A team of researchers in the department of Energy’s Oak Ridge National Laboratory has been taking efforts to reduce the number of errors in the analyses of the diagnostic images by the health professionals. These efforts are being taken so as to improve the understanding of the cognitive processes that are involved in the image interpretation.
The research work that has been published in the Medical Imaging Journal has the potential to enhance the health outcomes for around hundreds of thousands of American women that are affected by breast cancer every year. Breast cancer is considered as the second leading death cause in women and the early detection of the condition is quite critical for effective treatment.
Detecting the presence of disease early is most vital to the situation, the ORNL-led team, which included Hong-Jun Yoon, Gina Tourassi and Folami Alamudun, as well as Paige Paulus of the Aerospace, Mechanical, and Biomedical Engineering Department of the University of Tennessee, found that analyses of the mammograms by radiologists were often affected by contextual bias, or diagnostic history. As such, an experiment which targeted recording the eye movements of the radiologists at different skill levels to help infer the contextual bias involved in every individual inferences of the mammograms and images.
The team of researchers further calculated the deviation in the research context of several image categories, including the images that show the cancer and also those that will be easier or more difficult to further decipher. Furthermore, the researchers were able to learn how the eye movements of the researcher participants that differed from one mammogram to other mammogram.