Fully automated nipple detection in digital breast tomosynthesis

Area of Science:

• Medical imaging
• Radiology
• Computer-aided diagnosis

Background:

• Digital breast tomosynthesis (DBT) offers 3D imaging to improve mammography for dense breasts.
• Nipple location is a key landmark in DBT for breast alignment and lesion localization.
• Accurate nipple detection is essential for robust DBT image analysis.

Purpose of the Study:

• To develop an automated nipple detection algorithm for digital breast tomosynthesis (DBT) images.
• To address the challenge of detecting both visible and invisible nipples.
• To enhance the utility of DBT in clinical practice through improved image registration and lesion analysis.

Main Methods:

• Proposed a novel algorithm for nipple detection in DBT images.
• Incorporated analysis of fibroglandular tissue and breast area changes to detect invisible nipples.
• Validated the method on 138 DBT images.

Main Results:

• The algorithm achieved a mean Euclidean distance of 3.10±2.58mm for nipple detection.
• Demonstrated high accuracy in locating nipples, essential for image registration.
• The method effectively handles both visible and invisible nipple cases.

Conclusions:

• Automatic nipple detection in DBT images is crucial for image registration and lesion classification.
• The proposed algorithm provides an accurate and efficient solution for nipple localization in DBT.
• This method supports computer-aided detection, improving the efficiency of DBT image analysis.