Significance of DPP
According to Denis three-column theory [8], burst fractures mainly involve the middle column, which is located in the posterior part of the vertebral body. Because both sides of the posterior part of the vertebral body connect to the pedicle, burst fractures are often combined with unilateral or bilateral fractures at the junction of the pedicle and the vertebral body [9,10]. Deng [7] et al. analysed the relationship between the posterior wall of the vertebral body and the pedicle in burst fractures and pointed out that the incidence of incomplete or complete pedicle fractures was as high as 81%. Hu [11] et al. reported that the incidence of thoracolumbar fractures combined with pedicle rupture exceeded 60%. Anekstein et al also pointed out that, in most cases, the pedicle was separated from the vertebral body in a burst fracture [12]. Our clinical practice has also found that lumbar burst fractures are often associated with unilateral or bilateral pedicle rupture. The above features become an openable "window" into the spinal canal for a lateral decompression.
Posterior fixation, including fractured vertebrae, has been shown to possess significant advantages over traditional instrumentation, but not in fractured segments. For example, Dick et al. reported that combined pedicle screw fixation at the fractured vertebrae had better biomechanical stability during flexion, torsion, and compressive stresses[13].The internal fixation including fractured vertebrae can also optimize the distribution of stress of the instrumentation in the spine, improve the fixational efficiency, and reduce the incidence of break and loose of screws compared with internal fixation excluding fractured segment [14,15]. Moreover, the insertion of pedicle screws in the fractured vertebrae can be used to reconstruct the pivot for the reduction of the fracture, which helps to restore the displaced bone fragments and the collapsed endplate to achieve the correction of local kyphosis and anterior body height compression, to release the compression of the spinal canal and to decrease the loss of correction in long-term follow-up[16,17]. Also, other studies have shown that screws placed at the level of the fracture do not increase the operation time and surgical trauma [18]. Based on the above advantages of internal fixation including fractured vertebrae,and the process of inserting screws at the burst fractured level must be completed through a ruptured or unruptured pedicle. Therefore, the implanting screws via the ruptured pedicle which is rebuilt to the anatomic position after finishing spinal canal decompression with “pedicle-plasty” strategy, becomes a feasible design.
Based on the above mentioned characteristics of burst fractures with unilateral or bilateral pedicle fractures and the advantages of internal fixation including fractured vertebrae, DPP for the canal decompression and spinal reconstruction of burst fractures becomes a possible option. To clarify the important structures and anatomical relationships adjacent to the surgical approach, this study obtained measurements on CT imaging, determined the surgical anatomy and performed the surgical procedures on cadaveric specimens.
Anatomical value of CT measurement
With the rapid development of modern CT imaging technology and the widespread application of spiral CT and 3D reconstruction technology, the related structures of bone, joints and spine can be clearly displayed in the coronal, sagittal and transverse positions. Furthermore, the reconstructed 3D images can be rotated 360 degrees along any axis to distinctly display the three-dimensional surface structure. It is helpful for clinicians to comprehensively observe the anatomical structures and lesions of the joints and spine using local and whole perspectives [19,20]. Meanwhile, the measurement and accuracy of CT imaging on the various anatomical structures, including the bones, joints and spine, have been confirmed by a large number of studies [21,22,23,24,25], and in vivo, CT scanning can avoid the atrophy of bone structure and muscle tissue, which are caused by dehydration, muscle stiffness and loss of organic matter in cadaver specimens after long-term placement[26]. Therefore, the use of the measuring techniques in CT imaging to carefully measure the local important structures involved in DPP of the lumbar spine can provide important anatomical data for supporting the development of this technique.
While taking the measurements, we found that the DPP of the 5th lumbar vertebrae could not be performed due to occlusion of the iliac crest. Combined with literature reports, the fracture incidence of the 5th lumbar vertebrae is extremely low in clinical practice. For example, Sansur CA et al. showed that burst fractures of the lower lumbar accounted for approximately 1% of lumbar fractures [27], and Han-Dong Lee et al. noted that the incidence of fractures present at the 5th lumbar vertebrae is only 1.4% of the fractures in the lower lumbar vertebrae [28]. Therefore, the related parameters at the level of the 5th lumbar vertebrae were not measured in our study.
The determination of the osteotomy surface is the basis for measuring the OA, TLOS and DM in this study. The osteotomy surface is determined as the plane formed by the outer edge of the articular process and the widest cross-sectional diameter of the spinal canal. Because this plane is long and oblique, it is conducive to the healing of the osteotomy surface. Moreover, this surface is located near the midpoint of the line connecting the outer wall of the pedicle and the inner wall of the articular process. It is not easy to cause pedicle splitting and fracture of the superior articular process during the osteotomy, and the risk of nerve injury is also minimal for the wide spinal canal that is present in this plane. While taking the measurements, we found that the change trends of OA for men and women were the same, and it gradually increased from the 1st lumbar vertebrae to the 4th lumbar vertebrae. Each segment was larger in men than in women.The variation trend of the DM of each segment was consistent with the OA. In males,the TLOS increased from the 1st lumbar vertebra to the 2nd lumbar vertebra, and then decreased downwards in males.While,in females,the TLOS decreased from the 1st lumbar vertebra to the 2nd lumbar vertebra,then increased at 3rd lumbar vertebra,decreased again at 4th lumbar vertebra. each segment was larger in males than in females, and there was a significant difference between them.
At present, there has been no report that have determined the measurements of SD. Our measurements show that the 1st lumbar vertebra is the shortest, and the 4th lumbar vertebrae is the longest. The measurement gradually increases from the 1st to the 4th lumbar vertebrae. The sagittal length of the osteotomy surface is obtained by measuring SD to avoid excessive upwards or downwards dissection during the surgery, which may damage the posterolateral branch of the lumbar nerve and the accompanying intertransverse vessels that pass through the transverse process from front to back.
Anatomical value of the cadaveric specimen measurements
CT scans and three-dimensional reconstruction have unique advantages in displaying the bone anatomy, but there are certain limitations in displaying the anatomical relationship of the blood vessels, nerves and muscle tissue. Therefore, we chose to observe and measure the above structures in cadaveric specimens.
To protect the muscles of the lower back as much as possible, such as the erector spinae (lumbar iliocostalis, longissimus muscle and spinous muscle), transversospinales and the intertransverse process muscle group, and to reduce the occurrence of postoperative lumbar syndrome and ensure the decompression efficiency of the spinal canal[29], many researchers have improved the posterior midline approach in surgeries of the lumbar spine. Watkins first proposed an approach through the quadratus lumborum and lumbar iliocostalis space in 1959[30]. However, in the lower lumbar spine, to satisfactorily expose the posterolateral part of the lumbar spine, the muscles attached to the iliac wing need to be stripped or osteotomized. Therefore, the Watkins approach is still associated with significant trauma. Later, the Wiltse approach was introduced, which uses the space between the multifidus and the longissimus, in 1968. The intermuscular space is clearly demarcated without any obvious distribution of the blood vessels and nerves. After mild dissection, the posterolateral structure of the lumbar spine is easily exposed, and the Wiltse approach is widely used in lumbar fusion surgery [31]. In recent years, Weaver reported a new approach through the longissimus and lumbar iliocostalis space, known as the LIMP approach, which is an intermediate approach. The posterolateral structure of the lumbar spine can also be clearly exposed by this approach, while this incision is more lateral than the WILTSE approach [32]. In our study, according to the anatomical characteristics of the back muscles and the distance of the osteotomy to the midline, as measured by CT, we chose the Watkins approach below the 4th lumbar vertebral plane during the DPP procedure because the iliocostalis and the longissimus originate from the sacrum and iliac crest in the form of tendons. Above this plane, the space between the two muscles is clearer, and the LIMP approach is more appropriate.
At present, it is generally believed that spinal canal occupation with nerve defects is an important indication for spinal canal decompression[33,34], and the main surgical techniques can be recommended, including anterior, posterior, and combined approaches[35,36,37,38,39].In our study, according to the characteristics of burst fractures with pedicle rupture, spinal canal decompression was performed by DPP. The lateral “decompression window” of this technique is composed of the upper and lower intervertebral foramen of the fractured segment and of the space that is produced after the removal of the pedicle to the lateral side. This space is approximately oval, and it can completely expose the rear of the fractured vertebral body, the upper and lower intervertebral discs, the lower margin of the posterior part of the upper vertebrae and the upper margin of the posterior part of the lower vertebrae. The widest part of APDD is at the level of the pedicle, and the narrowest part is at the level of the lower part of the intervertebral foramen. Based on the OA that is measured by CT, it is convenient to reach the posterior part of the fractured vertebral body and then perform the necessary procedures, such as fracture removal or pushing the bone fragments forward into the vertebral body. There are two pairs of export nerve roots, a pair of branches of lumbar artery and the lumbar veins in the “decompression window”. To obtain a clear surgical field of vision, the export roots can be pulled moderately to the side and bipolar coagulation can be used for haemostasis of the branches of the lumbar artery, lumbar veins, and internal spinal venous plexus.
The lumbar plexus is composed of parts of the anterior branches of the 12th thoracic and the 4th lumbar nerve roots and the anterior branches of the 1st to 3rd lumbar nerve roots. The relative position of the arrangement of the lumbar plexus and the nerve roots on the side of the vertebral body and the pedicle is as follows. The nerve roots are inside and ventral, the lumbar plexus is outside and dorsal, and the nerve roots go down towards the nerve plexus from the inside to the outside at a certain angle[40]. During DPP, the pedicle need be pushed to the outside and turned over to expose the lateral spinal canal. Because the lumbar plexus is located at the posterolateral site of the thoracolumbar and lumbar vertebral bodies, closes to the outer wall in the base of the pedicle, and is fixed by many ligaments, the pedicle may injure the lumbar plexus if it is pushed too much, resulting in iatrogenic nerve damage[41].In our study, we found that the lumbar plexus at the level of the 1st lumbar vertebrae is relatively relaxed, the lateral distance is the largest, and it is relatively short at the levels of the 2nd to 4th lumbar vertebrae. Hence, the use of real-time neuropotential monitoring in surgery can reduce the risk of traction injury of the lumbar plexus.
The performance of DPP on cadaveric specimens
In the process of entering the lateral border of the superior articular process through the space between the longissimus and the lumbar iliocostalis, there are many terminal vessels of the lumbar segmental arteries distributed around the articular processes and transverse processes, such as the artery of the interarticular process, the superior artery of the articular process, the inferior artery of the articular process, and the communicating artery[42]. These arteries arise from the dorsal branches of the lumbar arteries, which extend from front to back in the medial space between the transverse processes. Sometimes, it is difficult to find the bleeding vessel during surgery because the abovementioned arteries are terminal blood vessels and their location greatly varies. To stop the bleeding, ligation or cauterization of the dorsal branch of the lumbar arteries between the transverse processes may be effective.
Determining the osteotomy angle of each segment before surgery can reduce the risk of pedicle splitting and the dural injury during surgery. The average osteotomy angle of each segment was preliminarily calculated, which can be used as a reference for the osteotomy, by the CT measurements in our study. However, due to the large individual differences, it is of important significance to measure the specific osteotomy angle according to each individual by using CT imaging before surgery. Piezosurgery is a relatively new technique for bone cutting in oral surgery and orthopaedic surgery that can precisely and safely select a specific section of the mineralized bone structures but does not damage the neurovascular tissue. Therefore, the use of this unit can also reduce the occurrence of the abovementioned risks intraoperatively [43].
The internal vertebral venous plexus is a complex network located at the epidural space in the spinal canal, it is close to the inner wall of the spinal canal, and it is mainly composed of anterior vertebral plexuses, posterior vertebral venous plexuses, lateral vertebral venous plexuses and the communicating branch [44]. This venous plexus is abundant and without venous valves, and blood can flow in both directions [45]. The distal venous return is blocked, and this venous plexus tends to become more varicose and filled after spine injury because of the occupation from fractured fragments. In our study, it was very important to expose the venous plexus at the anterior wall of the spinal canal and around the pedicle, and then haemostasis was performed by compression and bipolar coagulation before the decompression, especially for dealing with the anterior venous plexus. Otherwise, a lot of bleeding would have occurred in the surgical field, leading to difficulties and risks during surgery.
There are still some limitations in this study. First, the bone cement injected into the spinal canal was simulated the occupation caused by bone fragments. Second, the osteotomy at the junction between the pedicle and the vertebral body was simulated the pedicle fracture. The above situation was certain difference from the actual condition of burst fracture in the clinic. Third, the sample size was limited by the number of cadaver specimens donated. Although our experimental design has some deviation, the results obtained from the test still can be provided a reference for the deep clinical study.