With advances in cloud computing and artificial intelligence (AI), scale and accuracy can be improved without sacrificing one over the other.
This white paper discusses cloud computing and artificial AI technologies in detail and explores how each can assist radiologists in their day-to-day operations while simultaneously addressing some of the commonly believed myths about cloud tech and AI.
.gif?width=800&name=4DFlow%20(10).gif)
Misconceptions About Cloud Technology
Despite significant improvements in speed, scalability, and applicability, many radiologists still hesitate to adopt cloud computing technology.
Consider the following insights acquired by the Arterys Radiology Market Study:
“Our vendors have consistently provided us with cutting-edge solutions that fit our requirements.”
“Happy with product, too expensive to change.”
“Training and interoperability issues are very complex.”
“Don't think we'll switch in the future because our vendors have consistently provided us with cutting-edge solutions that fit our requirements.”
Research from analytics firm Grand View Research projects a changing future:
“The global radiology as a service market size was valued at USD 1.1 billion in 2020 and is projected to grow at a compound annual growth rate (CAGR) of 20.3% from 2021 to 2028. Increasing demand for advanced, low-cost cloud-based medical imaging services is the key driving factor for the market.
…Further, the growing number of medical images and the low availability of radiologists around the globe are projected to propel the market growth over (the) forecast years.”
Still, there are many widely believed myths and misconceptions that prevent radiologists from considering and using cloud technology to their advantage. We'll look at the top three:
|
|
Migration and integration difficulties are often brought up when cloud technology is being considered. But what is the basis for concern?
Enterprise software thought leader Charles Phillips highlights the primary problem with the migration process:
“Migrating an entire organization to the cloud isn’t as simple as moving resources from one server to another. It can be quite a challenge for inexperienced IT staff to grasp the cloud-based concepts of configuring and testing their resources throughout the migration process.
…Many IT workers who are already well versed in virtualization will need additional training to be able to navigate cloud environments successfully. Even some of the most basic day-to-day IT tasks will require new knowledge to handle properly in a cloud environment”
There is also the issue of data migration:
“Data migration is an important issue because archive technologies and storage media are the fastest changing PACS components, and the new technologies offer efficiencies and capabilities that can reduce PACS costs and improve productivity. These new technologies also offer more scalability, greater compliance with regulations, and increased uptime and reliability. Data migration is essential not only to those who invested in early PACS technologies, but also to those acquiring PACS technologies today.”
Cloud data migration services offered by Amazon, Google, Microsoft, and others are vastly different from traditional PACS DICOM migrations. Instead of clogging networks during off hours over a very long period of several months, data is ingested into cloud data centers using special data appliances in a matter of weeks.
Radiology-specific platforms have been developed that address these and other radiology specific concerns. Arterys, for example, offers a seamless integration process with the following workflows and systems:
- Existing on-premises PACS
- Dictation software
- Other health information systems (HISs)
- Emergency medical/health records (EMR/EHR)
Because Arterys requires zero on-premises hardware, there is no need for migration (since data is sent to Arterys on-demand), which allows for the platform to become a natural extension of a radiologist's existing day-to-day responsibilities and tasks.
|
Unlike traditional technology systems requiring a sizable upfront investment to purchase necessary hardware and software, there are usually no significant capital expenses associated with setting up the cloud.
In general, cloud technology solutions are priced on a per-license/per-unit basis with annual or monthly subscriptions.
Cloud technology really shines for radiologists from all the money not wasted elsewhere. Several industry publications have outlined numerous cost savings associated with a cloud-powered radiology workflow:
“...anytime, anywhere access to current and prior studies afforded by the cloud platform improves patient care in part by reducing the need for re-scans; (due to our cloud system our) re-scan rate for acutely transferred patients has plummeted from 10% to 15% to a mere 1% to 2%.” (Source: Radiology Business)
“...the hospital system saves about $38 each time an exam is sent to its destination through the cloud rather than via FedEx ® overnight courier.” (Source: Radiology Business)
“Improved collaboration via the cloud reduces the spend on burning and shipping CDs, a common complaint among enterprise managers. When you look at the cost to generate a single CD (as much as $15 each), combined with a 25% reading failure rate and the fact most PCs today do not have CD readers, using a web browser or zero-footprint viewer makes the most sense.” (Source: Diagnostic Imaging)
“Both Faith Regional and ARA [Austin Radiological Association] have determined that cloud storage has reduced their costs per terabyte… Quick scaling made possible by the cloud allowed ARA to cut back significantly on local archives… producing a 35 percent savings in our first year.” (Source: Imaging Technology News)
“Republic County Hospital [a rural hospital in North Central Kansas] selected its own PACS, a cloud-based system that ultimately cut its price per [imaging] study by 50 percent… freed itself from buying new equipment every few years and improved the software’s ease of use for clinicians.” (Source: Healthcare IT News)
In-house vs. Outsource IT Considerations
One additional detail clinics and hospitals should also consider is how their cloud infrastructure will be supported and maintained, assuming they don’t already have an IT solution in place.
The two main options are hiring an IT expert and outsourcing the responsibility to an IT company, with outsourcing typically being more advantageous.
Some of the key differences are:
| In-house IT | Outsourced IT |
| Relies on small staff of varying expertise to troubleshoot all issues | Helpdesk available 24/7/365 |
| Not generally abreast of industry best practices and can incur workflow management issues | Ongoing maintenance, security, and strategy handled by experts |
| IT managers average $150,000 per year | Average monthly per-user cost of $85-150 |
|
While bandwidth-related concerns still exist in some rural areas, advances in internet connectivity have allowed most non-urban regions to have access to the broadband connections needed to tap into cloud computing and handle heavy processing loads.
Well engineered cloud systems can provide an experience that matches on-premises solutions, but exceeds them, especially when you consider the distributed nature of radiology today. The user experience is optimized for bandwidth and latency challenges that existing on-premises solutions simply can’t match.
Some areas can take advantage of government-funded programs like the Alabama Department of Public Health's Telehealth program. There are also alternative types of internet access now available, including peer-to-peer connections and high-speed cellular networks.
For most individuals and clinics, however, broadband availability is no longer a problem. This allows radiologists and telehealth to take full advantage of the many benefits associated with cloud technology, including
- Essentially unlimited storage that is easy to scale
- Higher speeds of data transfer allow faster and easier access to large, high-definition images from anywhere within a browser
- Increased security and data backup/recovery in the case of a server crash and/or a malicious cyber-attack (essential for HIPAA compliance)
- Fewer technical human resources are required, further reducing costs.
How Does Artificial Intelligence (AI) Enhance the Benefits of Cloud Technology for Radiologists?
With an understanding of how cloud technology improves a radiologist’s ability to scale the volume of their work, let’s turn our attention to the use of AI for improved quality of work.
According to a study from the Journal of the American College of Radiology, the adoption of AI growing rapidly in the field of radiology:
- 30% of surveyed radiologists are using AI in their practice
- Most radiologists use AI to improve their interpretation of medical images
- 20% of practices not using AI right now plan to purchase AI tools within the next 1-5 years
However, within that same study, there was some skepticism expressed about the utility of AI by a large number of radiologists:
“Along with larger entities, about 20% of practices with fewer than five members said they use AI. Out of those not deploying the technology, 80% said they ‘see no benefit.’
One-third of respondents said they cannot justify the expense, or that the purchasing decision was beyond their control. Some expressed concerns about decreased productivity.
More than 70% of respondents hold no plans to pay for AI, while 20% see themselves doing so within the next five years.”
With much of the hesitancy behind AI adoption being due to a lack of familiarity with the benefits it provides to radiology practices, even those that serve multiple subspecialties, we are going to take a look at some of those benefits.
Benefit #1: AI improves existing radiologist workflows
Many radiologists are under the misconception that AI serves as an eventual mechanism for replacing their expertise. Instead, AI is an extension of radiologists' daily activities. The goal is to make their work easier so they can see more patients without sacrificing the quality of care.
This is best explained in a white paper published in 2017 by data management company CommVault titled “Will Machine Learning Replace Radiologists?”
“The goal for healthcare systems is to focus on using machine learning for tasks that the technology can complete better or faster than humans, while using radiologists to perform those tasks which machines are not capable of doing.
Radiologists who understand machine learning and embrace integrating it into their processes to increase quality and efficiency - as well as capitalize on the additional time gained to provide more value to patients and providers - will become highly valuable and sought after.”
In effect, this allows for a single workflow to be used by all radiologists where they:
“...can spend less time verifying cases of non-disease conditions and spend more time spotting dangerous conditions.”
Although AI can be deployed in on-premises software, AI works far better with cloud solutions compared to an on-premises PACS solution. When AI has to be integrated into the latter, a large amount of infrastructure is required to support machine learning analysis. Larger data sets have to be accommodated by the physical addition of databases.
With a cloud solution, the data capacity is nearly infinite in terms of scalability and elasticity, so daily business activities will never be interrupted due to data capacity issues.
Benefit #2: AI allows physicians to practice at peak performance despite escalating work volume, spending less time on system-related tasks while delivering the best patient care possible
Expanding upon the previous point, AI saves radiologists time in analyzing medical images.
The Radiological Society of North America (RSNA) cites three recently conducted experiments where AI verifiably improved productivity and efficiency for working radiologists.
Example #1:
In one radiology study, a deep-learning algorithm was trained with an extensive collection of anonymously-collected chest X-rays to detect abnormal X-rays from ordinary ones.
In addition to the higher accuracy generated by the algorithm, the average expert radiologist opinion took only 2.7 days to form compared to 11.2 days without the use of the AI system. (Source: Radiological Society of North America).
Example #2:
In the October 2018 issue of Radiology, an AI algorithm was trained over tens of thousands of high-definition digital mammograms to accurately quantify the amount of dense breast tissue a patient has (an independent risk factor for cancer).
In the first five months following its use in the clinic, the algorithm was able to get a 94% rate of acceptance from the interpreting radiologist, and results were generated in three seconds or less.
Example #3:
In another example quoted directly from an RSNA article, AI is shown to help improve patient safety:
“ Research shows that AI can improve image acquisition and quality while reducing radiation dose. Studies have shown that AI can automatically detect motion on an image and determine whether the patient should be imaged again before leaving. It can remove artifacts and reveal the true tissue contrast underneath.”
Example #4
At Arterys, we’ve been able to achieve similar results using our proprietary 4D Flow technology, which was featured by Signify Research:
“4D Flow uses cloud-based image processing technology to provide visualization and quantification of blood flow on cardiac MRI studies. With 4D Flow, cardiac MRI exam times can be significantly reduced from typically 60 to 90 minutes to around 10 minutes, which increases the efficiency, throughput of the hospital’s MRI service and improves the patient experience with free breathing exams."
Example #5
An independent study demonstrated that a machine learning algorithm built into the Arterys Cardio AI clinical application improves physician efficiency:
Cardio AI provides physicians with automated and quantitative cardiac MR image analysis and customizable reporting that reduces tedious and operator-dependent manual tasks and saves up to 25 minutes per study (36% faster segmentation).
Example# 6
Additionally, a clinical study of Cardio AI's automated measurements would reduce the time burden of manual segmentation for physicians:
Through the use of deep learning, Cardio AI provides automated, editable ventricular segmentations based on cine cardiac MRI images that are as accurate as segmentations performed manually by expert physicians.
Arterys is the world’s first internet platform for medical imaging. We want to transform healthcare through radiology. The Arterys platform is 100% web-based, AI-powered, and FDA-cleared unlocking simple solutions that require only a web browser. The company was the first to receive FDA clearance for a cloud-based product with artificial intelligence and currently has 7 FDA clearances. By making imaging diagnostics quantitative, intelligent, and available, Arterys seeks to improve the lives of millions of patients.
Benefit #3: AI can improve diagnostic outcomes for patients
Radiology Partners’ associate chief medical officer of clinical AI, Dr. Nina Kottler, describes how AI has led to practical improvements in patient healthcare:
“...the software improved patient care and outcomes…we identified 2.4% more intracranial hemorrhages and 4.4% more pulmonary embolisms that we were missing without the AI software.
Even though most of the missed findings were subtle, several of those cases progressed to significant findings, including a patient with a barely visible intracranial hemorrhage that returned with massive enlargement of the hemorrhage requiring surgical decompression.
Another positive, (Kottler) notes, is a decrease in the cost of care to the health system. The use of these triage algorithms decreased length of stays and improved ED throughput. In addition, ‘We saw radiologists’ efficiency shoot up by 11 or 12%. The reason: the algorithms are highly specific.’”
Conclusion
Many radiology practices have already begun to adopt cloud technology, AI, or a combination of both to enhance their workflows and allow their physicians to practice the best possible medicine for the greatest number of patients.
If you want to leverage the benefits of secure cloud platforms and AI within a single vendor while automating tedious manual processes such as measurements, directions, and other assessments involved in medical imaging, consider the Arterys platform.
Arterys is a fully cloud-native medical imaging platform that effortlessly integrates into existing PACS, RIS, dictation systems, and EHR (electronic health record) databases. Clinics benefit from multiple AI tools that empower physicians to make accurate diagnoses at scale without wasting additional time.
Some of our notable key features include:
- The use of multiple FDA-cleared algorithms per study for accurate diagnostic decision making within mere seconds
- Zero on-premises hardware required: Access our platform anytime, anywhere with just an Internet connection and a browser
- Complete PHI protection while working on the same patient study and images, whether across the world or in the next hospital room
- Commercially available in +100 countries and counting
- Cloud-computing architecture that’s fully optimized for speed and performance, allowing you to load and fully interact with a file of any size
- All of the actionable insights you need from your medical images are available in a single workstation
- No up-front costs, no IT staff necessary -- you come to us for your software issues, and we fix them for you
