A computational framework for patient-specific breast modelling with anatomically informed Cooper's ligament networks

Cooper's ligaments tether the dermis to the pectoral fascia and play a key role in breast suspension, but they are still poorly represented in finite element (FE) models used for surgical planning. This work proposes a patient-specific computational framework that explicitly incorporates anatomically informed Cooper's ligament networks into three-dimensional breast FE models.

The pipeline combines automatic magnetic resonance imaging (MRI)-based tissue segmentation with deep learning, cadaveric anatomical observations, and multi-component FE simulations in Abaqus. A synthetic three-dimensional ligament network is derived from published anatomical descriptions and cadaveric dissection, and modelled as beam elements embedded within volumetric soft tissue domains (skin, adipose, and glandular tissue) equipped with Neo-Hookean hyperelastic constitutive laws. Three ligament configurations (absent, simplified radial, and anatomically distributed) were compared under gravity while varying ligament stiffness between 0.1 and 10 MPa.

The anatomically distributed network reduced the maximum displacement from 30.88 mm to 19.05 mm, shifted internal stress from the skin envelope to the ligamentous scaffold, and showed a nonlinear stiffness-displacement response with limited benefit beyond approximately 2 MPa.

Ligament architecture has a major influence on predicted breast biomechanics and should not be reduced to a coarse radial pattern. The proposed framework is compatible with standard clinical MRI and FE tools and provides a basis for extending patient-specific breast models towards more reliable preoperative planning for reconstruction.