Milea Timbergen

132 Table 2. Properties of the acquisition protocols of the 203 T1-weighted MRI sequences in the dataset. Property Number % Magnetic field strength 1T 20 10 1.5T 167 82 3T 16 8 Manufacturer Siemens 93 46 Philips 79 39 General Electrics 27 13 Toshiba 4 2 Setting (Unit) Mean Std. Min Max Slice thickness (mm) 4.66 1.45 1.0 11.0 Repetition time (ms) 619 533 0.0 4620 Echo time (ms) 14 7 2.0 94.0 T, tesla; Std, standard deviation; mm, millimetre; ms, milliseconds Figure 2 . Segmentations on various MRI sequences before and after applying image registration in a des- moid-type fibromatosis case. The arrows are at the same position in each image and point at two details where the (mis)alignment is evident. (1) Original T1-weighted (T1w) MRI; (2) Original T2w-MRI; (3) Registered T2w-MRI; (4) Original T1w post-contrast MRI ; (5) Registered T1w post-contrast MRI Differential diagnosis The performance of models 1-6 for the differential diagnosis is shown in Table 3. Model 1, based on volume, showed little predictive value (mean AUC of 0.69). Model 2, based on age and sex, performed better (mean AUC of 0.86). Model 3, based on T1w-MRI, had a mean AUC of 0.79, thus performing worse than age and sex, but better than volume alone. Model 4, combining the T1w-MRI, age, and sex, showed little improvement in terms of mean AUC (0.88) over model 2. Addition of a T2w-MRI, i.e. model 5, or T1 post-contrast MRI, i.e. model 6, both with or without FatSat, both yielded a minor overall improvement over model 3 (mean AUC of 0.84 and 0.84, respectively). These observations were confirmed by the ROC curves in Figure 3. The models using either only non-FatSat or FatSat scans, both for the T2w and T1w post-contrast MRI, faired similar, see Supplemental Table 2. 5