Aim To assess if a contrast medium improved diagnostic yield of endodontic radiographs.
Methodology Forty-five extracted mandibular premolar teeth were radiographed in bucco-lingual and mesio-distal planes. Access cavities were prepared, pulpal tissue extirpated and Ultravist® contrast medium introduced under pressure. Radiographs were retaken and the teeth cleared following perfusion with India ink. Three examiners assessed all the films for: number of roots, number of root canals, curvature of root and/or root canal, presence of lateral canals, presence of a single foramen or apical delta and the presence or absence of canal obstructions. The examiners’ interpretations were compared with the anatomy revealed by clearing.
Results Kappa scores were calculated for each of the examiners, for each set of radiographs, to assess the level of intra- and inter-examiner agreement. Only moderate agreement was found throughout (κ = 0.40–0.61). For multiple root canals a false-positive result was significantly more likely with contrast (P < 0.05). The use of contrast did not significantly improve the sensitivity of diagnosis of lateral canals or a single apical foramen. Contrast significantly increased the risk of falsely perceiving lateral canals (P < 0.002). Overall there was no statistically significant difference in the overall assessment of the anatomy of the root canals using contrast or plain radiographs (P > 0.2).
Conclusions Plain film radiographs confidently predict the presence of root or canal curvature but apical anatomy was assessed accurately in only 46% of cases. Plain radiographs were insensitive in assessing the number of root canals present, the presence of lateral canals and the occurrence of canal obstructions. The use of Ultravist® contrast medium to improve diagnosis of root canal morphology of premolars is not supported.