RSNA 2017, part 2: Invent new radiology paradigms with enterprise imaging and AI

 

Explore. Invent. Transform. If you read my post yesterday, you know this is the theme for the RSNA (Radiological Society of North America) 2017 Annual Meeting taking place this week at McCormick Place in Chicago. With it, the organization is challenging radiologists and other imaging professionals to investigate and advance the field through innovation — all in an effort to positively impact patient care. This blog series will examine how Health IT — and specifically an enterprise imaging strategy — can help you satisfy each of the “Explore, Invent and Transform” objectives outlined in the theme for RSNA 2017.

Today, we’ll take a closer look at Invent.

Use AI to improve accuracy

Today, radiologists and other imaging professionals are expected to be more productive, accurate and clinically valuable — all while dealing with shrinking reimbursements. This environment requires radiologists to take a critical look at their craft and invent new ways to address emerging challenges. One of the hottest topics on how to reinvent imaging has been through use of artificial intelligence (AI). In fact, Tuesday’s Plenary Session at RSNA 2017 focused on the potential of AI and machine learning in the field of radiology.

Recent headlines have touted how AI is more accurate at diagnosing certain types of cancers than radiologists. However, the more realistic application of AI in imaging is not as a replacement of a human physician, but as a means to augment their work. For example, AI has the potential to shift the imaging paradigm by enhancing the workflow efficiency of imaging professionals while simultaneously improving patient care and throughput.

The primary way AI supports this effort is by analyzing and comparing vast amounts of data and presenting it to physicians in a concise and easily digestible format. Data overload is one of the biggest challenges facing physicians today. There’s just way too much patient information to sift through and much of it resides in disconnected silos. This includes DICOM and non-DICOM images, exam and procedure reports, lab values and more.

The siloed imaging environment that currently exists in many healthcare facilities makes it difficult, if not impossible, for a clinician to manually locate and search through imaging data to identify what is clinically relevant and actionable. AI can help automate this discovery process.

For example, AI can help identify potential findings on a CT scan or MRI — not only by analyzing the image itself, but by cross-referencing it with other elements of the patient history that could impact the part of the anatomy scanned. AI can also help alert the radiologist to other potentially pertinent health information (e.g. prior related DICOM or non-DICOM images, medication lists, prior imaging reports and lab results, etc.) and identify key indicators that could help produce a faster, more accurate diagnosis.

(Re)invent the radiology paradigm

However, AI isn’t a plug-and-play solution. For the technology to deliver any value, access to vast quantities of patient data and images is needed to feed the AI learning algorithms. Providing unified access to this type of data is no easy task, given the fragmented nature of many imaging environments.

That’s where enterprise imaging comes in. An enterprise imaging platform based on open, vendor neutral technology (e.g. VNA, enterprise viewer, etc.) can provide a basis for centralizing and consolidating images from all corners of the healthcare enterprise (e.g. radiology, cardiology, gastroenterology, surgery, ophthalmology, dermatology, etc.). This not only provides a foundation for interoperability, but also a shared pool of imaging data that AI applications can tap into.

Further value can be derived from AI if non-imaging data (e.g. clinical documents, reports, admission info, etc.) are also added to this collective data pool. Other healthcare content services, such as enterprise content management (ECM) can be added to the equation and integrated with the EHR to provide a complete and universal view of all patient content — both structured and unstructured.

The potential for AI in imaging is real. However, AI will deliver no value without an enterprise strategy for aggregating and managing all patient content. Therefore, any effort to reinvent the radiology paradigm needs to start with content first.

Ken Congdon

Ken Congdon

Ken Congdon is a content marketing manager at Hyland. His mission is to develop engaging content that educates healthcare providers and payers about potential solutions to their most pressing content management challenges. By helping healthcare organizations identify and address information management weaknesses, waste can be minimized, workflow streamlined and overall patient care improved. Ken joined Hyland after a two-year stint as content marketing manager at Lexmark Healthcare. Prior to that, Ken spent 12 years as a healthcare technology journalist, most notably as Editor In Chief of Health IT Outcomes. Ken received his bachelor’s degree in journalism from Duquesne University.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may also like...