News | Oncology Diagnostics | November 18, 2019

Study: AI Found to Reduce Bias in Radiology Reports

The Centre for Advanced Research in Imaging, Neuroscience and Genomics study found that AI powered CAD software can reduce hedging and defensive reporting statements in radiologist reports

 Oxipit ChestEye AI

November 18, 2019 — A study by the Centre for Advanced Research in Imaging, Neuroscience and Genomics (CARING) found that AI powered CAD software can reduce hedging and defensive reporting statements in radiologist reports, resulting in clearer and more actionable diagnosis descriptions. The study was undertaken using Oxipit ChestEye AI imaging suite. Moreover, the study found that nearly 80 percent of the reports generated by Oxipit ChestEye were as accurate as the radiologists' reports and were more accurate in 5 percent of analyzed patient cases.

The study led by Dr. Vasanth Venugopal aimed to identify elements of hedging in clinical radiology reports and evaluate the accuracy and clinical validity of AI-generated reports of chest X-rays. A set of 297 retrospective chest X-ray images was used in the study. The images were analyzed by Oxipit ChestEye and AI generated reports compared to the ones produced by a radiologist. Final report comparison and diagnosis validation was conducted by a radiologist with 9 years of clinical experience. The full study titled “Judging the Accuracy and Clinical Validity of Deep Learning-Generated Test Reports” will be presented at the RSNA conference in Chicago this year.

“Subjectivity is part of human nature. Sometimes radiologists might be fairly confident that there is no pathology in the analyzed case, but they choose to include inconclusive vague diagnosis statements to hedge against potential mistakes or legal threats. On the contrary, AI produces clear, unbiased reports that can help to reassure radiologist in their opinion. A combination of both produces clearer, actionable reports to make treatment decisions, as well as helps to avoid unnecessary procedures resulting from evasive and non-committal reports,” notes Oxipit’ Chief Medical Officer Dr. Naglis Ramanauskas.

Oxipit ChestEye imaging suite encompasses a fully automatic computer aided diagnosis (CAD) platform which supports 75 radiological findings, covering over 90 percent of radiological cases presented to radiologists on a daily basis. ChestEye produces a standardized preliminary text report that incorporates all the radiologically relevant information present in a chest X-ray image, speeding up case description and minimizing the risk of overlooked secondary findings. In February Oxipit ChestEye received CE certification. Internal platform trials showed 30 percent time saved per patient case and reduced error rate by up to 50 percent.

“Oxipit ChestEye was designed as a productivity and second-opinion tool to aid the work of radiologists. The evaluation that nearly 80 percent of preliminary reports generated by our software were clinically accurate to be deemed as final reports, and in 5 percent cases - even more accurate than the ones produced by a radiologist, is an inspiring validation of current capabilities of AI diagnostics,” Ramanauskas said.

In the words of Ramanauskas, even two radiologists can have differing opinions over a particular X-ray report.

“Therefore aiming for 100 percent correlation to radiologist reports is unrealistic. However, with constant improvement of deep learning algorithms and innovations in the medical imaging field, a radiologist working with the help of AI can help to reduce interpersonal subjectivity and greatly increase the quality of the reports while helping the radiologists become more efficient," Ramanauskas said.

For more information: www.caring-research.com 

Related Content

Image courtesy of GE Healthcare

Feature | Mobile C-Arms | July 08, 2020 | By Melinda Taschetta-Millane
Moblie C-arms have seen several advances over the past de
At the American Association of Physicists in Medicine (AAPM) 2019 meeting, new artificial intelligence (AI) software to assist with radiotherapy treatment planning systems was highlighted. The goal of the AI-based systems is to save staff time, while still allowing clinicians to do the final patient review. 
Feature | Treatment Planning | July 08, 2020 | By Melinda Taschetta-Millane
At the American Association of Physicists in Medicine (AAPM) 201
 Many patients with severe coronavirus disease 2019 (COVID-19) remain unresponsive after surviving critical illness. Investigators led by a team at Massachusetts General Hospital (MGH) now describe a patient with severe COVID-19 who, despite prolonged unresponsiveness and structural brain abnormalities, demonstrated functionally intact brain connections and weeks later he recovered the ability to follow commands

Getty Images

News | Coronavirus (COVID-19) | July 08, 2020
July 8, 2020 — Many patients with severe coronavirus disease 2019 (...
Several drivers will contribute to the growth of the teleradiology market in terms of penetration, revenue and read volumes over the next five years

Getty Images

Feature | Teleradiology | July 08, 2020 | By Arun Gill
Last year was a record year for the global...
A 3-D ultrasound system provides an effective, noninvasive way to estimate blood flow that retains its accuracy across different equipment, operators and facilities, according to a study published in the journal Radiology.

Volume flow as a function of color flow gain (at a single testing site). For each row the color flow c-plane and the computed volume flow are shown as a function of color flow gain. The c-plane is shown for four representative gain levels, whereas the computed volume flow is shown for 12–17 steps across the available gain settings. Flow was computed with (solid circles on the graphs) and without (hollow circles on the graphs) partial volume correction. Partial volume correction accounts for pixels that are only partially inside the lumen. Therefore, high gain (ie, blooming) does not result in overestimation of flow. Systems 1 and 2 converge to true flow after the lumen is filled with color pixel. System 3 is nearly constant regarding gain and underestimates the flow by approximately 17%. Shown are mean flow estimated from 20 volumes, and the error bars show standard deviation. Image courtesy of the journal Radiology

News | Ultrasound Imaging | July 01, 2020
July 1, 2020 — A 3-D ultrasound
R2* maps of healthy control participants and participants with Alzheimer disease. R2* maps are windowed between 10 and 50 sec21. Differences in iron concentration in basal ganglia are too small to allow visual separation between patients with Alzheimer disease and control participants, and iron levels strongly depend on anatomic structure and subject age. Image courtesy of Radiological Society of North America

R2* maps of healthy control participants and participants with Alzheimer disease. R2* maps are windowed between 10 and 50 sec21. Differences in iron concentration in basal ganglia are too small to allow visual separation between patients with Alzheimer disease and control participants, and iron levels strongly depend on anatomic structure and subject age. Image courtesy of Radiological Society of North America

News | Magnetic Resonance Imaging (MRI) | July 01, 2020
July 1, 2020 — Researchers using magnetic...