dioxane (5% wt/vol.) to produce a viscous
solution. The polymer solution was poured into a mold, which
 TISSUE ENGINEERED HEART VALVES 13

allowed rapid heat transfer. The polymer solution was rapidly

cooled to 25C, to freeze the dioxane and, thereby, precip-
itate the polymer. The composite was kept frozen for 30 to 60
minutes before solvent removal. The dioxane was removed by
freeze drying or by leaching into cold water. After solvent
removal, a porous foam of polymer was produced. The pore
size and distribution could be controlled by varying the freez-
ing process, polymer concentration, or by the addition of
porogens (e.g., sodium chloride). As with PHOH, we produced
a porous film (porosity, 80%; pore size, 80 180 m; thickness,
200 500 m).
The environmental scanning electron microscopic examina-
tion of both unseeded scaffold materials revealed a highly
porous surface, which is important for potential cell attach-
ment onto the polymer.6 Thus, we generated porous, three
dimensional polymers for tissue engineering of heart valves
(Figure 1, A and B).

Experimental Setting

In our current experimental setting, we used a pulmonary

and an aortic homograft to accurately recreate the anatomic
structure of a human heart valve. The homograft specimen
came from organ donors who could not be accepted for heart
transplantation. Furthermore, the homograft specimen used
could not be used for human cardiac surgery because of a
positive microbiological test. Harvest and preparation of the
homografts were performed under strict guidelines of the lab-
oratory for homografts at the German Heart Institute Berlin.
The distal part of the allograft was oversewn, and the coronary
arteries were ligated. Using a standard syringe connected to a
continuous air flow, the homograft was distended so that the
leaflets remained in a closed position. Continuous air flow was
used to keep the leaflets in a closed position and to guarantee
the distension of the sinus of Valsalva during the scanning
process. At this point, the air pressure inside of the homografts
was not measured. The whole experimental setup was then
placed into a computed tomographic (CT) scanner.
The flow chart for fabrication of an individually adapted
heart valve is shown in Figure 2. This workflow consists of the Figure 1. A: Porous surface of the polymeric scaffold material
following steps: (poly-3-hydroxyoctanoate-co-3-hydroxyhexanoate [PHOH]). B: Po-
rous surface of the polymeric scaffold material (poly-4-hydroxybu-
Image acquisition (e.g., with a CT scanner) tyrate [P4HB]).
Three-dimensional reconstruction of the shape of the valve
and the vessel using stereolithography (SLA 250, 3D Systems,
Valencia, CA) Belgium) was used for the conversion of the segmented image
Fabrication of a stereolithographic plastic model (Acrylat, data into a stereolithography (STL) model. Within this file
3D Systems) format, the surface of the valve was encoded as a list of
Fabrication of the polymeric scaffold using a stereolitho- triangles. The stereolithography machine was fed these data
graphic model and a plastic model was fabricated (Figure 3).11 In addition, we
fabricated a negative cast from the ventricular side of the
Rapid Prototyping Process homograft for the subsequent scaffold fabrication process. This
experiment was done for one human aortic and one pulmo-
The basis for the construction of optimum heart valves in nary homograft.
respect to their hemodynamic function is the exact spatial
reconstruction of the individual shape. Therefore, in the first
Fabrication of Heart Valve Scaffold from a Stereolithographic
step, the prepared valve was scanned with a CT scanner, a
Plastic Model
Tomoscan M (Philips, Eindhoven, Netherlands). Slice distance
was 2 mm, and the pixel size within all slices was 0.1 0.1 The stereolithographic plastic models were then used to
mm. The image data were visualized, and three dimensionally fabricate a polymer heart valve scaffold. The scaffolds were
reconstructed using a software tool developed by us (AMIRA- constructed from PHOH and P4HB by wrapping the polymer
ANAPLAST). The planing tool Mimics (Materialise, Leuven, films around the plastic models and forming a trileaflet heart
14 SODIAN ET AL.
Figure 2. Flow chart for the fabrication of an individually adapted
trileaflet heart valve. CT, computed tomographic; 3D, three dimen-
sional; STL, stereolithography.
valve by a thermal processing technique as previously de-
scribed.12 The trileaflet heart valve scaffolds were fabricated
from a single piece of porous PHOH and P4HB foam, which
required no additional suturing. Afterward, the scaffolds were
smoothly pressed into a negative cast and allowed to crystal-
lize at 4C for 24 hr. Thus, we fabricated heart valve scaffolds
that resembled the complicated structure of the human allo-
grafts and measured the height, length, and inner diameter of
the homografts, plastic models, and polymeric scaffolds.
Figure 3. Three dimensional reconstruction of the human homografts. Right: 3D surface of the aortic root in the AMIRA system. Left:
planing of the fabrication of the structure in the Mimics system (pulmonary homograft).
Functional Testing of the Heart Valve Scaffolds
Analysis of the heart valve scaffolds included functional
testing in a pulsatile bioreactor system under subphysiological
and supraphysiological flow (100 4,500 ml/min) and pressure
(5300 mm Hg) conditions.13 The testing device was com-
pletely transparent to facilitate observation of the proper func-
tion of the polymeric heart valve scaffold. Direct pressures
were measured with a digital pressure measurement device
(Baxter, Pressure Monitoring Kit, Irvine, CA) distal and proxi-
mal to the polymeric heart valve scaffold (Figure 4).
Results
Reconstruction of a Human Allograft
Using data derived from x-ray computed tomography linked
to the rapid prototype technique of stereolithography, we were
able to construct plastic models with the exact valvular design
of human aortic and pulmonary homografts (Figure 5, A and
B). These stereolithographic models were then used as physi-
ologic patterns to generate scaffolds for the tissue engineering
of heart valves.
Fabrication of Heart Valve Scaffolds
Using stereolithographic plastic models, we were able to
manufacture a three-dimensional, biodegradable scaffold from
thermoplastic poly-3-hydroxyalkanoate-co-3-hydroxyhexano-
ate (PHOH) and poly-4-hydroxybutyrate (P4HB), which were
used in prior studies for the tissue engineering of heart
valves.7,8,14 This material was molded into a whole trileaflet
heart valve scaffold requiring no suture materials or additional
stent. The heart valve scaffolds resembled the complex anat-
 TISSUE ENGINEERED HEART VALVES 15

Figure 4. Direct pressure measurement distal and proximal to the

polymeric heart valve scaffold: (1) testing device with heart valve
scaffold, (2) air driven pump, (3) direct pressure measurements
distal to the scaffold, (4) direct pressure measurements proximal to
the scaffold, (5) Marquette Transport Display.

omy of a human valvular homograft, which included the sinus

of Valsalva and the coronary arteries (Figure 6). In addition, the
polymeric scaffolds only showed minor differences in height,
length, and inner diameter (less than 1 mm) compared with
the homografts and stereolithographic plastic models.

Heart Valve Scaffold Function

Evaluation of the scaffold function showed a mild stenosis
and regurgitation in all heart valve scaffolds (Table 1). All
scaffolds opened and closed synchronously in a pulsatile flow
bioreactor under normal and supranormal flow and pressure
conditions.
Figure 6. Stereolithographic models and heart valve scaffolds
from PHOH (A) and P4HB (B). A: Seen from the ventricular side of
the pulmonary homograft. B: Side view of the reconstructed aortic
root, including the sinus of Valsalva and the coronary arteries.

Discussion and allografts.1719 The major problem with these substitute

In tissue engineering, polymer scaffolds are used as three
dimensional matrices for autologous cells to develop perma-
nent tissue for replacement of an area of defect or injury.15 The
major goal of this approach is to replace failed tissue with
living tissue that is designed and constructed outside the pa-
tients body to restore and improve tissue function.16 The
newly fabricated device consists of autologous tissue that the-
oretically has the ability to grow, remodel, and repair.
Currently, in the field of cardiovascular surgery, heart valves
and vessels are replaced with totally artificial substitutes such
as mechanical valves, synthetic vascular prostheses, or non-
living processed tissue such as glutaraldehyde fixed xenografts
Figure 5. Stereolithographic model of the aortic root. Right: Ven-
tricular side of the aortic valve. Left: Aortic side of the aortic valve,
including sinus of Valsalva and coronary arteries.
devices is that they consist of nonliving foreign body material
without the ability of normal viable tissue to grow or regener-
ate. To create viable and functional tissue, with no risk of
immunogenicity and rejection would be an ideal solution for
heart valve and vascular replacement, especially in patients
with congenital heart defects.
Because of the limitations associated with current valve
substitutes, our laboratory has focused on tissue engineering of
heart valves. In subsequent experiments in Dr. Vacantis and
Dr. Mayers laboratory, we were able to fabricate a whole
trileaflet tissue engineered heart valve in vitro and in vivo.7,8,14
In these experiments, we could demonstrate appropriate tissue
formation, as well as biomechanical and biochemical proper-
ties of the tissue engineered constructs. Although the function
of the tissue engineered heart valves was satisfying, the accu-
rate reconstruction of the physiologic valve design, including
coronary arteries and the sinus of Valsalva, remain a significant
issue for the field of tissue engineering of heart valves. For this
reason, we applied rapid prototyping techniques for the fabri-
cation of heart valve scaffolds, to recreate the complex ana-
tomic structure of a human aortic and pulmonary valve.
In our experiment, we were able to reconstruct a human
homograft by using data derived from x-ray computer tomog-
raphy linked to the rapid prototyping technique of stereolithog-
raphy. Because of the thermoplastic and elastic properties of
PHOH and P4HB, it was possible to mold a heart valve
scaffold that resembled the anatomic structure of a human
homograft.
Furthermore, we present a detailed technical description of
a method we have developed to advance our current scaffold
16 SODIAN ET AL.

Table 1. Direct Pressure Measurements of the Heart Valve ditionally, the fabrication of a custom made scaffold could be
Scaffolds demonstrated in our current experiment.
Direct
Direct Pressure References
Pressure Measurements
Measurements Distally to 1. Shinoka T, Ma PX, Shum-Tim D, et al: Tissue engineered heart
Proximally to the Heart valves. Autologous valve leaflet replacement study in a lamb
Inspiration/ Frequency Stroke the Heart Valve model. Circulation 94(Suppl II): II164 II168, 1996.
Expiration in Strokes/ Volume (ml) Valve Scaffolds 2. Vacanti JP, Langer R: Tissue engineering: the design and fabrica-
(Harvard Min (Harvard (Harvard Scaffolds (mm (mm Hg tion of living replacement devices for surgical reconstruction
Apparatus) Apparatus) Apparatus) Hg SD) SD) and transplantation. Lancet 354(Suppl I): 3234, 1999.
3. Sodian R, Sperling JS, Martin D, Mayer JE, Vacanti JP: Tissue
30/70 50 10 4.5 0.7 4.5 0.7 engineering of a trileaflet heart valve: Early in vitro experiences
30/70 50 25 10 10 with a combined polymer. Tissue Eng 5: 489 493, 1999.
30/70 50 50 30.5 0.7 27.5 3.5 4. Schoen FJ, Mitchell RN, Jonas RA: Pathological considerations in
30/70 50 100 50 2.8 43.5 2.1 cryopreserved allograft heart valves. J Heart Valve Dis 4: 72
30/70 50 200 109 1.4 106 1.4 76, 1995.
30/70 50 300 174 2.8 153 2.8 5. Nanda N, Helmcke FR, Bourge RC: Long-term function of cryo-
30/70 50 400 227.5 3.5 200 preserved aortic homografts. J Thorac Cardiovasc Surg 106:
30/70 50 500 278.5 12 240.5 0.7 154 166, 1993.
6. Sodian R, Hoerstrup SP, Sperling JS, et al: Evaluation of biode-
gradable, three dimensional matrices for tissue engineering of
heart valves. ASAIO J 46: 107110, 2000.
design for tissue engineering of heart valves that opened and 7. Sodian R, Hoerstrup SP, Sperling JS, et al: Early in vivo experience
closed synchronously under subnormal and supranormal flow with tissue-engineered trileaflet heart valves. Circulation.
and pressure conditions, with a maximum gradient of 20 102(Suppl 3): III22III29, 2000.
mm Hg under systemic systolic pressure conditions. On the 8. Sodian R, Hoerstrup SP, Sperling JS, et al: Tissue engineering of
heart valves: In vitro experiences. Ann Thorac Surg 70: 140
basis of this initial technical description, we believe that in situ
144, 2000.
reconstruction of a human heart valve and direct polymer 9. Williams SF, Peoples OP: Biodegradable plastics from plants.
stereolithography seem feasible. In addition, the material prop- Chemtech 26: 38 44, 1996.
erties have been evaluated in several in vitro and in vivo 10. Mooney DJ, Breuer CK, McNamara K, Vacanti JP, Langer R:
experiments and showed appropriate mechanical strength for Fabricating tubular devices from polymers of lactic and glycolic
acid for tissue engineering. Tissue Eng 1: 107118, 1995.
surgical application.7,8,14 However, the scaffold materials that 11. Cohen L, Cohen I: Finite element methods for active contour
we use are both biodegradable and biocompatible and cells models and balloons for 2D and 3D images. IEEE Trans Pattern
attached onto the polymer grew into a porous system that Recognition Machine Intelligence 15: 11311147, 1993.
finally formed viable tissue in vitro and in vivo.8,20 12. Sodian R, Sperling JS, Martin DP, et al: Fabrication of a trileaflet
Although we are encouraged by early in vitro and in vivo heart valve scaffold from a polyhydroxyalkanoate biopolyester
for use in tissue engineering. Tissue Eng 6: 183188, 2000.
results, tissue engineering of heart valves is still at an early 13. Hoerstrup SP, Sodian R, Sperling JS, Vacanti JP, Mayer JE Jr: New
stage of development and numerous issues remain to be ad- pulsatile bioreactor for in vitro formation of tissue engineered
dressed. In addition to the biological problems such as cell heart valves. Tissue Eng 6: 7579, 2000.
origin, cell conditioning, and guided tissue development for 14. Vacanti JP: The in vitro construction of a tissue engineered bio-
prosthetic heart valve. Eur J Cardiothorac Surg 11: 493 497,
tissue engineered constructs, an important area of future work
1997.
will include the development of polymers to ultimately mini- 15. Langer R, Vacanti JP: Tissue engineering. Science 260: 920 926,
mize the residual polymer fragments in our tissue engineered 1993.
devices. In addition, we are attempting to reconstruct a human 16. Niklason LE, Gao J, Abbott WM, et al: Functional arteries grown
heart valve under in situ conditions. Therefore, active experi- in vitro. Science 284: 489 493, 1999.
17. Vongpatanasin W, Hillis D, Lange RA: Prosthetic heart valves.
mental research is currently under way to develop technolo- N Engl J Med 335: 407 416, 1996.
gies to visualize and reconstruct a moving heart valve and, 18. Jamieson WR: Modern Cardiac Valve devices-bioprostheses and
finally, fabricate a custom made heart valve scaffold using mechanical prostheses: State of the art. J Cardiac Surg 8: 89 98,
stereolithography. 1993.
Although further in vitro and in vivo studies are required to 19. Schoen FJ, Levy RJ, Piehler HR: Pathological considerations in
replacement cardiac valves. Cardiovasc Pathol 1: 29 52, 1992.
evaluate the biological function of the newly created heart 20. Hoerstrup SP, Sodian R, Daebritz S, et al: Functional living trileaf-
valve scaffolds, the feasibility of using rapid prototyping tech- let heart valves grown in vitro. Circulation 102(Suppl 3): III44
niques to reconstruct a complex anatomic structure and, ad- III49, 2000.