Knowledge of the three-dimensional (3D) morphology of root canals is import
ant for successful endodontic treatment. The objective of the present study
was to determine the 3D root canal axis mathematically. Two views (mesiodi
stal and buccolingual) of digitized images were taken from extracted natura
l human teeth. Geometric reconstruction to standardize projection geometry
was conducted on images. Because 90-degree turnaround image pairs are Monge
images of a given root canal, these Monge images were positioned using pho
togrammetric methods. Each well-ordered axis pair of a given root canal was
put into a common coordinate system resulting in 3D polynomial function of
the actual root canal. On the basis of the results gained using 10 samples
evaluated with the Friedman statistical test, this description seems to be
reproducible. The 3D representation of the root canal may help the clinici
ans in choosing the optimal instruments and shaping techniques. The root ca
nal axis that is described by the 3D function forms a basis for determinati
on of curvature values and torsion values in each of the axis points. Evalu
ating these values may also yield a new type of classification.