Endovascular Irradiation From β-Particle–Emitting Gold Stents Results in Increased Neointima Formation in a Porcine Restenosis Model
Background—Recent studies have shown that ionizing radiation reduces neointima formation after balloon angioplasty and stent implantation in experimental models of restenosis and first clinical trials. The objective of this study was to determine the dose distribution of a new β-particle–emitting radioactive gold stent and to evaluate the dose-dependent vascular response in the coronary overstretch pig model.
Methods and Results—Sixteen Göttinger minipigs underwent placement of 11 nonradioactive and 36 β-particle–emitting stents with activity levels of 10.4±0.6, 14.9±2.4, 22.8±1.3, 35.8±2.8, and 55.4±5.3 μCi of 198Au. Three months after implantation, the percent area stenosis, neointimal thickness, neointimal area, and vessel injury were analyzed by quantitative histomorphometry. The lifetime radiation doses at a depth of 1 mm were 3.3±0.2, 4.7±0.5, 7.2±0.4, 11.4±0.9, and 17.6±1.7 Gy for the different activity groups. No dose-response relationship was observed in the radioactive stents with respect to percent area stenosis (P=0.297), mean neointimal thickness (P=0.82), or mean neointimal area (P=0.65). Significantly lower neointima formation and less luminal narrowing was seen in the control group than in the β-particle–emitting stents (P<0.001). Multilinear regression analysis revealed that only radioactivity made a significant independent contribution to the degree of percent area stenosis (P<0.001).
Conclusions—Neointima formation in pigs is markedly increased by β-particle–emitting stents with 198Au as the radioisotope. This study provides evidence that dosages of 3 to 18 Gy of low-dose-rate β-particle irradiation via endovascular stents cause pronounced luminal narrowing in the animal model at 3 months.
Coronary stent placement is increasingly used to improve the early outcome1 and reduce late restenosis after PTCA.2 3 However, substantial expansion of the arterial lumen comes at the expense of a potent stimulus for neointimal proliferation leading to stent restenosis, which occurs in 20% to 40% of patients after intracoronary stent placement.1 2 3 A systemic or local antiproliferative therapy used in conjunction with a stent is thought to function by inhibiting the neointima formation that causes late lumen loss after stent implantation. The concept of using local radiation to control neointima formation is based on the assumption that ionizing radiation suppresses smooth muscle cell proliferation and thereby inhibits neointima formation. This therapy has been applied to benign situations such as keloid4 or heterotopic bone formation.5 In general, 2 different strategies can be followed in terms of intravascular brachytherapy, using either high-dose-rate catheter-based systems with β- (eg, 90Y)6 or γ- (eg, 192Ir)7 8 emitters or low-dose-rate irradiation with radioactive stents (eg, 32P).9 10 The benefits of using β-emitters include their ease of use and practical integration into current catheterization laboratory procedures. Radiation energy from β-sources diminishes rapidly with distance and therefore does not require special radiation shielding in the catheterization laboratory. Endovascular radiation therapy from radioactive sources has been shown to effectively inhibit neointima formation under experimental conditions9 10 11 12 13 and in first clinical studies.8 Hehrlein et al13 described the use of intra-arterial radioactive Palmaz-Schatz stents produced by particle bombardment in a cyclotron. His results clearly indicated that low-dose radioactive stents potently inhibit smooth muscle cell proliferation. In a subsequent study, these results were confirmed with a pure β-particle–emitting stent.10 Others were able to demonstrate the efficacy of β-particle–emitting stents to reduce neointima formation in a porcine coronary restenosis model.9
Coronary stents with a 5-μm-thick gold coating have recently been introduced in routine clinical practice. Through neutron irradiation, 197Au can be converted to 198Au, a β-particle emitter that can be used for endovascular brachytherapy.
The objective of the present study was to investigate the dose-response effects of β-particle irradiation from a 198Au stent with respect to neointima formation using the coronary overstretch pig model.
Stainless steel stents 8 mm long, with a 5-μm-thick gold coating and multicellular design (InFlow Dynamics) (Figure 1⇓), were activated by irradiation in the Forschungsreaktor München (Technische Universität München, Garching, Germany). The irradiation time varied according to the required 198Au activity. Neutron beam radiation caused 197Au atoms to be converted into 198Au by the reaction 197Au+n=198Au. The nuclide 198Au has a half-life of 2.7 days and decays by β-particle emission, with the following major emissions: β (98.6%), with 0.96 MeVmax and 0.314 MeVmean; and γ (95.5%), with 0.412 MeV. The activity of each stent was calculated by measuring the 198Au 412-keV γ-line with a Ge(Li) detector coupled to a Canberra GENIE γ-ray spectroscopy system. The 198Au stents were implanted at mean activity levels of 10.4, 14.9, 22.8, 35.8, and 55.8 μCi. The control stents were identical to the radioactive stents, except that they were not subjected to neutron irradiation. Staff and implanting physicians always used protective goggles during implantation, and contact time between the stent and the gloved hands of the physician was kept at a minimum (<20 seconds).
Dosimetry Measurement Using Radiochromic Films
Depth doses for soft tissue equivalent to the material surrounding an activated stent were measured by means of radiochromic films (Gafchromic Dosimetry Media, type MD-55, ISP Technologies Inc). These films acted as dosimeters and simultaneously as an absorbing medium. The total thickness amounted to 0.25 mm. To obtain a dose-calibration curve, films were irradiated with 137Cs γ-radiation with doses from 5 to 80 Gy and 90Sr β-radiation with doses from 5 to 40 Gy. Calibration and measuring films were scanned by means of a CCD flatbed scanner (Ultrascan 5000, Vexcel) with a pixel size of 20×20 μm and 16-bit data depth. The measuring films were evaluated along the gray-value-versus-dose-calibration curve. The depth-dose distribution was measured with a stack of 10 layers of radiochromic films placed in a U-shaped formation around the stent. Consequently, the dose values given later are to be considered mean values averaged over a thickness of 0.25 mm in distances to the stent surface of 0.125, 0.375, 0.625 mm, etc.
This study was performed with Göttinger minipigs (32 to 57 kg, Ellegaard Göttingen Minipigs ApS, Dalmose, Denmark). Animal research was conducted according to guidelines of the institutional review board and animal use committees. A total of 16 Göttinger minipigs underwent random placement of 48 gold stents in the left anterior descending, left circumflex, or right coronary artery. Animals were medicated with aspirin 500 mg and amiodarone 150 mg IV. Under general anesthesia, a 7F sheath was placed retrogradely in the left carotid artery, and heparin 100 U/kg IV was administered. The stents were implanted according to the randomization protocol with the guiding catheter used as a reference to obtain a 1.2:1 balloon-to-vessel ratio compared with the baseline vessel diameter. The implanting physician was blinded to the type of stent radioactivity. The stents were hand-crimped on PTCA balloons as used for coronary angioplasty (Cruiser II, Nycomed Amersham, or Viva, Boston Scientific Corp), with balloon diameters applied from 2.5 to 3.5 mm. Placement was achieved with a single balloon inflation at 6 to 18 atm for 30 seconds. Angiography was completed after implantation to confirm patency of the stent and side branches as well as to assess for stent migration and intraluminal filling defects. Animals were allowed to recover and were returned to care facilities. They underwent follow-up coronary angiography 12 weeks after implantation, and were again deeply anesthetized, fully heparinized, and euthanized with an overdose of pentobarbital.
Each angiogram was evaluated for evidence of stent migration, intraluminal filling defects, side-branch occlusion, lumen narrowing, and distal coronary flow characteristics by standard criteria for quantitative coronary angiography. The baseline mean lumen diameter, balloon-inflated stent, and postimplant and follow-up coronary artery minimal lumen diameters (in mm) were measured from within the stented segment in nonoverlapped and nonforeshortened views with the guiding catheter used for image calibration. The acute balloon-to-vessel ratio (minimal stent balloon-inflated diameter/baseline lumen diameter) was calculated from these data for each stented vessel.
Immediately after euthanasia, the heart and ascending aorta were excised and the coronary arteries were perfusion-fixed with 6% neutral buffered formalin at 100 mm Hg for 15 minutes via the aortic stump. The stented coronary arteries were dissected from the epicardial surface, left in formalin overnight, and thereafter embedded in methylmethacrylate. Sections 8 μm thick were obtained from 5 different levels of the stented coronaries with a stainless steel carbide knife and stained with hematoxylin-eosin and van Gieson’s elastica stains. All segments were evaluated by computer-assisted histomorphometry. The segment most narrowed was then used for further calculations. Therefore, the data presented are composed of the worst levels obtained from each individual stent. The cross-sectional area of each section was measured with digital morphometry (NIH Image 1.59 for quantitative analysis) to determine the areas within the internal elastic lamina (IEL) and vessel lumen. The percent area stenosis was then defined as IEL−lumen area/IEL×100. The area of the neointima was determined by subtracting the area of the lumen from the IEL. Neointimal thickness extending perpendicular to the lumen was measured at each strut site. The severity of stent-induced vascular injury was graded according to the method of Schwartz et al.14 Neointimal and medial cell densities were measured in 5 randomly chosen 0.1-mm2 areas close to the stent struts of both the β-particle–emitting and the control stents at ×200 light magnification. The number of neocapillaries within the neointima was counted. The density was determined as number of neovascularizations per neointimal area and compared between radioactive and control stents.
Data are presented as mean±SD. Angiographic data at stent implantation, cell densities for the control and radioactive stents, and morphometric data between the different groups of β-particle–emitting stents were compared by the Kruskal-Wallis nonparametric ANOVA test with post hoc test for multiple comparisons. Percent area stenosis, neointimal thickness, and neointimal area between all radioactive and control stents were compared by Mann-Whitney rank sum test. Multiple linear regression analysis was applied to assess the impact of injury and radioactivity (taken as the independent variables) and their interaction (injury score × radioactivity) on percent area stenosis (taken as the dependent variable). Binary variables (value 0 or 1) denoting radioactive treatment were added to the regression equation to evaluate the significance of the treatment. Significance was established at P<0.05, except for Bonferroni-corrected Mann-Whitney tests (P<0.01 [0.05/5]).
γ-Ray spectroscopy revealed that >99.8% of the total stent activity was due to 198Au. In addition, 51Cr, 187W, and 65Zn contributed to the total stent activity with 0.1%, 0.006%, and 0.008%, respectively, with half-lives of 7.7 (51Cr), 23.7 (187W), and 244.3 days (65Zn). 55Fe and 59Fe were not detectable.
Figure 2⇓ is a plot of the calculated lifetime radiation dose at radial distances of 0.125, 0.375, 0.625, 0.875, 1.125, and 1.375 mm from the surface of a 46.5-μCi (1720×103-Bq), 3.3-mm-diameter stent. Minimum (cell area) and maximum (strut intersections) doses were averaged over 6×6 pixels measured at 8 different positions. At depths of 0.625, 0.875, and 1.125 mm, a uniform dose is delivered to the vessel. At depths <0.375 μm, however, the radiation distribution is highly heterogeneous, with maximum and minimum doses of ≈260 and ≈140 Gy, respectively, delivered to the vessel 0.125 mm from the struts. At a distance ≥1.375 mm outward from the stent, the radiation dose was below the detection threshold of the dosimeter film. On the basis of these results and the known activity of each stent at implant, the lifetime radiation dose at a depth of 1 mm was calculated for the different stent activity groups (Table⇓).
Forty-seven of 48 stents were successfully implanted in the left anterior descending (n=16), left circumflex (n=16), and right (n=15) coronary arteries of 16 pigs. One stent was lost in the aorta before the stent was advanced into the coronary system. One animal died of reasons unrelated to stenting at day 18. Another animal developed pneumonia 10 weeks after stent implantation, returned at day 73 (10.4 weeks) for follow-up angiography, and was euthanized. The remaining animals survived for 12.9±0.3 weeks (12.3 to 13.6 weeks) without complications.
Baseline mean lumen diameter and balloon-inflated and post–stent inflation minimal lumen diameters were not significantly different between the groups. Quantitative analysis of the coronary angiograms at implantation revealed a mean balloon-to-artery ratio of 1.16±0.11, with a range of 1.08 to 1.20. There were no significant differences between the radioactive and control stents. Typical edge effects were not observed within the group of the radioactive stents.
Coronary histology of the control stents demonstrated a neointima that consisted of a dense, circumferentially arranged population of smooth muscle cells on the lumen surface with less well-organized spindle-shaped cells, rare multinuclear giant cells, and some neovascularization in the region of the stent struts. A neointima with haphazardly placed cells near the stent struts and moderately increased extracellular matrix but appearance otherwise similar to that of control stents was seen in almost all radioactive stents. No reduction in neointimal (P=0.179), medial (P=0.07), or adventitial (P=0.159) cell density was seen in arteries with β-particle–emitting stents compared with control stents. There was some inflammatory reaction at the site of both the β-particle and control stent struts. The inflammatory reaction consisted of groups of histiocytic cells adjacent to the struts with occasional multinucleated foreign-body giant cells. In contrast, eosinophils, plasma cells, and neutrophils were not prominent. Quantitative differences within different treatment groups were not observed. A considerably thickened adventitia was observed in most of the radioactive stents. The vessel morphometry for the β-particle–emitting and control stents is summarized in the Table⇑. Microscopic examination revealed proliferative neointimal responses and lumen stenosis of various magnitudes in all groups. Comparison of the different β-particle–emitting stent groups revealed no significant differences for mean percent area stenosis (P=0.297, Figure 3A⇓), mean neointimal thickness (P=0.82), and mean neointimal area (P=0.65). In the control group, however, a significantly lower neointimal formation and less luminal narrowing was seen than in the β-particle–emitting stents (P<0.001, Figures 3B⇓ and 4⇓, top and bottom). Quantitative differences with respect to neointimal thickness, percent area stenosis, and neointimal area within the different sections from 1 stent were <5%, demonstrating a rather even vessel wall response to the stent implantation. There was a trend toward greater neovascularization in the radiated group (10.1±3.92 versus 14.01±10.08). However, this did not reach the level of significance (P=0.504).
To assess the contributions of injury score, radioactivity, and their interaction term (injury score×radioactivity) to percent area stenosis, (dependent variable) multiple linear regression analysis was applied. According to this model, only radioactivity had a significant positive correlation to the percent area stenosis (P<0.001 for radioactivity, P=0.229 for injury score, and P=0.082 for the interaction term).
The aim of the study was to describe the dose-response effects of a low-dose-rate β-particle–emitting 198Au stent in the coronary overstretch pig model. However, no dose-response relationship was found for stents implanted with activities ranging from 5 to 55 μCi with respect to a reduction of neointima formation, and all radioactive stents revealed more severe neointimal responses and greater luminal narrowing than the control stents.
Several other β-emitters, such as 32P and 90Sr/90Y, have been used in different animal models and first clinical studies that revealed feasibility and efficacy to reduce neointima formation.6 9 10 In general, β-emitters show a rapid dose falloff versus distance because of the large number of low-energy electrons in their respective spectra that have concomitant short ranges.15 32P has a maximum energy of 1.7 MeV and a mean energy of 0.69 MeV; the corresponding properties for 90Sr/90Y are 2.27 MeV and 0.97 MeV, respectively.15 Compared with these 2 isotopes, 198Au has far lower transition energies (0.96 MeV maximum energy and 0.312 MeV mean energy) and therefore has the greatest dose falloff. On the basis of the maximum energy and the dose measurements performed, the tissue range of 198Au β-particles is <1.5 mm, compared with 5 to 6 mm with 32P.15 Taking these basic physical properties into consideration, the radioactive 198Au gold stent of the design used should fulfill the desired criteria for endovascular brachytherapy in the animal model to reach neointima, media, and adventitia with sufficient dose distribution.15
In the pig model, the greatest number of proliferating cells 2 to 3 days after angioplasty was found primarily in the adventitia surrounding the injured artery.16 In the same model 1 week after angioplasty, cell proliferation was reduced in the media and adventitia, with the greatest number of proliferating cells found in the neointima. Myofibroblasts in the adventitia proliferate after angioplasty and may migrate into the neointima, where they appear as smooth muscle cells.16 Thus, hypothetically, intravascular irradiation is effective in inhibiting neointima formation, if a significant reduction of proliferating adventitial cells can be achieved.16 With high-dose-rate brachytherapy, doses ≤20 Gy targeted at a distance of 2 mm from the source18 appear to be efficacious in lowering or even preventing restenosis.17 18 As a rule in radiobiology, however, the biological effect of a given dose is lowered if the dose rate is reduced and the overall exposure time is increased.19 This implies that with low-dose-rate endovascular irradiation via radioactive stents, the applied dosages have to be considerably higher to be effective. Because a multicellular stent design was used, the maximal distance from strut to strut for vessel diameters ≤3 mm was ≤550 μm. On the basis of the results from the dose measurements, this approach allowed application of doses of 50 to 100 Gy to the adventitia and neointimal tissue (mean distance from strut to adventitia, 200 to 400 μm) in our highest-activity group. Thus, an effective prevention of restenosis was expected, but the opposite effect was seen.
Because the implantation characteristics and vascular injury scores were comparable in all groups, these somewhat unexpected findings cannot be attributed to higher vascular injury in the irradiated groups. Furthermore, the results from the multiple linear regression analysis showed that only radioactivity made a significant independent contribution to the percent area stenosis. This lends additional support to the hypothesis that irradiation itself caused the pronounced neointima formation.
Several preclinical studies have demonstrated that intracoronary radiation reduces neointima formation when applied immediately after vessel injury.7 9 10 11 13 However, some found an accentuated neointima response after radiation.20 21 Most of the studies showing significant efficacy performed a 4-week follow-up, and rates of proliferation of smooth muscle cells were usually not assessed. Only limited data on long-term studies after radiation therapy are available. At least 1 recently completed study, conducted by Virmani et al,21 evaluated the long-term dose-response effects of 32P stents with activities of 0 to 12 μCi in a proliferative double-injury pig model. For stents with activities ≥1 μCi, a significant increase was found in mean neointimal area, percent restenosis, and neointimal thickness compared with nonradioactive stents.21 Furthermore, for stents with activities from 1 to 12 μCi, no dose-response relationship was found.21 There are considerable differences between our study and the double-injury atherosclerosis model with 6 months of follow-up used by Virmani et al. Despite these differences, the principle of an increased proliferative response after endovascular irradiation using low-dose-rate brachytherapy is seen in both studies. However, the underlying mechanisms remain unclear. It is probable that the radiation applied was insufficient to inhibit smooth muscle cell proliferation. In addition, the irradiation-induced delay in endothelialization facilitated fibrin-platelet deposition, and the resulting thrombus promoted cell proliferation and matrix protein production. This aspect is important in the development of restenosis, especially in the porcine animal model.
Catheters and radioactive stents deliver radiation at different dose rates. Catheter studies using 192Ir or 90Sr/90Y delivered the dose between 3 minutes and 1 hour, and dose rate varied from ≈0.2 Gy/h to ≈2.5 Gy/min.6 7 The dose-rate range of importance in radiotherapy extends from 0.1 Gy/h to several Gy/min. In this range, the fraction of cells killed by a given dose decreases as the dose-rate is reduced, principally because of the repair of sublethal damage.19 With radioactive stents, the dose rates are some 100-fold lower than high-dose-rate brachytherapy. However, Laird et al11 demonstrated that 32P-labeled stents with an initial activity as low as 0.14 μCi, which resulted in a dose rate of <0.001 Gy/h, inhibited neointima formation in porcine arteries. Thus, low-dose radiation may impair cell proliferation without producing cell death. These results, in conjunction with the present study, further point toward a complex interaction between low-dose-rate β-particle–emitting stents and cellular wall vascular elements.
The results observed in this experimental model may not sufficiently reflect the pathological mechanisms that occur in human restenosis after angioplasty. Because other experiments and clinical trials using isotopes with different physical properties have been shown to effectively reduce neointima formation, these results cannot be applied toward low-dose-rate endovascular irradiation in general. In our study, no antiplatelet or anticoagulation regime was used after the initial injury and treatment procedure. Thus, endovascular irradiation was insufficient to inhibit smooth muscle cell proliferation, and an irradiation-induced delay in endothelialization facilitated fibrin-platelet deposition, promoting cell proliferation and matrix protein production. Therefore, the contribution of insufficient inhibition of smooth muscle cell proliferation and local thrombus formation for increased neointima formation must be considered.
Low-dose-rate β-particle irradiation via endovascular stents with activities from 5 to 55 μCi causes increased neointima formation and greater luminal narrowing in an experimental animal model of restenosis. This study provides evidence for a complex interaction between radiation dose, dose rate, and vessel wall cellular elements. These results suggest that implanting radioactive 198Au gold stents, which deliver dosages from 3 to 18 Gy to the coronary artery targeted at a distance of 1 mm from the stent surface, provokes rather than reduces neointima formation.
- Received November 11, 1999.
- Revision received February 14, 2000.
- Accepted February 18, 2000.
- Copyright © 2000 by American Heart Association
Schömig A, Kastrati A, Mudra H, et al. Four-year experience with Palmaz-Schatz stenting in coronary angioplasty complicated by dissection with threatened or present vessel closure. Circulation. 1994;90:2716–2724.
Fingeroth RJ, Ahmed AQ. Single dose 6 Gy prophylaxis for heterotopic ossification after total hip arthroplasty. Clin Orthop. 1995;317:131–140.
Verin V, Urban P, Popowski Y, et al. Feasibility of intracoronary β-irradiation to reduce restenosis after balloon angioplasty: a clinical pilot study. Circulation. 1997;95:1138–1144.
Mazur W, Ali MN, Khan MM, et al. High dose rate intracoronary radiation for inhibition of neointimal formation in the stented and balloon-injured porcine models of restenosis: angiographic, morphometric, and histopathologic analyses. Int J Radiat Oncol Biol Phys. 1996;36:777–788.
Carter AJ, Laird JR, Bailey LR, et al. Effects of endovascular radiation from a β-particle-emitting stent in a porcine coronary restenosis model: a dose-response study. Circulation. 1996;94:2364–2368.
Hehrlein C, Stintz M, Kinscherf R, et al. Pure β-particle-emitting stents inhibit neointima formation in rabbits. Circulation. 1996;93:641–645.
Laird JR, Carter AJ, Kufs WM, et al. Inhibition of neointimal proliferation with low-dose irradiation from a β-particle–emitting stent. Circulation. 1996;93:529–536.
Waksman R, Robinson KA, Crocker IR, et al. Endovascular low-dose irradiation inhibits neointima formation after coronary artery balloon injury in swine: a possible role for radiation therapy in restenosis prevention. Circulation. 1995;91:1533–1539.
Hehrlein C, Gollan C, Donges K, et al. Low-dose radioactive endovascular stents prevent smooth muscle cell proliferation and neointimal hyperplasia in rabbits. Circulation. 1995;92:1570–1575.
King SB, Williams DO, Chougule P, et al. Endovascular radiation to reduce restenosis after coronary balloon angioplasty: results of the Beta Energy Restenosis Trial (BERT). Circulation. 1998;97:2025–2030.