OBJECTIVE We sought to test the feasibility and technical ease of a newly designed nitinol-based modified esophageal stent and its effects on preventing postcaustic stricture in mongrel dogs and to try to explain the result at the molecular level. METHODS Twenty-four dogs were included in this controlled study. Stenosis index (wall thickness/intraluminal diameter), pathologic features, hydroxyproline quantities, esophageal compliance, and biomechanics were compared between the injured but unstented and stented dogs. Transforming growth factor beta1, Sma/Mad (Smad)3, and Smad7 mRNA expression and protein levels in esophageal tissue were detected by means of reverse transcriptase-polymerase chain reaction and Western blotting, respectively. RESULTS The modified esophageal stent was able to be placed and retrieved successfully and conveniently and was not only intact but there was also no macroscopic esophageal mucosal injury after the stent removal 4 months later. In comparison with the injured but unstented group, esophageal compliance, biomechanics, and the stenosis index were significantly better in the stented group. Histopathologic study revealed that collagen bundles were thinner and its orientation tended toward a regular and parallel pattern. Transforming growth factor beta1 and Smad3 mRNA expression and protein levels increased and Smad7 mRNA expression and protein levels decreased significantly in esophageal tissue in the stented group. These variables showed no statistically significant difference 2 months after stent removal. CONCLUSIONS The modified esophageal stent might be a promising stent in preventing stricture formation after corrosive esophageal burns clinically.