IBM Examines The Application Of AI For Breast Cancer Screenings

As reported by ZDNet, a recent study conducted by IBM Research in collaboration with Sage Bionetworks, Kaiser Permanente Washington Health Research Institute, as well as the University of Washington School of Medicine, has unveiled how merging machine learning algorithms and evaluations by radiologists could enhance the general precision of breast cancer screenings.

 

As the institute’s researcher Stefan Harrer states, the mammogram screenings which are normally employed by radiologists to facilitate premature detection of breast cancer relies highly upon the expertise of the radiologist in being able to visually identify the cancer signs which may not always be accurate. This, he states is where the need arises of aiming to enhance the approach through the fusion of Artificial Intelligence for diminishing the rate of a false cancer diagnosis as well as to diminish the rate of overlooking cancer detection. 

 

The research incorporated over 310,800 anonymous mammograms as well as clinical data from Kaiser Permanente Washington (KPWA) and Sweden’s  Karolinska Institute (KI). Out of the merged datasets around 166,500 examinations were facilitated by Karolinska Institute from 6,800 women, where 780 were cancer positive; while the leftover 144,200 examinations were facilitated by KPWA from 85,500 women, where 941 were cancer positive.

 

Harrer also added that what made the study momentous was its size. He claimed that the study empowered the team to design a guarded ecosystem, providing researchers admission to datasets which were initially not available for research activities.