Flat panel CT offers an isotropic spatial resolution which is about two-fold higher than multidetector CT [1].

Imaging of the complex bony and neurovascular anatomy of the skull base profits from high spatial resolution. While some preclinical studies showed no difference in imaging the skull base between MDCT and FPCT [6] or superiority of MDCT over FPCT [2]. Other preclinical studies showed superiority of FPCT compared to MDCT, regarding target registration error for intraoperative navigation in skull base surgery [1], or delimitability of anatomical maxillofacial and anterior skull base structures [14, 20]. Our study showed superiority of two standard FPCT protocols, with different radiation dose, over a standard MDCT-protocol, systematically assessing clinically relevant structures in the skull base of 10 whole head preparations.

The 20s FPCT protocol (20s scan time) yielded best delimitability of bony skull base structures compared with 10s FPCT - which holds half the scan time and half the radiation exposure of the 20sFPCT – and the MDCT protocol. Statistically significant difference in delimitability of nearly all anatomical structures was seen in anterior and central skull base for 20s FPCT and 10s FPCT compared to MDCT, while in posterior skull base differences were not statistically significant between FPCT and MDCT.

Skull base foramina of the middle cranial fossa transmit important neural and vascular structures. The foramen rotundum connects the middle cranial fossa with the pterygopalatine fossa, the second branch of the trigeminal nerve (maxillary nerve) and emissary veins run through it [11].

Asymmetric widening of the foramen rotundum may be a sign of tumoral spread along the maxillary nerve.

The Foramen venosum (Foramen of Vesalius) is an inconsistent communication between the middle cranial fossa and the scaphoid fossa. It transmits a dural sinus, which connects the cavernous sinus with the pterygoid venous plexus [13], its incidence is reported to be 70–80% [11, 16] which is in line with an incidence of 70% in our cohort.

Asymmetry of the Foramen venosum may be due to e.g. carotid cavernosus fistula, tumor and neurofibromatosis [16] or caused by embryological confluence with the foramen ovale.

Foramen ovale forms the communication between the middle cranial fossa and the infratemporal fossa. It carries the third branch of the trigeminal nerve (mandibular nerve), it may contain an accessory meningeal branch of the internal maxillary artery which supplies the trigeminal ganglion (gasserian ganglion) and in the absence of the inconsistent canaliculus innominatus – as in our cohort - it transmits the lesser superficial petrosal nerve, which originates from the tympanic branch of the glossopharyngeal nerve and additionally holds fibers from the facial nerve.

In our cohort foramen ovale was well delimited from neighboring skull base foramina, e.g., foramen lacerum, which sometimes lacks its lateral wall and then communicates with foramen ovale [11].

Foramen spinosum was present bilaterally in all our preparations. It communicates as the foramen ovale between middle cranial fossa and fossa infratemporalis and holds the middle meningeal branch of the external carotid artery, the middle meningeal vein and the recurrent branch of the mandibular nerve [11].

Hypoplasia or absence of the foramen spinosum exists in case of an aberrant middle meningeal artery. This situation arises either due to a fault in embryological development of the stapedial artery, which originates as a branch of the second aortic arch and therefore the internal carotid artery. If the communication between the stapedial artery and the external carotid artery fails to evolve during embryological development, aberrant middle meningeal artery originates from the ophthalmic artery and courses through the superior orbital fissure.

The other cause for an aberrant middle meningeal artery is a persistent stapedial artery which transmits through the tympanic cavity, the facial nerve canal, and the facial hiatus (sulcus of major petrosal nerve) to become the middle meningeal artery.

Clinically important structures in the anterior skull base are the anterior and posterior ethmoidal canals which transmit the anterior ethmoidal artery, vein and nerve and the posterior ethmoidal ethmoidal artery, vein and and nerve respectively. The course of these arteries is variable and therefore important regarding paranasal surgery.

Radiation dose of FPCT has been mentioned by several authors [9, 23].

Struffert et al. showed [26] that FPCT can have a significant dose reduction compared to MDCT standard protocol if collimation is used in FPCT, resulting in the same effective dose of 0.2mSv for 10s FPCT and MDCT, while the effective dose of the 20s FPCT is 0.4mSv.

Therefore, the FPCT with a scan time of 10s can be considered comparable to the MDCT protocol regarding radiation dose in our study.

Reduction of dose results in reduction of signal to noise (SNR) which is in line with the qualitative image rating in our study.

Additionally, contrast to noise (CNR) measurements showed decreased ratios for 10s FPCT and MDCT compared to 20s FPCT, nevertheless all FPCT and MDCT images held diagnostic image quality.

In a recent global survey of usage patterns and the role of intraoperative neuronavigation nearly 25% of skull base surgeons reported using neuronavigation in all cases, main image modalities used were magnetic resonance imaging (MRI) and CT in over 56% of cases [7]. Most indications were complex sinonasal, extended skull base and reoperation cases.

To achieve an accurate navigation, correct registration of CT data to the head of the patient is pivotal. Registration can be done with externally fixed systems (fiducials) or surface registration (e.g., laser or optical) at which externally fixed systems yield the highest accuracy [4, 17]. Taeger et al. showed that a combination of flat-panel volume CT and electromagnetic navigation is highly precise, however there was no significant difference in fiducial registration error (FRE) using a MDCT data set versus a conventional FPCT data set [27].

Nonetheless, FPCT carries in contrast to MDCT the advantage of being available as a mobile device, and therefore being suitable for intraoperative scanning, which allows for adaption of navigation to surgery-related anatomical changes, or robot assisted stereotactic surgery [28].

Comparative studies regarding skull base CT imaging, dealing with image quality and resolution, were predominantly performed in explanted skull base preparations or 3D printing models.

In our study on whole-head preparations, we show that clinically relevant skull base structures are significantly better delimitable using FPCT compared to MDCT.

Regarding the radiation-reduced 10s FPCT protocol, there were only 4 of 18 evaluated anatomical structures significantly inferior delimitable in comparison to the 20s FPCT protocol, while image quality was better than that of MDCT.

In conclusion, 10s FPCT protocol serves as a substantiated tool whenever a high spatial resolution imaging of the skull base is needed.