Several techniques have been used to image the nasolacrimal system, providing functional (dacryoscintigraphy) or morphologic (dacryocystography, CT dacryocystography [CTD]) information. Using gadopentetate dimeglumine-diluted solution injected into the lacrimal canaliculus or instilled into the conjunctival sac, we compared the sensitivity of MR dacryocystography (MRD) with that of CTD.Eleven healthy volunteers and 25 patients affected by primary epiphora (21 patients) or postsurgical recurrent epiphora (four patients) underwent MRD after the topical administration of contrast media or cannulation of the lacrimal canaliculus. The MR imaging findings were compared with irrigation and CTD data. All patients underwent surgical treatment (dacryocystorhinostomy), which served as a standard of reference for confirming the MRD findings.The topical administration of contrast-enhanced saline solution and the injection of contrast-enhanced saline solution after cannulation were always well tolerated. In healthy volunteers, outflow of contrast media was always revealed by MRD. Eight (32%) of 25 patients with epiphora had stenosis proximal to the lacrimal sac revealed by MRD, whereas 17 (68%) of 25 showed a dilated lacrimal sac and nasolacrimal duct stenosis, as confirmed by surgical findings. The findings of MRD after the topical administration of contrast medium and MRD after cannulation of the lacrimal canaliculus were comparable with irrigation or CTD data for all patients except one.In patients with epiphora, MR imaging performed after the topical administration of diluted contrast material can reveal stenosis of the lacrimal apparatus and can be added to the standard orbital imaging protocol when lacrimal system involvement is suspected.