Background contextIn adolescent idiopathic scoliosis (AIS), identification of curve patterns and determination of fusion levels are most important for correcting deformity, improving trunk balance, and saving motion segments. Lenke 1A scoliosis is the most common type of AIS, accounting for around 16% of all AIS; however, how to determine the lower instrumented vertebrae in this common curve type is unclear. PurposeThe aim was to classify Lenke 1A curves according to lumbar axial plane analyses to determine optimal distal fusion level selection. Study DesignThis was a retrospective study. Patient SampleThis study included 69 consecutive patients with AIS (13 males and 56 females) of Lenke Type 1A curve who underwent posterior correction and fusion of the thoracic spine between 2001 and 2013 in a single center. Outcome MeasuresCoronal, sagittal, and axial parameters were measured from plain radiographs that were obtained at initial medical examination of the patients. MethodsCoronal and sagittal plane and whole spine segmental vertebra rotations from thoracic 1 to lumbar 5 were evaluated by using Drerup method. As a result of analysis of axial plane, all patients with Lenke 1A curves were divided into three groups depending on lumbar vertebral rotation. In Group I, the rotation of lumbar vertebral rotation was accepted as neutral. In Group II, the direction of lumbar vertebral rotation was same with other vertebrae in the main curve. In Group III, the rotation of lumbar vertebral rotation had opposite direction with vertebrae in the main curve. ResultsIn Group I curves, the mean position of lower end vertebrae (LEVs) was more frequently at T11, neutral vertebra (NV)-last touched vertebra (LTV) at T12, and stable vertebra (SV) at L2. In Group II curves, the mean position of LEV was more frequent at L1, LTV at L2, NV at L3, and SV at L4, whereas in Group III curves, the median position of LEV-NV-LTV was frequently at T12 and that of SV at L1. Then, Group I to III curves between Group II curves showed the gap differences of NV-LEV, SV-LEV to be significantly different. Similarly, the SV was not more than two segments distal from LEV in Group I and Group III but more than two to three segments from LEV in Group II. ConclusionsOur analysis suggested that not all Lenke 1A curves yield satisfactory outcome with the same fusion extend although a high percentage of the patients with Lenke 1A curves have shown satisfactory outcome with NV fusion. Thus, it seems that some Lenke 1A curves may require fusion to SV.