Advancing Tech in Radiography: AI and Virtual Grids

Advancing Technology in General Radiography

• Advancements include virtual grids, dose reduction strategies, robotics, augmented reality, remote scanning, photon-counting, environmental sustainability, artificial intelligence (AI), and patient care improvements.

Virtual Grids

• Virtual grids use detector system software to estimate scatter areas and reduce their impact on the image.
• Carestream Digital Radiography (DR) systems have virtual grid capability (Smartgrid) that improves chest x-ray radiographs to look similar in quality to those taken with an anti-scatter grid.
• Virtual grids allow for lower exposure values, resulting in lower patient dose.

Artificial Intelligence (AI) in Radiography

• AI is commonly used to assist with image interpretation and diagnostics.
• AI post-processing tools can be used for bone suppression and tube and line visualization.
• AI is defined as the theory and development of computer systems that perform tasks requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
• AI is an umbrella term for technologies incorporating machine-based decision-making.
• Previous radiology terms include Computer-Aided Detection (CAD).
• AI is only as good as the data used to train it, following the principle of "garbage in, garbage out."

AI Learning Types

• Supervised learning: AI is taught by experts.
• Unsupervised learning: AI is taught by data, requiring large datasets.
• Reinforcement learning: Feedback is provided to the AI model.
• Deep learning: Complex datasets and training tools train the AI machine.

Potential AI uses in Radiography

• Distinguish anatomy/pathology/image noise and patterns
• Radiotherapy planning
• Workflow management
• Streamline services/appointments
• Image acquisition
• Assist with patient positioning

Dose Reduction

• Radiation in medical imaging must offer a significant benefit compared to potential risks.
• Flat-panel detectors can improve image quality while lowering radiation exposure.
• Dynamic Z-axis tracking in helical scanning reduces unnecessary radiation by automatically correcting the X-ray beam position.
• In cardiac imaging, prospective triggered gating technology reduces X-ray exposure time by activating the X-ray only during critical phases of the cardiac cycle, with studies indicating potential reductions of up to 70%.

Dark Field Radiography

• Dark field radiography is an innovative technology focusing on the properties of X-ray waves to improve soft tissue visualization.
• It is primarily used in research with cadavers, phantoms, and animals.
• Dark field radiography shows promise in detecting fractures and pulmonary pathologies.

Radiography Robotics

• Radiography robotics are applicable for medical uses with high payloads.

Augmented Reality (AR) Usage

• Augmented Reality enhances a radiologist's ability to navigate complex anatomical structures, making interventional procedures accurate and less invasive.
• AR is being used for medical education to visualize complex anatomical structures or simulate clinical scenarios.
• Interactive Virtual Reality (VR) environments are available for patients in MRI to explore before visiting, reducing anxiety and claustrophobia.

Remote Scanning

• Remote scanning workflow splits tasks between on-site and remote radiographers.
• Onsite, patient manager.
• Remote, exam manager
• Remote Magnetic Resonance (MR) scanning involves a radiographer not being in the control room.
• Remote setup can occur from home, office, or a scanning hub.
• Remote scanning is used in the USA, Australia, Germany, Denmark, Belgium, and China and fundamentally changes how radiographers work.

Photon Counting

• Photon Counting Computed Tomography (PCCT) detectors register individual photons.
• The Photon Counting Detector ignores lower energy photons and uses photons over a specific threshold.
• PCCT can eliminate noise artifacts from images
• PCCT first went into clinical use in 2021
• Multi-energy, spectral detection allows the scanner to determine tissue composition by separating photons into energy levels above the low threshold.
• Multi-energy spectral detection enables the removal of beam-hardening artifacts common in Computed Tomography (CT).
• Multi-energy spectral detection allows for contrast to be specifically detected, reducing radiation dose by removing the pre-contrast scan.

Environmental Sustainability

• Radiology departments, especially interventional radiology, produce high waste due to single-use items.
• Radiography consumes a lot of electricity.
• Equipment relies on high energy supplies, even when idle.
• Patients void contrast media into the water supply, which can contaminate drinking water.
• More research, guidance, and regulation are needed to move radiography into a more sustainable future.
• Lower the carbon footprint of healthcare by reducing unnecessary imaging and turning off scanners when not in use
• During scanning, maximize scan efficiency and reduce the amount of time on standby

Factors that improve AI in Radiography

• Robust validation of AI tools
• More interdisciplinary research in medical imaging and radiotherapy
• Regulatory frameworks for AI products
• More involvement of service-users in the design and implementation of AI tools
• Effort from the industry to help healthcare professionals understand and adopt AI solutions
• Clearer accountability in medico-legal frameworks
• Tailored, evidence-based educational provisions for using AI technology
• Clearer career pathways and role expansion for radiographers into a future where AI will be central

Recommendations when using AI in Radiography

• AI should have clear decision-making processes
• Radiographers should have enough understanding of AI so tools are correctly used
• Quality assurance and quality control of AI is essential in healthcare, and this can be done by radiographers
• AI tools in clinical use are decision support systems, meaning responsibility is retained by the practitioner performing the examination
• Radiographers need understanding of AI so they can explain it to other clinicians and patients
• Radiographers should engage with Artificial Intelligence (AI) as it is a powerful tool that can benefit radiology