Breast cancer happens to be the most commonly diagnosed cancer among women across the globe. Earlier detection is one of the first steps to prevention as successful intervention in early stages could be the difference between life and death. Researchers have been building prediction models based on artificial intelligence and IBM says that its new AI models can predict breast cancer with “radiologist-level accuracy.”
A team of IBM researchers has published a research around a new artificial intelligence model which can predict the development of malignant breast cancer in patients within the year. The AI is capable of making these predictions at rates that are comparable to human radiologists.
It’s said to be the first algorithm of its kind to make decisions by using both the imaging data as well as a comprehensive patient’s health history. So it can limit the chance of a bad diagnosis by making connections between what it sees in the imagery and the patient’s overall health history which would contain information about things such as thyroid function and iron deficiencies.
With this data, the model was able to correctly predict the development of breast cancer in 87 percent of the cases that it analyzed. It was also able to correctly predict 77 percent of non-cancerous cases. Radiologists could one day use AI models like these to confirm or deny positive breast cancer cases.
The model was tested against 71 different cases which radiologists had initially deemed to be “non-malignant,” but the patients were ultimately diagnosed with breast cancer within the year. The AI model was able to correctly identify breast cancer in 48 percent of the 71 cases.