We use cookies to understand how you use our site and to improve your experience. This includes personalizing content and advertising. To learn more, click here. By continuing to use our site, you accept our use of cookies. Cookie Policy.

Features Partner Sites Information LinkXpress hp
Sign In
Advertise with Us
IBA-Radcal

Download Mobile App




AI Algorithms Enable Highly Accurate and Cost Effective Medical Image Diagnosis

By HospiMedica International staff writers
Posted on 13 Apr 2022

Medical imaging is an important part of modern healthcare, enhancing precision, reliability and development of treatment for various diseases. More...

Artificial intelligence (AI) has also been widely used to further enhance the process. However, conventional medical image diagnosis employing AI algorithms require large amounts of annotations as supervision signals for model training. To acquire accurate labels for the AI algorithms – radiologists, as part of the clinical routine, prepare radiology reports for each of their patients, followed by annotation staff extracting and confirming structured labels from those reports using human-defined rules and existing natural language processing (NLP) tools. The ultimate accuracy of extracted labels hinges on the quality of human work and various NLP tools. The method comes at a heavy price, being both labor intensive and time consuming.

Now, an engineering team at the University of Hong Kong (HKU, Hong Kong) has developed a new approach “REFERS” (Reviewing Free-text Reports for Supervision), which can cut human cost down by 90%, by enabling the automatic acquisition of supervision signals from hundreds of thousands of radiology reports at the same time. It attains a high accuracy in predictions, surpassing its counterpart of conventional medical image diagnosis employing AI algorithms. The innovative approach marks a solid step towards realizing generalized medical artificial intelligence.

For training REFERS, the research team used a public database with 370,000 X-ray images, and associated radiology reports, on 14 common chest diseases, including atelectasis, cardiomegaly, pleural effusion, pneumonia and pneumothorax. The researchers managed to build a radiograph recognition model using only 100 radiographs that attained 83% accuracy in predictions. When the number was increased to 1,000, their model exhibited an amazing performance with an accuracy of 88.2%, which surpassed its counterpart trained with 10,000 radiologist annotations (accuracy of 87.6%). When 10,000 radiographs were used, the accuracy stood at 90.1%. In general, an accuracy level of above 85% in predictions is useful in real-world clinical applications.

REFERS achieves the goal by accomplishing two report-related tasks, i.e., report generation and radiograph–report matching. In the first task, REFERS translates radiographs into text reports by first encoding radiographs into an intermediate representation, which is then used to predict text reports via a decoder network. A cost function is defined to measure the similarity between predicted and real report texts, based on which gradient-based optimization is employed to train the neural network and update its weights. As for the second task, REFERS first encodes both radiographs and free-text reports into the same semantic space, where representations of each report and its associated radiographs are aligned via contrastive learning.

"AI-enabled medical image diagnosis has the potential to support medical specialists in reducing their workload and improving the diagnostic efficiency and accuracy, including but not limited to reducing the diagnosis time and detecting subtle disease patterns,” said Professor Yu Yizhou, leader of the team from HKU’s Department of Computer Science under the Faculty of Engineering. “We believe abstract and complex logical reasoning sentences in radiology reports provide sufficient information for learning easily transferable visual features. With appropriate training, REFERS directly learns radiograph representations from free-text reports without the need to involve manpower in labeling.”

“Compared to conventional methods that heavily rely on human annotations, REFERS has the ability to acquire supervision from each word in the radiology reports. We can substantially reduce the amount of data annotation by 90% and the cost to build medical artificial intelligence. It marks a significant step towards realizing generalized medical artificial intelligence,” said Dr. Zhou Hong-Yu, the paper’s first author.

Related Links:
University of Hong Kong


Gold Member
12-Channel ECG
CM1200B
Hand-held Hemoglobin Analyzer
DiaSpect Tm
Pediatric Mask
Respire SOFT
Multi-Chamber Washer-Disinfector
WD 390
Read the full article by registering today, it's FREE! It's Free!
Register now for FREE to HospiMedica.com and get access to news and events that shape the world of Hospital Medicine.
  • Free digital version edition of HospiMedica International sent by email on regular basis
  • Free print version of HospiMedica International magazine (available only outside USA and Canada).
  • Free and unlimited access to back issues of HospiMedica International in digital format
  • Free HospiMedica International Newsletter sent every week containing the latest news
  • Free breaking news sent via email
  • Free access to Events Calendar
  • Free access to LinkXpress new product services
  • REGISTRATION IS FREE AND EASY!
Click here to Register








Channels

Artificial Intelligence

view channel
Image Credit: Shutterstock

Explainable AI Examines Sepsis Mortality Risk and Emergency Care Delays

Preventing sepsis-related deaths in emergency departments (EDs) remains challenging because patient deterioration can result from multiple rapidly evolving and interconnected factors. These include triage... Read more

Patient Care

view channel
Image: A view inside the custom virtual reality environment used in the study: a calm room with a teddy bear, soft colors and large windows looking out over the sea. (Image Credit: Walaa S. Abu Rukbah et al., Virginia Commonwealth University; CC-BY 4.0, https://creativecommons.org/licenses/by/4.0/)

Virtual Reality Program May Reduce Anxiety Before Pediatric Surgery

Presurgery anxiety affects an estimated 50% to 75% of children undergoing surgery and can complicate anesthesia induction, increase postoperative pain, and prolong recovery. Sedation is often used to manage... Read more

Health IT

view channel
Image Credit: Shutterstock

Telemonitoring Program Improves Quality of Life After Heart Failure Discharge

Heart failure is a chronic condition associated with high mortality, frequent hospitalizations, and reduced quality of life. Patients face particularly high risks of readmission and death after discharge,... Read more

Point of Care

view channel
Image: Jessica Ramella-Roman, professor of biomedical engineering at Florida International University, holding the PPRIM (Portable Preterm Imaging) device designed to detect early signs of preterm labor risk. (Photo courtesy of Florida International University)

Handheld Imaging Device May Help Identify Preterm Birth Risk Earlier

Preterm birth, defined as delivery before 37 weeks of pregnancy, can lead to serious neonatal complications, including neurological, developmental, and respiratory problems. However, identifying pregnancies... Read more

Business

view channel
Image: Stryker has acquired ZuriMED, adding the FiberLocker System, a rotator cuff augmentation technology designed to strengthen repairs, to its Sports Medicine portfolio. (Photo courtesy of ZuriMED)

Stryker Completes Acquisition to Broaden Shoulder Surgery Offerings

Stryker (Portage, MI, USA) has completed its acquisition of ZuriMED (Zurich, Switzerland), a privately held developer of the FiberLocker System, expanding its Sports Medicine portfolio with a commercialized... Read more
Copyright © 2000-2026 Globetech Media. All rights reserved.