Evaluation of a fluorescence feedback system for guidance of laser angioplasty

Abstract

Laser-induced fluorescence spectroscopy (LIFS) may be capable of guiding laser angioplasty by discriminating normal and atherosclerotic artery and by determining catheter-tissue environment. Previous optical multichannel analyzer based LIFS systems have been expensive and cumbersome. To simplify LIFS, a system based on photomultiplier tubes was developed and evaluated. Tissue fluorescence was induced by a helium cadmium laser (wavelength = 325 nm, power = 0.2-0.5 mW), collected by clinical multifiber laser angioplasty catheters and directed through one of two filters (10 nm bandpass, 380 nm or 440 nm peak transmission) to a photomultiplier tube. An LIFS ratio was defined as the relative intensity at 380:440 nm after calibration with an elastin fluorescence spectrum; 157 coronary artery cadaveric specimens were evaluated spectroscopically and histologically. To evaluate the utility of LIFS to optimize catheter position by determining catheter-tissue contact, by determining saline dilution of blood, and by orienting eccentric multifiber catheters a new variable, the total fluorescence intensity (TFI) was defined as the sum of arterial fluorescence intensities at 380 nm and 440 nm. TFI was recorded in vitro through multifiber catheters from 20 arterial specimens in vitro in blood and evaluated as a function of the catheter-to-tissue distance (d) over a range from 0 to 400 mu. Defining normal specimens as those with an intimal thickness < or = 200 mu, and atherosclerotic as those with an intimal thickness > 200 mu, 47/50 (94%) normal and 85/107 (79%) atherosclerotic specimens were correctly classified using a threshold LIFS ratio of 2.0. Mean (+/- SE) normal ratio was 1.76 +/- 0.02 and mean atherosclerotic ratio was 2.78 +/- 0.08 (P < or = 0.01). The classification accuracy of atherosclerotic specimens increased with intimal thickness so that 95% of atherosclerotic specimens (69/73) with intimal thickness > or = 400 mu were correctly classified. TFI was capable of determining catheter-tissue contact as maximal TFI was recorded with the catheter in contact with the tissue (d = 0 mu) and decreased markedly with distance (to 52 +/- 6% at d = 100 mu, 19 +/- 4% at d = 200 mu, and 3 +/- 1% at d = 300 mu). TFI was recorded from ten arterial specimens in blood/saline mixtures ranging in hematocrit from 0% (saline) to 50% (whole blood). TFI was capable of detecting saline hemodilution of blood as TFI decreased markedly at higher hematocrits such that TFI could only by recorded at hematocrits < 10% for catheter-to-tissue distances > or = 300 mu. TFI was recorded through ecentric multifiber catheters from 25 arterial specimens and eval-uated as a function of the degree of catheter-tissue overlap. TFI was capable of detecting maximal catheter-tissue overlap as TFI correlated linearly with the area (A) of overlap (TFI = 1.12 A + .07, r = 0.92). By discriminating atherosclerotic from normal tissue and by confirming catheter-tissue contact and saline hemodilution, fluorescence feedback should minimize irradiation of normal tissue and/or blood and enhance the safety and efficacy of laser angioplasty.