Biomechanical interpretations of bone adaptation in biological reconstructions following bone tumors would be crucial for orthopedic oncologists, particularly if based on quantitative observations. This would help to plan for surgical treatments, rehabilitative programs and communication with the patients. In particular, outcomes of the Capanna technique, which combines bone allograft and vascularized fibula autograft, lead to stable and durable reconstructions [1, 2], and different remodeling patterns have been described [3] as a response to mechanical loading. However, there are several events that are not understood and require a biomechanical interpretation, as the evolution patterns can evolve towards conditions that threaten the strength of the reconstruction. We aimed to (i) analyze the biomechanical adaptation of a femoral reconstruction after Ewing sarcoma, in terms of morphological and densitometric evolution of bone from CT data, internal loads acting on the bone during movement, mechanical competence of the reconstruction, and (ii) relate in-progress bone resorption to the mechanical stimulus induced by different motor activities. Eight CT datasets of a patient (8 yrs at surgery using the Capanna technique) during 76-month follow-up were available. The evolution of bone morphology, density and moments of inertia was quantified. At the last control, the patient underwent gait analysis (walking, chair rise/sit, stair ascent/descent, squat). We created a multiscale musculoskeletal and finite element model from CT scans and motion analysis data at the end of follow-up, using state-of-the-art modeling workflows [4, 5], to analyze muscle and joint loads, and to compare the mechanical competence of the reconstructed bone with the contralateral limb, in the current real condition and in a possible revision surgery that removed proximal screws. Although there were no reconstruction complications and osteo-fusion with intense remodeling between allograft and autograft was shown, there was a progressive decrease in allograft cortical thickness and density. There were strategies of muscle coordination that led to differences in joint loads between limbs more marked in more demanding motor activities, and generally larger in the contralateral limb. The operated femur presented a markedly low ratio of physiological strain due to load-sharing with the metal implant, particularly in the lateral aspect. A possible revision surgery removing the three most proximal screws would help restore a physiological strain configuration, while the safety of the reconstruction would not be threatened. We suggest that bone resorption is related to load-sharing and to the internal forces exerted during movement, and the mechanical stimulus should be improved by adopting modifications in the surgical treatment and by promoting physical therapy aimed at specific muscle strengthening