12932
Blog
Tia Albright

How Healthcare Leaders are Using AI to Address Radiology Backlogs

As the population ages and medical technology advances, there’s a greater need and demand for imaging than ever before. The challenge: This rise in demand comes at a time of radiology workforce shortages.

A study published in Clinical Imaging found that nearly half of responding facilities experienced radiology vacancies and 68% had unreported imaging exams. By 2034, these shortages are expected to double.

At the physician level, radiologists are paying the price in burnout, which will only continue to impact already existing shortages. Becker’s Hospital Review reported nearly half of U.S. radiologists are at retirement age and an expected shortage of 42,000 radiologists by 2033. 

These backlogs are also detrimental to patients, leading to longer wait times for results that could adversely affect their level of care.

AI is a powerful solution to help alleviate radiology backlogs. By rapidly triaging critical cases and bumping them to the top of the worklist, read times can be significantly reduced, sometimes as much as 90%.  

Here’s what radiology leaders have to say about the ability of AI to address the backlog challenge.

Embrace an enterprise-wide AI platform

With the United States shifting more to an integrated delivery network model, the demand on radiologists has increased. 

Dushyant Sahani, MD, Chair of the Department of Radiology at UW Medicine, explained that with the imaging demands from various facilities, an enterprise-specific solution can help standardize health system operations.

When speaking of a previous cross-sectional backlog during a staffing shortage, Dr. Sahani emphasized that “what we suffer from due to backlogs is lowered morale across our faculty, and the potential of missing important findings — even unexpected findings. Case in point, incidental pulmonary embolisms which can significantly impact patient outcomes. Brain hemorrhages sometimes come in as an outpatient exam, and therefore might be at risk for delayed care.”

To address their backlog, UW Medicine implemented Aidoc to help prioritize urgent exams, particularly in flagging suspected critical findings, like pulmonary embolisms and intracranial hemorrhages. 

Improve efficiency and reduce mental fatigue

John Borsa, MD, Chair of Radiology at St. Luke’s Health System experienced backlogs and staff shortages at his hospital, and his main goal was to find “any solution that could help what limited resources they have [left] be more efficient and get through…their day’s work with less mental fatigue.” 

This is when they brought in Aidoc, to improve efficiency, relieve administrative burden and ultimately help reduce provider fatigue while driving better outcomes.

“With our shortage of radiologists, it’s been a game changer as far as triaging patients,” said Dr. Borsa.

Leverage AI for faster care coordination

For AI to work, Alexander McKinney, MD, Chair of Radiology at University of Miami Health System, emphasized “it has to help us work smarter, faster…and be patient-centric [in order to] get patients to where they need to be.” 

Why? Because according to Dr. McKinney nearly 50% of radiologists are already burned out.

An AI tool can help alleviate this pain point by quickly prioritizing potentially life-threatening findings, and empowering care teams with real-time alerts and communication tools to ensure patients get the timely treatment they need. 

This results in a better experience all-around. “[Patients are] happy, the referring providers are happy. We’re happier that we’re not letting people walk out the door with a serious condition,” said Dr. McKinney.

Prioritize ordering the right exams that can be analyzed by AI

When it comes to backlogs, one factor is the ordering patterns of clinicians. Sriram Mannava, MD, President of Columbus Radiology, noted two operational items, “One is excessive ordering in the Emergency Department for acute indications — mainly head bleeds, trauma and [then the diagnosis and workup of chest pain for pulmonary embolisms]…trauma and chest pain are the two most common indications for [patients that present] to the ER.” 

Dr. Mannava shared that the advantage of the right AI platform, like Aidoc, is that it can alleviate these operational challenges. 

“Aidoc’s [ICH and C-spine] algorithms essentially cover the first part of a trauma workup… for any trauma or [fallen patients],” said Dr. Mannava. “Then the chest pain workup, which is essentially ruling out PE, is covered by [PE and RibFx] algorithms…and so a large part of our clinical workflow is covered by [these]; it’s basically [pre-screening] those patients [with] AI.”

How Aidoc helps address the radiology backlog

Traditional AI in medical imaging is limited by rigid protocol constraints, analyzing only what it’s explicitly programmed to find based on the ordered study. These artificial “blinders,” combined with excessive backlogs, increase the risk of critical findings being missed or overlooked – delaying patient care. 

Aidoc’s aiOS™ intelligent orchestration overcomes protocol-based limitations by leveraging both text- and image-based AI at the DICOM and pixel level. It identifies the best image slice, analyzes all present anatomy—including partial views—and runs all clinically relevant algorithms simultaneously. This enables awareness of incidental and suspected critical findings, even those outside the original scan protocol.

By presenting suspected cases with urgent findings within the worklist, the aiOS™ ensures that the right findings surface at the right time, helping to improve patient outcomes without disrupting workflows or increasing clinician workload.

Set up a demo to experience Aidoc’s aiOS™ in action.

Editors’ note: Some quotes have been edited for length and clarity.

Explore the Latest AI Insights, Trends and Research

Tia Albright