MIT’s AI Accurately Identifies Breast Cancer Like A Radiologist Would

Artificial intelligence has shown to be of great use in medical applications and Massachusetts Institute of Technology is giving us yet another example of that. It has developed an automated model with the Massachusetts General Hospital that can identify breast cancer as accurately as a radiologist would.

Studies suggest that one in eight women in the United States are affected by this diseases. One of the reasons for the development of this disease is dense breast tissue which around 40 percent of women in the U.S. are said to have. The tissue not only makes mammogram screen difficult but it also increases the risk of breast cancer in these women.

The AI model developed by MIT and MGH can identify dense breast tissue in mammograms as accurately as a trained radiologist. The model is trained on thousands of digital mammograms and can thus identify the different types of breast tissue. Researchers have found a 90 percent correlation to the diagnosis made by expert radiologists.

Adam Yala, one of the authors on the model’s paper, says that the AI takes less than a second per image and it can be scaled cheaply throughout hospitals. They are now working on finding out how this model can be used in other healthcare applications.

This isn’t the only AI model developed for detecting breast cancer. Google has recently claimed that its AI has a 99 percent accuracy in detecting metastatic breast cancer.

Adnan Farooqui
Categories: Medical
Tags: artificial intelligence