This retrospective study was performed at our tertiary center, approved by the local ethical committee (IRB: 2022-P2-055–01, 2020-P2-061–01), with written informed consent obtained from all patients. The patients included in this study sought medical attention at the Department of Otorhinolaryngology Head and Neck. The decision of whether to perform HRCT or U-HRCT was made through a collaborative determination between the clinicians and the patients after thorough communication. Throughout the scanning process, we strictly adhered to ethical requirements, ensuring patient privacy protection and minimizing radiation exposure.

The inclusion criteria were as follows: (1) patients hospitalized due to suspected otosclerosis between October 2020 and November 2022; (2) no history of chronic otitis media or mastoiditis on the affected side, ear surgery, trauma, or other related ear conditions; (3) pure tone audiometry finding of an air–bone gap more than 10 dB; (4) preoperative HRCT and/or U-HRCT examination at our hospital; (5) stapedotomy was performed with the detailed plan as previously described [18], and the fixation of the stapes and the movement of the malleus and incus were observed intraoperatively; and (6) postoperative hearing significantly improved. The exclusion criteria included the following: (1) CT evidence of middle ear mastoiditis, cholesteatoma, or tympanic sclerosis; (2) malformation of ossicles or suspected congenital stapes fixation on CT images; (3) retrofenestral otosclerosis on CT images; and (4) for patients who underwent both U-HRCT and HRCT, there was an interval of more than 6 months between the two examinations.

There were 255 ears that showed signs of otosclerosis on CT. Among them, 181 underwent surgery, and 74 did not. Finally, a total of 85 patients (85 ears) with IFO were included, including 48 right operated ears and 37 left operated ears. There were 60 females and 25 males, with an average age of 42.3 ± 12.0 years. Twenty ears underwent both U-HRCT and HRCT, while 20 ears underwent only U-HRCT, and 45 ears underwent only HRCT (Fig. 1).

Inclusion process and diagnostic results of research objects

Patients were scanned using an ultra-high-resolution CT scanner (Ultra3D, LargeV). The scanning range was from the apex of the petrous bone to the mastoid tip. The parameters were set as follows: 100–110 kVp; 120–180 mAs; field of view, 65 mm × 65 mm; and isotropic 0.1 mm. The scan comprised 370 layers, and the exposure time was 20 s for each side.

Patients were scanned using either a 64-channel CT scanner (Brilliance, Philips Healthcare) or a 256-channel CT scanner (Revolution, GE Healthcare). The acquisition parameters were as follows: 100–140 kV; 120–200 mA; matrix, 512 × 512; field of view, 180–220 mm × 180–220 mm; collimation, 16 or 64 × 0.625; slice thickness, 0.67 mm; slice spacing, 0.33 mm; pitch, 0.6 mm; and bone algorithm reconstruction.

Fenestral otosclerosis was defined by the occurrence of the foci at the external wall of the otic capsule with the fissula ante fenestram, the round window, the oval window, or the facial canal involved. Retrofenestral otosclerosis was defined when the foci were located more medially within the otic capsule [8, 19]. IFO was defined as the presence of fenestral otosclerosis without concurrent retrofenestral involvement. The otospongiotic phase was diagnosed by a notable reduction in bone density (Fig. 2). The otosclerotic phase was diagnosed when the density of the foci increased, making it challenging to differentiate from the normal otic caupusle; this phase was further defined by irregular shapes in the corresponding sites (Fig. 3) and an otic capsule thickness exceeding 2.3 mm [8, 19].

Typical U-HRCT imaging findings of fenestral otosclerosis in the otospongiotic phase. a Decreased bone density in the right fissula ante fenestram (thick arrow) and thickening of the right annular ligament (thin arrow). b Uniform bone density in the left fissula ante fenestram (thick arrow), and the left annular ligament is clearly shown with linear soft tissue density (thin arrow)

Typical imaging findings of fenestral otosclerosis in the otosclerotic phase. HRCT (a) shows no definite positive signs. U-HRCT (b) shows no obvious reduction in the density of fissula ante fenestram (thick arrow), increased density of adjacent annular ligament, and slightly thick stapes footplate (fine arrow)

The original HRCT and U-HRCT images were imported into RadiAnt DICOM Viewer for multiplaner reconstruction. This included standardized axial/coronal images, reconstructed parallel/perpendicular to the horizontal semicircular canals, as well as double oblique reformations of the stapes. The reconstructed images maintained the original slice thickness. The maximum diameter of the lesion was measured at the double oblique reformations of the stapes (Fig. 4).

Fenestral otosclerosis. The maximum diameter of the lesion is 0.93 mm

Two general radiologists, with 8 and 12 years of experience respectively, independently evaluated the images. Multiplanar reconstructions were performed using the original images by the radiologists themselves. In the event of a discrepancy, consensus was reached through discussion. Simultaneously, two neuroradiologists, with 8 and 13 years of experience respectively, also evaluated the images, resolving any disagreements through consensus. All radiologists were aware of the main clinical symptom-hearing loss but were kept blind to the side of the symptom, the side chosen for surgery, and intraoperative findings. All radiologists reviewed images from both sides, but only the data corresponding to the surgical side were used for statistical analysis.

Statistical analysis was performed using the SPSS 19.0 software. Cohen’s kappa test was employed to measure the level of agreement between each two independent observers, with the strength of agreement interpreted as follows: slight (0.00–0.20), fair (0.21–0.40), moderate (0.41–0.60), good (0.61–0.80), and excellent (0.81–1.00). Differences between the groups were assessed with the chi-squared test. A p value of less than 0.05 was considered statistically significant.