AIM: Previous experiments using our in vivo tissue engineering chamber (TEC) model demonstrated that adipose flap was spontaneously generated without the need for adipocyte or stem cell implantation. The purposes of the present study are to clarify 1) the reproducibility of this method to create adipose flaps, 2) the time-course of adipogenesis, and 3) the long-term stability of the adipose flap generated. METHODS: The chambers that afforded a protected space for tissue growth were implanted into the groins of rabbits. A vascular pedicle as the vascular source of newly formed tissue, a collagen sponge as a scaffold, and platelet-rich plasma (PRP) and fibroblast growth factor (bFGF) as growth factors were contained within the chamber. There were three experimental groups according to the implantation period of the chamber; Group 4 w, Group 8 w, and Group 12 w (n = 5 in each group). RESULTS: The percent volumes of the combined adipose/pedicle tissue compared with the total volume of the generated tissue were 14.8% (0.437 cm3/2.96 cm3), 47% (0.87 cm3/1.85 cm3) and 80% (1.82 cm3/2.27 cm3) in Groups 4 w, 8 w, and 12 w, respectively. When a 12-week adipose flap was transferred outside the chamber on its vascular pedicle and retained for a further five months, it became more like mature adipose tissue and had increased fat density. CONCLUSION: Adipose flaps were spontaneously generated in vivo in TECs at 12 weeks with reproducibility and showed long-term stability outside the chamber following pedicle transfer. The tissue-engineered adipose flap will contribute to breast reconstruction and augmentation without donor-site morbidity.