Just a couple of days back, Script All DNA Technologies had conducted a poll regarding integrating AI with RIS, and how that could support radiologists in making final decisions and diagnoses. Participants were asked if they would be apprehensive / somewhat concerned / perfectly OK with AI impacting radiologists’ decisions, especially with the controversies surrounding AI bias. 70% of the participants responded that they were somewhat concerned, 10% were downright apprehensive, and 20% were perfectly OK having AI support radiologists.

Even though the participants who responded that they were concerned/apprehensive are justified in their thinking, the fact remains that more and more radiology centers are opting to have AI integrated in their workflow. Using AI allows for more dynamic and scalable solutions, providing radiologists with additional support and data to make more informed, deeper, and relevant decisions regarding patient health.

According to the study entitled ‘Artificial intelligence in radiology’ released by Ahmed Hosy, Chintan Parmer et al, in Nature Reviews Cancer, AI methods excel at ‘automatically recognizing complex patterns in imaging data and providing quantitative, rather than qualitative, assessments of radiographic characteristics’. Another study, ‘Stand-Alone Artificial Intelligence for Breast Cancer Detection in Mammography: Comparison with 101 Radiologists’ by Rodriguez-Ruiz et al, found that ‘the performance of the AI system was statistically noninferior to that of the average of the 101 radiologists’. The AI system had a 0.840 (95% confidence interval [CI] = 0.820 to 0.860) area under the ROC curve and the average of the radiologists was 0.814 (95% CI = 0.787 to 0.841). In layman language, this basically means that AI can substantially maximize a radiologist’s decision-making ability when integrated with the PACS/RIS system. However, this by no means indicates that AI will replace radiologists, but yes, radiologists that utilize AI will be able to deliver greater value.

How does AI support a radiology workflow?

AI has the ability to take over simple and repetitive tasks that radiologists usually need to do. This gives the radiologist more free time for consultations and conversations with patients, resulting in increased value and efficiency. The Department of Imaging at University of Antwerp is using an AI-integrated radiology workflow for a few years. With the help of AI, the radiologists are able to detect intracranial hemorrhages; AI identifies high-density lesions and highlights them to the radiologists so they can easily spot the abnormalities. This drastically reduces the chances of a radiologist missing out on the lesions with the naked eye and causing a judgmental error.

What does integrating AI with PACS/RIS look like?

Interested in learning how an AI-integrated radiology workflow will function? It all depends on how seamless the integration is, without any downtime or heavy IT support. Those imaging centers using AI will have a smoother experience in terms of generating results. This is because the AI software is always going to be on, functioning in the background, and collecting, analyzing, and categorizing data for the radiologists before the day even begins. In addition to this, AI-integrated system will flag the urgent cases first, ensuring that cases are dealt with on a priority-basis.

There are significant benefits to integrating AI with RIS/PACS. In addition to what has already been outlined, having an integrated workflow results in faster, more accurate, consistent, and patient-centric diagnosis that has been completed within defined-parameters, keeping urgency in mind. AI is no doubt critical in radiology and will pave the way for radiologists in the future by redefining collaborative patient care alongside deeper medical capability.