Development and Surface Indentation Assessment of an MRI-Based Subject-Specific Hand-Forearm Finite Element Model: A
Songhao Chen1, Lei Jia2, Shuang Ren3,4,5
1School of Mechanics and Engineering Science, Peking University, Beijing 100871, China.
Background:
Excessive or poorly distributed contact loading of the hand and forearm can cause pain, functional limitation, skin breakdown, and deep soft-tissue injury, motivating reliable quantification of local soft-tissue deformation and stress. Although anatomically detailed finite element models of the hand and forearm have been developed separately, their integration within a continuous subject-specific hand-forearm model remains limited.
Methods:
This preliminary single-subject study developed an MRI-based subject-specific finite element model of the hand and forearm from one healthy adult male volunteer. The model incorporated the complete anatomical structures of the hand and forearm, together with the humerus, which provided proximal attachment sites for the forearm muscles and the extrinsic muscles of the hand. Selected muscle-tendon paths were refined using published hand-wrist musculoskeletal parameters. Skin-, fat-, and muscle-thickness fields were quantified using an anatomy-informed thickness-estimation framework centered on a modified bidirectional ray-projection (mBRP) procedure. Representative indentation sites were selected by thickness fields. A material combination selected from the P1 and P2 analyses was evaluated unchanged at the held-out P3 site, followed by exploratory post hoc screening at P3; agreement was assessed using root-mean-square error (RMSE), normalized root-mean-square error (NRMSE), coefficient of determination (R2), and relative absolute peak-force error.
Results:
The model comprised 179 components, including bones, muscles, ligaments, tendons, cartilage, skin, and subcutaneous fat, with 120,563 tetrahedral elements and 56,944 pentahedral wedge elements. All five anthropometric measurements differed by less than 5% from the male P50 reference values. In total, 27 of the 29 comparable muscle-belly lengths fell within published ranges or agreed with available single-specimen values. The mBRP-derived skin-thickness measurements differed from the prescribed values by only 0.00-1.67%. The best-fitting combinations were Skin2-Fat2 at P1 (R2 = 0.972) and Skin3-Fat3-Muscle3 at P2 (R2 = 0.973). The Skin3-Fat2-Muscle3 combination retained good agreement at the held-out P3 site (R2 = 0.877). In comparison, the P2-optimal Skin3-Fat3-Muscle3 combination showed substantially poorer agreement at P3 (R2 = 0.344). Subsequent exploratory screening of all 27 combinations at P3 identified Skin2-Fat2-Muscle3 as the closest site-specific fit (R2 = 0.983).
Conclusions:
Within this preliminary single-subject study, the model showed anatomical consistency and reproduced the passive indentation responses at the three investigated sites. Cross-site transferability was material-combination dependent: the jointly selected P1-P2 combination retained good agreement at P3, whereas the P2-specific optimum did not. This study demonstrates the technical feasibility of an individualized hand-forearm modeling workflow but does not establish universally transferable tissue parameters or population-level generalizability. Evaluation in additional participants is required.
