Can AI Solve the Challenges Radiologists are Facing Today?

Nearly half (49%) of radiologists report signs of burnout1, no doubt in part owing to radiology’s increased complexity in recent years, including a five-fold jump in the number of MRI images produced per scan2. And, in a sign that radiology departments are being pushed to their limits with huge volumes of data and not enough professionals to hire, an estimated 40 million diagnostic errors involving imaging occur annually worldwide3

Artificial intelligence (AI) has the potential to address some of these challenges. Advanced analytical solutions currently available or under development could help inform critical steps of the patient journey and provide support for HCPs in handling their ever-rising workloads. The ultimate goal among those creating these new applications is to support radiologists in their mission to provide a timely and accurate diagnosis for their patients.  

Three examples of how AI can make an impact in radiology:

Triage Algorithms

In diagnostic imaging, AI triage algorithms have already shown their potential to flag suspected pathologies for prioritized reading of cases that may need urgent diagnosis and treatment. This has been demonstrated in conditions such as large vessel occlusions, intracranial hemorrhages, pulmonary emboli, and fractures. This ability can be translated into a simple optimization of the reading workflow. Traditionally, radiologists often look at patient studies chronologically, commonly focusing on the ones that come in earliest. Triage AI algorithms can now prioritize the worklist and alert the radiologist if there is a suspected finding needing immediate attention. This can reduce turnaround time, and the speed of delivery of the result to the referring doctors who in turn can manage patients in a more timely manner.  

Workflow Support Algorithms

Another area of promise is departmental workflow focused AI algorithms. Using AI in these applications could help facilitate how patients are scheduled, studies are acquired, reports are generated, and findings are communicated to patients and doctors. AI has the ability to automate parts of these processes with the potential to increase efficiency while reducing workloads.

Quantifying Previously Qualitative Findings

A third way AI algorithms can help relieve burden on radiologists is in automatically quantifying findings that previously were mostly qualitatively assessed. For example, in multiple sclerosis (MS), radiologists would traditionally qualitatively assess these according to the size and estimated amount of plaques present. New AI tools can potentially quantify more precisely and add numbers to those assessments, which could help the neuro-radiologist in comparing results with prior studies and reporting on therapy effects. This could provide benefits for the neurologist taking care of MS patients. This kind of quantification of data also could transform the way radiology findings are reported and even has the potential to impact treatment paradigms and guidelines down the road. 

For Dr. Ryan Lee, Associate Professor at Thomas Jefferson University and Radiology Chair at Einstein Healthcare Network in Philadelphia, AI has great potential for guiding radiologists in new ways. 

“I do not think we have come anywhere close to seeing what all the use cases are for these algorithms. It is just the beginning,” says Dr. Lee.  “As a radiologist, if I need to evaluate a CTPA of the chest for pulmonary embolus, for example, it is helpful and comforting to have an AI algorithm alert me there may be a possible finding. In the generalist radiology workflow, these AI algorithms could have a tremendous amount of impact outside of the specialty that maybe the radiologist did not regularly practice in.”

Bayer’s Radiology Digital Solutions organization is launching its AI digital platform, CalanticTM Digital Solutions, to help relieve some of the pressure on physicians like Dr. Lee and on overburdened and overstretched hospitals. The company describes the platform as an orchestrated suite of AI radiology solutions that will focus on the quality of care today and help transform radiology tomorrow. The cloud-based marketplace, with service line AI apps, are integrated into the radiologists’ workflow. The Calantic Viewer integrated into the PACS provides the opportunity for the radiologist to visualize findings of the algorithms and accept or reject them, prior to storage in PACS. 

The Calantic Digital Solutions platform has several components designed to simplify and make the radiology workflow more efficient, with access to tools and AI applications that support diagnosis, workflow, and quantification. The intuitive functionality is tailored for specific disease areas starting with Neuro and Thoracic apps. The approach will help to tackle the radiology challenges from a disease management angle throughout the patient journey with additional areas to be covered in the near future. 

Stroke affects 15 million people each year4 and is a leading cause of serious long-term disability and death in the US5. Every 10-minute delay may negatively affect the patient outcome6, so early diagnosis is of critical importance. There are several applications on the Calantic Marketplace with triage studies that potentially demonstrate pathologies that can lead to a stroke. New AI tools can be used in triage, identifying patients with signs of Large Vessel Occlusion (LVO) or Intercranial Hemorrhage (ICH) who need rapid intervention and supporting the radiologists and treating physician in the decision making. They also support the prioritization of these cases. 

According to Dr. Ankur Sharma, Sr. Director Medical and Clinical Affairs Digital Lead, Radiology at Bayer, Calantic Digital Solutions has been developed to address the challenges burdening radiologists from a disease management angle, by combining orchestrated AI apps and clinical expertise for specific anatomical regions.   

“We are really committed to making this vision come true, with integrated solutions aiming to aid in the complex decision-making process throughout the patient’s journey,” Dr. Sharma said.
He continued, “It is important to always think about patients. Simply put, the timing and precision of results matter, and we are committed to providing radiologists with solutions offering them support to do that.”

Dr. Lee added, “I think AI is going to revolutionize the radiologist’s impact on the patient’s journey.  And I think it is the job of the early adopters and AI visionaries to really get their teams excited and to show how it can help referring physicians down the line.”

The future of radiology is closer than you think, visit to learn more.


1 Baggett SM, Martin KL. Medscape Radiologist Lifestyle, Happiness; Burnout Report 2022. February 18, 2022.
2 McDonald RJ, Schwartz KM, Eckel LJ, et al. The effects of changes in utilization and technological advancements of cross-sectional imaging on radiologist workload. Acad Radiol. 2015 Sep;22(9):1191-8. doi: 10.1016/j.acra.2015.05.007. Epub 2015 Jul
22. PMID: 26210525.
3 Itri JN, Tappouni RR, et al. Fundamentals of Diagnostic Error in Imaging. RadioGraphics 2018 38:6, 1845-1865.
4 Stroke, Cerebrovascular accident. World Health Organization. Accessed June 17, 2022.
5 Stroke Facts. The Centers for Disease Control and Prevention. Page last reviewed: April 5, 2022. Accessed June 17, 2022.
6 American Stroke Association. Even short delays in the ER may reduce the lifespan of stroke survivors [press release]. March 11, 2021.

Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.


Featured Whitepapers

Featured Webinars