The three major radioisotopes largely used in clinical trials for brachytherapy in the vasculature are iridium192 (Ir92), strontium and its breakdown product yttrium 90 and phosphorous. Ir92 is a gamma emitter and the other two are beta emitters, which may have some clinical importance. They have significantly varying half-lives. Gamma Ir92, essentially the only gamma source used for intravascular radiation for restenosis, has a half-life of 74 days. Strontium has a 28 year half-life, but yttrium has a 64 hour half-life. Phosphorous has a 14 day half-life. Table 1 shows the properties of the major radioisotypes used for intravascular brachytherapy.

The issues of whether to use gamma or beta radiation continue to be somewhat controversial. While there are significant differences in the physics and biological effects, it is unclear whether the differences have clinical relevance. Both forms can be used as low energy or high energy sources. Low energy sources are largely used for intravascular application. Both have exponential decay. Gamma has better depth penetration. Most importantly, gamma is not attenuated by calcium or metal such as stents, which may be an advantage. But, that is also a disadvantage in handling because gamma radiation must be shielded in tungsten safes and considerable care needs to be used in the catheterization laboratory to avoid radiation exposure. Yet, despite this, gamma radiation is relatively easy to use. Beta is severely attenuated by calcium, metal, and plastic, an advantage in handling as it can be placed in plastic containers that can be handled at the bedside by the interventional cardiologist.

Noteworthy is the exponential decay because both the beta and gamma sources have a very sharp drop off in radiation radially from the site of the source. The drop off is exponential but after several millimeters the dose becomes extremely low. The Ir92 has a much greater depth of penetration at any given level following the source. A dose volume histogram looking at the dose delivered to the amount of tissue being radiated showed that gamma penetrates more deeply and thus a higher dose will reach the adventitia when the same dose is delivered at the vessel lumen by both gamma and beta sources.

Surprisingly little is known about how radiation actually works. A study by Waxman in a swine model of cellular proliferation showed a decrease in cell proliferation at 3 days. At 7 days cell densities were the same between the control groups and two doses of radiation (14 Gy and 28 Gy). The effect on remodeling was the most remarkable result. The vessel size increased with both doses, and largely explains the preservation of lumen size in this animal model.

A fairly steep dose-response relationship has been shown in numerous animal studies. A study by Mazur shows a decrease in intimal proliferation in the area of the stenosis in the porcine model at 15 Gy and 28 Gy. Notably, at lower doses of radiation there was a nonsignificant increase in intimal hyperplasia and wall thickness, suggesting that low dose radiation perhaps has some potentially adverse effect. This finding has been seen in a number of other studies.

Remarkably, the experimental studies (gamma, beta, beta stent, beta balloon, external radiation) have been positive using intravascular applications, with the exception of only 2 of the 7 studies with external radiation being positive. Further study is needed to determine why external radiation was not successful. In the area of restenosis it is unusual to see this much uniformity of results, with the reduction in restenosis larger than reductions reported for pharmacological therapy. These studies are summarized in Table 2.

Table 2. Experimental studies of intravascular brachytherapy. (Faxon 2000)

De novo lesion studies

The Condado study from Venezuela used an Ir192 guide wire at a dose of 20 Gy. In the 21 patients studied, there was a 27% restenosis rate. The late loss index (LLI) was 0.19 mm, indicating probable beneficial effect. The BERT trial used graded doses of strontium, with an average of 14 Gy, in 82 patients in the United States and Canada. The overall restenosis rate was 24% and the LLI of 0.08 was surprisingly low. The Verin pilot study of 18 patients from Switzerland with strontium 90 at a dose of 3 Gy had a 50% restenosis rate. The ARREST pilot study of 25 patients with an Ir192 dose of 12 Gy had a 47% restenosis rate and a LLI of 0.7 mm.

The PREVENT trial with p32 is the only randomized trial in native circulation to date. Patients with de novo lesions that were stentable (about 50% received stents) were randomized on a 3:1 ratio to radiation or control. In the 72 patients, the reduction in restenosis was 42%. The restenosis reduction was 81% when considering only the center portion of the radiated area. Table 3 summarizes these studies.

Table 3. Summary of clinical studies of intravascular radiation. (LLI, late loss index) (Faxon 2000)

Dose may be very important. The studies that were negative (with restenosis rates similar to those expected with balloon angioplasty alone) used radiation doses in the range of 3-12 Gy. The studies that were positive (due to a lower than expected restenosis rate) used doses of 14-20 Gy.

Restenosis studies

The results from the instent restenosis studies have been positive and consistent, in contrast to the mixed results in de novo lesions (Table 4). The Scripps study showed a marked 69% reduction in restenosis in the 24 patients randomized to Ir192, compared to the 28 randomized to placebo (p=0.01). Three-year follow-up data show this benefit is preserved, with a 47% difference between groups (p<0.05). The mixed patient group may be relevant. Some had in stent restenosis and some had restenosis without a stent and received a stent at the time of radiation.

Table 4. Summary of clinical studies using radiation in coronary instent restenosis. (LLI, late loss index) (Faxon 2000)

The WRIST (Washington Radiation for In-Stent Restenosis Trial) trial included 120 patients with in stent restenosis. About 40% had stents placed within stents, a so-called stent sandwich. In WRIST the restenosis rate was 48% (or 58% depending on the QCA system used), and was 16% in the control group (or 19%) on follow-up. The LLI was 0.9 mm, and 0.36 mm in the control group on follow-up.

The IVUS component of WRIST showed a substantial decrease in intimal hyperplastic volume following angioplasty with or without radiation due to either the ablative techniques used or plaque shifts. But, during follow-up, the Ir192 group showed virtually no change in the amount of intimal hyperplasia, while the placebo group showed a recurrence of intimal hyperplasia to the same degree seen at baseline.

The Gamma-I study had similar results, with a 56% reduction in restenosis. This study differs due to its larger and more varied patient population. The restenosis rate was nearly 60% in the placebo group and was about 36% at follow-up. Adverse clinical events during follow-up were similar to those in WRIST.

The START trial is the only beta radiation instent restenosis study. A 66% reduction in restenosis (from 40% to about 12%) within the stent itself was achieved. A 47% reduction was obtained for the entire area radiated. A 36% reduction in restenosis was obtained for the entire lesion segment including the edges of the radiated area. Clinically, this is important because an edge stenosis can result in clinical restenosis. Target lesion revascularization was reduced 42% in the radiation group compared to the placebo group, target vessel revascularization 34% and major adverse cardiac events 31%.

The ARTISTIC pilot study of 26 patients with Ir192 used the smallest radioactive guide wire available (14,000th of an inch) thus it can traverse any size vessel in any location. The in lesion restenosis rate was 20% and the in stent restenosis rate was 13.3%. The in lesion LLI was 0.12 mm and in stent LLI was 0.06 mm. The large randomized trial of about 350 patients is halfway to completion.

To summarize the clinical studies of instent restenosis, radiation reduced restenosis in the Scripps trial by 69%, in WRIST 67%, in GAMMA-I 56%, and in START 36%. The differing restenosis rates in these trials in the control groups indicate that the patient population differs. However, in every setting intravascular radiation significantly reduced the restenosis rate of in stent restenosis to perhaps an acceptable level. Therefore, it is likely that a device will be approved in the US by year end.

Radioactive stents are of great interest due to the vision of the stent itself preventing abnormal remodeling of the vessel and the radiation preventing intimal hyperplasia, making this perhaps the perfect tool to prevent restenosis. However, the results in the pilot studies of randomized trials have been disappointing. The Iris 1A and 1B studies from the US have a 31% and 32% restenosis rate, respectively. The Milan study, a dose finding study with escalating doses of radiation, had angiographic restenosis rates ranging from 40-55%.

Edge restenosis, which is largely increased intimal hyperplasia, may explain why radioactive stents have failed to fulfill their promise thus far. However edge restenosis is not unique to radioactive stents, with an occurrence rate of 5-29%, and with clinical trials reporting some degree of edge restenosis.

Another factor leading to the failure of radiation to prevent restenosis could be geographic miss, meaning that the radioactive wire is not placed to bridge the damaged area that the balloon has injured by at least 5 mm. Geographical miss can lead to a fairly highly rate of restenosis in the missed segment. Dosing may be another issue. In the GAMMA-I study a clear dose-response relationship can be seen, with the maximum benefit achieved at more than 14 Gy. The studies using 12 Gy or lower had minimal effect. Erbel found that low dose (12 Gy) beta radiation actually seemed to increase LLI (nearly 0.60 mm), supporting findings in animal studies. Low doses may not only lack benefit, but have a detrimental effect. Dose-response relationships and the minimally effective dose are becoming increasingly understood.

Total occlusion is a recently noted problem in radiation. In the WRIST trial, 30 of the 104 patients who returned for follow-up had total occlusion; a 12% incidence for the overall trial. Focal in stent stenosis was found in 12 patients, diffuse in stent stenosis in 7, and new lesion in 12 patients. In the GAMMA-1 trial there was no stent thrombosis in the first 30 days in the Ir192 or placebo groups. But, at the 9-month follow-up there was a stent thrombosis in 8 patients in the Ir192 group and none in the placebo group. An MI rate of 12% in the Ir192 group and 6.6% in the placebo group was found. Both studies are influenced by the use of an additional stent on top of the original stent. The START Trial had a very low incidence of stents within stents and had prolonged use of antiplatelet drugs for at least 6 months (an addition during the course of the study). START had a very low rate (0.4%) of early or late stent thrombosis.

Stenting without radiation has a 1.2% incidence of stent thrombosis. Stenting within a stent with radiation is associated with a 7-10% stent thrombosis rate. The most likely mechanism is lack of re-endothelialization, and therefore prolonged antiplatelet therapy is needed. The FDA now requires prolonged antiplatelet therapy for at least six months in all randomized trials in the US. Edge restenosis can be addressed by adequate coverage with the source, and reoccurrence by optimizing the dose.