The orbit was visible and assessable under the examined conditions for T1, T2 and TIRM sequences (see Figure 1), even in bilaterally implanted patients with the magnet in situ (Table 2). The eyeball, as well as the straight and oblique extraocular muscles, the optic nerve, the medial bony boundary to the ethmoid bone (lamina papyracea), the lateral boundary, the orbital floor, were clearly visible and assessable in the T1 and TIRM sequences. However, in some sequences, particularly T1, the orbital roof was not fully assessable, leading to a downgrade in visibility to category 2. Nevertheless, in DWI (diffusion weighted imaging) the orbit was not clearly assessable ipsi- and contralaterally especially in 90° and 120° positioned CI and in bilateral CI condition with magnet in situ, while in 135° positioned CI without magnet the orbit was accessible (see Figure 2).

The evaluation of the paranasal sinuses was conducted as follows: the maxillary, frontal, and ethmoid sinuses were assessed separately for each side. The ethmoidal cells were evaluated as a single unit, while the sphenoid sinus was considered a single unit, as an intersphenoid septum is not always present anatomically. As demonstrated in Table 2 and Figure 3, the included sequences allowed for a clear diagnostic evaluation of the maxillary and frontal sinuses, as well as the ethmoidal cells, except for DWI. The nasal cavity was also clearly visible in T1 and T2 sequences and unaffected by artifacts. However, the sphenoid sinus was partially obscured by artifacts, particularly in bilateral CI condition. After magnet explantation, the sphenoid sinus became assessable even in bilateral CI cases.

The clivus, an anatomical structure in the central skull base, was evaluated separately (Table 2), as it can be affected by space-occupying lesions in the sphenoid sinus (15) and by large nasopharyngeal tumors. This structure was only assessable, particularly in bilateral CI condition or at oblique angles ipsilaterally, after magnet explantation. DWI was only assessable as category 2 for unilateral 135° condition for the contralateral frontal sinus. Except for this, all other DWI images of the paranasal sinuses were not assessable due to the artifacts (category 3 and 4—see Table 2).

In cases of ipsilateral CI (with and without the magnet in situ), the nasopharynx was clearly assessable in the performed sequences [T1-MP-RAGE, T2-TSE, T1-TIRM (Table 2)]. On the contralateral side, the nasopharynx was evaluable in all conditions and sequences. However, with bilateral CIs and the magnet in situ, the nasopharynx was not fully assessable due to artifact overlap (see Figure 4). After magnet explantation, visibility was restored. DWI was not assessable at all for the nasopharynx with magnet in situ for ipsi- and contralateral CI condition in all three examined angular positions. The sphenopalatine foramen/pterygopalatine fossa was visible in the ipsilateral CI condition but not in the bilateral CI condition.

In all sequences, except for the DWI, the orbit was visible and assessable (category 1 or 2) under the examined conditions, including in the bilaterally CI condition with the magnet in place. However, the orbital roof was not fully assessable in T1 sequences, especially when the CI was positioned at a 90° angle. The literature emphasizes the importance of T1- and T2-weighted sequences for differentiating the diverse group of orbital lesions (4, 7). Nevertheless, several recent studies have highlighted the utility of DWI in orbital MRI (5–7) for distinguishing between various types of space-occupying lesions. In this case, removing the magnet is essential to make reliable conclusions regarding DWI findings.

The results of this study demonstrate the clear and reliable visibility of the maxillary sinus, frontal sinus, ethmoidal cells, and nasal cavity under the evaluated conditions and across the included T1- and T2-sequences. However, the upper ethmoidal cells were partially obscured by artifacts in cases of ipsilateral 90° CI positioning. For a complete evaluation of the ipsilateral portion of the sphenoid sinus in cases of unilateral or bilateral CI, magnet explantation is recommended to ensure a clear and comprehensive assessment.

In both benign and malignant tumors of the paranasal sinuses, MRI is essential not only for planning surgical resection but also for precise radiotherapy planning. Farina et al. (16), along with multiple earlier studies, emphasized that early-stage and recurrent tumors often present with non-specific or absent symptoms, making MRI crucial for identifying dural, skull base, orbital, pterygopalatine fossa, and masticator space infiltration (8, 16–18). Furthermore, DWI has become increasingly important in MRI imaging of inverted papillomas of the paranasal sinuses due to its ability to differentiate purely benign tumors from those that have undergone malignant transformation (19, 20). DWI imaging is not assessable for all included CI conditions with the magnet in situ in our study, and therefore explantation of the magnet before MRI is recommended in cases involving DWI.

Due to its specific anatomical location, the clivus was evaluated separately in this study and showed good visibility in T2 sequences. However, in ipsilateral and bilateral CI cases with the magnet in situ, streak-like artifacts interfered with imaging, particularly in T1-weighted sequences. Palsetia et al. (15) and Kunimatsu and Kunimatsu (21) highlighted the importance of T1-weighted imaging for evaluating the clivus, as clival pathologies are best assessed using non-fat-saturated, non-contrast T1-weighted sequences due to the clivus’s central location and fatty marrow content. Tumors of the sphenoid sinus and nasopharyngeal carcinomas can invade the clivus, significantly worsening the disease prognosis (15). Therefore, in such cases, magnet explantation is recommended to ensure sufficient visibility in T1-weighted imaging.

The findings indicate that an ipsilateral CI does not compromise the assessment of either the ipsilateral or contralateral nasopharynx. However, in bilateral CI cases, optimal visibility was only achieved after magnet explantation. As Baba et al. (13) highlighted, MRI is crucial for suspected angiofibroma cases, with preoperative embolization recommended before resection. This tumor, most common in young male adolescents, is the most frequent benign sinonasal tumor, making up 0.5% of all head and neck tumors (13, 22). It may involve the sphenopalatine foramen, pterygopalatine fossa, and skull base (13). Our study demonstrates that magnet explantation is necessary to exclude infiltration of these structures. Nasopharyngeal carcinomas, which are typically treated with a combination of chemotherapy and radiotherapy, require fusion imaging of CT and MRI scans for precise three-dimensional radiation planning (23, 24). In conclusion, to exclude nasopharyngeal masses such as sinonasal tract angiofibromas or nasopharyngeal carcinoma in bilateral CI patients, magnet explantation is strongly recommended.

As outlined in our previous study (3) and supported by the results of this study, artifact sizes are consistent with those reported in earlier studies from other manufacturers (1, 2, 25–27) and can, therefore, be generalized to other CI manufacturers.

A limitation of this study is the ethical decision not to administer contrast agents, meaning that T1 sequences were examined without gadolinium enhancement. Furthermore, the bandaging of the CI around the head rather than implanting the CI possibly resulting in poorer quality of MRI scans due to artifacts from the skin and subcutaneous tissue. Given that this is a pilot study and no comparable literature exists for the analysis of these specific anatomical regions, alternative MRI sequences may potentially reduce artifact overlap. This should be addressed in future research.