The use of fluorescence spectroscopy to discriminate atherosclerotic from normal tissue is limited by a lower sensitivity for calcified than noncalcified atherosclerotic plaque (65% vs. 93%, respectively). To evaluate plasma emission as a means to detect calcified plaque, 325 normal and atherosclerotic cadaveric aortic sites were irradiated through a 100-micron silica fiber in blood by a pulsed holmium laser (lambda = 2.1 microns, fluence = 4 J/mm2). A photodiode positioned near the proximal end of the fiber detected plasma emission during a laser pulse. Plasma emission was detected at 0% (0/110) of normal, 0% (0/107) of noncalcified atherosclerotic tissue, and 91% (98/108) of calcified atherosclerotic sites. Spectroscopic analysis confirmed the presence of calcium lines in the plasma emission from calcified atherosclerotic plaque. Although ablative fluences (greater than 3 J/mm2) were required for plasma generation, a single laser pulse ablated only to a depth of 67 +/- 16 microns in normal tissue. In an additional 10 calcified atherosclerotic sites, laser ablation was continued as long as plasma emission was detected. In all cases, plaque ablation was terminated before arterial perforation. Furthermore, the adjunctive use of plasma detection improved the accuracy of fluorescence spectroscopic classification of normal and atherosclerotic tissue. In conclusion, plasma detection has a high sensitivity (91%) and specificity (100%) for calcified atherosclerotic plaque and may be a useful adjunct for laser angioplasty guidance. Furthermore, plasma detection can be implemented both simply and inexpensively.