AI in Healthcare – news picked by GLI /07

Welcome to AI in Healthcare for November 2023. We’ve developed this special series to bring you carefully selected news, updates, and perspectives curated by the Graylight Imaging (GLI) team. In the least edition we delved into healthcare reimbursement policies and real-world AI in healthcare applications. Today we’d like to explore more on the very specific use case of AI technology in clinical environment and the capacity to handle imaging beyond radiology.

Leveraging AI for improved healthcare in Yale New Heaven

The rising quantity of imaging data has led to greater pressure on radiologists and their support personnel. This is an everyday challenge for many clinics and hospitals, and it was so for the Yale New Haven health system. They initiated an exploration of artificial intelligence technology within the framework of emergency room procedures. Part of this initiative involved determining the optimal approach for triaging patients presenting acute findings detected in head CT scans [1].

The algorithms operating in the background of hospital systems and prioritizing patients whom doctors should see first exist to some extent in the market. However, a significant challenge they face is the analysis of non-specific examinations, often without contrast enhancement. Also – besides clinical expectations, the biggest one was around the user experience. This technology was about saving time; therefore it could not be an additional step in the workflow.

Trust Me, I’m Artificial Intelligence

Initially, AI technology was introduced to support emergency radiologists and neuroradiologists. Its primary purpose was to assess head CT scans for potential issues and identify any cases where blood-related findings were suspected.

Results?

A substantial decrease in the response time occurred. The AI system did detect a head bleed in an outpatient. It led to a critical intervention that might have been missed otherwise. Additionally, the Yale New Heaven health system observed that radiologists started to feel more at ease knowing they had additional assistance in reviewing these cases. The utilization of AI extended from head CT scans to encompass various radiological applications. Of which we could include identifying pulmonary embolisms and coronary artery calcification.

In addition to tangible benefits, such algorithms working in the background exhibit tremendous potential for screening studies in different modalities and specializations.

AI and the capacity to handle imaging beyond radiology

One of the most significant advantages of AI in healthcare is its potential to handle imaging studies from a wide range of medical specialties. Speaking of different modalities and specializations, it is worth mentioning the latest research by Signify on enterprise imaging [2]. Radiology and cardiology are anticipated to maintain their prominent positions in the market for enterprise imaging opportunities. Nevertheless, according to Signify Research most radiology IT vendors are assessing approaches to utilize their current client bases to enter broader, multi-specialty agreements.

enterpriste imaging by speciality: signify research report

The report highlights that radiology IT vendors must adapt to changes in the market by aiming to create interoperable and optimized platforms suitable for technologies like the cloud and AI in healthcare.

Moreover, they should demonstrate their capacity to handle imaging beyond radiology and DICOM, which can be quite challenging. Expanding into other specialties requires careful attention to data standards and formats, including DICOM, video formats, JPEG, and various clinical formats that may not conform to DICOM. Additionally, the management of encounter-based workflows versus order-based workflows is a crucial consideration.

To meet these challenges, healthcare organizations are increasingly relying on enterprise imaging platforms, such as vendor-neutral archives (VNAs) and enterprise content management systems, which serve as central repositories for medical images and data from various specialties. These platforms act as a hub for AI integration and help streamline the entire healthcare workflow.