The goal of surgical treatment of adolescent idiopathic scoliosis (AIS) is to prevent further worsening of the deformity by fusion and to correct the deformity as much as possible on this basis, maintaining the balance of the spine and trunk. However, there is a conflict between fusion of the spine and preservation of spinal function, because too long a fusion may affect the motor segments of the spine and accelerate the degeneration of the residual motor segments; too short a fusion may result in an extended lateral bend (Adding on), an oversized residual bend, or trunk imbalance. However, there are many controversies about the specific criteria of selective fusion. The advantages of selective thoracic bend fusion: Newton et al. pointed out that for each additional fused segment distally from thoracic 10, the spine would lose 5% of flexibility. Danielsson et al. pointed out that fusion of long segments can cause degeneration of the adjacent disc below the fused segment and may accelerate the degeneration of the residual motion segment; therefore, the number of fused segments should be reduced as much as possible to ensure the orthopedic effect. At the same time, some long-term follow-up results show that the extension of fusion to the middle and lower lumbar spine will significantly increase the probability of lower back pain, therefore, selective thoracic fusion is important for the prevention of postoperative lower back pain. 2, the indications for selective thoracic fusion: So far, there is no consensus on the criteria or indications for selective thoracic fusion. From the literature, King type II scoliosis, Lenke type I-IV, PUMC type IIb1 and IIIa, and part of PUMC type IIc3 are all feasible for selective thoracic fusion. In 1983, King et al. first proposed selective thoracic fusion for King II idiopathic scoliosis, and reported 111 cases of King II idiopathic scoliosis with selective thoracic fusion, with only 2 cases requiring further management due to progression of the lumbar curvature. Therefore, it is recommended that selective fusion of the thoracic spine to the thoracic curvature stable vertebrae can be safely performed in this type of scoliosis, provided that the thoracic curvature does not exceed 80°. Thus, the mobile segment of the lumbar segment can be preserved, and the lumbar bend can be automatically corrected to achieve a new trunk balance. However, this concept is based on the analysis of cases treated with the Harrington Brace Bar System, a first-generation orthopedic device. Thus, when applying this concept to third-generation orthopedic endoprosthetic systems, the issue of lumbar bending loss or trunk imbalance became prominent and widely debated. Lenke et al. reported that in 50 patients with elective thoracic fusion using the C-D system with a follow-up of more than 2 years, postoperative loss of compensation occurred in 5 cases, 3 of which required revision and extension of the fusion to L4. Therefore, they concluded that elective thoracic fusion could be safely performed if the ratio of thoracic to lumbar Cobb angles, the ratio of vertex rotations, and the ratio of vertex excursions were greater than 1.2. However, if the lumbar bend is greater than 60° and the rotation exceeds II degrees; or if the parietal offset exceeds 4 cm, then a double bend to L4 should be fused. Richards et al [7] reported a group of King type II cases with selective thoracic fusion using the C-D or TSRH systems, and found that if the preoperative lumbar bend was greater than 40º after selective fusion of the thoracic spine They found that if the preoperative lumbar curve was greater than 40ºafter selective fusion of the thoracic spine, the lumbar curve would continue to be greater than the thoracic curve and the tilt between the lower lumbar spine and pelvis would persist. Therefore, selective thoracic fusion should be performed with caution when the lumbar curve exceeds 40°. Kalen et al. reported that in 46 patients who underwent elective thoracic fusion using the C-D system, no increase in lumbar curvature or trunk imbalance was observed. McCall et al [9] also reported that if the lumbar bend was greater than 45° and the flexibility index [lumbar flexibility (%) – thoracic flexibility (%)] was low (≤20 %), the likelihood of postoperative loss of compensation was significantly increased. Mason et al. found that if the lumbosacral angle was greater than 15°, or if the lumbar bend apical offset was > 2 cm, the likelihood of loss of compensation after elective thoracic fusion would be increased. In the 12 patients reported by Equuschus et al [11], all lumbar bends were less than 45°, the parietal rotation was within II degrees, and the offset was also less than 2 cm, with flexibility greater than 60%; and the ratio of thoracic bend to lumbar bend lateral bend was greater than 1.2 in all but one of the 12 patients, and the ratio of thoracic bend to lumbar bend parietal offset was greater than 1.2, with only one case of postoperative thoracolumbar segmental kyphosis. Patel et al. reported that selective fusion of the thoracic bend can be performed safely and effectively when the lumbar bend is less than 45°, the flexibility is greater than 70%, the parietal spine offset is less than 2 cm, the parietal spine rotation is within Ⅱ degrees, and there is no lordosis of the thoracolumbar segment, in order to preserve as many lumbar mobile segments as possible and improve the function of the lumbar spine. Dobbs et al. reported that in a group of 66 patients fixed with hook or pedicle nailing, all had a mean lumbar bend of less than 50°, a thoracolumbar bend Cobb ratio greater than 1.2, a parietal offset ratio greater than 1, and a trunk imbalance of 27% (trunk offset greater than 2 cm). Suk et al. reported 203 cases of selective thoracic fusion with a mean thoracic bend of 51° and a mean lumbar bend of 30°. At follow-up, trunk imbalance was observed in 5% (10/203) of cases; Adding-on phenomenon was observed in 8% (17/203) of cases, and these 17 cases were due to fusion to the second vertebra proximal to the neutral vertebra. 3. Effect of surgical approach on the outcome of selective thoracic bend fusion: Dobbs et al. compared the outcome of selective thoracic fusion using hook and pedicle nailing in Lenke type C cases and found that the proportion of trunk imbalance was 11.8% (4/34) with pedicle nailing compared to 40.6% (13/32) with hook fixation and that the number of lumbar residual bends was greater in the hook group than in the nail group. In a comparative study, it was noted that internal fixation screws were superior to internal fixation hooks in the self-correction of thoracic and lumbar curvatures, probably due to the better control of the distal fused vertebrae by pedicle screws. Suk et al. performed total pedicle nail fixation in 236 patients and issued a mean self-correction rate of lumbar curvature of 66% at 5 years of follow-up, with only 5% (10/203) cases of selective thoracic fusion with trunk imbalance. Lenke et al [15] reported in 1999 that the rate of self-correction of lumbar bending was better in the anterior approach (56%) than in the posterior approach (37%); in the same year, Betz et al [16] reported that there was no significant difference between the degree of self-correction of lumbar bending in the anterior and posterior approaches, both being 51%; Patel et al. Patel et al. conducted a comprehensive analysis of 132 patients with anterior and 44 patients with posterior elective thoracic bending fusion and pointed out that the surgical approach of elective thoracic bending fusion had no significant effect on the postoperative self-correction of lumbar bending. 4. Postoperative loss of compensation after selective thoracic bending fusion: Trunk loss of compensation or trunk imbalance after selective fusion refers to the deviation of the C7 spine from the sacral midline in the coronal plane or the progressive aggravation of the lumbar bending in the sagittal plane after selective fusion of the thoracic vertebrae. The causes of trunk imbalance include overcorrection of the thoracic bend, incorrect selection of the fused segment, poor flexibility of the lumbar bend, and large offset or heavy rotation of the lumbar bend parietal vertebrae. The skeletal development of the patient at the time of surgery also has an impact on the occurrence of trunk loss, and Edwards et al. showed that patients with lower skeletal maturity (Riser’s sign grade 0/I) were significantly more likely to develop trunk loss after surgery. Some authors have also attributed the postoperative loss of compensation with third-generation orthopedic devices to the de-rotation operation and the bracing below the concave side of the thoracic bend during the orthopedic procedure, suggesting that these operations increase the stress on the thoracolumbar and lumbar segments and cause the loss of compensation . However, McCance et al [19] compared the effect of the Harrington system with third-generation orthopedic systems on postoperative compensation after selective thoracic fusion and found that third-generation orthopedic devices had a tendency to increase postoperative compensation, but there was no significant difference. The percentage of postoperative loss of compensation was significantly increased. Newton et al [2] noted that when the thoracolumbar lordosis is greater than 10º the distal end needs to be fused to L2 or L3 to avoid postoperative sagittal imbalance (increased lordosis). In summary, for thoracolumbar double or triple bends in idiopathic scoliosis with small lumbar bends or good flexibility, i.e., King II, Lenke I-IV, and PUMC IIb1, IIIa, and some IIc3 types, if the Cobb angle and parietal offset ratio between thoracic and lumbar bends are greater than 1.2, the lumbar bend does not exceed 50°, the flexibility of the lumbar bend is above 65%, and the parietal rotation is within Selective thoracic fusion is feasible if the thoracic bend is within I degree. Measures to prevent selective thoracic bend fusion treatment for idiopathic scoliosis include: distal fixation with pedicle nailing as much as possible; excessive bracing of the concave side in the distal junctional area should be avoided; and the posterior ligamentous structures in the distal junctional area should be kept intact as much as possible. The residual Cobb angle of the thoracic bend should be minimized; the fusion should be moderately extended in cases where the posterior convexity of the thoracolumbar segment exceeds 10°. Newton et al. predicted patient satisfaction and surgeon acceptance by studying the DFQ (Deformity-Flexibility Quotient) and found that the lower the DFQ, i.e., the fewer residual degrees of lumbar bending and the more mobile lumbar segments retained, the more satisfied the patient and surgeon would be. Therefore, as a spine surgeon, selective thoracic bend fusion should be performed whenever possible in cases where there is an indication. However, preoperative communication with the patient and family must be adequate so that they fully understand the advantages of the procedure and the risks of possible reoperation.