OBJECTIVE: Semiautomatic computer-assisted planimetry is currently the standard morphometric technique in models of intimal thickening. We evaluated stereological point counting as an alternative method for the measurement of cross-sectional vascular areas (lumen, intima, and media) by comparing precision, efficiency, and variance components. METHODS: Sections from murine atherosclerotic aorta (n = 21), stented rabbit iliac arteries (n = 30), and porcine coronary arteries (n = 30) were analyzed at two institutes using both techniques. To determine reproducibility, porcine arteries were measured twice. RESULTS: Area measurements showed almost identical means and standard deviations for planimetry and stereology [e.g., intima (mm2): 0.10+/-0.11 vs. 0.12+/-0.12 (mouse), 0.60+/-0.16 vs. 0.60+/-0.15 (rabbit), and 1.92+/-1.52 vs. 1.97+/-1.61 (pig)]. Deming regression and Bland-Altman plots demonstrated a good agreement between both techniques that was not influenced by artery size. Both methods exhibited excellent repeatability, although planimetry (-0.18+/-0.27) was more precise than stereology (-0.17+/-0.47; variance, P < .01, Levene test). In addition, intraoperator variance (error inherent to the technique) was greater for stereology (1.6-15.8% vs. 4.8-33.5%), whereas interoperator variance (error between institutes) was very small for both methods (0.1-0.9% vs. 0.1-1.7%). Indeed, biologic variability was, by far, the most important variance component in all measurements (84-98% vs. 65-95%). Finally, stereology required 20% (mouse; P < .05) to 40% (pig; P < .001) less time to complete analysis. CONCLUSION: The quantification of vascular structures by planimetry and stereology yielded comparable results in all models of intimal thickening, but stereology proved to be less time-consuming. Therefore, study design may dictate the most appropriate choice of technique.
OBJECTIVE: Semiautomatic computer-assisted planimetry is currently the standard morphometric technique in models of intimal thickening. We evaluated stereological point counting as an alternative method for the measurement of cross-sectional vascular areas (lumen, intima, and media) by comparing precision, efficiency, and variance components. METHODS: Sections from murineatherosclerotic aorta (n = 21), stented rabbit iliac arteries (n = 30), and porcine coronary arteries (n = 30) were analyzed at two institutes using both techniques. To determine reproducibility, porcine arteries were measured twice. RESULTS: Area measurements showed almost identical means and standard deviations for planimetry and stereology [e.g., intima (mm2): 0.10+/-0.11 vs. 0.12+/-0.12 (mouse), 0.60+/-0.16 vs. 0.60+/-0.15 (rabbit), and 1.92+/-1.52 vs. 1.97+/-1.61 (pig)]. Deming regression and Bland-Altman plots demonstrated a good agreement between both techniques that was not influenced by artery size. Both methods exhibited excellent repeatability, although planimetry (-0.18+/-0.27) was more precise than stereology (-0.17+/-0.47; variance, P < .01, Levene test). In addition, intraoperator variance (error inherent to the technique) was greater for stereology (1.6-15.8% vs. 4.8-33.5%), whereas interoperator variance (error between institutes) was very small for both methods (0.1-0.9% vs. 0.1-1.7%). Indeed, biologic variability was, by far, the most important variance component in all measurements (84-98% vs. 65-95%). Finally, stereology required 20% (mouse; P < .05) to 40% (pig; P < .001) less time to complete analysis. CONCLUSION: The quantification of vascular structures by planimetry and stereology yielded comparable results in all models of intimal thickening, but stereology proved to be less time-consuming. Therefore, study design may dictate the most appropriate choice of technique.