Development of a breast reconstruction-specific computational model to predict shoulder function in breast cancer survivors.

Abstract

Purpose: Mastectomy and reconstruction are treatment components for many breast cancer patients, resulting in long-term shoulder dysfunction. Computational models specific to surgical approach would enable study of underlying mechanisms of shoulder dysfunction, but none exist. Our objectives were as follows: (1) develop and validate models representing lumpectomy, implant-based, and autologous flap-based reconstruction; and (2) determine how muscle contribution to hand acceleration during functional movements differs across models.

Methods: The upper limb model in OpenSim was scaled to force-generating properties and anthropometry of adult females. A 405-cc wrapping surface was placed beneath the pectoralis major muscle path representing subpectoral implant placement. For model validation, shoulder moment was predicted in five postures, with an external load applied equal to mean strength measured from a breast cancer patient cohort. Induced acceleration analysis was used to identify primary muscle contributors to hand acceleration during functional movements.

Results: Following model development, pectoralis major moment arm was reduced in the implant model compared to lumpectomy and flap models. Predicted shoulder moments fell within 1 standard deviation of experimental moments (i.e., external rotation: lumpectomy model, 15.1Nm; implant model, 14.1Nm; flap model, 17.5Nm; experimentally measured, 14.1Nm ± 5.4Nm; 13.0Nm ± 3.6Nm; 15.5Nm ± 5.3Nm, respectively), except horizontal abduction (all groups) and elevation (lumpectomy group), providing validation. Large shoulder muscles, including deltoid, infraspinatus, and subscapularis, were the largest contributors to hand acceleration. Pectoralis major was also identified, possibly relating to post-surgical functional deficits.

Conclusion: This work identified muscle moment arm changes for implant-based reconstruction. These models can be used to predict functional outcomes of differing reconstruction surgeries.