Introduction: Uveal melanoma is the most frequently occurring primary eye tumour. The aim of this research was to determine the difference in geometrical tumour measurements between ultrasound and magnetic resonance imaging (MRI), to compare MRI and ultrasound-based tumour models, and to evaluate the clinical implications of these differences for uveal melanoma patients.
Methods: Tumours of 42 uveal melanoma patients were segmented on T2-weighted MRI scans. Prominence, largest basal diameter (LBD), and second basal diameter (SBD) were calculated automatically and compared to clinical ultrasound measurements. Volumes of segmented MRI tumours were compared to the volumes of ellipsoid ultrasound-based models. Results: For the prominence, LBD, and SBD, the mean absolute differences were 0.8 ± 0.7 mm, 1.6 ± 1.3 mm and 1.6 ± 1.4 mm between the conventional ultrasound measurement and the MRI measurement. A significant difference was observed for the prominence between anteriorly and posteriorly located tumours (0.6 ± 0.5 mm versus 1.1 ± 0.8 mm, p = 0.02). Tumour volume on MRI was on average 70 mm³ smaller than the ultrasound tumour model volume, with a dice similarity coefficient of 0.81 ± 0.10.
Conclusion: The largest clinical impact of the use of MRI measurements may be observed in tumours with a prominence between 6 and 8 mm and a LBD between 14 and 18 mm. Furthermore, the use of MRI could add valuable shape information to proton beam therapy planning.