EnBloc Resection of a Chordoma of the Thoracic Spine by “L”-Shaped Osteotomy for Spinal Canal Preservation

Abstract

Background/Objectives: EnBloc resections of bone tumors of the spine are very demanding as the target to achieve a tumor-free margin specimen (sometimes impossible due to the extracompartimental tumor extension) is sometimes conflicting with the integrity of neurological functions and spine stability. Methods: The surgical treatment of a huge multi-level chordoma of the thoracic spine with unusual extension is reported. Anteriorly, the tumor widely invaded the mediastinum and displaced the aorta; on the left side, it expanded in the subpleuric region; posteriorly, it was uncommonly distant 13 mm from the posterior wall. Results: EnBloc resection is largely performed for primary bone tumors of the spine and many reports have been published concerning brilliant solutions to difficult issues of surgical anatomy. One of the major challenges is still the compatibility between oncological and functional requirements. Conclusions: Oncological staging, careful imaging analysis, a multidisciplinary surgical team, and utilization of the most recent technologies like navigation and robotics have made an oncologically appropriate EnBloc resection of a multi-level chordoma of the thoracic spine possible without affecting the continuity of the spinal canal and without any involvement of its content by an original “L”-shaped osteotomy.

1. Introduction

The surgical techniques known as enbloc resections were designed to resect bone tumors of the limb to replace amputations and are globally known as “limb salvage surgery” [1,2]. The possibility to spare the functions of a limb is allowed—in high-grade malignancies—by the local effect of neo-adjuvant chemotherapy (reduction in tumor volume, better definition of the margins) [3].

The same techniques were extended to low-grade malignancies and to aggressive benign tumors, not sensitive to chemotherapy, reducing the local recurrence rate [4]. The local and systemic outcome was correlated to the pathological evidence of the integrity and the thickness of the healthy tissue surrounding the tumor, the so called “margin” [1].

Bertil Stener became the pioneer of performing enbloc resection in the spine by scrupulously describing the planning of each surgical procedure tailored to the tumor location and extension in order to achieve a tumor-free margin resection [5]. Roy Camille [6], and later Tomita [7], popularized techniques of spondylectomy unrelated to the tumor extension and less respectful of the oncological validity of margins. As consequence of the spread of commitment to oncological principles, many surgeons started to propose—in the spine also—strategies to adapt the surgical technique to the tumor location and extension, with the primary target of achieving an appropriate margin [8,9,10,11,12,13]. Unfortunately, in many cases, this cannot be achieved without functional sacrifices possibly not accepted by the patient [14,15,16] or without including highly morbid steps, like resection of the involved aorta inside the tumoral mass and bypass [17,18]. In recent years, the literature on enbloc resection in the spine has been widely populated, with single reports of demanding cases and tips on techniques [19,20,21,21,22,23,23,24,25,26], some retrospective studies from single centers [27,28,29,30,31,32,33,34,35], and a systematic review [36]. An attempt to categorize the different surgical techniques of enbloc resection according to tumor extension was published by Tomita—the original paper, specifically dedicated to metastases [37]. More recently, based on the criteria that make an oncologically valid resection possible and according to tumor extension and location, a classification of possible strategies, including seven combinations of approaches, was recently proposed [38], making reference to the WBB (Weinstein–Boriani–Biagini) staging system [39]. According to this paradigm, when a tumor in the thoracic and lumbar spine is widely extended in the anterior and antero-lateral region of the spine (layer A according to WBB), an anterior approach is preconized first to grant the appropriate margin and a posterior approach second to release the content of the spinal canal and finalize the tumor removal. This strategy is staged as type 3. The tumor extension of a recently observed case of T9 to T11 chordoma was unusual: imaging studies showed a huge extension in the mediastinum but without involvement of the posterior part of the vertebral body. A type 3 resection strategy—not considered in the previously cited article [38]—was planned from the perspective of performing a resection without opening the spine canal (reducing the risks of neurological damage and of spine instability) and without violating the tumor margin.

2. Case Report

2.1. Clinical Findings

An adult female aged 56 was referred, complaining of lumbar pain for six months. Once a huge tumoral lesion was found arising from T9, a trocar CT scan-guided biopsy was performed through the left pedicle, which allowed for the diagnosis of chordoma.

At admission, the general conditions were excellent, and back pain was reported as being located in the thoracic area, with circumferential irradiation to the sternum. The pain was reported as interfering with sleep and was slightly reduced during daily activities, and no peripheral pain was reported. Opioid pain killers were effective.

2.2. Imaging Studies

The CT scan shows a lytic, moth-eaten bone lesion in the anterior part of the T9 vertebral body (Figure 1), with huge soft tissue mass better described by MRI as invading the mediastinum and contouring the aorta. The tumor tissue invades the anterior part of the T8 and T10 vertebral bodies without involvement of the disks (Figure 2a). The tumoral soft tissue expands on left side in the retro-pleural space up to the level of the cost-transverse joint in T9 (sector 8 to 4 according to the WBB staging system) (Figure 2b), while in T8 (Figure 2c) and T10 (Figure 2d) sectors 4 and 5 are not involved in the tumor.

A wide tumor-free zone can be seen from the posterior tumor margin and the posterior vertebral wall in the transverse image of all three vertebral bodies (layer A, only part of layer B, no layer C involvement). At the level of the largest tumor extension, the AP diameter of the T9 vertebral body was 28 mm, and the distance between the tumor and the posterior wall was 13 mm (Figure 2b).

2.3. Surgical Planning

To achieve tumor-free margins, and considering the tumoral extension, the surgical planning could include the following.

An antero-lateral approach to release the tumor mass from the mediastinal structures, namely from the aorta. A vascular surgeon should be part of the multidisciplinary team and be ready to perform a graft if the aorta must be resected and included in the tumor specimen, in case of involvement of the aortic wall inside the tumor mass.

A posterior approach to complete the circumferential release of the spine without any opening of the spinal canal and to finalize the enbloc resection. An “L”-shaped osteotomy has been designed, with the longest leg in the coronal plane, cutting between the posterior tumor margin and the posterior vertebral wall, and the shorter leg in the sagittal plane, to achieve an appropriate margin to the tumoral extension in the retropleural space facing sector 4. The costo-vertebral joint at T9 will therefore be included in the tumoral specimen.

The osteotomy lines were planned to achieve an appropriate tumor-free margin of the specimen and contemporary to save the continuity of the spinal canal without any involvement of its content. Navigation and robotic guidance are reliable and valuable tools to perform such demanding lines with a precision of few millimeters’ tolerance, as the distance between the tumor and the posterior wall is only 13 mm and the target is to leave at least 5 to 7 mm of healthy bone as a margin over the tumor.

According to WBB-based surgical planning [38], type 1 (single anterior approach) has no possibility of achieving a tumor-free margin specimen due to the eccentric extension into the retropleural space. A second posterior approach with the original “L” osteotomy can be staged as a further sub-type of type 3 [38] (Figure 3).

Pre-operative angiography and selective embolization of the T8, T9, and T10 segmental arteries can be useful to not cause ischemia of the lesion but to embolize those arteries stretched by the tumor growth, which are sometimes difficult to detect. A tear of these stretched arteries may not immediately cause bleeding but can produce dangerous hematomas in the early post-operative course when the arterial walls relax and blood pressure increases.

First step (I): Right lateral decubitus. Antero-lateral left trans-pleural approach to release the tumor mass from mediastinal structures, namely the descending aorta, and partial diskectomies at T7–T8 and T10–T11: a left posterolateral thoracotomy is performed at the level of the sixth intercostal space. Once the pleural cavity is opened, the neoplasm can be appreciated at the level of the T8, T9, and T10 bodies. The pulmonary ligament is dissected, and the parietal pleura is opened, isolating the mass. Finally, partial costectomy of the VII, VIII, IX, and X ribs is performed. Diskectomies complete the first surgical step.

Second step (II): Prone position. The posterior approach includes conventional lateral release of the vertebral bodies—once the posterior segment of three ribs is sacrificed—until the tumor mass is reached. “L-shaped” osteotomy from the T7–T8 to T10–T11 disk level and specimen removal are performed.

2.4. Surgical Procedure

The surgical procedure is illustrated in Figure 4, Figure 5, Figure 6, Figure 7, Figure 8, Figure 9, Figure 10 and Figure 11 and described in the corresponding legends. The distinctive step of this original procedure is the robot-guided coronal cut by a transcutaneous Tomita saw and the conversion to a sagittal cut by pivoting on a pedicle screw. From an oncological point of view, the precision of the robotic–assisted positioning of the saw allows one to achieve a tumor-free margin, and pivoting over the screw allows for resection of the tumor involved in the left posterior part of the vertebral body. From a mechanical point of view, this technique allows one to spare the wall of the spinal canal and the enbloc resection is performed without encroaching on the epidural space.

The original plan was to remove the screw together with the specimen, but during the osteotomy on the sagittal plane, the saw became stuck in the threads of the screw. For this reason, the screw was removed, and the resection was completed free–hand. This was possible because the saw became stuck during the final part of the resection, and when we were therefore already in a safe zone. This problem could be avoided by using a partially threaded screw.

2.5. Reconstruction

A sound posterior fixation by a pedicular screw and rods (three levels above and two levels below for a final T5-T12 construct) was performed using a PEEKReinforced Carbon Fiber system (Carboclear by Carbofix). This titanium-free material has the same—if not better—mechanical properties as titanium [40], but it is radiolucent and allows for the early detection of a recurrence and a better radiotherapy planning as delivers a significant reduction in artifacts [41], allowing for a better definition of the target and organs at risk [42]. In case of accelerated particles, radiotherapy is indicated, and the backscattering effect by PRCF implants is significantly reduced, resulting in a higher dose on the tumor and a lower irradiation of neighboring organs [43].

The anterior column loss only concerns the anterior half of bodies and disks, as the coronal osteotomy allowed us to save the posterior vertebral walls. To grant anterior stability, a segment of autologous vascularized fibula was inserted between the upper and lower bone surfaces left by the “L” osteotomy.

The feeding artery of the fibula was anastomized with the previously prepared intercostal artery during the first surgical step.

2.6. Pathologist’s Report

The diagnosis of chordoma was confirmed. The margin was defined “marginal” at the anterior part of the tumor, expanding into the mediastinum, represented by a thin layer of capsule and reactive tissue and partially covered by the pleura. The margin at the bone was calculated to be 3 to 5 mm, and intact. This was also defined as “marginal” (Figure 12).

2.7. Post-Operative Period and Follow Up

The patient was hospitalized for 19 days after surgery and the post-operative period was complicated by a pleural effusion which progressively decreased. After five months, she was re-admitted because of a post-surgical infection due to Staphylococcus Aureus that required a reintervention of debridement associated with long-term antibiotic therapy. Six months after the reintervention, no signs of infection or local recurrence were detected with MRI. The grafted fibula appeared stable with bone integration in progress, as shown by CT scan. The absence of local recurrence was confirmed at the last follow up, 15 months after the primary surgery.

3. Discussion

The surgical oncology of bone tumors requires the careful planning of enbloc procedures: the soft tissue dissection and the osteotomy lines must be decided by considering the distance from the tumor periphery, which anatomical structures can be released, and which to be sacrificed for tumor-free margins. The pioneer of enbloc resection who planned to achieve an appropriate margin was Bertil Stener [5]. He planned the surgical steps based on his own designs. Nowadays, high-resolution imaging studies allow surgeons to better understand the details of tumor extension and anatomical element involvement; 3D images and 3D-printed plastic specimens are extremely useful to visualize the surgical anatomy [44]. Newer technologies allow us to improve the accuracy of the resection, develop new surgical techniques, and reduce morbidity. Minimally invasive anterior antero-lateral approaches can substantially reduce the surgical morbidity of the approach and offer a better visualization of hidden areas like the contralateral vertebral body [45,46]. Navigation has become an indispensable tool to perform osteotomies at the right distance from the tumor in complex regions like the pelvis and spine [47,48]. Robotic-assisted steps during enbloc resection in the spine have already been reported [49]. The thread wire saw is the most frequently used tool to perform osteotomy or diskectomy, even if many surgeons prefer the bone scalpel, or the more current ultrasound scalpel [50]. A bone scalpel was used by Roy Camille to conclude osteotomy, reducing the risk of to damage the thecal sac by the uncontrolled use of a thread wire saw [6]. Protection of the thecal sac has been proposed by Tomita with manually managed special tools [7]. Recently, a safer device to protect the cord—connected to the rod—has been designed [51]. The thin thread wire saw remains today the more versatile tool and has been proposed as the key device of newly proposed techniques [52]. Also, in the case presented here, the thread wire has a very important role in an accurate spine osteotomy. In the case presented here, some peculiarities suggest surgical planning considerations not considered in the previously published classification [53]:

Location in the thoraco-lumbar region;

Huge anterior tumor extension (layer A) in close contact with the aorta and the cava;

Posterior part of the vertebral body (layer C) not involved by the tumor;

The tumor extended layers 5 to 8, fully saving the pedicles (sectors 4 and 9).

During the analysis of the imaging for the surgical planning, the possibility to save the canal circumference became evident, without obviously violating the tumor mass. There are two obvious consequences: significant reduction in the risk of neurological complications, and higher post-operative stability. However, the step to be solved was how to guide the thread wire saw along a non-linear path. The osteotomy had to run first on the coronal plane at an adequate distance from the posterior tumor margin for an oncologically appropriate resection, changing direction on the left side as the tumor expanded in the sub-pleural area at sector 4. A variant of the type 3 enbloc resection osteotomy [53] had to be planned to adapt the osteotomy line to the tumor extension and to save the canal circumference. The technical solution was to rely on navigation and robotic guidance to grant accuracy in the osteotomy lines with a few millimeters of tolerance, namely in the change in direction from the coronal plane to a sagittal plane. On the coronal plane, two K-wires, implanted using robotic assistance, guided the thread wire on the right coronal cut and a screw, positioned in the left pedicle of T9, acted as a pivot by which the saw could change direction to a sagittal plane to complete an “L-shaped” osteotomy. An anterior approach was decided as the first step to release the tumor mass from the aorta, aiming to leave a tumor-free margin. The sacrifice of the vessel and grafting had to be considered [17] if the aortic wall were to be found infiltrated or if the tumor were to be found completely surrounding the aorta. To this purpose, the vascular team participating in the operation were ready to perform a graft. Reconstruction of the continuity of the posterior vertebral wall and of the canal circumference allowed them to reduce the usual concern for the validity of the anterior support. A vascularized fibula was positioned abutting the preserved portion of the posterior wall, between the inferior endplate of T7 and the superior endplate of T11. A posterior construct was completed for the fixation T5 to T12. In the T8 and T10 pedicles, additional screws were implanted on the left side after tumor removal. Being manufactured in PRCF, it should not interfere with accelerated particle radiotherapy in case of local recurrence treatment.

4. Conclusions

This paper is intended to present an original technical solution to the performance of enbloc resection in a chordoma of the thoracic spine, unusually sparing the posterior part of the vertebral body, an area where most chordomas originate as best corresponding to the residual of notochord. The target of this technique is to spare the posterior wall and therefore the circumference of the spinal canal for the purpose of reducing the risk of neurological damage and to grant sounder stability. A marginal margin can be considered an acceptable result, but obviously the oncological outcome will be evaluated only in 5 years, as most local recurrence occurs later. During the anterior approach, the margin was achieved under direct visual control. All of the team were ready to perform a resection of the aorta and by-pass grafting in case of tumor involvement. This was not necessary, as a perimetral reactive tissue around the tumor, easily released from the aortic wall, was considered a safe margin. Navigation and robot-guided techniques made the “L” osteotomy possible, with a coronal leg performed at the best distance between the posterior tumor margin and the posterior vertebral wall and the shorter sagittal leg directed posteriorly into the pedicle at a distance from the costovertebral joint involved by the retropleural tumor extension. The key achievement was the idea to use a screw as a pivot to change the direction of the thread wire saw from the coronal to the sagittal plane.