Smart technologies built with artificial intelligence (AI) can be seen across many aspects of daily life. AI and smart technologies, such as machine learning and deep learning, work behind the scenes to create conveniences that consumers rely on every day.
AI technology can strengthen customer experiences by lowering operational costs, improving efficiency, boosting revenue, and better predicting customer behavior.[1] Greater AI adoption across industries, combined with innovations in AI, machine learning, big data analytics, and the Internet of things, has paved the way for highly sophisticated applications, specifically designed for environments like manufacturing, logistics, planning, and business operations. [2] AI applications are rapidly expanding across many industries, including healthcare.
Broadening AI applications in healthcare
GlobalData Healthcare’s 2022 report says AI is a key investment priority and will continue to attract industry investment across the next two years.[3] Widespread use of advanced technologies and AI in healthcare institutions helps to improve care, quality of service, and the efficiency of medical resources and staff. Radiology has effectively adopted these technologies, namely in image reconstruction with improvements in image quality, and in optimizing workflows with patient positioning, protocol assistance, and reporting.
“AI is firmly established as an essential tool for progress, and radiology is one of the best use cases for AI in the healthcare industry,” said Scott Miller, Chief Marketing Officer of Imaging at GE HealthCare. “AI use in diagnostic imaging is growing and has shown impressive accuracy and sensitivity in the identification of pathologies.”
Within medical imaging, the Food and Drug Administration (FDA) has authorized over 500 AI-enabled devices for clinical use—91 in 2022 alone.[4] From 2017 to 2018, the FDA has doubled their review of AI- and machine learning-enabled devices, and that number continues to grow. 4 The FDA authorized 115 submissions in 2021, an 83 percent jump from 2018.4 GE HealthCare is committed to facilitating the adoption of AI-enabled technologies into the radiology workflow. A leader in AI innovation in medical devices, the company already has 42 AI-enabled offerings authorized by the FDA.[5]
Expanding AI and deep-learning applications in medical imaging
Radiology departments are seeking tools that can improve efficiencies across the clinical workflow and reduce the cognitive burden on radiology staff. Across healthcare, radiology departments are rapidly incorporating innovative AI-enabled technology developed for medical imaging.
AI tools are helping radiologists process large volumes of imaging data with advanced image reconstruction algorithms, while helping to improve the consistency and accuracy of medical imaging and diagnostics. Several AI-based image analysis models, including deep learning, have been utilized to assist with image reconstruction. AI-based clinical imaging applications can offer significant advantages, such as the potential to help improve disease visualization with better image quality for more informed diagnoses.
“AI applications can deliver better image quality in many cases,” explained Miller, “and can be used to improve efficiencies as well as workflow. We’re continuing our commitment at GE HealthCare to leverage deep learning and other AI-based tools to improve image acquisition and reconstruction technologies to deliver better patient outcomes and to make sure that radiologists can work more efficiently and cost-effectively.”
Industry leaders like GE HealthCare continue to develop AI-enabled solutions to help radiologists tap into efficiencies through clinical data and AI integration, add precision to their reporting, and decrease the barriers to AI adoption.
“As radiology organizations adopt and integrate these AI-enabled tools into their imaging workflows,” Miller continued, “we help to reduce the burdens of implementation and overhead through a standardized, orchestrated approach that can significantly reduce the long-term cost of ownership for these investments.”
Enhancing image resolution with deep-learning imaging reconstruction
Deep learning, a subset of machine learning, has become an integral factor in AI solutions developed for medical imaging applications. Deep learning uses layers of information processing, gradually learning more and more complex representations of data, and then uses complex neural networks to replicate human intelligence. The convolutional neural network is considered state of the art in image analysis.7 It is used as an AI solution for advanced image reconstruction.
"We’re entering a new era in image resolution with deep-learning imaging reconstruction,” explained Miller. “Using deep-learning applications to acquire higher-resolution imaging helps to increase image quality and contributes to more confident interpretation.”
Powerful AI applications, such as deep learning for image acquisition and reconstruction, can be accessed by providers, often with a software upgrade and without the significant financial investment of replacing imaging systems. Improving accessibility to these tools can help deliver better outcomes for patients, offering them access to affordable quality imaging services.
Deep learning reconstruction is already used in several imaging modalities, such as X-ray, magnetic resonance imaging (MRI), and computed tomography (CT). It has also been recently integrated into other imaging modalities, such as positron emission tomography (PET)/CT.
Advancing X-ray imaging with radiology AI
More than 80 percent of all health system visits include an imaging exam, making radiology an essential part of diagnostics and healthcare. X-ray is the most prevalent imaging exam in radiology departments, accounting for 60 percent of all imaging studies in the USA.8
“As the gateway to patient care, X-ray is critically important,” said Andrew DeLaO, Chief Strategy and Marketing Officer for North America at GE HealthCare. “Getting consistent diagnostic clarity can enable clinicians to make a definitive diagnosis from the very first X-ray.”
AI-enabled X-ray image processing technology from GE HealthCare delivers sharp detail and consistent performance in X-ray, despite variations in exposure technique and challenging exam conditions. Clinicians can maintain dose efficiency and get high-quality images with AI algorithms that harness high-resolution power and advanced X-ray detectors, delivering outstanding clarity and extraordinary anatomical detail where it matters most.
This technology aids the imaging process by keeping the patient at the center, enabling the right image the first time and supporting the clinicians to make confident clinical decisions. Additionally, streamlining the X-ray workflow can lead to fewer repeat exams, exam rejects, and patient positioning errors, along with lower user variability, increasing throughput and capacity.
Growing deep-learning imaging reconstruction in MRI
Applying deep-learning algorithms to MR image reconstruction enables improvements that haven’t been possible using traditional reconstruction methods. Healthcare providers are using this technology to produce high-quality images with shorter scan times, overcoming the historical MR trade-offs between scan time and image quality.
GE HealthCare’s MRI deep-learning imaging reconstruction innovation uses an algorithm trained to eliminate image noise by leveraging MRI raw data with user-selectable signal-to-noise ratio improvement levels, achieving sharper, clearer, and accurate MR images. The availability of this unique, innovative technology is not limited to new MRI systems—it can be utilized on most existing installed systems.
“Deep learning reconstruction in MRI is changing patient care, fundamentally shifting the balance between image quality and acquisition time,” explained Ioannis Panagiotelis, Ph.D., Chief Marketing Officer for Magnetic Resonance Imaging at GE HealthCare. “More than 6 million patients have benefitted so far*, and our AI solutions in MRI continue to expand and broaden capabilities in diagnostics. We’re committed to delivering on the promise that there can be quality care for all through the democratization of clinical excellence.”
Advances in MRI technology and AI-based innovations in image reconstruction have enabled compelling results thus far. As these solutions are used more widely across different MRI applications—now available for all anatomies with 2D and 3D imaging including motion insensitive scan technique —they can provide clinicians with new insights into diseases and treatments.
Pairing deep learning with outstanding IQ in CT and PET/CT
Industry leaders are continuously striving to develop deep-learning solutions that can be applied in healthcare, specifically in radiology.7 In CT applications, GE HealthCare’s deep learning image reconstruction engine was trained using thousands of high-quality, filtered back projection images. Using this deep neural network results in high-quality images with reduced noise and preferred noise texture.
“GE HealthCare is leveraging deep learning across our portfolio to train image reconstruction algorithms that can create high-quality images and may be able to use lower radiation doses in CT as well as improving diagnostic performance in various clinical indications. For PET/CT, a new deep-learning algorithm designed to reproduce time-of-flight appearance and quantitative performance is already in clinical use in some regions, and is pending FDA clearance. This algorithm could be used in the future to reduce scan times, while preserving the image quality of a standard duration scan,” explained Sonia Sahney, Chief Marketing Officer for Molecular Imaging and CT at GE HealthCare.
Advances in AI-enabled CT image reconstruction techniques have significantly improved image quality and resolution, helping providers inform diagnoses with precision and improve patient outcomes. Additionally, deep learning technology can further enhance image processing capabilities in PET/CT. Since their introduction into routine clinical use, hybrid molecular imaging technologies such as PET/CT have advanced complex diagnostics and disease staging by improving specificity and quantitation.
As PET/CT technology evolves with forward-thinking molecular imaging innovations such as deep-learning-based image reconstruction, it enables broader access to advanced imaging technologies and novel AI techniques that can learn and replicate the enhanced image quality typically generated by specific hardware.
As a quantum leap in AI innovation, new deep-learning solutions are available via GE HealthCare’s game-changing digital PET/CT system.** The system’s imaging software uses data and AI able to process vast amounts of data and replicate the image quality produced by highly sophisticated PET/CT detector technology, helping to realize performance benefits without investing in more hardware or new equipment.
Leveraging innovation in AI to continuously improve patient outcomes
Innovative AI and deep learning solutions in medical imaging have the power to improve the quality of diagnostic imaging and efficiencies in radiology operations. The widespread utilization of advanced imaging technologies and AI tools supports providers’ efforts to deliver quality imaging services with powerful technologies to enhance diagnostic confidence. Expanding AI capabilities and applications across medical imaging has the potential to impact diagnostic accuracy and improve disease detection, thereby impacting patient outcomes across clinical specialties.
RELATED CONTENT:
View the on-demand presentation: Entering a new era in image resolution with deep-learning imaging reconstruction
Learn more about GE HealthCare’s deep-learning image processing and reconstruction applications:
- MR deep-learning image reconstruction application, AIR™ Recon DL
- CT deep-learning image reconstruction application, TrueFidelity
- PET/CT deep-learning image processing, Precision DL*
- X-ray advanced image processing, Helix™
DISCLAIMERS
Not all products or features are available in all geographies. Check with your local GE HealthCare representative for availability in your country.
*Calculated by installed base data with estimation 20 scans per day, 5.5 working day in a
week, fully start using AIR™ Recon DL 4 weeks after delivery, as of January
2023.
**Omni Legend and Precision DL are CE marked. Omni Legend is 510k-cleared by the US FDA. Precision DL is 510k-pending with the US FDA. Not available for sale in the United States. Clinical images shown was processed with Precision DL and obtained from an investigational device, limited by U.S. law to investigational use.
REFERENCES
[1] Yoh. Predicting consumer behavior with AI. Yoh.com. Published 2022. https://www.yoh.com/blog/predicting-consumer-behavior-with-ai. Accessed January 23, 2023.
[2] Lee D, Yoon SN. Application of artificial intelligence-based technologies in the healthcare industry: Opportunities and challenges. Int J Environ Res Public Health. 2021;18(1):271. doi:10.3390/ijerph18010271.
[3] GlobalData Healthcare. Digital transformation and emerging technology in the healthcare industry – 2022 edition. Pharmaceutical-technology.com. https://www.pharmaceutical-technology.com/comment/ai-investment-healthcare/. Accessed January 23, 2023.
[4] Reuter E. 5 takeaways from the FDA’s list of AI-enabled medical devices. Medtechdive.com. Published 2022. https://www.medtechdive.com/news/FDA-AI-ML-medical-devices-5-takeaways/635908/. Accessed January 23, 2023.
[5] FDA. Artificial intelligence and machine learning (AI/ML)-enabled medical devices. FDA.gov. Updated 2022. https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-aiml-enabled-medical-devices. Accessed January 23, 2023.
6 GE HealthCare. ClearReadTM CT. https://apps.gehealthcare.com/app-products/clearread-ct. Accessed January 23, 2023.
7 Soffer S, Ben-Cohen A, Shimon O, et al. Convolutional neural networks for radiologic images: A radiologist’s guide. Radiology. 2019;290(3). https://doi.org/10.1148/radiol.2018180547.
8 IMV Medical Information Division. 2019 X-ray/CR/DR benchmark report. IMVinfo.com. Pages 9, 37. https://imvinfo.com/product/2019-x-ray-dr-cr-benchmark-report/. Accessed January 23, 2023.
