BACKGROUND The Heidelberg retina tomograph, a scanning laser ophthalmoscope that uses confocal optics to provide high resolution of images, is able to scan the retina in three dimensions to obtain quantitative topographic measurements. The authors evaluated its usefulness for measuring macular lesions by determining the reproducibility of its topographic measurements at the macula. METHODS For each of ten healthy patients, the authors took five images with the patient's right eye undilated and five with the eye cyclopleged and dilated. As a measure of reproducibility, the standard deviation of height measurements for the same location at the macula was calculated for each patient and then the pooled standard deviation for all patients was calculated. The authors performed similar calculations for the mean depth within a contour line. RESULTS The pooled standard deviation for height measurements was 47.4 microns in undilated eyes and 36.0 microns in cyclopleged, dilated eyes. The authors obtained an extremely low standard deviation of 2.2 microns when the software calculated relative differences between measurements, such as the mean depth within a contour line. When the average of three height measurement values on 1 day was compared with the average of the three values on another day, the 95% confidence interval was +/- 58.7 microns for mean height values and +/- 3.7 microns for mean depth values within a contour line. CONCLUSIONS The authors obtained good reproducibility for height measurements with the Heidelberg retina tomograph and excellent reproducibility when the instrument calculated relative differences in height measurements. The authors recommend that patients, especially young patients, be dilated and cyclopleged to obtain lower variability of measurements. The scanning laser ophthalmoscope could potentially be used to quantify small changes in retinal lesions.