A machine learning algorithm can determine appropriate follow-up imaging based off of radiology reports, according to a new study published in the Journal of Digital Imaging. The technology may eventually be developed to automatically tell if a patient completed their appropriate follow-up exam.
“With the exception of specific screening programs (e.g., mammography), radiology departments often do not have a means to automatically track which patients have been recommended for follow-up imaging and, more significantly, may not know if patients have scheduled and completed the appropriate follow-up imaging study in a timely manner,” wrote first author Sandeep Dalal, and colleagues. “Even when the patient has completed the follow-up imaging study, this information is not typically explicitly recorded in any of the hospital systems, such as the electronic health record or radiology information system.”
Dalal, with Philips Research North America’s Clinical Informatics Solutions and Services, and colleagues tested a previously designed recommendation detection algorithm on 559 follow-up imaging recommendations from the “findings” or “impressions” sections of reports taken from a multi-hospital healthcare system. Three radiologists identified appropriate follow-up exams from those reports in order to establish a ground-truth.
The researchers then trained extremely randomized trees on “recommendation attributes,” study meta-data and text similarity of reports to produce the most likely follow-up exam for a radiologists’ recommendation.
Overall, the algorithm achieved a “reasonable accuracy” for automatically producing clinically appropriate follow-up studies. The radiologists’ inter-annotator F-score ranged from 0.853 to 0.868, compared to the 0.807 F-score achieved by the algorithm.
“The accuracy of the algorithm suggests that automated methods can be integrated into a follow-up management application to improve adherence to follow-up imaging recommendations,” the researchers wrote.
The authors did cite a number of limitations, including a reliance on longitudinal radiology reports which may prove difficult in tertiary care hospitals utilizing various referral methods. Despite the drawbacks, Dalal and colleagues believe their technique could improve patient care.
“The follow-up matching algorithm provides a means for radiologists and radiology department administrators to determine if patients have completed the clinically appropriate follow-up imaging study,” the authors wrote. “A system to proactively send reminders to referrers, primary care providers, and patients can ensure that patients receive the timely care they need.”
Four of the authors employed at Philips disclosed that this project, in conjunction with the University of Washington’s Department of Radiology, is part of an “industry-supported” master research agreement, but does not use an existing Philips product.
Matt joined Chicago’s TriMed team in 2018 covering all areas of health imaging after two years reporting on the hospital field. He holds a bachelor’s in English from UIC, and enjoys a good cup of coffee and an interesting documentary.