The future of medical imaging is being driven forward by the Artificial Intelligence (AI) and augmented intelligence. The trend can only be truly described as tectonic and medical imaging has come along at the right place, at the right time. Thanks to AI, medical imaging continues to get stronger, more efficient and faster because of AI. This includes deep learning, machine learning, natural language processing and neural networks,
A recurring theme of discussions amongst Radiologists appears to be a fear of AI as they fear they will somehow be replaced by it. However, Curtis Langlotz, MD, PhD, a professor of biomedical informatics and radiology and the director of the Center for Artificial Intelligence in Medicine and Imaging (AIMI Center) at Standford University has coined a famous quote:
“Artificial Intelligence will not replace radiologists…but radiologists who use AI will replace radiologists who don’t”.
The future of medical imaging will incorporate both machine plus human integration. The potential of AI’s of improves the ability to interpret the images. Additionally, radiologists are beginning to benefit from apps created for AI and testing apps that are currently in development. Like other industries, the healthcare sector looks to AI to gain insights into data which makes the information more actionable and useful. Machines are not able to think; however, they are programmed to learn and radiology is learning how AI will mould the future of healthcare.
The president of radiology’s leading professional association stated: “AI has the potential to enhance our profession and transform the practice of radiology worldwide,” President Vijay Rao, MD. “It will allow radiologists to spend more time on initiatives that will benefit both patients and physicians.” The first job would be to rebrand and recreate the reading rooms into digital data hubs which would create a space for total image care. The Radiologists need to be present with referrers, care-decision-making and patients alike. This will allow for Radiologists to become better consultants. If you are giving patients more data that is actionable, the physicians will become more an integral part of the care process.
Rao has visualised clinical teams gathering in the diagnostic data hub as a communal space for participation in virtually through video conferencing or in person to make decisions with various insights regarding the patients care. In this situation, radiologists could someday use AI to compare the current image findings with prior images from other social determinations and other departments. Surgical, lab results, biopsy findings, physical exams, health history, patient genomics and patient demographics can be incorporated and integrated in the analysis.
AI tools in the health sector utilises data more effectively, enhances image reconstruction and improve damaged or distorted images. Work-list prioritisation will help the sickest patients to receive mandatory care so they can receive diagnosis and care as soon as possible, even in settings where radiologists are not available immediately. AI is helping expose and explore EHR data so the patient receives a more holistic approach for the patient. To accomplish this, intuitive interfaces can be created based on the types of caregivers which eases fatigue in documentation.
One of the most important ways that AI can assist today’s radiologists is through the day-to-day workflow and improving it through AI. Mundane tasks that take a significant length of time could be improved through AI which means more time spent on helping patients and reading further studies. NLP can assist in these mundane tasks because it automates image segmentation which helps with measurement, comparing newer studies with older ones and labelling.
Overall, Artificial Intelligence in healthcare will positively impact the care of patients on an individual level. Improving insight, analysing longitudinal data and improving decision-making and insight as the evidence evolves allows for change and guideline advancement will further methods of treatment, prevention and diagnosis. AI can be expected to change clinical practice gradually by helping radiologists with better performances, reliability, consistency, and improved workflow. However, Radiologists are important in labelling training datasets and developing this knowledge from the image data. Even if image interpretation API jumps in leaps and bounds, radiologists will continue to play a key role in the detection of incidental findings and the diagnosis of rare diseases.
Now is the perfect time to move forward towards AI and define a strategy to progress towards more consultative approach to medicine.