Comparison of brachytherapy strategies based on dose-volume histograms derived from quantitative intravascular ultrasound

Abstract

Purpose. We present in this paper the comparison, by simulation, of different treatment strategies based either on β- or γ-sources, both with and without a centering device. Ionizing radiation to prevent restenosis is an emerging modality in interventional cardiology. Numerous clinical studies are presently being performed or planned, but there is variability in dose prescription, and both γ- and β-emitters are used, leading to a wide range of possible dose distributions over the arterial vessel wall. This paper discusses the potential merits of dose-volume histograms (DVH) based on three-dimensional (3-D) reconstruction of electrocardiogram (ECG)-gated intravascular ultrasound (IVUS) to compare brachytherapy treatment strategies.

Materials and Methods. DVH describe the cumulative distribution of dose over three specific volumes: (1) at the level of the luminal surface, a volume was defined with a thickness of 0.1 mm from the automatically detected contour of the highly echogenic blood-vessel interface; (2) at the level of the IVUS echogenic media-adventitia interface (external elastic lamina [EEL]), an adventitial volume was computed considering a 0.5-mm thickness from EEL; and (3) the volume encompassed between the luminal surface and the EEL (plaque + media). The IVUS data used were recorded in 23 of 31 patients during the Beta Energy Restenosis Trial (BERT) conducted in our institution.

Results. On average, the minimal dose in 90% of the adventitial volume was 37 ± 16% of the prescribed dose; the minimal dose in 90% of the plaque + media volume was 58 ± 24% and of the luminal surface volume was 67 ± 31%. The minimal dose in the 10% most exposed luminal surface volume was 296 ± 42%. Simulations of the use of a γ-emitter and/or a radioactive source train centered in the lumen are reported, with a comparison of the homogeneity of the dose distribution.

Conclusions. It is possible to derive DVH from IVUS, to evaluate the dose delivered to different parts of the coronary wall. This process should improve our understanding of the mechanisms of action of brachytherapy.

Introduction

Coronary artery diseases remain the major cause of death and disabilities in industrialized countries. The only revascularization procedure available up to 1977 was bypass surgery. Percutaneous transluminal coronary angioplasty (PTCA) introduced by Andreas Grüntzig [1] profoundly modified our therapeutic arsenal with a minimally invasive alternative. Presently, interventional cardiology consists of several techniques to cut, drill, scrape, burn, and otherwise remove atherosclerotic plaque [2]. With more than one million interventions undertaken per year worldwide, angioplasty is now a cornerstone therapy for coronary artery diseases. However, despite a high acute procedural success rate, the long-term benefit is hindered by the phenomenon of restenosis. Mechanisms involved in the restenosis process are the elastic recoil of the artery, local thrombus formation, vascular remodeling with shrinkage of the vessel, and exuberant healing process with neointimal cellular proliferation and matrix synthesis 3, 4, 5. Stent implantation minimizes elastic recoil and remodeling of vessels, and carefully controlled and randomized clinical trials have demonstrated a significant decrease in the rate of restenosis 6, 7, 8. However, stents increase the proliferative response of tissue to the intervention and, depending on the type of lesions treated, a significant restenosis rate of 15–50% remains the key limitation of transcatheter procedures. Restenosis is the subject of numerous investigations to further improve applicability and cost-effectiveness of angioplasty and to reduce the need of reinterventions. Virtually all attempts to limit restenosis with systemic drugs have failed, with the recent exceptions of abciximab, probucol, and cilostazol 9, 10, 11.

Some investigators have considered restenosis as an accentuation of the wound healing process associated with the trauma of angioplasty, and because radiotherapy had proved effective for the treatment of keloid formation and other nonmalignant diseases, radiation therapy for intravascular application was attempted. The therapy was introduced by Friedman et al. [12] early in 1964, for the prevention of atherosclerosis, and subsequent animal experiments demonstrated a reduction of intimal hyperplasia following endovascular irradiation. Waksman [13] has recently reviewed these early studies. Three clinical studies have been reported that confirmed a significant reduction in the restenosis rate using additional brachytherapy 14, 15, 16.

Currently, the vascular brachytherapy devices available for clinical trials are radioactive stents and catheter-based systems using a radioactive wire advanced with an afterloader, or radioactive seeds delivered with a hydraulic delivery system. Other systems based on radioactive balloons are in development. There is variability in the dose prescription, and both γ- and β-emitters are used. These variations lead to a wide range of dose distributions over the arterial vessel wall requiring a careful interpretation and comparison of the results of the ongoing studies. The typical dose prescription distance in the coronary arteries is in the range of 2 mm from the source axis. Because of the steep dose fall-off, particularly for β-emitters, accurate dosimetry requires precise knowledge of geometry.

In this paper, we describe a dosimetry evaluation tool for coronary brachytherapy based on three-dimensional (3-D) reconstruction of electrocardiogram (ECG)-gated intravascular ultrasound (IVUS) images. IVUS was developed to overcome the limitations of x-ray angiography. Its methodology and clinical applications have been reviewed extensively [17]. IVUS, by its tomographic approach, provides a mean for the evaluation of both lumen and vessel wall morphology. Assuming that the catheter containing the radioactive source is lying in the same position as the IVUS catheter, it is possible to measure the distance from the source to any vascular structure in one cross-sectional image, and to construct isodose plots. IVUS recordings can be performed with a constant speed motorized pull-back device (e.g., 0.5 mm/s), which permits the evaluation of the length of a stenosis. Recently, 3-D image reconstruction and analysis systems have been introduced that can be used for complete quantitative analysis of IVUS images 18, 19, 20, 21. However, image artifacts that result from cyclic changes in coronary dimensions and from the movement of the IVUS catheter in the arterial lumen limit the accuracy of the 3-D boundary detection systems [22]. This problem led to the development of a new approach in our institution. To limit cyclic movement artifacts, we use an ECG-gated image acquisition workstation that controls a dedicated pull-back device. Feasibility, reproducibility, and improvement in the quantitative parameters analyzed have been reported recently 23, 24. The complete 3-D data set of the coordinates of the automatically detected lumen corresponding to the highly echogenic blood–vessel interface, and of the echogenic media–adventitia interface can be used for dosimetry evaluation.

Dose-volume histograms (DVH) are used everyday in radiotherapy to condense the large body of information of the complete 3-D dose distribution data into a plot graphically summarizing the radiation distribution throughout the target volume and the anatomical structures of interest 25, 26. We have recently reported preliminary data on the methodology to compute DVH for coronary brachytherapy from 3-D IVUS data [27].