High-Resolution Single Tooth MRI With an Inductively Coupled Intraoral Coil-Can MRI Compete With CBCT?

Objectives The aims of this study were to quantify T1/T2-relaxation times of the dental pulp, develop a realistic tooth model, and compare image quality between cone-beam computed tomography (CBCT) and high-resolution magnetic resonance imaging (MRI) of single teeth using a wireless inductively coupled intraoral coil. Methods T1/T2-relaxometry was performed at 3 T in 10 healthy volunteers (283 teeth) to determine relaxation times of healthy dental pulp and develop a realistic tooth model using extracted human teeth. Eight MRI sequences (DESS, CISS, TrueFISP, FLASH, SPACE, TSE, MSVAT-SPACE, and UTE) were optimized for clinically applicable high-resolution imaging of the dental pulp. In model, image quality of all sequences was assessed quantitatively (contrast-to-noise ratio) and qualitatively (visibility of anatomical structures and extent of susceptibility artifacts using a 5-point scoring scale). Cone-beam computed tomography served as the reference modality for qualitative assessment. Statistical analysis was performed using 2-way analysis of variance, Fisher exact test, and Cohen kappa. Results In vivo, relaxometry of dental pulps revealed T1/T2 relaxation times at 3 T of 738 +/- 100/171 +/- 36 milliseconds. For all sequences, an isotropic resolution of (0.21 mm)(3) was achieved, with acquisition times ranging from 6:19 to 8:02 minutes. In model, the highest contrast-to-noise ratio values were observed for UTE, followed by TSE and CISS. The best image/artifact quality, however, was found for DESS (mean +/- SD: 1.3 +/- 0.3/2.2 +/- 0.0), FLASH (1.5 +/- 0.3/2.4 +/- 0.1), and CISS (1.5 +/- 0.4/2.5 +/- 0.1), at a level comparable to CBCT (1.2 +/- 0.3/2.1 +/- 0.1). Conclusions Optimized MRI protocols using an intraoral coil at 3 T can achieve an image quality comparable to reference modality CBCT within clinically applicable acquisition times. Overall, DESS revealed the best results, followed by FLASH and CISS.