Comparison of CT and adjusted MRI for evaluating paranasal sinuses surgical key landmarks*

Background: Sinus CT is the imaging technique of choice for planning endoscopic sinus surgery (ESS). Although MRI has a better soft tissue demonstration, it is not commonly used for ESS due to suboptimal bone demonstration. We hypothesised that adjustment of certain MRI parameters, would allow better demonstration of bones and enable the surgeon to adequately identify surgical landmarks. Methodology: Twenty patients identified as candidates for ESS underwent CT and adjusted MRI exams of the paranasal sinuses (40 in total). rhinologist and a neuroradiologist independently compared and graded 46 bony structures (23 on each side) in each patient's CT and MRI. Overall, 920 anatomical structures were graded by each observer (1840 structures in total). Statistical analysis included overall and per variable grading distribution for each observer, and overall agreement. Results: MRI images were equal, or superior to CT for assessing paranasal anatomy in 66.8% and 86.4% of structures evaluated by the rhinologist and neuroradiologist, respectively. Overall agreement between observers (77%) was moderate. Conclusion: The rhinologist prefers CT demonstration of bony structures, while the neuroradiologist prefers MRI. Still, with the MRI protocol used in this study, according to both, most bony structures are well demonstrated by MRI.


Introduction
The paranasal sinus complex is a labyrinth of air-filled cells separated by a thin bony septae. Paranasal sinuses structure varies greatly among patients. Additionally, their structure may differ between the right and left side of the same patient, rendering accurate orientation essential for operative precision.
When considering endoscopic sinus surgery (ESS), a sinus computed tomography (CT) scan is indicated for two main reasons: first, the CT scan demonstrates disease presence and its extent, thereby helping the surgeon decide whether surgery is indicated. Second, due to the complex anatomy of the paranasal sinuses and their proximity to vital structures, such as the brain, eyes and carotid artery, the surgeon uses the CT scan intraoperatively as a roadmap for the surgical steps and to prevent severe com-plications such as blindness, brain damage or massive bleeding.
Moreover, the CT scan is considered the imaging of choice since it provides a clear demonstration of the bony septae and hence contributes to precision and safety during surgery.
A unique feature of recent decades is that patients independently obtain information about possible risks from various sources (television, internet, etc.). One of the main drawbacks of CT is the radiation associated with acquiring images. Although CT technology has changed significantly over the years with lower dose exposures noted (1) , many patients still fear radiationemitting devices. For many patients who may be candidates for ESS, the preoperative CT scan is not their first. nor likely to be their last CT exam, hence the fear of developing a neoplastic disease and cataract can lead patients to become overly cauti-ous during their presurgical assessment (2)(3)(4)(5)(6) .
Another well-known disadvantage of CT is poor soft tissue demonstration. In contrast to CT, magnetic resonance imaging (MRI) provides excellent soft tissue demonstration and does not involve ionizing radiation (7,8) . However, the disadvantages of MRI include suboptimal demonstration of bony structures, high cost, longer examination time and lower accessibility (7,8) .
To overcome these obstacles, the authors hypothesize that by improving certain MRI parameters a better demonstration of the bony septae of the nose and paranasal sinuses can be achieved to a degree where, at least in certain and highly selected indications, the surgeon would be able to rely on MRI preoperatively and during surgery. If achieved, the surgeon will be able to clearly visualize both hard and soft tissue features while reducing patients' exposure to radiation.

Patients and setting
This prospective study was reviewed and approved by the local institutional ethics committee. The study was performed during 2019-2020. Patients with chronic rhinosinusitis and without other comorbidities who were candidates for ESS at Assuta Medical Centers were included in the study. Revision cases were excluded.
The patients were referred to a routine CT scan of the paranasal sinuses at the radiology department as part of the routine management. Following the CT scan, the patients signed an informed consent and underwent MRI of the paranasal sinuses.

Imaging protocol
The MRI protocol was developed by experienced neuroradiologist (J.L.), a rhinologist/endoscopic surgeon (R.L.) and an MRI technician to enable a usable demonstration of the nose, septae and bony structures of the paranasal sinuses during endoscopic surgery.
To establish the protocol, the rhinologist selected 23 specific nose and paranasal bony structures to serve as key anatomical landmarks for orientation during surgery. These bony elements included intranasal and paranasal sinus structures, and framework-related bony elements that separate the nose and paranasal sinuses from adjacent vital structures: the brain, orbit and blood vessels (Table 1). Next, the neuroradiologist and the rhinologist evaluated various MRI sequences by comparing pairs of CT and MRI scans of four patients until the optimal protocol was agreed upon.

Image analysis
Following optimization of the MRI protocols, the rhinologist and the neuroradiologist observed and evaluated the CT and MRI images of each of the study participants independently. Each structure was first identified on the CT scan, which served as the reference image. Then, the same structure was located on the MRI image and graded on a scale of 1 to 4 as follows: 1) not visible at all on MRI, 2) demonstrated less than in CT (CT superior), 3) demonstrated the same as in CT, 4) demonstrated better than in CT (MRI superior). Overall, 920 anatomical structures were graded by each observer.

Statistical analysis
The grades provided by the rhinologist and neuroradiologist for the comparison of pairs of CT and MRI images were summarised using absolute frequency, relative frequency, and confidence intervals for proportion by specialty (rhinologist/neuroradiologist) and by anatomical structures. Confidence intervals for the proportion of "MRI same or superior" (i.e., pooled grades 3 and 4) were also computed. The strength of agreement between the rhinologist and the neuroradiologist was calculated using
According to the rhinologist's observations, CT was superior to Both the rhinologist and neuroradiologist did not report any image comparisons that were grade 1, i.e., "not visible on MRI" ( Figure 1).
An example case demonstrating a lamina papyracea defect on MRI compared with CT is illustrated in Figure 3F.
Kappa analysis to evaluate the magnitude of agreement between the evaluations of the rhinologist and the neuroradiologist showed that overall agreement was 77.6% (Kappa =0.41; Table 5).

Discussion
CT is currently regarded as a mandatory scan prior to, and during ESS (9)(10)(11) . CT is considered better at demonstrating bony structures and remains the gold standard for imaging sinus pathology, while MRI is usually used as a complementary scan only in selected cases. In a study by Hähnel et al. (12) , CT and MRI of patients with inflammatory paranasal sinus disease were evaluated subjectively by two neuroradiologists (without rhinologist) and confirmed the common opinion that CT is superior to MRI for planning ESS.
The advancement of CT imaging has led to reduced long-term cancer risk following the use of this technology (1,13). Nonetheless, there is direct evidence from epidemiologic studies that the ionizing radiation doses delivered by some CT scanners are in ranges linked to increased risk of cataract and malignant tumour development. This evidence is convincing and applies to both adults and children (2)(3)(4)(5)(6) . Cone Beam Computed Tomo-graphy (CBCT), a low-dose volumetric imaging technique, is a good alternative to conventional CT as an imaging technique of bony structures in children as well as in adults (7,14,15) . CBCT of the paranasal sinuses delivers the same radiation dosage as an X-ray examination of the chest (16)(17)(18) . According to Lechuga et al. (19) , chronic sinusitis is an emerging indication for CBCT in the anterior skull region; however, conventional CT shows superior soft tissue differentiation with apparently more uniform, accurate and clear images (20) .
Compared with a CT scan, MRI has a known high sensitivity for demonstrating inflammatory processes, malignant tissue, brain and orbital structures (21)(22)(23)(24) . In the study reported here, our novel MRI protocol was shown to be equal, or superior to CT for assessing paranasal anatomy in 66.8% and 86.4% of structures  (25,26) . However, when considering to use MRI, some other disadvantages must be taken into account, including the need for insurance preauthorization due to increased costs, lower availability and longer examination time. Yet, in cases in which the use of CT scans prior to ESS appears questionable, and given the complete absence of ionizing radiation, MRI enables the assessment of radiosensitive populations such as children, pregnant women, and patients with repeated exposure to radiation.
The risks of using gadolinium-based contrast agents (GBCA) should also be addressed due to their popular use in diagnostic imaging. Intravenous administration of GBCA is associated with contrast deposition in neuronal tissues that is unrelated to renal or hepatobiliary dysfunction (27,28) . Despite uncertainties and while additional studies are needed, it seems reasonable to recommend using GBCA only when clinically necessary and at   and lateral lamellae (black circles) are generally sharper on a CT image, they can also be clearly seen on an MRI image.