The use of metabolic measurements in angiographic regression trials