Evaluation of Temporal Bone Image Quality with Low-Dose Photon-Counting CT in Cadaveric Heads

BACKGROUND AND PURPOSE: Photon-counting computed tomography (PCCT) is an emerging technology that offers improved spatial resolution and noise reduction at lower radiation doses than conventional CT; however, the effect of dose reduction on clinically significant image quality has not been well established. Following the ALARA (As Low As Reasonably Achievable) principle, this study evaluated temporal bone image quality across decreasing radiation doses to identify the minimum dose required to maintain diagnostic integrity. MATERIALS AND METHODS: Seven cadaveric heads underwent temporal bone PCCT imaging at four tube current-time products: 290, 75, 25, and 11 mAs. All scans were acquired at 120 kV, 0.55 pitch, and a 0.5 second rotation time and reconstructed with 0.2-mm slice thickness in axial and coronal planes for left and right temporal bones. Three board-certified radiologists independently assessed prespecified, randomized temporal bone landmarks using a 5-point Likert scale to determine clinical utility (1-nondiagnostic; 5-excellent). Mixed-effects linear regression with random intercepts for patient, reader, and landmark estimated mean scores by dose. Pairwise comparisons were Tukey-adjusted, and one-sided FDR-adjusted t-tests evaluated whether predicted mean scores exceeded 3 (clinically sufficient). RESULTS: At 290 and 75 mAs, predicted mean scores were 4.81 (95% CI: 4.42-5.21) and 4.21 (95% CI: 3.82-4.61), respectively, demonstrating good/excellent image quality for all evaluated landmarks and exceeding threshold for clinical sufficiency (FDR-adjusted p<0.0001). At 25 mAs, predicted mean scores were 2.80 (95% CI: 2.41-3.19), falling in the poor/sufficient range, and at 11 mAs, were rated the lowest at 2.09 (95% CI: 1.70-2.48), scoring below the clinical sufficiency threshold (p=0.99). Overall, mean Likert image quality scores decreased with progressive mAs reduction, remaining technically interpretable but below the threshold for clinical sufficiency at 25 mAs. All pairwise comparisons were statistically significant (Tukey-adjusted p<0.0001). Interreader agreement (Krippendorff's α) ranged from 0.301 to 0.657 across landmarks. CONCLUSION: PCCT maintains clinically sufficient temporal bone image quality at tube current-time products as low as 75 mAs, while doses of 25 mAs and below fall below the threshold for reliable clinical use. Per ALARA, this threshold optimizes safety via significant dose reduction while maintaining the high image quality necessary for complex temporal bone anatomy.