Removing Skin From MRI Images May Help Clinicians Spot Breast Cancer
An algorithm that segments the skin makes it easier to see the tumors underneath.
The early diagnosis of breast cancer is essential for successful treatment. Imaging techniques such as mammography and ultrasound are commonly used to detect breast tumors, but MRI has become increasingly important in daily practice due to its higher sensitivity. In a study published April 18 in Scientific Reports, researchers developed an algorithm that might further improve the ability of clinicians to identify breast tumors based on MRI images.
The segmentation algorithm automatically removed the skin from MRI images, making the underlying tumors easier to see than in the original images. Skin thickness analysis, which improves the retention of small vessels and mammary glands and increases the accuracy of skin segmentation, plays a critical role in the success of the algorithm. Overall, the algorithm takes less than two minutes to segment the skin in each case, compared with an average of five hours for manual segmentation by physicians and radiologists.
According to the authors, this is the first reliable skin segmentation algorithm for dynamic contrast-enhanced MRI. In the future, their proposed algorithm may provide clinicians with quick and accurate anatomical information for diagnosing and monitoring breast cancer.