Magnetom skyra system

For the IPMN group, MRI examinations were performed using a 1.5 T scanner (Magnetom Avanto Systems, Siemens Healthineers). T2-weighted MRCP three-dimensional (3D) images were acquired through sampling perfection with application-optimized contrast using different flip angle evolution (SPACE). Imaging relied on a repetition time (TR) of 2500 ms and an echo time (TE) of 700 ms, as well as a slice thickness (ST) of 1 mm with a trigged breathing technique.
For the control group, MRI examinations were performed using a 1.5 T scanner (Magnetom Avanto Systems) or a 3 T scanner (Magnetom Skyra Systems, Siemens Healthineers). For the 1.5 T scanner, T2-weighted MRCP 3D images were acquired through SPACE (TR/TE = 4000–4700/705 ms, ST 1.3 mm) with a trigged breathing technique. On the 3 T scanner, T2-weighted MRCP 3D images were acquired through SPACE (TR/TE = 3000–4600/700 ms, ST 1.3 mm) with a trigged breathing technique.

We also analyzed challenges associated with the heterogeneity of protocols (protocol trees) and assessed the similarity between different protocol trees on protocol and sequence level. Therefore, we computed the overlap coefficient for different protocol trees (for all body regions) and the sequences used within the protocols. However, since the sequence name can be adapted individually, it is not suitable for the objective characterization of a protocol step, i.e., the parameterized MRI sequence. Therefore, different strategies were recently developed to categorize and standardize MRI sequence descriptions, which can be used as a meaningful, standardized feature for machine learning applications [11 , 37 , 38 ]. We used a heuristic rule–based approach [11 ] that generates a standardized sequence name based on the MR acquisition parameters of a sequence (e.g., TR, TE, imaging technique, imaging orientation). We show the effect and benefit of sequence name standardization on assessing sequence heterogeneity across different protocol trees.
We evaluated the protocols from three different MRI scanners of the same radiology practice. Two scanners (scanner 1 and 2) are located at the same site, and one scanner (scanner 3) is located at a different site. All MRI scanners are 3 T MAGNETOM Skyra systems (Siemens Healthcare, Erlangen, Germany) with the same software baseline.

Studies were performed with a 3-Tesla scanner (Siemens MAGNETOM Skyra System, Siemens Healthcare, Erlangen, Germany). DTI data were acquired using a single shot echo-planar imaging (EPI) sequence (echo time [TE] = 91 ms; repetition time [TR] = 9900 ms; field-of-view [FOV]: 256 mm; slices: 72, slice thickness: 2 mm; slice gap: 0 mm; 30 gradient directions with b = 1000 s/mm² and 1 with b = 0 s/mm²).