Carestream’s CS 9600: Artificial Intelligence Removes Metal Artifacts in CBCT

At the Greater New York Dental Meeting 2024, Dr. Scott Ganz presented an overview of the Carestream CS 9600 device—a cone-beam computed tomography (CBCT) system—and its accompanying software, which utilizes artificial intelligence algorithms to reduce metal artifacts in 3D images.

Context for clinical practice

The essence of the problem

Artifacts related to metal restorations, dental implants, and orthodontic appliances remain one of the key challenges in interpreting CBCT scans: beam hardening, scatter, and associated distortions reduce the readability of periapical zones, the implant-bone contact area, and the thin cortical plate.

The role of AI in artifact correction

The presentation emphasized that modern machine learning algorithms and neural network models can enhance visualization by automatically identifying and correcting typical artifact patterns. This can increase the diagnostic value of images for planning implant treatment, endodontic diagnosis, and assessing bone volume near metal structures.

Features and clinical benefits

• Improved visualization of periapical zones — the reduction of streak and block artifacts around metal objects allows for better assessment of periapical lesions and the condition of the cortical bone.
• More accurate implant planning — the reduction of artifacts facilitates adequate assessment of bone volume and density in the implant placement area.
• Workflow optimization — the integration of AI software can accelerate the acquisition of “cleaned” images and simplify communication with patients and specialists during multidisciplinary planning.

Limitations and caveats

Despite the benefits, it is important to consider the following points:

• Residual distortions: algorithms do not guarantee complete removal of all artifacts; in some situations, they may create new structural distortions or smooth out fine details.
• Validation of results: it is necessary to compare original and processed DICOM series and document image transformations for clinical verification.
• Does not replace clinical assessment: AI correction is an auxiliary tool; final diagnosis must be based on clinical data, other imaging methods, and clinical experience.
• Standardization and quantitative parameters: grayscale density values (HU) in CBCT are not absolute and can be further distorted by processing—overreliance on quantitative metrics after artifact correction should be avoided.

Recommendations for practicing dentists

Integration into clinical protocol

• Test the algorithm on a control sample of clinical cases with known findings before relying on it for routine diagnostics.
• Store the original unprocessed DICOM data together with the processed series for audit and comparative evaluation.
• Train the team in image interpretation to recognize typical side effects of MAR (metal artifact reduction) to avoid diagnostic errors.

Verification and documentation

When using AI processing, it is important to document the software version and processing parameters in the patient’s medical record. When planning surgical interventions, it is recommended to verify key dimensions and anatomical landmarks on unprocessed images or if alternative imaging data is available.

Expert commentary

As a consulting radiologist or dental diagnostician, I note that the integration of AI-based MAR algorithms is a logical step in the evolution of 3D imaging. However, the key to safe and effective use remains strict validation protocols and critical evaluation of results. Practicing dentists benefit from collaborating with radiologists and manufacturers to understand the limitations of specific MAR implementations and to optimize scanning parameters (FOV, exposure, positioning) to minimize artifacts at the acquisition stage.

Inference

Dr. Scott Ganz’s presentation at GNYDM 2024 introduced a combined solution: the Carestream CS 9600 device paired with AI-powered software to mitigate the impact of metal artifacts on imaging. Such technologies have the potential to enhance the diagnostic value of CBCT images in clinical scenarios involving metal restorations and implants, but require careful integration into clinical practice, validation, and retention of raw data for quality control.