The Future of Radiology Made Possible
Canon Medical Systems Corporation
Meaningful Innovation. Made Possible.
Challenges and opportunities in radiology for today and tomorrow explored through the eyes of world-leading experts
Towards Integrated Diagnostics: A Solution to Evolving Healthcare Challenges in Diagnostic Imaging
Professor Marco Essig , University of Manitoba , Canada.
Integrated diagnostics, which involves the interdisciplinary assessment and coordination of diagnostic results, is an evolving concept that addresses the rising challenges in healthcare and improve quality of care. In the rapidly changing landscape of diagnostic imaging, healthcare professionals are grappling with challenges arising from increased workload, workforce shortages, and the growing complexity of clinical questions.
The demand for diagnostic services has increased by nearly 50% in the past five years driven by a surge in the need for CT and MRI scans, which has been catalyzed further by the COVID-19 pandemic and its role in reshaping patient interactions.
Prof. Marco Essig, Chair of the Radiology Department at the University of Manitoba, Canada, summarized the impact: “I'm heading a fairly large organization with more than fifty sites that provides our entire diagnostic service in the province of Manitoba. Managing the escalating workload poses a significant bottleneck, urging the need for innovative solutions,” he said. “At the same time, we also have to look at sustainability. There’s ecological sustainability, where we investigate reducing the energy consumption and waste reduction. There is economic sustainability which we need to look at the optimal use of equipment. Finally, we must look at social sustainability. We need to take care of our staff. We need to work on recruitment, training, and retention of our healthcare staff.”
The traditional approach of building additional scanner capacity is neither feasible nor cost-effective. Instead, the focus has shifted towards optimizing workflow and building a stronger IT backbone for a more integrated diagnostic system to enhance efficiency.
Integration of diagnostic information with functional imaging
Beyond morphology, diagnostic imaging is increasingly tasked with providing functional information. Technological advancements have led to the introduction of functional imaging tools, with the aim of providing a holistic view of an individual patient's health, facilitating early disease detection, efficient treatment planning, and streamlined post-operative follow-ups. Functional Magnetic Resonance Imaging, diffusion imaging, and fusion imaging are becoming essential tools across a range of indications, from trauma assessments to dementia follow-ups and post-stroke evaluations. The integration of these tools into diagnostic workflows requires robust data processing capabilities, often driven by artificial intelligence (AI). AI and machine learning play pivotal roles in managing vast amounts of integrated data. These technologies help recognize patterns, standardize interpretations, and identify when additional testing is necessary.
The role of Integrated Diagnostics in future healthcare
Integrated Diagnostics goes even further by incorporating data from radiology, pathology, laboratory results, genetic information, and more, in order to gain a comprehensive understanding of an individual patient. This approach aims to enhance diagnostic oversight, enabling early detection, precise treatment decisions, and improved post-treatment monitoring. Through integrated diagnostics, healthcare systems can achieve better efficiency, reduced costs, and improved patient outcomes. The greater clinical impact also ensures increased visibility of diagnostic disciplines in the clinical environment, with radiologists as the diagnosticians.
“Integrated Diagnostics is one of the futures that we are working on, or have to work on,” said Prof. Essig. “We need a change in workflow to integrate all the information along with a structured and integrated evaluation for specific diseases.”
Successful implementation of Integrated Diagnostics necessitates robust data management and communication systems, such as cloud-based diagnostic dashboards, specified diagnostic meetings, and enhanced communication with referring physicians. Establishing a platform for seamless and timely communication ensures that diagnostic information is effectively translated into individualized diagnoses, treatment plans, and follow-ups without diagnostic delays.
Prof. Essig shared how automation has helped workflow in his center: “Max Rady College of Medicine is a big stroke center and receives multiple patients with acute stroke and a treatment decision must be made fast. Patients get perfusion imaging either by CT or by MR,” he said. “We have introduced an automatic processing tool which produces maps that clearly show where the acute infarction is in the patient. This pre-processed information is sent to the care team automatically. The care team, which includes the Stroke Neurologist, the Interventional Radiologist and the Neuroradiologist, gets this information sometimes even before the patient leaves the CT department and can make a treatment decision in a timely manner.”
In summary, the evolving landscape of diagnostic imaging demands a paradigm shift towards Integrated Diagnostics. By addressing current challenges through optimized workflows, harnessing technological advancements, and embracing sustainability practices, the healthcare industry can pave the way for a future in which Integrated Diagnostics enhances efficiency, improves patient outcomes, and ensures the sustainability of diagnostic practices on a global scale.
Radiology At a Crossroads: Navigating the Synergy between Human Expertise and Artificial Intelligence
Radiology stands at a pivotal juncture. How do radiologists stay relevant in view of revolutionary transformation of the healthcare landscape from technological advancements and Artificial Intelligence (AI)?
领英推荐
“It is obvious that AI is not going to replace Radiologists. If you look at the hype cycle of deep learning - the technology that's now behind most AI - it was at the top of the hype cycle of emerging technologies in the year 2018. In 2019, we saw things like the rise of robot radiologists. However, a few years later, in 2022, it already dropped to the bottom of the cycle to the ‘trough of disillusionment’,” said Prof. Mathias Prokop, Professor of Radiology and Chairman of the Department of Medical Imaging at Radboud University Medical Center in the Netherlands. “We have to realize that AI systems are trained by specific training data sets and AI performance is best in cases where the data is dense. It is really difficult to train AI in rare diseases where data is limited – performance usually drops.”
Nonetheless, given the rapid evolution of AI in radiology, it is conceivable that, in due course, it will surpass human capabilities in certain diagnostic tasks. AI systems, as demonstrated in studies, outperform radiologists in detecting conditions like pneumonia, TB, pneumothorax, and nodules. Initially, using AI can help non-expert readers to function at the same level of more experienced readers. However, there may come a point where our interaction with AI will make the results worse, and AI achieves superior performance by itself, without any human interference. This sparks contemplation about a future where AI becomes fully autonomous in clinical practice, requiring a reshape of the traditional role of radiologists.
As technology propels radiology into uncharted territories, the synergy between human radiologists and AI emerges as a central theme. AI, a potent tool in radiology, enhances diagnostic capabilities by learning from specific datasets. However, the challenge lies in cultivating effective collaboration between radiologists and AI to maximize the strengths of both human expertise and machine efficiency. The future success of radiology hinges on striking a delicate balance where AI complements, rather than supplants, human expertise.
Shift from hardware to software dominance
Traditionally, the focus in radiological exams has revolved around hardware enhancements. However, a paradigm shift is occurring, with software now taking center stage. The history of CT is an illustration of this shift. “CT had an increase in performance which was exponential in the years from its beginning in the 1970s until roughly 2010,” said Prof. Prokop. “And if you look at the performance in terms of speed versus resolution, it doubles more or less every two years. However, since 2010, that hasn’t really changed much, even with photon counting CT. The reason for this is that hardware development is no longer influencing the performance of these scanners as much as it has in the past.”
Nowadays, software advancements play a pivotal role in enhancing imaging capabilities. Ultra-high-resolution imaging and deep-learning reconstructions underscore the transformative impact of software innovation on diagnostic quality, surpassing the limitations of hardware improvements. The integration of AI into radiology transcends mere diagnostic enhancement: It extends to optimizing workflow and patient care. AI-driven software automates scanning processes, reduces acquisition times, and enhances image quality. This paradigm shift signifies the potential for AI to streamline tasks, making diagnostic imaging more efficient and cost-effective.
Prof. Prokop described some of the current improvements: “A three-dimensional deep learning reconstruction really improves the sharpness of images of a regular scanner quite dramatically without the need for super expensive hardware,” he said. “The images are corrected automatically. Of course, you can apply it also on a photon counting detector scanner and it’s probably going to become even better. We see similar things in MR where we can shorten the acquisition time or increase the resolution of these systems by using deep learning reconstructions and in some cases cut down the reconstruction of the acquisition time by a factor of ten without losing image quality.”
He continued with a similar example that illustrates the advantages of the software solution for iodine maps, namely subtraction, over the hardware approach based on dual energy.
Radiologists’ role in shaping future healthcare
Radiology is not excluded from the rising pressures in healthcare. “As radiologists, we have really succeeded in improving patient outcomes. Where we have not yet succeeded is reducing cost. In addition, team experience now is one of the biggest issues. We have seen our patients wanting more care, which puts a lot of workloads on the team. We have to focus on those two things - lowering costs and improving team experience. We need to remove the work that adds little value and no satisfaction. And I think what really helps with this is AI,” Prof. Prokop emphasized. “AI can help us ideally create time for doing more meaningful work and actually also help lower costs by using our diagnostic imaging in a better way”.
One example is AI's potential to revolutionize screening processes, promising faster, cost-effective procedures that are less reliant on highly qualified personnel. Innovative applications pave the way for radiologists to play an increasingly active role in early diagnosis in addition to screening. Radiologists may also become more active in the treatment phase, shifting from surgical procedures to less invasive treatments. In AI-supported radiological interventions, such as liver tumor ablation, AI optimizes treatment outcomes by ensuring precise margin assessments and reducing recurrence rates, underscoring its therapeutic potential.
Summary
Radiology finds itself at a crossroads, navigating the delicate intersection of human expertise and AI advancements. As the profession embraces the exponential growth of AI, the key lies in collaboration, adaptation, and leveraging technology to shape a promising future. Radiologists must evolve with the changing landscape, embracing AI as a collaborative partner rather than a replacement, ensuring that the essence of human touch and expertise remains integral to the field's progress.
Watch the full webinar, The Future of Radiology Made Possible, recording here:
Program
Learning Objectives