Measurement of vessel width on fundus photographs.

Many factors play a role in determining the width of retinal vessels, based on fundus photographs. We have tried to estimate their influence by comparing experimentally six different measuring methods, taking into account film and developer, intra- and interindividual reproducibility of results under short- and long-term conditions, and training in measuring technique. Our method of choice was tenfold projection of the negatives on a fine-grained wax layer screen and use of a very narrow marker line to be aligned with the vessel borders. Width determinations on pictures of good definition, yet low contrast, repeated by a trained observer on consecutive days, lay with a 95% probability within an interval of about 9 micron on the retina. The width so determined is, of course, relative and only a rough approximation to the absolute width, because of photographic bias, interobserver difference, and uncertainly of the refractive power of the eye. The method appears to be especially suitable for the study of intraindividual variations; however, because of photographic bias, conclusions should be drawn only when based on several pictures.