Coronary angioplasty is sometimes thought to be insufficiently validated considering the considerable development it has undergone and its place in modern cardiological practice. Nevertheless, several randomised clinical trials comparing angioplasty with medical therapy in stable angina (ACME) and with surgical treatment in stable and unstable angina (RITA) have provided more scientific support for the technique. The serious perioperative complications have become rare, the limiting factor being restenosis which is responsible for a large number of clinical recurrences and the reappearance of documented myocardial ischaemia. It is therefore logical to make restenosis the first objective of evaluation of PTCA. There are two possible approaches to this problem. The first relies on automatic quantitative operator-independent angiography as a gold standard. However, this method is methodologically complex, technically fastidious and only takes into consideration the anatomical appearances, the correlations with clinical outcome and prognosis of which are poor. It allows measurement of the amplitude of the process which is an unquestionable advantage, but it is only a partial view of the problem. The second method considers that only stenosis causing ischaemia is significant and that the criterion of evaluation should be the rate of new events and that the necessity of repeated attempts at revascularisation is the criterion of failure of the method. This overlooks the possibility of an anti-restenosis drug producing clinical results independents of its anatomical effect. Both methods have their advantages and drawbacks, which necessitates using them both in all trials of new tools or new molecules designed to prevent restenosis.