Definition and measurement of restenosis after successful coronary angioplasty: implications for clinical trials.

Angiographic restenosis represents the most established measure of long-term outcome in most prospective clinical trials of coronary angioplasty (PTCA). The accuracy of assessing this endpoint is of utmost importance. The purpose of this article is to propose guidelines for the use of coronary angiography in this setting. First, the cineangiograms must be of high technical quality and performed in a high proportion of consecutive patients in follow-up under controlled study conditions that are reproducible. Second, computer-assisted quantitative coronary angiographic analysis is essential to minimize interobserver and intraobserver variability in stenosis measurement between successive studies. The following recommendations are presented for quantitative coronary angiographic analysis. Because biplane orthogonal views cannot always be performed both at baseline and at follow-up, stenosis measurement in the single-plane, most severe view often constitutes the most consistent and practical approach. The edge-detection method is still much more reproducible and accurate than densitometry and should be the preferred method of analysis. Measurement of reference diameter by the interpolated method is more objective than measurement by the user-defined approach and should be used whenever possible. Finally, measurements of absolute minimum diameter and percent diameter stenosis are both important in the assessment of outcome in clinical trials. Absolute minimum diameters are independent of variations in reference diameter, and the extent of reduction in minimum diameter between the immediate postangioplasty and follow-up angiograms, when expressed in dichotomous or continuous fashion, accurately defines the extent of vessel wall hyperplasia as an endpoint. On the other hand, vessel size corresponds in general to the size of myocardium subserved, and absolute changes do not take into account this physiologic fact. Therefore defining restenosis in terms of significant reduction in percent diameter stenosis is also a useful approach because of its clinical relevance. Thus clinical restenosis requires that a successfully dilated segment (< 50% diameter stenosis) show a > or = 50% diameter stenosis at follow-up angiography with, in addition, a meaningful degree of change, that is, exceeding 2 SDs of observer variability in quantitative measurements which, in our experience, translates into > or = 15% difference between early postangioplasty and follow-up angiography measurements.