Feasibility of Dose Reduction in the Context of Preoperative Diagnostics in Cochlear Implant Surgery With a Photon-Counting Detector CT and Deep Learning-Supported Denoising.
Authors
Affiliations (6)
Affiliations (6)
- Department of Otolaryngology-Head & Neck Surgery, University of Tübingen Medical Center, Tübingen, Germany.
- Department of Otolaryngology and Communication Enhancement, Boston Children's Hospital.
- Department of Otolaryngology, Harvard Medical School, Boston, MA.
- Department of Diagnostic and Interventional Radiology.
- Institute for Clinical Epidemiology and Applied Biometry, University of Tübingen Medical Center.
- Institute of Clinical Anatomy and Cell Analysis, University of Tübingen, Tübingen, Germany.
Abstract
Photon-counting detector CT (PCD-CT) with deep learning-supported denoising can significantly reduce the radiation dose for cochlear implant (CI) planning without compromising the accuracy of cochlear duct length (CDL) measurements. Optimal electrode placement in CI surgery requires detailed cochlear anatomy from CT scans, but reducing radiation exposure is critical. This study explores PCD-CT with denoising algorithms to lower doses while preserving diagnostic accuracy. Four body donors without inner ear malformations were scanned using PCD-CT at 100%, 50%, 25%, 10%, and 5% dose levels. Images were denoised with ClariAce, a deep learning algorithm, and CDL was measured using OTOPLAN software. Neurotologists compared the results to manual segmentations. Statistical analyses evaluated accuracy across dose levels, with Bland-Altman plots assessing systematic errors. Automatic segmentation succeeded across all doses but showed increased failure below 50%. At 100% and 50% doses, CDL measurements closely matched the gold standard, with minor deviations (eg, -0.17 mm at 50%). Below 50%, CDL underestimation increased (-1.25 mm at 25% and -4.0 mm at 5%). Denoising improved segmentation but minimally affected CDL accuracy at low doses, where manual segmentation performed better. PCD-CT enables significant dose reduction for CI planning, with reliable CDL accuracy down to 50%. Deep learning denoising enhances image quality but is less effective below 50%, necessitating manual segmentation. These findings align with ALARA principles and suggest further refinement of AI algorithms for lower-dose applicability in CI diagnostics.