To define the criteria of posterior selective thoracic fusion in patients with adolescent idiopathic scoliosis. Methods By reviewing the medical records and roentgenograms of 17 patients with adolescent idiopathic scoliosis who un-derwent posterior selective thoracic fusion

the curve type

Cobb angle

apical vertebral rotation and translation

trunk shift

and thoracolumbar kyphosis were measured and analyzed. Results There were 17 King type Ⅱ patients （PUMC type: Ⅱb1 13

Ⅱc3 4）. The coronal Cobb angle of thoracic curve be-fore and after operation were 56.9°and 21.6° respectively

the mean correction rate was 60.1%. The coronal Cobb angle of lumbar curve before and after operation were 34.8° and 12.1° respectively

and the mean spontaneous correction rate was 64.8%. At final follow-up

the coronal Cobb angle of thoracic and lumbar curve were 23.5° and 15.2° respectively

there were no significant changes in the coronal Cobb angle

apical vertebral translation and rotation compared with that after operation. One patient had 12° of thoracolumbar kyphosis after operation

no progression was noted at final follow-up. There was no trunk decompensation or deterioration of the lumbar curve. In this group

3.9 levels were saved compared with fusing both the th-oracic and lumbar curves. Conclusion Posterior selective thoracic fusion can be safely and effectively performed in King type Ⅱ patients with a mo-derate and flexible lumbar curve

which can save more mobile segments and at the same time can maintain a good coronal and sagittal balance.