- Email: [email protected]
A subgroup analysis of the scripps coronary radiation to inhibit proliferation poststenting trial
Int. J. Radiation Oncology Biol. Phys., Vol. 42, No. 5, pp. 1097–1104, 1998 Copyright © 1998 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/98/$–see front matter
PII S0360-3016(98)00281-8
●
Clinical Investigation A SUBGROUP ANALYSIS OF THE SCRIPPS CORONARY RADIATION TO INHIBIT PROLIFERATION POSTSTENTING TRIAL PAUL S. TEIRSTEIN, M.D.,* VINCENT MASSULLO, M.D.,* SHIRISH JANI, PH.D.,* JEFFREY J. POPMA, M.D.,† GARY S. MINTZ, M.D.,† ROBERT J. RUSSO, M.D., PH.D.,* RICHARD A. SCHATZ, M.D.,* ERMINIA M. GUARNERI, M.D.,* STEPHEN STEUTERMAN, M.S.,* DAVID A. CLOUTIER, B.S.,* MARTIN B. LEON, M.D.,† AND PRABHAKAR TRIPURANENI, M.D.* *Divisions of Cardiovascular Diseases and Radiation Oncology, Scripps Clinic, La Jolla, CA, and †Division of Cardiology, Washington Hospital Center, Washington, DC Introduction: In the Scripps Coronary Radiation to Inhibit Proliferation Poststenting (SCRIPPS) Trial, 192Ir significantly reduced angiographic, ultrasonographic, and clinical endpoints of restenosis. The objective of this analysis was to quantitate the impact of patient, lesion and technical characteristics on late angiographic outcome. Methods: Patients with restenotic, stented coronary lesions were randomized to receive either 192Ir or placebo sources. Late luminal loss and loss index were calculated for several patient subgroups, including patients with diabetes, in-stent restenosis, multiple previous percutaneous transluminal coronary angioplasty (PTCA) procedures, longer lesion lengths, saphenous vein grafts, small vessel diameters, and minimum dose exposures < 8.00 Gy. Two-factor analysis of variance was used to test for an interaction between patient characteristics and treatment effect. Results: In the treated group, late loss was particularly low in patients with diabetes (0.19 mm), in-stent restenosis (0.17 mm), reference vessel diameters < 3.0 mm (0.07 mm), and patients who received a minimum radiation dose to the entire adventitial border of at least 8.00 Gy. The loss index in each of these subgroups was similarly low at 20.02, 0.03, 20.02, and 0.03, respectively. By 2-factor analysis of variance, a significant interaction between subgroup characteristic and treatment effect (late loss) was found in patients with in-stent restenosis (p 5 0.035), and patients receiving a minimum dose of 8.00 Gy to the adventitial border (p 5 0.009). Conclusion: In this pilot study, patient characteristics associated with a more aggressive proliferative response to injury appeared to confer an enhanced response to radiotherapy. Furthermore, a dose threshold response to 192 Ir was found with an enhanced response occurring when the entire circumference of the adventitial border was exposed to at least 8.00 Gy. © 1998 Elsevier Science Inc.
INTRODUCTION
patient, lesion, and technical characteristics on late angiographic outcome, and use this information to generate hypotheses that can be tested in future clinical trials.
Numerous animal trials (1–12) and several early-phase clinical trials (13–21) indicate a promising role for ionizing radiation in the treatment of coronary restenosis. Despite this limited early success, numerous basic questions have been raised. Most importantly, effective radiation delivery techniques and appropriate patient selection criteria must be defined. The Scripps Coronary Radiation to Inhibit Proliferation Poststenting (SCRIPPS) Trial was a pilot, doubleblind randomized study comparing 192Ir to placebo in restenotic, stented coronary lesions. The results of this trial demonstrated significant benefit in clinical, angiographic, and ultrasonographic endpoints of restenosis (14). Although the trial size was small, a subgroup analysis of angiographic endpoints may provide clues that can impact clinical trial design and aid in radiation device development. The objective of the present analysis was to quantitate the impact of
The SCRIPPS Trial was a double-blind randomized trial comparing 192Iridium with placebo sources. The clinical trial was approved by the institution’s Human Subjects and Radiation Safety Committees. Patient inclusion criteria required a target lesion in a restenotic coronary artery that either already contained a stent or was a candidate for stent placement. Previous target lesion intervention was required to occur . 4 weeks prior to study entry. The reference vessel was required to be between 3 and 5 mm in diameter and the target lesion # 30 mm in length. Patients were excluded if the coronary revascularization procedure was
Reprint requests to: Paul S. Teirstein, M.D., Scripps Clinic, Division of Cardiovascular Diseases, SW206, 10666 N. Torrey Pines Road, La Jolla, CA 92037.
Acknowledgments—The authors acknowledge the collaboration of Krishnan Suthanthiran of Best Industries. Accepted for publication 6 July 1998.
METHODS
1097
1098
I. J. Radiation Oncology
●
Biology
●
Physics
unsuccessful, a suboptimal result was achieved, a stent was implanted as an unplanned, emergency procedure, or the target lesion undergoing stent placement contained angiographic evidence of thrombus. Procedure Preprocedural medications included aspirin 325 mg, intracoronary nitroglycerin, and intravenous heparin to maintain an activated clotting time of . 300 s. If the lesion was not previously stented, single or, if required, tandem coronary stenting (Johnson & Johnson Interventional Systems, Warren, NJ) was performed. If stents had been previously placed, redilatation was undertaken and, in many such patients, additional stents were placed within the original stent as required to optimize the angiographic result. Intravascular ultrasound was then performed using a 3.2-Fr catheter (CVIS, Inc., Sunnyvale, CA) with a motorized pull-back device at 0.5 mm/s to interrogate the stented vessel segment. The intravascular ultrasound examination ensured that an optimal final result had been achieved and allowed assessment of lesion geometry for dosimetry calculation by the radiation therapist (see below). A 4-Fr infusion catheter (USCI, Inc., Bellireca, MA) was then inserted to span the stented region. Patients were randomized to receive a 0.03-inch ribbon (Best Industries, Springfield, VA) containing either 192Iridium (192Ir) sealed sources at its tip or a ribbon containing placebo, inactive sources. In this protocol, one of two devices were selected dependent on the lesion length. Discrete lesions were treated with a 19-mm ribbon containing 5 3-mm long sources, each separated by 1 mm. Longer lesions were treated with a similar ribbon containing 9 sources totaling 35 mm in length. All personnel, except one physics member from the Division of Radiation Oncology and one research nurse from the Division of Cardiology who were not involved in endpoint analysis, were blinded to the randomization code. The catheterization laboratory was then cleared of all personnel, except the radiation oncologist who inserted the study ribbon into the infusion catheter. Radiation sources were left in place for 20 – 45 min as required to administer the prescribed radiation dose. Sources were then removed by the radiation oncologist and placed into an adequately shielded container. Femoral sheaths were removed 2– 4 h after the procedure and the patient discharged the following morning taking aspirin 325 mg daily indefinitely and, if new stents were implanted, ticlopidine 500 mg daily for 2 weeks. Dosimetry 192 Iridium dosimetry was calculated in the following manner. A series of tomographic sections, obtained by intravascular ultrasound using a motorized pullback apparatus, were scanned and measurements were performed along the length of the stent. The distance between the center of the ultrasound catheter (equivalent to the “source” position) and the adventitial border (the “target”) was measured every 1 mm along the stented segment of the artery.
Volume 42, Number 5, 1998
A maximal, as well as minimal, source-to–target distance was determined. The radiation oncologist and physicist combined this information with the specific activity of the radioactive sources to determine a dwell time that provided 8.00 Gy to the target farthest from the radiation source, provided that no more than 30.00 Gy was delivered to the target closest to the radiation source.
Quantitative angiographic analysis After removal of radiation sources, intracoronary nitroglycerin was administered and final angiography performed in two orthogonal projections. At follow-up examination, identical angiographic techniques were used. All procedural and follow-up cineangiograms were forwarded to the Washington Hospital Center Angiographic Core Laboratory for analysis by observers who were blinded to the treatment allocation. Selected serial cineframes, obtained from two unforeshortened projections and matched for position within the cardiac cycle using side-by–side projectors, were digitized with a cinevideo converter using the contrast-filled catheter as the calibration standard. Reference vessel diameter, as well as minimal luminal diameter within the axial length of the stent, and minimal luminal diameter along the axial length of the radiation sources were determined using a validated edge detection program (CMS, MEDIS) (22) at baseline, after the procedure, and at follow-up. These measurements were used to determine serial relative % diameter stenoses within the stent alone and at the stent border (beyond the stent but still covered by the radiation sources). Acute luminal gain was defined as the final minimal luminal diameter minus the initial minimum luminal diameter (mm), and the late luminal loss was defined as the final minimum luminal diameter minus the follow-up minimum luminal diameter. The loss index was defined as the late luminal loss divided by acute luminal gain. Binary restenosis was defined as a $ 50% diameter stenosis at follow-up.
Subgroup analysis and statistics Six subgroups were prospectively specified by randomizing each patient from 1 of 8 randomization tables stratified by lesions with 1 stent spanned by 5 sources or 2 tandem stents spanned by 9 sources, in-stent restenosis or no previous stent, and vein graft or native lesion location. Other subgroups were retrospectively identified. These included diabetes, number of previous percutaneous transluminal coronary angioplasty (PTCA) procedures, reference vessel diameter, and minimum dose delivered to the advential border. For the analysis of continuous data, Mann–Whitney U tests were used to assess differences between the two treatment groups. The results are expressed as mean 6 SD with concomitant test probabilities. Discrete data were compared by chi-square or, where appropriate, Fisher exact probability tests. A 2-factor analysis of variance (ANOVA) was conducted to test for an interaction between various patient and lesion characteristics and treatment effect.
Scripps coronary radiation trial
●
P. S. TEIRSTEIN et al.
1099
Table 1. Baseline clinical and angiographic characteristics
Diabetes (%) In-stent restenosis (%) Vessel diameter , 3.0 mm (%) Saphenous vein target (%) Lesion length . 15 mm (%)
192
Placebo (n 5 28)
Ir (n 5 24)
p
39.3 64.3 71.4 28.6 53.6
29.2 75 66.6 25 66.7
0.637 0.593 0.944 0.980 0.499
RESULTS Between March and December of 1995, 55 patients were randomized and 52 patients had angiograms eligible for analysis; 28 were assigned to placebo and 24 to active radiation. Three patients were not included in this analysis because follow-up angiograms were not obtained (1 cardiac death in a placebo patients), early stent thrombosis occurred (1 192Ir patient) or technically could not be analyzed (1 patient). Baseline clinical and angiographic characteristics were similar for both groups (Table 1). Many study patients had one or more risk factors for restenosis, such as diabetes, in-stent restenosis, smaller diameter vessels, saphenous vein graft target, and longer lesion length. As reported previously, angiographic indices of restenosis were markedly different between placebo and treated patients (14). For the entire group, treatment with 192Ir suppressed late loss (1.03 vs. 0.38 mm, p 5 0.025), loss index (0.6 vs. 0.12, p 5 0.002) and the dichotomous ($ 50% diameter stenosis) angiographic restenosis rate (53.6% vs. 16.7%, p 5 0.01). The effect of diabetes on angiographic outcome is depicted in Fig. 1. Diabetic patients treated with radiation had very little late loss and a loss index of essentially zero, compared to a somewhat less pronounced effect in nondiabetic patients. To the right of each graph, however, is the result of a 2-factor analysis of variance, which tests for an interaction between treatment effect and the patient characteristic of diabetes. Statistically, there was only a slight trend towards a difference in treatment effect in diabetic vs. nondiabetic patients; with a p value of 0.113 for loss index. The difference between diabetic and nondiabetic patients is highlighted, however, by the very high loss index of 0.80 found in placebo diabetic patients compared to 0.50 in placebo nondiabetic patients. Therefore, despite a more aggressive restenotic process in diabetics, radiotherapy was extremely effective in suppressing late luminal loss. Patients with in-stent restenosis (Fig. 2) also demonstrated an enhanced response to radiation therapy compared to that in patients not previously stented. There was very little suppression of late loss in patients without previous stents, but a profound reduction in late loss and loss index in patients with previous stents. Here, the 2-factor analysis of variance for late loss is significant, indicating a statistically better treatment effect in patients with in-stent restenosis. Patients with multiple previous PTCA procedures of the
Fig. 1. The effect of diabetes on treatment with 192Ir. For this and subsequent figures, late loss is depicted for both placebo and 192Ir patients on the top of the figure and loss index is depicted at the bottom of the figure. Data pertaining to patients with diabetes is displayed on the right side of the figure and that from patients without diabetes is displayed on the left. To the right of each figure is the p value from the 2-factor analysis of variance (ANOVA), which tests for an interaction between the patient characteristic of diabetes and treatment effect. Diabetic patients treated with radiation had very little late loss and a late loss index of essentially zero, compared to a less pronounced effect in nondiabetic patients. The 2-factor ANOVA, however, is not significant, demonstrating only a slight trend toward a difference in treatment effect in diabetic vs. nondiabetic patients (p 5 0.113).
target lesion (Fig. 3) appeared to have a slightly better treatment effect than patients with only one previous PTCA. However, the differences here were small and not supported by the 2-factor analysis of variance. Similarly, longer lesions (Fig. 4) were associated with a slightly more pronounced treatment effect, but the differences between longer and shorter lesions were minor and not significant. Also, patients with lesions located in a saphenous vein graft (Fig. 5) appeared to have less of a treatment effect than patients with native coronary targets, but the 2-factor analysis here was also not significant. Perhaps most interestingly, vessel diameter appeared to be related to treatment effect. Lesions located within vessels , 3.0 mm in diameter (as measured by off-site quantitative angiography) demonstrated a profound treatment effect, with very little late loss and a loss index of essentially zero (Fig. 6). Conversely, less treatment effect was observed in vessels larger than 3 mm. The 2-factor analysis of variance here showed a strong trend towards a difference in treatment effect for loss index with a p value of 0.07. After observing the radiotherapy efficacy could be dependent on vessel diameter, we evaluated the interaction be-
1100
I. J. Radiation Oncology
●
Biology
●
Physics
Fig. 2. The effect of in-stent restenosis on treatment with 192Ir. There is very little suppression of late loss in patients without previous stents, but a profound reduction in late loss and loss index in patients with previous stents. The 2-factor ANOVA is significant, indicating a statistically better treatment effect in patients with in-stent restenosis.
tween vessel diameter and radiation dosimetry. Our protocol used intravascular ultrasound measurements of source– target distance to deliver 8.00 Gy to the adventitial border
Fig. 3. The effect of multiple previous PTCA procedures on treatment with 192Ir. Patients with multiple previous PTCA procedures of the target lesion had a slightly better treatment effect than patients with only one previous PTCA. However, the differences were small and the 2-factor ANOVA was not significant.
Volume 42, Number 5, 1998
Fig. 4. The effect of lesion lengthy on treatment with 192Ir. Longer lesions were associated with a slightly more pronounced treatment effect, but the differences between longer and shorter lesions were minor and the 2-factor ANOVA was not significant.
farthest from the source, provided no more than 30.00 Gy was delivered to the adventitial border closest to the source. Our radiation source was not centered in the artery, and usually lay against one side of the vessel. In large diameter vessels, if an eccentrically placed source delivered 30.00 Gy to the wall closest to the source, the wall furthest from the source might be exposed to less than 8.00 Gy. Therefore, in many large-diameter vessels, our protocol mandated that the far wall be exposed to significantly less than the 8.00 Gy believed required to inhibit cellular proliferation. The mean source-to–target distances and radiation doses in vessels greater than and less than 3.0 mm are shown in Fig. 7. The mean longest source-to–target distance in largediameter vessels was 3.52 mm, compared to only 3.11 mm in small-diameter vessels. This corresponded to a mean minimum dose of only 6.94 Gy in large-diameter vessels compared to a mean minimum dose of 7.80 Gy in smalldiameter vessels (p , 0.05). Thus, small-diameter vessels were more likely to receive the desired minimum dose of 8.00 Gy. It is noteworthy to compare treatment effect in lesions exposed to a minimum dose of 8.00 Gy vs. lesions exposed to a minimum dose of , 8.00 Gy (Fig. 8). When the entire circumference of the adventitial border was exposed to at least 8.00 Gy, there was a profound treatment effect with very little late loss and a very low loss index. However, if a portion of the vessel wall was exposed to , 8.00 Gy, radiotherapy was, on average, not effective. Here, the 2-factor analysis of variance shows a significant difference in
Scripps coronary radiation trial
Fig. 5. The effect of saphenous vein graft (SVG) lesion location on treatment with 192Ir. Patients with lesions located in a saphenous vein graft appeared to have less of a treatment effect than patients with lesions in native coronary arteries, but the 2-factor ANOVA was not significant.
treatment effect for late loss between the two subgroups and a strong trend for loss index. We tested the possibility that all of the differences in treatment effect noted in the various subgroups were simply due to differences in vessel diameter and, therefore, dosimetry, by comparing vessel diameters and minimum dose exposures in each of the subgroups (Table 2). As expected, mean vessel diameters were larger and corresponding minimum doses lower in patients with vessels . 3 mm and saphenous vein grafts. Also, patients with only one previous PTCA had somewhat larger diameters vessels than patients with multiple previous PTCAs. Interestingly, however, the mean vessel diameter was nearly identical in diabetic vs. nondiabetic patients, as well as patients with in-stent restenosis vs. no previous stents, and patients with long and short lesion lengths. The corresponding mean minimum doses for these subgroups were also similar. Therefore, the different treatment effects described earlier for these subgroups were probably not related to differences in vessel diameter or dosimetry.
●
P. S. TEIRSTEIN et al.
1101
Fig. 6. The effect of vessel diameter on treatment with 192Ir. Lesions located within vessels , 3.0 mm in diameter demonstrated a profound treatment effect, with very little late loss and a late loss index of essentially zero. Conversely, less treatment effect was observed in vessels larger than 3 mm. The 2-factor ANOVA demonstrated a strong trend toward a difference in treatment effect for loss index (p 5 0.07).
liferative entities, such as keloid formation and heterotopic ossification, have been found to require similar minimum dose thresholds to achieve acceptable rates of treatment success (23–25). Our results are also consistent with many
DISCUSSION Two important concepts emerge from this subgroups analysis. First, there appears to be a relationship between efficacy and minimum dose exposure. We found that, on average, an adequate treatment effect required that a minimum dose of at least 8.00 Gy be delivered to the entire circumference of the adventitial border. Other benign pro-
Fig. 7. Differences in source to target distances and radiation doses in vessels greater than and less than 3.0 mm. The mean longest source to target distance in vessels $ 3.0 mm was 3.52 mm, compared to only 3.11 mm in vessels , 3.0 mm. This corresponds to a mean minimum dose of only 6.94 Gy in the larger diameter vessels, compared to a mean minimum dose of 7.80 Gy in the smaller diameter vessels.
1102
I. J. Radiation Oncology
●
Biology
●
Physics
Fig. 8. The effect of minimum dose exposure on treatment with 192 Ir. When the entire circumference of the adventitial border was exposed to 8.00 Gy, there was a profound treatment effect with very little late loss and a very low loss index. However, if a portion of the vessel wall was exposed to , 8.00 Gy, treatment with 192Ir was, on average, not effective. The 2-factor ANOVA demonstrates a significant difference in treatment effect for late loss between the two subgroups (p 5 0.009) and a strong trend for loss index (p 5 0.081).
of the early, preclinical studies that demonstrated efficacy at similar dose thresholds in animal models of restenosis (1– 8, 10, 11). One of the principles of radiation biology is that efficacy is proportional to minimum dose exposure and toxicity is proportional to maximum dose exposure. Our dosimetry algorithm was strongly influenced by concerns regarding the potential toxicity of higher exposures (. 30.00 Gy)
Table 2. Mean vessel diameter and mean minimum various patient subgroups
Diameter , 3 mm Diameter $ 3 mm Native SVG Previous PTCA 5 1 Previous PTCA $ 2 Diabetic Nondiabetic No previous stent In-stent restenosis Length , 15 mm Length 15–30 mm * p , 0.05.
192
Ir dose in
Mean vessel diameter (mm)
Mean minimum dose (Gy)
2.6* 3.5 2.7* 3.5 3.1 2.7 2.9 2.9 3.0 2.9 2.9 3.0
780* 694 776* 678 735 773 728 761 734 757 752 750
Volume 42, Number 5, 1998
at the vessel wall closest to the radiation source. Therefore, in larger diameter vessels, our protocol required a reduction in the minimum dose delivered to the vessel wall farthest from the source to ensure , 30.00 Gy was delivered to the vessel wall closest to the source. Additional studies are needed to determine if a noncentered radiation delivery system can safely provide 8.00 Gy to the far wall of largediameter vessels, even if this requires delivering more than 30.00 Gy to the near wall. Alternatively, a source-centering system may be capable of reducing dose heterogeneity and, thus, increase efficacy without increasing toxicity. Another important concept to emerge from this subgroups analysis is that certain specific patient and lesion characteristics appear to confer increased responsiveness to radiotherapy. 192Ir was particularly effective in inhibiting proliferation in restenotic lesions of patients with in-stent restenosis, small-diameter vessels, and diabetes mellitus. It is noteworthy that each of these patient subgroups is traditionally associated with high restenosis rates (26 –33). For example, previous studies have documented higher loss index rates in diabetic patients (34) and, in the present study, diabetic patients in the placebo arm had a striking loss index of 0.8, among the highest ever reported and consistent with our diabetic patient population, which was strongly enriched with other risk factors for restenosis. These higher restenosis rates likely result from an increase in cellular proliferation in response to the vessel injury that occurs during angioplasty procedures. Thus, cell division within the target lesion is probably increased in these patient subgroups. These data are interesting when one considers that the mechanism of radiation therapy’s effect is believed to be a double-stranded break in the cell’s DNA, inhibiting the G2 and M phases of cell division (35). It is well known that different tissues differ in the response to radiotherapy. This is described by the law of Bergonie and Triboundeau, which holds that the radiosensitivity of a tissue is directly proportional to its mitotic activity and inversely proportional to its degree of differentiation (36, 37). Thus, we may be fortunate to find that radiotherapy proves most effective for our most clinically challenging patients, especially those with highly proliferative restenotic processes. The results of this subgroup analysis may have an impact on clinical trial design. Our results may encourage future clinical trials directed at specific patient subgroups at increased risk for restenosis; in particular, de novo lesions in patients with diabetes and small-diameter vessels. Our results may also have implications for radiotherapy deliver device design. The finding of a reduced treatment effect in larger diameter vessels, especially when a portion of the adventitial border is exposed to , 8.00 Gy, suggests a minimum threshold of acceptable dose prescription and exposes a potential limitation of the radiation delivery system used in this trial. If a maximum exposure limit of 30.00 Gy to any part of the vessel wall is proven inviolable, the some form of centering (or at least partially centering)
Scripps coronary radiation trial
system will likely be required for successful treatment of larger diameter vessels, even when using a gamma emitter. Alternatively, the maximum limit of 30.00 Gy used in this study might be reassessed. Perhaps a higher dose can be tolerated by the vascular wall closest to the source, allowing higher dose exposures at the target farthest from the sources. Delivering higher doses of gamma radiation to the target should, however, only be undertaken with caution. A limit of 30.00 Gy has been proven safe (with 12 months clinical follow-up) in this trial; however, in another trial, Condado et al. documented the formation of two aneurysms (one a pseudoaneurysm) when a noncentered 192Ir source was used to deliver up to 92.00 Gy to the wall closest to the radiation source (13). Although 92.00 Gy is significantly higher than the highest dose used in this study, it underscores the point that radiotherapy can cause dose-related adverse effects and,
●
P. S. TEIRSTEIN et al.
1103
therefore, should only be used at the lowest possible effective dose (37). Study limitations There are several important limitations to this subgroups analysis. Many of the subgroups evaluated were not prospectively identified. The number of patients entered into this trial was small and, therefore, the subgroup sizes were even smaller. This resulted in very wide confidence intervals. It is critical, therefore, to view our results and the hypotheses generated as preliminary concepts requiring further testing in larger clinical trials. Nevertheless, the SCRIPPS Trial was a pilot trial and, as such, this analysis can help identify subgroups worthy of future study and provide technical insights that might aid in device development and improve the success of future trials.
REFERENCES 1. Wiedemann J, Leavy J, Amols H, Homma S, Di Tullio M, Sherman D, Apfelbaum M, Schwartz A, Weinberger J. Intracoronary irradiation acutely impairs endothelial and smooth muscle function as assessed by intravascular ultrasound. (Abstr.) Circulation 1992;86(Suppl. I):I–188. 2. Shefer A, Eigler NL, Whiting JS, Litvack FI. Suppression of intimal proliferation after balloon angioplasty with local beta irradiation in rabbits. (Abstr.) JACC 1993;21(Suppl. A):185A. 3. Wiedermann JG, Marboe C, Amols H, Schartz A, Weinberger J. Intracoronary irradiation markedly reduces restenosis after balloon angioplasty in a porcine model. J Am Coll Cardiol 1994;23:1491–1498. 4. Waksman R, Robinson KA, Crocker IR, Smith R, Cipolla GD, King SB III. Effects of endovascular irradiation in a swine model of restenosis after angioplasty. (Abstr.) J Am Coll Cardiol 1994;February:1A– 484A. 5. Mazur W, Ali MN, Dabaghi SF, Cristead C, Abukhalil J, Paraside P, DeFelice CA, Schulz D, Berner BM, Fajardo LF, French BA, Raizner AE. High-dose rate intracoronary radiation suppresses neointimal proliferation in the stented and ballooned model of porcine restenosis. (Abstr.) Circulation 1994;90:I-652. 6. Wiedermann JG, Marboe C, Amols H, Schwartz A, Weinberger J. Intracoronary irradiation markedly reduces neointimal proliferation after balloon angioplasty in swine: Persistent benefit at 6-month follow-up. J Am Coll Cardiol 1995;25: 1451–1456. 7. Waksman R, Robinson KA, Crocker IR, Gravanis MB, Cipolla GD, King SB III. 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. 8. Waksman R, Robinson KA, Crocker IR, Gravanis MB, Palmer SJ, Wang C, Cipolla GD, King SB III. Intracoronary radiation before stent implantation inhibits neointima formation in stented porcine coronary arteries. Circulation 1995;92:1383– 1386. 9. Hehrlein C, Gollan C, Do¨nges K, Metz J, Riessen R, Fehsenfeld P, von Hodenberg E, Ku¨bler W. Low-dose radioactive endovascular stents prevent smooth muscle cell proliferation and neointimal hyperplasia in rabbits. Circulation 1995;92: 1570 –1575. 10. Verin V, Popowski Y, Urban P, Belenger J, Redard M, Costa M, Widmer M-C, Rouzaud M, Nouet P, Grob E, Schwager M, Kurtz JM, Rutishauser W. Intra-arterial beta irradiation pre-
11.
12.
13.
14.
15.
16.
17. 18.
19.
20.
vents neointimal hyperplasia in a hypercholesterolemic rabbit restenosis model. Circulation 1995;92:2284 –2290. Waksman R, Robinson KA, Crocker IR, Wang C, Gravanis MB, Cipolla GD, Hillstead RA, King SB III. Intracoronary low-dose b-irradiation inhibits neointima formation after coronary artery balloon injury in the swine restenosis model. Circulation 1995;92:3025–3031. Laird JR, Carter AJ, Kufs WM, Hoopes TG, Farb A, Nott SH, Fischell RE, Fischell DR, Virmani R, Fischell TA. Inhibition of neointimal proliferation with low-dose irradiation from a b-particle-emitting stent. Circulation 1996;93:529 –536. Condado JA, Waksman R, Gurdiel O, Espinosa R, Gonzalez J, Burger B, Villoria G, Acquatella H, Crocker IR, Seung KB, Liprie SF. Long-term angiographic and clinical outcome after percutaneous transluminal coronary angioplasty and intracoronary radiation therapy in humans. Circulation 1997;96: 727–732. Teirstein PS, Massullo V, Jani S, Popma JJ, Mintz GS, Russo RJ, Schatz RA, Guarneri EM, Steuterman S, Morris NB, Leon MB, Tripuraneni P. Catheter-based radiotherapy to inhibit restenosis after coronary stenting. N Engl J Med 1997;336: 1697–1703. Bottcher HD, Schopohl B, Liermann D, Kollath J, Adamietz IA. Endovascular irradiation—A new method to avoid recurrent stenosis after stent implantation in peripheral arteries: Technique and preliminary results. Int J Radiat Oncol Biol Phys 1994;29:183–186. Liermann DD, Boettcher HD, Kollatch J, Schopol B, Strassman G, Strecker EP, Breddin KH. Prophylactic endovascular radiotherapy to prevent intimal hyperplasia after stent implantation in femoro-popliteal arteries. Cardiovasc Intervent Radiol 1994;17:12–16. Steidle B. Preventive percutaneous radiotherapy for avoiding hyperplasia of the intima following angioplasty together with stent implantation. Strahlenther Onkol 1994;170:151–154. Baim DS, Fischell T, Weissman NJ, Laird JR, Marble SJ, Ho KK. Short-term (1 month) results of the IRIS feasibility study of a beta-particle emitting radioisotope stent. (Abstr.) Circulation 1997;96(Suppl):1206. Bonan R, Arsenault A, Tardif J-C, Tanguay J-FL, Crepeau J, Gallo R, Gosselin G, Joyal M, Lucier G, Crocker IR. Beta energy restenosis trial, Canadian arm. (Abstr.) Circulation 1997;96(Suppl):1208. King SB III, Williams DO, Chougule P, Klein L, Waksman R, Robinson KA, Anderberg K, Scruby J, Crocker IR. Intracoro-
1104
21.
22.
23.
24. 25. 26.
27.
28.
I. J. Radiation Oncology
●
Biology
●
Physics
nary beta radiation inhibits late lumen loss following balloon angioplasty: Results of the BERT-1 trial. (Abstr.) Circulation 1997;96(Suppl):1211. Giedd KN, Amols H, Marboe CC, Knapp FF, Weinberger J. Effectiveness of a beta-emitting liquid-filled perfusion balloon to prevent restenosis. (Abstr.) Circulation 1997;96(Suppl): 1215. Van der Zwet PMJ, Reiber JHC. A new approach for the quantification of complex lesion morphology: The gradient field transform; basic principles and validation results. J Am Coll Cardiol 1994;24:216 –224. MacLennan I, Keys HM, Evarts CM, Rubin P. Usefulness of post-operative hip irradiation in the prevention of heterotopic bone formation in a high risk group of patients. Int J Radiat Oncol Biol Phys 1984;10:49 –53. Ollstein RN, Siegel HW, Gillooley JF, Barsa JM. Treatment of keloids by combined surgical excision and immediate postoperative x-ray therapy. Ann Plast Surg 1981;7:281–285. Malaker A, Ellis F, Paine CH. Keloid scars: A new method of treatment combining surgery with interstitial radiotherapy. Clin Radiol 1976;27:179 –183. Kastrati A, Scho¨mig A, Elezi S, Schu¨hlen H, Dirschinger J, Hadamitzky M, Wehinger A, Hausleiter J, Walter H, Neumann F-J. Predictive factors of restenosis after coronary stent placement. J Am Coll Cardiol 1997;30:1428 –1436. Foley DP, Melkert R, Serruys PW. Influence of coronary vessel size on renarrowing process and late angiographic outcome after successful balloon angioplasty. Circulation 1994;90:1239 –1251. de Groote P, Bauters C, McFadden EP, Lablanche JP, Leroy F, Bertrand MB. Local lesion-related factors and restenosis after coronary angioplasty. Circulation 1995;91:968 –972.
Volume 42, Number 5, 1998
29. Stein B, Weintraub WS, Gebhart SSP, et al. Influence of diabetes mellitus on early and late outcome after percutaneous transluminal coronary angioplasty. Circulation 1995;91:979 – 989. 30. Baim DS, Levine MJ, Leon MB, Levine S, Ellis SG, Schatz RA. Management of restenosis within the Palmaz–Schatz coronary stent (the U.S. multicenter experience). Am J Cardiol 1993;71:364 –366. 31. Kuntz RE, Safian RD, Carrozza JP, Fishman RF, Mansour M, Baim DS. The importance of acute luminal diameter in determining restenosis after coronary atherectomy or stenting. Circulation 1992;86:1827–1835. 32. Reimers B, Mcussa I, Akiyama T, Tucci G, Ferraro M, Martini G, Blengino S, Di Mario C, Colombo A. Long-term clinical follow-up after successful repeat percutaneous intervention for stent restenosis. J Am Coll Cardiol 1997;30:186 – 192. 33. Yokoi H, Kimura T, Nakagawa Y, Nosaka H, Nobuyoshi M. Long-term clinical and quantitative angiographic follow-up after the Palmaz-Schatz stent restenosis. (Abstr.) J Am Coll Cardiol 1996;27(Suppl. A):224A. 34. Carrozza JPJ, Kuntz RE, Fishman RF, Baim DS. Restenosis after arterial injury caused by coronary stenting in patients with diabetes mellitus. Ann Intern Med 1993;118:344 –349. 35. Hall EJ. Radiobiology for the radiologist. 4th ed. Philadelphia, PA: J.B. Lippincott Co., 1997. p. 192. 36. Dalrymple GV, Gaulden ME, Kollmorgan GM, Vogel HH. Medical radiation biology. Philadelphia, London, Toronto: W.B. Saunders Co., 1997. 37. Watson RM. Radiation exposure: Clueless in the cath lab, or sayonara ALARA. Cathet Cardiovasc Diagn 1997;42:126 – 127.
PII S0360-3016(98)00281-8
●
Clinical Investigation A SUBGROUP ANALYSIS OF THE SCRIPPS CORONARY RADIATION TO INHIBIT PROLIFERATION POSTSTENTING TRIAL PAUL S. TEIRSTEIN, M.D.,* VINCENT MASSULLO, M.D.,* SHIRISH JANI, PH.D.,* JEFFREY J. POPMA, M.D.,† GARY S. MINTZ, M.D.,† ROBERT J. RUSSO, M.D., PH.D.,* RICHARD A. SCHATZ, M.D.,* ERMINIA M. GUARNERI, M.D.,* STEPHEN STEUTERMAN, M.S.,* DAVID A. CLOUTIER, B.S.,* MARTIN B. LEON, M.D.,† AND PRABHAKAR TRIPURANENI, M.D.* *Divisions of Cardiovascular Diseases and Radiation Oncology, Scripps Clinic, La Jolla, CA, and †Division of Cardiology, Washington Hospital Center, Washington, DC Introduction: In the Scripps Coronary Radiation to Inhibit Proliferation Poststenting (SCRIPPS) Trial, 192Ir significantly reduced angiographic, ultrasonographic, and clinical endpoints of restenosis. The objective of this analysis was to quantitate the impact of patient, lesion and technical characteristics on late angiographic outcome. Methods: Patients with restenotic, stented coronary lesions were randomized to receive either 192Ir or placebo sources. Late luminal loss and loss index were calculated for several patient subgroups, including patients with diabetes, in-stent restenosis, multiple previous percutaneous transluminal coronary angioplasty (PTCA) procedures, longer lesion lengths, saphenous vein grafts, small vessel diameters, and minimum dose exposures < 8.00 Gy. Two-factor analysis of variance was used to test for an interaction between patient characteristics and treatment effect. Results: In the treated group, late loss was particularly low in patients with diabetes (0.19 mm), in-stent restenosis (0.17 mm), reference vessel diameters < 3.0 mm (0.07 mm), and patients who received a minimum radiation dose to the entire adventitial border of at least 8.00 Gy. The loss index in each of these subgroups was similarly low at 20.02, 0.03, 20.02, and 0.03, respectively. By 2-factor analysis of variance, a significant interaction between subgroup characteristic and treatment effect (late loss) was found in patients with in-stent restenosis (p 5 0.035), and patients receiving a minimum dose of 8.00 Gy to the adventitial border (p 5 0.009). Conclusion: In this pilot study, patient characteristics associated with a more aggressive proliferative response to injury appeared to confer an enhanced response to radiotherapy. Furthermore, a dose threshold response to 192 Ir was found with an enhanced response occurring when the entire circumference of the adventitial border was exposed to at least 8.00 Gy. © 1998 Elsevier Science Inc.
INTRODUCTION
patient, lesion, and technical characteristics on late angiographic outcome, and use this information to generate hypotheses that can be tested in future clinical trials.
Numerous animal trials (1–12) and several early-phase clinical trials (13–21) indicate a promising role for ionizing radiation in the treatment of coronary restenosis. Despite this limited early success, numerous basic questions have been raised. Most importantly, effective radiation delivery techniques and appropriate patient selection criteria must be defined. The Scripps Coronary Radiation to Inhibit Proliferation Poststenting (SCRIPPS) Trial was a pilot, doubleblind randomized study comparing 192Ir to placebo in restenotic, stented coronary lesions. The results of this trial demonstrated significant benefit in clinical, angiographic, and ultrasonographic endpoints of restenosis (14). Although the trial size was small, a subgroup analysis of angiographic endpoints may provide clues that can impact clinical trial design and aid in radiation device development. The objective of the present analysis was to quantitate the impact of
The SCRIPPS Trial was a double-blind randomized trial comparing 192Iridium with placebo sources. The clinical trial was approved by the institution’s Human Subjects and Radiation Safety Committees. Patient inclusion criteria required a target lesion in a restenotic coronary artery that either already contained a stent or was a candidate for stent placement. Previous target lesion intervention was required to occur . 4 weeks prior to study entry. The reference vessel was required to be between 3 and 5 mm in diameter and the target lesion # 30 mm in length. Patients were excluded if the coronary revascularization procedure was
Reprint requests to: Paul S. Teirstein, M.D., Scripps Clinic, Division of Cardiovascular Diseases, SW206, 10666 N. Torrey Pines Road, La Jolla, CA 92037.
Acknowledgments—The authors acknowledge the collaboration of Krishnan Suthanthiran of Best Industries. Accepted for publication 6 July 1998.
METHODS
1097
1098
I. J. Radiation Oncology
●
Biology
●
Physics
unsuccessful, a suboptimal result was achieved, a stent was implanted as an unplanned, emergency procedure, or the target lesion undergoing stent placement contained angiographic evidence of thrombus. Procedure Preprocedural medications included aspirin 325 mg, intracoronary nitroglycerin, and intravenous heparin to maintain an activated clotting time of . 300 s. If the lesion was not previously stented, single or, if required, tandem coronary stenting (Johnson & Johnson Interventional Systems, Warren, NJ) was performed. If stents had been previously placed, redilatation was undertaken and, in many such patients, additional stents were placed within the original stent as required to optimize the angiographic result. Intravascular ultrasound was then performed using a 3.2-Fr catheter (CVIS, Inc., Sunnyvale, CA) with a motorized pull-back device at 0.5 mm/s to interrogate the stented vessel segment. The intravascular ultrasound examination ensured that an optimal final result had been achieved and allowed assessment of lesion geometry for dosimetry calculation by the radiation therapist (see below). A 4-Fr infusion catheter (USCI, Inc., Bellireca, MA) was then inserted to span the stented region. Patients were randomized to receive a 0.03-inch ribbon (Best Industries, Springfield, VA) containing either 192Iridium (192Ir) sealed sources at its tip or a ribbon containing placebo, inactive sources. In this protocol, one of two devices were selected dependent on the lesion length. Discrete lesions were treated with a 19-mm ribbon containing 5 3-mm long sources, each separated by 1 mm. Longer lesions were treated with a similar ribbon containing 9 sources totaling 35 mm in length. All personnel, except one physics member from the Division of Radiation Oncology and one research nurse from the Division of Cardiology who were not involved in endpoint analysis, were blinded to the randomization code. The catheterization laboratory was then cleared of all personnel, except the radiation oncologist who inserted the study ribbon into the infusion catheter. Radiation sources were left in place for 20 – 45 min as required to administer the prescribed radiation dose. Sources were then removed by the radiation oncologist and placed into an adequately shielded container. Femoral sheaths were removed 2– 4 h after the procedure and the patient discharged the following morning taking aspirin 325 mg daily indefinitely and, if new stents were implanted, ticlopidine 500 mg daily for 2 weeks. Dosimetry 192 Iridium dosimetry was calculated in the following manner. A series of tomographic sections, obtained by intravascular ultrasound using a motorized pullback apparatus, were scanned and measurements were performed along the length of the stent. The distance between the center of the ultrasound catheter (equivalent to the “source” position) and the adventitial border (the “target”) was measured every 1 mm along the stented segment of the artery.
Volume 42, Number 5, 1998
A maximal, as well as minimal, source-to–target distance was determined. The radiation oncologist and physicist combined this information with the specific activity of the radioactive sources to determine a dwell time that provided 8.00 Gy to the target farthest from the radiation source, provided that no more than 30.00 Gy was delivered to the target closest to the radiation source.
Quantitative angiographic analysis After removal of radiation sources, intracoronary nitroglycerin was administered and final angiography performed in two orthogonal projections. At follow-up examination, identical angiographic techniques were used. All procedural and follow-up cineangiograms were forwarded to the Washington Hospital Center Angiographic Core Laboratory for analysis by observers who were blinded to the treatment allocation. Selected serial cineframes, obtained from two unforeshortened projections and matched for position within the cardiac cycle using side-by–side projectors, were digitized with a cinevideo converter using the contrast-filled catheter as the calibration standard. Reference vessel diameter, as well as minimal luminal diameter within the axial length of the stent, and minimal luminal diameter along the axial length of the radiation sources were determined using a validated edge detection program (CMS, MEDIS) (22) at baseline, after the procedure, and at follow-up. These measurements were used to determine serial relative % diameter stenoses within the stent alone and at the stent border (beyond the stent but still covered by the radiation sources). Acute luminal gain was defined as the final minimal luminal diameter minus the initial minimum luminal diameter (mm), and the late luminal loss was defined as the final minimum luminal diameter minus the follow-up minimum luminal diameter. The loss index was defined as the late luminal loss divided by acute luminal gain. Binary restenosis was defined as a $ 50% diameter stenosis at follow-up.
Subgroup analysis and statistics Six subgroups were prospectively specified by randomizing each patient from 1 of 8 randomization tables stratified by lesions with 1 stent spanned by 5 sources or 2 tandem stents spanned by 9 sources, in-stent restenosis or no previous stent, and vein graft or native lesion location. Other subgroups were retrospectively identified. These included diabetes, number of previous percutaneous transluminal coronary angioplasty (PTCA) procedures, reference vessel diameter, and minimum dose delivered to the advential border. For the analysis of continuous data, Mann–Whitney U tests were used to assess differences between the two treatment groups. The results are expressed as mean 6 SD with concomitant test probabilities. Discrete data were compared by chi-square or, where appropriate, Fisher exact probability tests. A 2-factor analysis of variance (ANOVA) was conducted to test for an interaction between various patient and lesion characteristics and treatment effect.
Scripps coronary radiation trial
●
P. S. TEIRSTEIN et al.
1099
Table 1. Baseline clinical and angiographic characteristics
Diabetes (%) In-stent restenosis (%) Vessel diameter , 3.0 mm (%) Saphenous vein target (%) Lesion length . 15 mm (%)
192
Placebo (n 5 28)
Ir (n 5 24)
p
39.3 64.3 71.4 28.6 53.6
29.2 75 66.6 25 66.7
0.637 0.593 0.944 0.980 0.499
RESULTS Between March and December of 1995, 55 patients were randomized and 52 patients had angiograms eligible for analysis; 28 were assigned to placebo and 24 to active radiation. Three patients were not included in this analysis because follow-up angiograms were not obtained (1 cardiac death in a placebo patients), early stent thrombosis occurred (1 192Ir patient) or technically could not be analyzed (1 patient). Baseline clinical and angiographic characteristics were similar for both groups (Table 1). Many study patients had one or more risk factors for restenosis, such as diabetes, in-stent restenosis, smaller diameter vessels, saphenous vein graft target, and longer lesion length. As reported previously, angiographic indices of restenosis were markedly different between placebo and treated patients (14). For the entire group, treatment with 192Ir suppressed late loss (1.03 vs. 0.38 mm, p 5 0.025), loss index (0.6 vs. 0.12, p 5 0.002) and the dichotomous ($ 50% diameter stenosis) angiographic restenosis rate (53.6% vs. 16.7%, p 5 0.01). The effect of diabetes on angiographic outcome is depicted in Fig. 1. Diabetic patients treated with radiation had very little late loss and a loss index of essentially zero, compared to a somewhat less pronounced effect in nondiabetic patients. To the right of each graph, however, is the result of a 2-factor analysis of variance, which tests for an interaction between treatment effect and the patient characteristic of diabetes. Statistically, there was only a slight trend towards a difference in treatment effect in diabetic vs. nondiabetic patients; with a p value of 0.113 for loss index. The difference between diabetic and nondiabetic patients is highlighted, however, by the very high loss index of 0.80 found in placebo diabetic patients compared to 0.50 in placebo nondiabetic patients. Therefore, despite a more aggressive restenotic process in diabetics, radiotherapy was extremely effective in suppressing late luminal loss. Patients with in-stent restenosis (Fig. 2) also demonstrated an enhanced response to radiation therapy compared to that in patients not previously stented. There was very little suppression of late loss in patients without previous stents, but a profound reduction in late loss and loss index in patients with previous stents. Here, the 2-factor analysis of variance for late loss is significant, indicating a statistically better treatment effect in patients with in-stent restenosis. Patients with multiple previous PTCA procedures of the
Fig. 1. The effect of diabetes on treatment with 192Ir. For this and subsequent figures, late loss is depicted for both placebo and 192Ir patients on the top of the figure and loss index is depicted at the bottom of the figure. Data pertaining to patients with diabetes is displayed on the right side of the figure and that from patients without diabetes is displayed on the left. To the right of each figure is the p value from the 2-factor analysis of variance (ANOVA), which tests for an interaction between the patient characteristic of diabetes and treatment effect. Diabetic patients treated with radiation had very little late loss and a late loss index of essentially zero, compared to a less pronounced effect in nondiabetic patients. The 2-factor ANOVA, however, is not significant, demonstrating only a slight trend toward a difference in treatment effect in diabetic vs. nondiabetic patients (p 5 0.113).
target lesion (Fig. 3) appeared to have a slightly better treatment effect than patients with only one previous PTCA. However, the differences here were small and not supported by the 2-factor analysis of variance. Similarly, longer lesions (Fig. 4) were associated with a slightly more pronounced treatment effect, but the differences between longer and shorter lesions were minor and not significant. Also, patients with lesions located in a saphenous vein graft (Fig. 5) appeared to have less of a treatment effect than patients with native coronary targets, but the 2-factor analysis here was also not significant. Perhaps most interestingly, vessel diameter appeared to be related to treatment effect. Lesions located within vessels , 3.0 mm in diameter (as measured by off-site quantitative angiography) demonstrated a profound treatment effect, with very little late loss and a loss index of essentially zero (Fig. 6). Conversely, less treatment effect was observed in vessels larger than 3 mm. The 2-factor analysis of variance here showed a strong trend towards a difference in treatment effect for loss index with a p value of 0.07. After observing the radiotherapy efficacy could be dependent on vessel diameter, we evaluated the interaction be-
1100
I. J. Radiation Oncology
●
Biology
●
Physics
Fig. 2. The effect of in-stent restenosis on treatment with 192Ir. There is very little suppression of late loss in patients without previous stents, but a profound reduction in late loss and loss index in patients with previous stents. The 2-factor ANOVA is significant, indicating a statistically better treatment effect in patients with in-stent restenosis.
tween vessel diameter and radiation dosimetry. Our protocol used intravascular ultrasound measurements of source– target distance to deliver 8.00 Gy to the adventitial border
Fig. 3. The effect of multiple previous PTCA procedures on treatment with 192Ir. Patients with multiple previous PTCA procedures of the target lesion had a slightly better treatment effect than patients with only one previous PTCA. However, the differences were small and the 2-factor ANOVA was not significant.
Volume 42, Number 5, 1998
Fig. 4. The effect of lesion lengthy on treatment with 192Ir. Longer lesions were associated with a slightly more pronounced treatment effect, but the differences between longer and shorter lesions were minor and the 2-factor ANOVA was not significant.
farthest from the source, provided no more than 30.00 Gy was delivered to the adventitial border closest to the source. Our radiation source was not centered in the artery, and usually lay against one side of the vessel. In large diameter vessels, if an eccentrically placed source delivered 30.00 Gy to the wall closest to the source, the wall furthest from the source might be exposed to less than 8.00 Gy. Therefore, in many large-diameter vessels, our protocol mandated that the far wall be exposed to significantly less than the 8.00 Gy believed required to inhibit cellular proliferation. The mean source-to–target distances and radiation doses in vessels greater than and less than 3.0 mm are shown in Fig. 7. The mean longest source-to–target distance in largediameter vessels was 3.52 mm, compared to only 3.11 mm in small-diameter vessels. This corresponded to a mean minimum dose of only 6.94 Gy in large-diameter vessels compared to a mean minimum dose of 7.80 Gy in smalldiameter vessels (p , 0.05). Thus, small-diameter vessels were more likely to receive the desired minimum dose of 8.00 Gy. It is noteworthy to compare treatment effect in lesions exposed to a minimum dose of 8.00 Gy vs. lesions exposed to a minimum dose of , 8.00 Gy (Fig. 8). When the entire circumference of the adventitial border was exposed to at least 8.00 Gy, there was a profound treatment effect with very little late loss and a very low loss index. However, if a portion of the vessel wall was exposed to , 8.00 Gy, radiotherapy was, on average, not effective. Here, the 2-factor analysis of variance shows a significant difference in
Scripps coronary radiation trial
Fig. 5. The effect of saphenous vein graft (SVG) lesion location on treatment with 192Ir. Patients with lesions located in a saphenous vein graft appeared to have less of a treatment effect than patients with lesions in native coronary arteries, but the 2-factor ANOVA was not significant.
treatment effect for late loss between the two subgroups and a strong trend for loss index. We tested the possibility that all of the differences in treatment effect noted in the various subgroups were simply due to differences in vessel diameter and, therefore, dosimetry, by comparing vessel diameters and minimum dose exposures in each of the subgroups (Table 2). As expected, mean vessel diameters were larger and corresponding minimum doses lower in patients with vessels . 3 mm and saphenous vein grafts. Also, patients with only one previous PTCA had somewhat larger diameters vessels than patients with multiple previous PTCAs. Interestingly, however, the mean vessel diameter was nearly identical in diabetic vs. nondiabetic patients, as well as patients with in-stent restenosis vs. no previous stents, and patients with long and short lesion lengths. The corresponding mean minimum doses for these subgroups were also similar. Therefore, the different treatment effects described earlier for these subgroups were probably not related to differences in vessel diameter or dosimetry.
●
P. S. TEIRSTEIN et al.
1101
Fig. 6. The effect of vessel diameter on treatment with 192Ir. Lesions located within vessels , 3.0 mm in diameter demonstrated a profound treatment effect, with very little late loss and a late loss index of essentially zero. Conversely, less treatment effect was observed in vessels larger than 3 mm. The 2-factor ANOVA demonstrated a strong trend toward a difference in treatment effect for loss index (p 5 0.07).
liferative entities, such as keloid formation and heterotopic ossification, have been found to require similar minimum dose thresholds to achieve acceptable rates of treatment success (23–25). Our results are also consistent with many
DISCUSSION Two important concepts emerge from this subgroups analysis. First, there appears to be a relationship between efficacy and minimum dose exposure. We found that, on average, an adequate treatment effect required that a minimum dose of at least 8.00 Gy be delivered to the entire circumference of the adventitial border. Other benign pro-
Fig. 7. Differences in source to target distances and radiation doses in vessels greater than and less than 3.0 mm. The mean longest source to target distance in vessels $ 3.0 mm was 3.52 mm, compared to only 3.11 mm in vessels , 3.0 mm. This corresponds to a mean minimum dose of only 6.94 Gy in the larger diameter vessels, compared to a mean minimum dose of 7.80 Gy in the smaller diameter vessels.
1102
I. J. Radiation Oncology
●
Biology
●
Physics
Fig. 8. The effect of minimum dose exposure on treatment with 192 Ir. When the entire circumference of the adventitial border was exposed to 8.00 Gy, there was a profound treatment effect with very little late loss and a very low loss index. However, if a portion of the vessel wall was exposed to , 8.00 Gy, treatment with 192Ir was, on average, not effective. The 2-factor ANOVA demonstrates a significant difference in treatment effect for late loss between the two subgroups (p 5 0.009) and a strong trend for loss index (p 5 0.081).
of the early, preclinical studies that demonstrated efficacy at similar dose thresholds in animal models of restenosis (1– 8, 10, 11). One of the principles of radiation biology is that efficacy is proportional to minimum dose exposure and toxicity is proportional to maximum dose exposure. Our dosimetry algorithm was strongly influenced by concerns regarding the potential toxicity of higher exposures (. 30.00 Gy)
Table 2. Mean vessel diameter and mean minimum various patient subgroups
Diameter , 3 mm Diameter $ 3 mm Native SVG Previous PTCA 5 1 Previous PTCA $ 2 Diabetic Nondiabetic No previous stent In-stent restenosis Length , 15 mm Length 15–30 mm * p , 0.05.
192
Ir dose in
Mean vessel diameter (mm)
Mean minimum dose (Gy)
2.6* 3.5 2.7* 3.5 3.1 2.7 2.9 2.9 3.0 2.9 2.9 3.0
780* 694 776* 678 735 773 728 761 734 757 752 750
Volume 42, Number 5, 1998
at the vessel wall closest to the radiation source. Therefore, in larger diameter vessels, our protocol required a reduction in the minimum dose delivered to the vessel wall farthest from the source to ensure , 30.00 Gy was delivered to the vessel wall closest to the source. Additional studies are needed to determine if a noncentered radiation delivery system can safely provide 8.00 Gy to the far wall of largediameter vessels, even if this requires delivering more than 30.00 Gy to the near wall. Alternatively, a source-centering system may be capable of reducing dose heterogeneity and, thus, increase efficacy without increasing toxicity. Another important concept to emerge from this subgroups analysis is that certain specific patient and lesion characteristics appear to confer increased responsiveness to radiotherapy. 192Ir was particularly effective in inhibiting proliferation in restenotic lesions of patients with in-stent restenosis, small-diameter vessels, and diabetes mellitus. It is noteworthy that each of these patient subgroups is traditionally associated with high restenosis rates (26 –33). For example, previous studies have documented higher loss index rates in diabetic patients (34) and, in the present study, diabetic patients in the placebo arm had a striking loss index of 0.8, among the highest ever reported and consistent with our diabetic patient population, which was strongly enriched with other risk factors for restenosis. These higher restenosis rates likely result from an increase in cellular proliferation in response to the vessel injury that occurs during angioplasty procedures. Thus, cell division within the target lesion is probably increased in these patient subgroups. These data are interesting when one considers that the mechanism of radiation therapy’s effect is believed to be a double-stranded break in the cell’s DNA, inhibiting the G2 and M phases of cell division (35). It is well known that different tissues differ in the response to radiotherapy. This is described by the law of Bergonie and Triboundeau, which holds that the radiosensitivity of a tissue is directly proportional to its mitotic activity and inversely proportional to its degree of differentiation (36, 37). Thus, we may be fortunate to find that radiotherapy proves most effective for our most clinically challenging patients, especially those with highly proliferative restenotic processes. The results of this subgroup analysis may have an impact on clinical trial design. Our results may encourage future clinical trials directed at specific patient subgroups at increased risk for restenosis; in particular, de novo lesions in patients with diabetes and small-diameter vessels. Our results may also have implications for radiotherapy deliver device design. The finding of a reduced treatment effect in larger diameter vessels, especially when a portion of the adventitial border is exposed to , 8.00 Gy, suggests a minimum threshold of acceptable dose prescription and exposes a potential limitation of the radiation delivery system used in this trial. If a maximum exposure limit of 30.00 Gy to any part of the vessel wall is proven inviolable, the some form of centering (or at least partially centering)
Scripps coronary radiation trial
system will likely be required for successful treatment of larger diameter vessels, even when using a gamma emitter. Alternatively, the maximum limit of 30.00 Gy used in this study might be reassessed. Perhaps a higher dose can be tolerated by the vascular wall closest to the source, allowing higher dose exposures at the target farthest from the sources. Delivering higher doses of gamma radiation to the target should, however, only be undertaken with caution. A limit of 30.00 Gy has been proven safe (with 12 months clinical follow-up) in this trial; however, in another trial, Condado et al. documented the formation of two aneurysms (one a pseudoaneurysm) when a noncentered 192Ir source was used to deliver up to 92.00 Gy to the wall closest to the radiation source (13). Although 92.00 Gy is significantly higher than the highest dose used in this study, it underscores the point that radiotherapy can cause dose-related adverse effects and,
●
P. S. TEIRSTEIN et al.
1103
therefore, should only be used at the lowest possible effective dose (37). Study limitations There are several important limitations to this subgroups analysis. Many of the subgroups evaluated were not prospectively identified. The number of patients entered into this trial was small and, therefore, the subgroup sizes were even smaller. This resulted in very wide confidence intervals. It is critical, therefore, to view our results and the hypotheses generated as preliminary concepts requiring further testing in larger clinical trials. Nevertheless, the SCRIPPS Trial was a pilot trial and, as such, this analysis can help identify subgroups worthy of future study and provide technical insights that might aid in device development and improve the success of future trials.
REFERENCES 1. Wiedemann J, Leavy J, Amols H, Homma S, Di Tullio M, Sherman D, Apfelbaum M, Schwartz A, Weinberger J. Intracoronary irradiation acutely impairs endothelial and smooth muscle function as assessed by intravascular ultrasound. (Abstr.) Circulation 1992;86(Suppl. I):I–188. 2. Shefer A, Eigler NL, Whiting JS, Litvack FI. Suppression of intimal proliferation after balloon angioplasty with local beta irradiation in rabbits. (Abstr.) JACC 1993;21(Suppl. A):185A. 3. Wiedermann JG, Marboe C, Amols H, Schartz A, Weinberger J. Intracoronary irradiation markedly reduces restenosis after balloon angioplasty in a porcine model. J Am Coll Cardiol 1994;23:1491–1498. 4. Waksman R, Robinson KA, Crocker IR, Smith R, Cipolla GD, King SB III. Effects of endovascular irradiation in a swine model of restenosis after angioplasty. (Abstr.) J Am Coll Cardiol 1994;February:1A– 484A. 5. Mazur W, Ali MN, Dabaghi SF, Cristead C, Abukhalil J, Paraside P, DeFelice CA, Schulz D, Berner BM, Fajardo LF, French BA, Raizner AE. High-dose rate intracoronary radiation suppresses neointimal proliferation in the stented and ballooned model of porcine restenosis. (Abstr.) Circulation 1994;90:I-652. 6. Wiedermann JG, Marboe C, Amols H, Schwartz A, Weinberger J. Intracoronary irradiation markedly reduces neointimal proliferation after balloon angioplasty in swine: Persistent benefit at 6-month follow-up. J Am Coll Cardiol 1995;25: 1451–1456. 7. Waksman R, Robinson KA, Crocker IR, Gravanis MB, Cipolla GD, King SB III. 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. 8. Waksman R, Robinson KA, Crocker IR, Gravanis MB, Palmer SJ, Wang C, Cipolla GD, King SB III. Intracoronary radiation before stent implantation inhibits neointima formation in stented porcine coronary arteries. Circulation 1995;92:1383– 1386. 9. Hehrlein C, Gollan C, Do¨nges K, Metz J, Riessen R, Fehsenfeld P, von Hodenberg E, Ku¨bler W. Low-dose radioactive endovascular stents prevent smooth muscle cell proliferation and neointimal hyperplasia in rabbits. Circulation 1995;92: 1570 –1575. 10. Verin V, Popowski Y, Urban P, Belenger J, Redard M, Costa M, Widmer M-C, Rouzaud M, Nouet P, Grob E, Schwager M, Kurtz JM, Rutishauser W. Intra-arterial beta irradiation pre-
11.
12.
13.
14.
15.
16.
17. 18.
19.
20.
vents neointimal hyperplasia in a hypercholesterolemic rabbit restenosis model. Circulation 1995;92:2284 –2290. Waksman R, Robinson KA, Crocker IR, Wang C, Gravanis MB, Cipolla GD, Hillstead RA, King SB III. Intracoronary low-dose b-irradiation inhibits neointima formation after coronary artery balloon injury in the swine restenosis model. Circulation 1995;92:3025–3031. Laird JR, Carter AJ, Kufs WM, Hoopes TG, Farb A, Nott SH, Fischell RE, Fischell DR, Virmani R, Fischell TA. Inhibition of neointimal proliferation with low-dose irradiation from a b-particle-emitting stent. Circulation 1996;93:529 –536. Condado JA, Waksman R, Gurdiel O, Espinosa R, Gonzalez J, Burger B, Villoria G, Acquatella H, Crocker IR, Seung KB, Liprie SF. Long-term angiographic and clinical outcome after percutaneous transluminal coronary angioplasty and intracoronary radiation therapy in humans. Circulation 1997;96: 727–732. Teirstein PS, Massullo V, Jani S, Popma JJ, Mintz GS, Russo RJ, Schatz RA, Guarneri EM, Steuterman S, Morris NB, Leon MB, Tripuraneni P. Catheter-based radiotherapy to inhibit restenosis after coronary stenting. N Engl J Med 1997;336: 1697–1703. Bottcher HD, Schopohl B, Liermann D, Kollath J, Adamietz IA. Endovascular irradiation—A new method to avoid recurrent stenosis after stent implantation in peripheral arteries: Technique and preliminary results. Int J Radiat Oncol Biol Phys 1994;29:183–186. Liermann DD, Boettcher HD, Kollatch J, Schopol B, Strassman G, Strecker EP, Breddin KH. Prophylactic endovascular radiotherapy to prevent intimal hyperplasia after stent implantation in femoro-popliteal arteries. Cardiovasc Intervent Radiol 1994;17:12–16. Steidle B. Preventive percutaneous radiotherapy for avoiding hyperplasia of the intima following angioplasty together with stent implantation. Strahlenther Onkol 1994;170:151–154. Baim DS, Fischell T, Weissman NJ, Laird JR, Marble SJ, Ho KK. Short-term (1 month) results of the IRIS feasibility study of a beta-particle emitting radioisotope stent. (Abstr.) Circulation 1997;96(Suppl):1206. Bonan R, Arsenault A, Tardif J-C, Tanguay J-FL, Crepeau J, Gallo R, Gosselin G, Joyal M, Lucier G, Crocker IR. Beta energy restenosis trial, Canadian arm. (Abstr.) Circulation 1997;96(Suppl):1208. King SB III, Williams DO, Chougule P, Klein L, Waksman R, Robinson KA, Anderberg K, Scruby J, Crocker IR. Intracoro-
1104
21.
22.
23.
24. 25. 26.
27.
28.
I. J. Radiation Oncology
●
Biology
●
Physics
nary beta radiation inhibits late lumen loss following balloon angioplasty: Results of the BERT-1 trial. (Abstr.) Circulation 1997;96(Suppl):1211. Giedd KN, Amols H, Marboe CC, Knapp FF, Weinberger J. Effectiveness of a beta-emitting liquid-filled perfusion balloon to prevent restenosis. (Abstr.) Circulation 1997;96(Suppl): 1215. Van der Zwet PMJ, Reiber JHC. A new approach for the quantification of complex lesion morphology: The gradient field transform; basic principles and validation results. J Am Coll Cardiol 1994;24:216 –224. MacLennan I, Keys HM, Evarts CM, Rubin P. Usefulness of post-operative hip irradiation in the prevention of heterotopic bone formation in a high risk group of patients. Int J Radiat Oncol Biol Phys 1984;10:49 –53. Ollstein RN, Siegel HW, Gillooley JF, Barsa JM. Treatment of keloids by combined surgical excision and immediate postoperative x-ray therapy. Ann Plast Surg 1981;7:281–285. Malaker A, Ellis F, Paine CH. Keloid scars: A new method of treatment combining surgery with interstitial radiotherapy. Clin Radiol 1976;27:179 –183. Kastrati A, Scho¨mig A, Elezi S, Schu¨hlen H, Dirschinger J, Hadamitzky M, Wehinger A, Hausleiter J, Walter H, Neumann F-J. Predictive factors of restenosis after coronary stent placement. J Am Coll Cardiol 1997;30:1428 –1436. Foley DP, Melkert R, Serruys PW. Influence of coronary vessel size on renarrowing process and late angiographic outcome after successful balloon angioplasty. Circulation 1994;90:1239 –1251. de Groote P, Bauters C, McFadden EP, Lablanche JP, Leroy F, Bertrand MB. Local lesion-related factors and restenosis after coronary angioplasty. Circulation 1995;91:968 –972.
Volume 42, Number 5, 1998
29. Stein B, Weintraub WS, Gebhart SSP, et al. Influence of diabetes mellitus on early and late outcome after percutaneous transluminal coronary angioplasty. Circulation 1995;91:979 – 989. 30. Baim DS, Levine MJ, Leon MB, Levine S, Ellis SG, Schatz RA. Management of restenosis within the Palmaz–Schatz coronary stent (the U.S. multicenter experience). Am J Cardiol 1993;71:364 –366. 31. Kuntz RE, Safian RD, Carrozza JP, Fishman RF, Mansour M, Baim DS. The importance of acute luminal diameter in determining restenosis after coronary atherectomy or stenting. Circulation 1992;86:1827–1835. 32. Reimers B, Mcussa I, Akiyama T, Tucci G, Ferraro M, Martini G, Blengino S, Di Mario C, Colombo A. Long-term clinical follow-up after successful repeat percutaneous intervention for stent restenosis. J Am Coll Cardiol 1997;30:186 – 192. 33. Yokoi H, Kimura T, Nakagawa Y, Nosaka H, Nobuyoshi M. Long-term clinical and quantitative angiographic follow-up after the Palmaz-Schatz stent restenosis. (Abstr.) J Am Coll Cardiol 1996;27(Suppl. A):224A. 34. Carrozza JPJ, Kuntz RE, Fishman RF, Baim DS. Restenosis after arterial injury caused by coronary stenting in patients with diabetes mellitus. Ann Intern Med 1993;118:344 –349. 35. Hall EJ. Radiobiology for the radiologist. 4th ed. Philadelphia, PA: J.B. Lippincott Co., 1997. p. 192. 36. Dalrymple GV, Gaulden ME, Kollmorgan GM, Vogel HH. Medical radiation biology. Philadelphia, London, Toronto: W.B. Saunders Co., 1997. 37. Watson RM. Radiation exposure: Clueless in the cath lab, or sayonara ALARA. Cathet Cardiovasc Diagn 1997;42:126 – 127.