The market for radiology AI solutions increased substantially in 2021, correlating to increased demand for radiology services, and a bottleneck in available physicians. AI solutions are being considered and implemented to increase efficiency, reduce staffing burdens and improve healthcare delivery.
Arterys recently conducted a study of 30 independent radiology groups, representing a variety of practice types and patients. These study participants were mostly directly involved in their radiology practices' overall IT strategy and decision-making. There were a number of notable trends, but in this article we’d like to focus on the adoption of AI in these radiology practices.
AI can transform a radiology practice. Almost 57% of respondents in our 2021 study have fully adopted an AI application into their clinical care.
This survey shows that increased productivity and accuracy are the factors driving the integration of AI into independent radiology practices.
Productivity improvement is one of the main drivers of AI adoption into radiology practice, with 27% of respondents adopting AI into their workflows for enhanced efficiency.
Without AI support, many of the tasks performed by radiologists are tedious and time-consuming. As the volume of study data to review increases, radiologists need a more effective, modern user experience that streamlines complex workflows. For example, manually measuring the size of lesions or scoring images requires prolonged focus and is prone to error.
The efficiency of existing workflows is also hindered by radiologists multi-tasking between reading images, searching for relevant clinical data, and navigating across multiple applications. The slower pace of image analysis and interpretation limits the number of patients a given radiologist can cater to and decreases overall productivity, financial reimbursement, and quality of patient care. With the integration of AI into existing PACS, radiologists can benefit from faster, more simplified and data driven workflows that streamline image interpretation and increase overall productivity.
17% of survey respondents chose to adopt AI to increase the accuracy of results. As you can see, the respondents primarily don't use AI for increased accuracy, but instead a multitude of other reasons such as increasing efficiency and automating manual tedious tasks. Medical image analysis and interpretation are subjective processes, even for the most experienced radiologists. AI algorithms can standardize aspects of these processes such that radiologists can make fast and accurate diagnostic decisions.
Besides improvements to productivity and accuracy, survey respondents also adopted AI into their practices for reasons such as:
Some respondent groups adopted AI into their radiology practice for more specific reasons, such as improving the detection of lung nodules or breast cancer.
The survey's data shows greater AI adoption for larger-sized radiology groups: all respondents from radiology groups that conduct over 1 million exams per year have adopted AI into their practice. However, only 25% of groups that perform 250,000 to just under 500,000 exams per year have integrated AI into their workflows.
In contrast, 43% of independent radiology groups which have not adopted AI into their workflows mention financial constraints as the main limitation. Other reasons for not adopting AI include:
Even with the transformative potential of AI, some practices have been reluctant to adopt AI into independent radiology practices. Of the 13 survey respondents who have not adopted AI into their radiology practices, 54% are not considering the option. However, 46% of respondents are actively looking into how to adopt AI into their practices, with improved disease detection and exam turn-around time as the top motivating factors. Other reasons include:
How do independent radiology groups envision AI integrating into their practices?
As the AI market grows and technology advances, there remain many possibilities for radiology AI applications. However, the immediate focus for most independent radiology groups within the next five years is mainly around automation of:
Automation can significantly increase the bandwidth of radiologists by shortening the time spent on image interpretation. For example, radiologists spend at least 3-4 seconds per image to meet the workload demand for an eight-hour day.
AI can help redeem that time and help radiologists address pressing patient needs. An internal study revealed that our Cardio AI product dramatically improves the rate of LV/RB function segmentation by 25 minutes, demonstrating the remarkable capability of AI to streamline diagnostic decision-making and simplify study workflows. Image analysis driven by machine learning (ML) also dramatically boosts radiologist productivity by decreasing the time spent manually analyzing images from the typical five minutes to only 20 seconds with ML support.
Another key priority for radiology groups is improved disease detection to minimize false positives, especially for complex disease conditions such as cancer.
Arterys has worked with academic and industry leaders to understand the radiology AI landscape and identify relevant solutions for improved radiologist productivity. Your radiology practice can benefit from AI solutions that enhance clinical effectiveness and differentiated service. Our platform is the world’s first and only internet medical imaging product powered by AI and cleared by the FDA. Your radiology practice can benefit from the superior performance and cost of a transformative AI solution.
Let us show you how!
Arterys: May 02, 2023
Arterys, a Tempus company, has announced a US and EU distribution agreement with Therapixel, a medical imaging software company providing AI...
Arterys: Apr 19, 2023
The 2023 Society of Breast Imaging (SBI) Breast Imaging Symposium is a must-attend event for radiologists and healthcare professionals involved in...
Arterys: Feb 27, 2023
There are a lot of radiology conferences and events in March, and you can connect with the Arterys team at a number of them! Six specific events in...
Arterys