Reproducibility and repeatability of volumetric measurements for olfactory bulb volumetry: which method is appropriate? An update using 3 Tesla MRI

Rationale and objectives: The aim of this study was to compare different sequences for olfactory bulb volumetry using 3-T magnetic resonance imaging, evaluating reproducibility, repeatability, and systematic biases.

Materials and methods: Twenty-two volunteers underwent 3-T magnetic resonance imaging of the frontal skull base in this prospective study. Imaging included constructive interference in steady state (CISS), T2-weighted (T2w) three-dimensional (3D) sampling perfection with application-optimized contrasts using different flip-angle evolutions, and T2w two-dimensional (2D) turbo spin-echo sequences. Two observers independently performed two olfactory bulb volumetric studies per bulb and sequence. Intraobserver and interobserver reliability was assessed using intraclass correlation coefficients. For the evaluation of reproducibility, concordance correlation coefficients were determined, and for repeatability and systematic biases, Bland-Altman plots were analyzed.

Results: Intraclass correlation coefficient analysis of the specialized observer yielded almost perfect results for intraobserver reliability (0.94, 0.85, and 0.93 for the CISS, T2w 3D, and T2w 2D sequences, respectively). For the less experienced observer, the results were 0.86 0.78, and 0.74 for the CISS, T2w 3D, and T2w 2D sequences, respectively. Interobserver reliability showed almost perfect agreement for all sequences (0.92, 0.86, and 0.86, respectively). The CISS sequence yielded the highest concordance correlation coefficient (0.84), precision (0.85), and accuracy (0.99). Bland-Altman plot analyses revealed the lowest repeatability coefficients for the T2w 2D sequence. Volumetric measurements of T2w 2D imaging showed systematically lower volumetric results compared to the CISS sequence (-22.7%) and the T2w 3D sequence (-8.3%).

Conclusions: Comparison of three imaging sequences for olfactory bulb volumetry yielded the best values for the CISS sequence in terms of intraobserver and interobserver reliability, reproducibility, accuracy, and precision. Given that even less experienced observers achieve almost perfect results, the CISS sequence is recommended for olfactory bulb volumetry.