Artificial Intelligence (AI) has the capability to provide radiologists with tools to help improve their productivity and decision making, possibly leading to quicker diagnosis and improved patient outcomes. As evidenced by the great number of vendors entering the market, it is initially deploying as a diverse collection of assistive applications and tools. These are allowing radiologists to augment, quantify and stratify the information available to them and has the promise to provide major opportunities to enhance and augment the radiology reading and richness of the resulting reports. It is also improving access to medical record information with the goal of helping to give radiologists more time to think about what is going on with patients, diagnose more complex cases, collaborate with patient care teams, and perform more invasive procedures.
Deep Learning algorithms in particular have promise to transform the foundation for decision making and workflow, as these types of algorithms have the ability to “learn” by example to execute a task as well as interpret new data. Therefore, it may be possible by deploying AI in radiology workflows to assist health delivery organizations to realize key operational and clinical outcomes such as:
- Helping to improve productivity of clinical workflows utilizing imaging
- Assisting in lowering the risk of “negative” clinical consequences associated with delays in radiologist reading, interpreting and reporting
- Empowering care teams to easily view radiology work product, accelerate clinical decision making, and streamline workflows, helping to result in an improved patient experience and outcomes
In the context of the radiologist themselves, the intelligent workflow and clinical assistant capabilities can assist radiologists to be more:
Productive: Through the automation and prioritization of tasks and data feeds
Quantitative: By providing applications and tools to semi-automatically or automatically extract and quantify information
Precise: By ensuring the right information is available, filtered and presented to support the diagnosis, as well as ensuring the repeatability of any quantification processes
For example, GE Healthcare demonstrated at the recent Society of Imaging Informatics for Medicine annual conference, that when AI is applied to workflow it can assist with exam selection. And through its Edison™ applications, allocating a patient to the right scanner, protocoling, and patient positioning, as well as dose monitoring. In the interpretation of exams, GE Healthcare has also shown how the results of AI applications can be used to prioritize worklists. This can help to ensure the “right read” is being performed by the “right radiologist” at the “right time”. In the case of GE Healthcare’s Centricity™ Universal Viewer Smart Reading Protocols, it can also learn the radiologists’ hanging preferences, saving time in setting-up future reads. In support of both productivity and quantification, other companies such as iCAD have shown how AI based tools can segment out, label and measure anatomies, as well as areas of interest such as lesions and even provide “screening reads.”
In support of precision, the same tools can create more precise, repeatable measurements to better compare and track clinical studies. Deep Learning enabled visualization systems can also interpret context, searching out relevant imaging studies, clinical notes and other patient data that may exist in other “pools of information” dispersed throughout the healthcare enterprise. This will help assist in providing a more accurate diagnosis by showing the radiologist a more complete context and picture of the patient, the reason for the exam, and assist them in quickly separating relevant from irrelevant information.
In summary, AI shows promise to enable health delivery organizations to provide faster, more effective care to their patients. Is your organization prepared for deploying AI in your radiology workflows to benefit from these new and advanced developments?