Objective To summarize the technical points and clinical efficacy of pedicle subtraction osteotomy (PSO) in tunneling and to explore the related complications of this technique. Methods A total of 67 cases of old vertebral fractures of the thoracolumbar region from June 2012 to June 2017 were collected. According to the inclusion and exclusion criteria, a total of 41 cases were included in the study. There were 19 males and 22 females; aged 37-67 years, mean 60.1±12.7 years; 15 cases of non-surgical treatment after trauma, 13 cases of failure after surgery and 13 cases of osteoporosis. Injury segment: 9 cases of T11, 22 cases of T12, 8 cases of L1, 2 cases of LS. Preoperative patients were diagnosed by X-ray, CT plain and 3D reconstruction combined with MRI. There were 23 cases of intractable back pain, 16 cases of lower extremity root pain, and 2 cases of intermittent claudication. Patients were divided into the traditional PSO treatment group (21 cases) and modified PSO treatment group (20 persons) according to the random number method. The traditional group were treated with the egg shell technique, and the improved group were treated with tunnel forming technology. The procedure was divided into four steps: exposure (step 1), nail placement and resection of the posterior column complex (step 2), vertebral osteotomy (step 3), orthopedics and bone grafting (step 4). The operation time, bleeding volume and complications of each step were compared between the two groups. The clinical efficacy was evaluated using the visual analogue scale (VAS) score and the Oswestry disability index (ODI). The X-ray spine Cobb angle was measured to evaluate the Keloid deformity correction, and the bone graft fusion was observed by CT examination. Results All patients were followed up for 12 to 24 months. The total operation time of the traditional group was 273.3±21.1 min, and the total operation time of the modified group was 178.1±12.5 min, the difference between the two groups was statistically significant (t=8.981, P=0.0019). The differences between the two groups in steps 2 and 3 were statistically significant (t2=4.614, P2=0.036; t3=9.089, P3=0.020) . The difference in the total bleeding volume was statistically significant (t=8.529, P=0.011). The differences in the bleeding volume between the two groups in step 2 and step 3 were statistically significant (t2=11.933, P2=0.016; t3=6.972, P3=0.013). The Cobb angles of the traditional group before surgery, 1 week after surgery and half year after surgery were 40.2°±8.9°, 12.5°±6.8°, 10.4°±2.5°, respectively. The Cobb angles of the modified group before surgery, 1 week after operation and half year after surgery were 39.5°±6.3°, 10.4°±3.5°, 9.5°±1.9°, respectively. The differences in the Cobb angle between the preoperative, postoperative 1 week and postoperative half year were statistically significant(Fmodified group=189.573, Pmodified group=0.021; Ftraditional group=194.699, Ptraditional group=0.029). The bone fusion time in the osteotomy area was 3-6 months, with an average of 4.8 months. The VAC and ODI scores of half-year post operation of the traditional group were 2.1±0.3 and 34.1±4.3, and the improved group were 2.2±1.1 and 28.3±6.8, respectively, and the difference was not statistically significant. In the improved group, there were 1 case of over-excavation and 1 case of over-underexcavation (2/20), which were corrected in time during operation. In the traditional group, 4 cases (4/21) of dural tear occurred during the operation, and were repaired in time. Bone fusion was obtained half a year later. No clinical deaths, and no cases of surgical infection occured. Conclusion Tunnel forming technology is an alternative surgical procedure for treating old fractures of the thoracolumbar region with PSO, which can shorten the operation time, reducd intraoperative bleeding and reduce surgical complications. Key words: Osteotomy; Orthopedic procedures; Spinal fractures