- Email: [email protected]
En Bloc Resection of Thoracic Tumors Invading the Spine: A Single-Center Experience
En Bloc Resection of Thoracic Tumors Invading the Spine: A Single-Center Experience Xiao-Miao Zhang, MD,* Ludovic Fournel, MD,* Audrey Lupo, MD, PhD, Emelyne Canny, MD, Antonio Bobbio, MD, Salom e Lasry, MD, Jean-Franc¸ois Regnard, MD, PhD, Frederic Sailhan, MD, and Marco Alifano, MD, PhD ˇ
Background. Vertebral involvement by a thoracic tumor has long been considered as a limit to surgical treatment, and despite advances, such an invasive operation remains controversial. The aim of this study was to characterize a single-center cohort and to evaluate the outcome, focusing on survival and complications. Methods. We retrospectively reviewed the data of all patients operated on for tumors involving the thoracic spine in an 8-year period. En bloc resection was generally performed by a double team involving thoracic and orthopedic surgeons. Distant follow-up was recorded for oncologic and functional analysis. Results. There were 31 patients operated on. An induction therapy was administered in 20 patients. Spinal resection (mostly including ‡2 vertebral levels) was combined with lobectomy in 48.3% of the patients, and osteosynthesis was required in 22 patients. We observed no in-hospital death and a major complications rate of 32.3%, including 5 patients with early neurologic com-
plications. There were 61.3% primary lung carcinomas, 12.9% extrapulmonary primaries, 9.7% metastases, and 16.1% benign tumors. Mean follow-up was 32.1 months. The 5-year overall survival rate was 81.3% in the entire cohort and 75.0% in patients with a malignant tumor. Occurrence of an early postoperative major complication was the only factor significantly associated with shorter overall survival (p [ 0.03). The 5-year disease-free survival rate was 37.0% in malignancies. Delayed complications occurred in 35.5% of patients, including persistent neurologic deficit in 12.9%, instrumentation migration in 19.4%, and local infection in 12.9%. Conclusions. En bloc resection of spinal thoracic tumors offers long-term survival and few recurrences in highly selected patients but is associated with significant delayed mechanical or infectious complications.
V
operative period [4–11]. However, the selection and eligibility criteria, as well as the type of multimodal treatment used in these different cohorts, are very heterogeneous, and a real consensus has not yet been reached. The aim of this study was to characterize a singlecenter cohort and to evaluate the short-term and longterm outcomes of resected thoracic tumors involving the spine in the setting of a multidisciplinary approach including thoracic surgeons, spine surgeons, medical oncologists, and pathologists. A secondary objective was to analyze the occurrence of adverse events related to such heavy surgical procedures, focusing on delayed complications and their management.
ertebral involvement by a thoracic tumor has long been considered as a limit to surgical treatment, and rightly so, because of the complexity of the posterior mediastinum anatomy and the poor short-term and longterm prognosis that was usually associated with the resection of such tumors [1–3]. With the progress of surgical techniques, the improvement of preoperative workups, and especially, the evolution of multidisciplinary care, including systemic treatments and radiotherapy for malignant lesions, some previously disqualified patients are now considered eligible for curative-intent surgical procedure. In this regard, some centers reported their experience and favorable oncologic results, although mostly restricted to the early post-
(Ann Thorac Surg 2019;108:227–34) Ó 2019 by The Society of Thoracic Surgeons
Accepted for publication Feb 8, 2019. *Drs Zhang and Fournel contributed equally to this work. Address correspondence to Dr Alifano, Department of Thoracic Surgery, Paris Center University Hospital, 27, rue de Faubourg Saint Jacques, 75014 Paris, France; email: [email protected].
Ó 2019 by The Society of Thoracic Surgeons Published by Elsevier Inc.
The Video can be viewed in the online version of this article [https://doi.org/10.1016/j.athoracsur.2019.02.019] on http://www.annalsthoracicsurgery.org.
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2019.02.019
GENERAL THORACIC
Departments of Thoracic Surgery, Pathology, and Orthopedic Surgery, Paris-Center University-Hospital, Assistance PubliqueHopitaux de Paris, Paris-Descartes University, Paris, France; and Department of Thoracic Surgery, Shanghai General Hospital affiliate to Shanghai Jiao Tong University, Shanghai, China
228
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Ann Thorac Surg 2019;108:227–34
Table 1. Preoperative Patients’ Characteristics
GENERAL THORACIC
Variables Sex Male Female Smoking Current Former >3 months No Major comorbidity Chronic obstructive pulmonary disease Ischemic heart disease Previous cancer Neoadjuvant treatment Radiotherapy Chemotherapy Radiochemotherapy None Response to neoadjuvant therapy Partial response Stable disease
Table 2. Perioperative Course and Pathologic Analysis No. (%) (n ¼ 31) 23 (74.2) 8 (25.8) 14 (45.2) 5 (16.1) 12 (38.7) 4 (12.9) 4 (12.9) 3 (9.7) 1 (3.2) 5 (16.1) 14 (45.2) 11 (35.5) (n ¼ 20) 11 (55) 9 (45)
No. ¼ number.
Material and Methods Ethics We retrospectively reviewed clinical records of all patients operated on in a curative intent from 2011 to 2018 in our Department of Thoracic Surgery at Cochin Hospital for a tumor with vertebral extension or development (at least at imaging and macroscopic intraoperative assessment). This study was performed in accordance with principles outlined in the Declaration of Helsinki and in agreement with French laws on biomedical research. The Comit e d’Ethique de la Recherche Clinique–Soci et e Franc¸aise de Chirurgie Thoracique et Cardio-Vasculaire approved the study (CERC-SFCTCV-2018-3-6-11-33-4-folu).
Preoperative Workup and Multimodal Management In addition to a physical examination, the preoperative workup generally included thoracoabdominal and cerebral computed tomography (CT), bronchoscopy, 18fluodeoxyglucose positron emission tomography, and evaluation of respiratory function by spirometry. The use of magnetic resonance imaging was systematic in the presence of close contact with the intervertebral foramina or medullary canal. A preoperative biopsy to obtain the histologic type of the tumor was most often requested, especially in patients who were likely to receive neoadjuvant therapy. All patients were discussed beforehand case-by-case at a multidisciplinary thoracic oncology meeting. The eligibility for surgical resection mostly relied on the patient’s general condition, the histologic type of the tumor, and its
Variables Type of pulmonary resection Lobectomy Segmentectomy Wedge Type of vertebral resection (largest resection considered) Total vertebrectomy Hemivertebrectomy Partial (transverse process and vertebral body wedge resection) Number of operated-on vertebral levels 1 2 3 Number of operated-on rib levels 0 1 2 3 4 Early complications Pneumonia Neurologic deficit Hemorrhage Early reintervention Mechanical ventilation Other Lung carcinoma Squamous cell carcinoma Adenocarcinoma Large cell neuroendocrine carcinoma Non-small cell carcinoma Sarcoma Osteosarcoma Ewing sarcoma Synovial sarcoma Other malignant tumor B3 thymoma Benign tumor Neurinoma Hemangioma pT status T2a,b T3 T4 pN status N0 N1 N2
No. (%) (n ¼ 31) 15 (48.3) 4 (12.9) 6 (19.4)
3 (9.7) 19 (61.3) 9 (29.0)
16 (51.6) 14 (45.2) 1 (3.2) 9 4 6 9 3
(29) (12.9) (19.4) (29) (9.7)
5 (16.1) 5 (16.1) 1 (3.2) 1 (3.2) 6 (19.4) 7 (22.6) 19 (61.3) 8 (25.8) 6 (19.4) 1 (3.2) 4 (12.9) 6 (22.6) 4 (12.9) 1 (3.2) 1 (3.2) 1 (3.2) 1 (3.2) 5 (16.1) 4 (12.9) 1 (3.2) 19 (61.3) 5 (26.4) 7 (36.8) 7 (36.8) 19 (61.3) 16 (84.2) 1 (5.3) 2 (10.5) (Continued)
Ann Thorac Surg 2019;108:227–34
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Table 2. Continued Variables
No. (%) (n ¼ 31)
R status R0 R1 Presence of vascular or lymphatic emboli Yes No
31 29 2 19 6 13
(100) (90.6) (9.4) (61.3) (31.6) (68.4)
No. ¼ number.
tumor dose of 45 Gy, avoiding, if possible, the lung hilum and uninvolved mediastinum. Response to induction therapy was evaluated according to Response Evaluation Criteria In Solid Tumors (RECIST) criteria, and a stable or responsive disease was required to confirm surgical indication, whereas a progressive disease was considered as a contraindication.
Technical Considerations The rules reported below were generally followed (Video): En bloc resections were performed by a multidisciplinary team including orthopedic and thoracic surgeons. In case of primary lung cancer, lobectomy associated with systemic lymph node dissection was the operation of choice, with sublobar resection and pneumonectomy being performed in case of necessity (poor cardiorespiratory function, or conversely, lobectomy not allowing a satisfactory oncological resection). Sublobar resections were usually Fig 1. Vertebral mapping of resected tumors. Each thoracic (T) spinal level is labeled with the number (N) and mean size of tumors involving it. In the entire series, some tumors involved more than 1 level.
GENERAL THORACIC
anatomical borders, which should anticipate the possibility of complete resection. The general policy of our team is to prescribe induction treatments for primary lung cancer involving the spine, especially in the case of superior sulcus tumors. For other primaries, decision for induction was based on anticipated sensitivity of the known malignancy to radiotherapy and chemotherapy. Induction in non-small cell lung cancer (NSCLC) consisted of platinum-based chemoradiotherapy with a
229
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Ann Thorac Surg 2019;108:227–34
GENERAL THORACIC
230
Fig 2. Overall survival (OS) in (A) all study patients, (B) non-small cell cancer (NSCLC) primary lung tumors, and (C) benign (curve 1) versus malignant tumors (curve 2). Disease-free survival (DFS) in (D) all study patients, (E) primary lung tumors, and (F) benign (curve 1) versus malignant tumors (curve 2). The vertical line indicates the 95% confidence interval at 60 months.
performed in extrapulmonary primaries invading both spine and lung. Posterolateral thoracotomy was performed with the patient positioned in Paulson lateral decubitus and lying anteriorly at 45 degrees on the operating table. The vertebral resection method depended on the degree of tumor invasion. In cases of infiltration of the posterior arch of the rib with contact of the transverse process (even if minimally infiltrated), the transverse process was resected; in case of larger infiltration of transverse process, intervertebral foramen (or foramina), vertebral lamina, or vertebral body or disk at maximum 1 cm within the vertebral midline, an hemivertebrectomy was performed. In case of more extensive involvement of vertebral body, a total vertebrectomy was performed. Hemivertebrectomy and total vertebrectomy could both be performed in more than 1 spinal level, 3 being the maximum number accepted. Spinal resection and stabilization was performed with a posterior midline incision without changing the patient’s position. After the pedicle screws were positioned, a laminectomy of the involved side was
performed. In case of total vertebrectomy, bilateral instrumentation was associated with a titanium cage filled with fragments of free bone graft. Nerve roots and accompanying segmental vessels were exposed and ligated inside the spinal canal to prevent cerebrospinal fluid leakage. These maneuvers allowed gentle traction of the dural sac and en bloc extraction of the specimen through the thoracotomy. Spinal rods were then implanted and secured on both sides.
Statistical Analysis Demographic data and major comorbidities of the patients were collected and analyzed as well as the radiologic evaluation of the tumor, the type of surgical intervention and associated multimodal treatment, the immediate operative follow-up, and pathologic features. Descriptive statistics are expressed as mean SD or median (range) for quantitative variables and as frequencies (percentages) for categorical variables. The follow-up was performed by interrogation of referring or family physicians and city registers. Survival rates,
Ann Thorac Surg 2019;108:227–34
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Table 3. Risk Factors Analysis Variable
p Value 0.4
60.7 (21.1–100.3) 73.4 (62.1–84.8) 0.9 70.8 (54.1–85.7) 74.9 (61.3–88.6) 0.6 74.8 (55.7–94.1) 46.2 (28.7–63.7) 0.69 74.5 (57.0–91.9) 72.7 (59.4–86.0) 0.39 64.8 (43.6–86.1) 39.2 (9.9–68.5) 0.09 74.9 (63.6–86.1) 16.9 (0.0–36.8) 0.11
14 (45.2%), and 3 levels in 1 (3.2%) and required spinal osteosynthesis in 22 patients (71.0%; Fig 1). The median number of resected ribs was 2 (range, 0 to 4). There were no in-hospital deaths. In-hospital stay was uneventful in 15 patients (48.4%), with a mean postoperative duration of stay of 13.9 9.5 days. Major complications occurred in 10 patients (32.3%), with 5 cases of early postoperative neurologic disorders (including phrenic nerve palsy), in particular. Notably, 1 patient who underwent repeat resection of a cervicothoracic compressive schwannoma (after initial medullary decompression performed urgently in another institution some weeks previously) experienced early postoperative spinal shock with persistent paraplegia. We also recorded 1 patient with a dural wound, followed by symptomatic cerebrospinal fluid leakage into the right pleura and pneumocephalus, with recovery at postoperative day 20 after medical therapy, including intravenous acetazolamide. One patient underwent a repeat operation at postoperative day 1 for diffuse bleeding that led to a major hemothorax. Other complications were minor and included medically treated events such as a urinary tract infection or an intestinal subocclusive syndrome. Univariate analysis did not reveal any factors significantly associated with occurrence of major complication.
Pathologic Analysis
70.5 (45.2–95.8) 14.0 (10.1–17.9) 0.03 38.9 (20.2–57.6) 81.3 (73.4–89.2)
CI ¼ confidence interval.
including non–cancer-related deaths, were calculated by the Kaplan-Meier method and compared by log-rank tests.
Results Study Population Characteristics Thirty-one patients who met the inclusion criteria were identified in the study period. Main patients’ characteristics are reported in Table 1. At the preoperative evaluation, 26 patients exhibited a malignant tumor, including 19 NSCLCs, whereas the 5 remaining patients had a benign disease. An induction therapy was performed in 20 patients, allowing partial response in 11 (55%) or stable disease in 9 (45%) at reevaluation. In patients who received radiotherapy, the median irradiation dose was 45 Gy (range, 30 to 56 Gy).
Operative Procedure and In-Hospital Course As summarized in Table 2, a pulmonary resection was performed in 80.6% of patients and was associated with spinal resection for en bloc tumor removal. Vertebral resection involved 1 level in 16 patients (51.6%), 2 levels in
Pathologic analysis of resected specimens revealed a majority of primary lung carcinomas (61.3%) and 5 benign tumors (16.1%; Table 2). Focusing on the T classification of the 19 NSCLCs, we observed a postinduction downstaging in 63.2% of patients. There were 2 patients with mediastinal lymph nodes involvement. In 1 squamous cell carcinoma of the lung and 1 benign tumor, resection was classified R1 because of positive margins at the foraminal level. Analyzing the response to neoadjuvant treatment, we observed a mean percentage of viable tumor cells as 17.1% 23.1% of the whole tumor area among all postinduction specimens, with complete response (no residual disease) in 4 patients (20%).
Prognosis and Delayed Complications Mean follow-up was 32.1 23.7 months. During the study time frame, 4 patients died, including 1 non–cancerrelated death 2 months after hospital discharge. All patients who died were smokers and had NSCLC. A Kaplan-Meier analysis showed that 5-year and 7-year survival rates were similar in each subgroup and equal to 81.3% (95% confidence interval [CI], 59.9% to 92.7%) in the entire population, 75.0% (95% CI, 48.9% to 90.4%) in those with malignant tumors, and 61.3% (95% CI, 30.6% to 85.4%) in those with NSCLC (Figs 2A–C). All patients with a benign disease were alive during follow-up (range, 24 to 83 months). Log-rank tests revealed that occurrence of major complications in the early postoperative period (p ¼ 0.03), analyzed as a binary categorical variable, was the only factor significantly associated with shorter overall survival (OS). There was no significant impact on OS of such factors as age, chronic obstructive pulmonary
GENERAL THORACIC
Chronic obstructive pulmonary disease Yes No Age 60 years >60 years Radiologic response to induction therapy Partial response Stable disease Number of resected vertebral levels 1 >1 Pathologic N status N0 Nþ Pathologic R status R0 R1 Viable tumor cells after induction therapy 10% >10% Occurrence of early major complication Yes No
Mean Survival Estimates Months (95% CI)
231
232
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Ann Thorac Surg 2019;108:227–34
Table 4. Case Series of Patients With Long-Term Complications
Patient 1
GENERAL THORACIC
2
3
Type of Vertebral Resection
Late/Chronic Complication
Postoperative Time to Diagnosis of Complication (months)
Management
24
Instrumentation replacement
10
Extension of osteosynthesis
11
Local debridement and coverage with latissimus dorsi muscle flap Rehabilitation
T1-T2 Pedicle screw loosening and hemivertebrectomy þ 3 displacement ribs T1-T2 total Traumatic migration of spinal vertebrectomy þ 3 ribs instrumentation and neurologic deficit Secondary acute deep infection C7-T4 laminectomy þ T2 Persistent T6 paraplegia hemivertebrectomy þ 2 ribs Chronic restrictive respiratory failure
0
0
Chronic noninvasive ventilation Local debridement and coverage with paraspinal muscle flap Local debridement and coverage with trapezius muscle flap Rehabilitation Superficial debridement and vacuum therapy Rehabilitation
4
Instrument replacement
4
T2-T3 hemivertebrectomy Superficial and deep infection
57
5
T2-T3 Wound disunion and deep infection hemivertebrectomy þ 1 rib T5-T6 hemivertebrectomy Lower-back sensitive neurologic deficit T2 hemivertebrectomy Asymptomatic pedicle screw loosening superficial wound infection T3-T4 partial Mixed chronic respiratory failure with vertebrectomy phrenic nerve palsy T3-T4 hemivertebrectomy Asymptomatic pedicle screw migration out of the vertebra with major increase of kyphosis T1–3 hemivertebrectomy Asymptomatic dural screw migration without cord compression T1-T2 partial C8 sensitive deficit vertebrectomy
19
6 7 8 9
10 11
disease status, completeness of resection, or pN status (Table 3). Relapse of the disease occurred in 7 patients (22.6%) during the follow-up period, with 3 (42.9%) thoracic and 4 (57.1%) extrathoracic first recurrences. The 5-year disease-free survival (DFS) rates were 50.5% (95% CI, 30.0% to 74.8%) in the whole population, 37.0% (95% CI, 19.5% to 67.2%) in those with malignant tumors, and 61.1% (95% CI, 30.8% to 87.7%) in those with NSCLC (Figs 2D–F). At late follow-up, 11 patients (35.5%) experienced single or multiple distant complications. Permanent neurologic dysfunction occurred in 4 patients (12.9%) despite rehabilitation, and a displacement of the osteosynthesis hardware occurred in 6 patients (19.4%). Among these patients, half required a hospital readmission and reoperation for instrumentation replacement and spine stabilization. The occurrence of this late complication was significantly higher in patients with chronic obstructive pulmonary disease (p ¼ 0.02) and tumors with nerve root involvement (p ¼ 0.002). One of the 5 patients who experienced early postoperative neurologic complications completely recovered from a C8-T1 sensitive deficit. Four patients (12.9%) required readmission for late-onset
0 5
10 0
Surveillance Rehabilitation
infections, consisting of 1 superficial and 3 organ-space infections contiguous to the spinal instrumentation. At the study end point, all of the patients, except for those with late-onset infections, were successfully and conservatively treated, without implant removal, particularly. The history and management of late or persistent complications is detailed in Table 4.
Comment Despite advances in preoperative work-up and multimodal approach, surgical selection of patients with tumors invading the thoracic spine remains challenging because of the delicate regional anatomy, potential functional impact, or prognostic considerations. In the present study, we aimed to address these issues and showed that careful preoperative workup and doubleteam (thoracic and orthopedic) en bloc resection could allow safe and reliable surgical management, with no inhospital mortality and high long-term survival rates, even when excluding benign tumors. Such satisfactory oncologic results underline the paramount role of patient selection, neoadjuvant/adjuvant treatments, and surgical approach.
Ann Thorac Surg 2019;108:227–34
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
233
Table 5. Prognosis of Patients After Vertebral Procedures for Resection of Thoracic Malignancies in Relevant Studies Induction Treatment (%)
Chadeyras [4] Yokomise [6] Anraku [7]
32 6 23
66 100 100
NSCLC NSCLC NSCLC
31 33 48
13 0 17
65/24 68/68 58/40
Schirren [8]
20
30
NSCLC
40
20
z75/47
Fadel [13]
54
50
NSCLC
55
9
47/31
Provost [15]
30
97
Ewing sarcoma
27
7
z82/61
Mody [20]
32
78
Malignancy
56
6
z60/40
Collaud [23]
48
98
NSCLC
12
z73/61
CT ¼ chemotherapy;
No. ¼ number;
Major Histologic Complication Rate Rþ Type (%) (%)
38 (prolonged intensive care unit stay)
2-Year/5-Year Survival Rates (%)
NSCLC ¼ non-small cell lung carcinoma;
As shown in Table 5, the heterogeneity of published surgical cohorts of chest malignancies invading the spine led to a wide range of long-term survivals, suggesting that significant biases could have influenced our results, which slightly exceed the upper limit of the range [4, 6, 12, 13]. Like most authors in this research field, we acknowledge the bias of analyzing a small number of patients with few events occurring; thus, a part of these good survival results could be attributed to hazard. With respect to patient selection, we performed an extensive preoperative workup, similar to other teams, to avoid intraoperative discovery of unresectable disease and decrease the rate of incomplete resections [9,12–17]. Despite accurate preoperative imaging, the en bloc pattern of all resections, and a large subset of patients who received an induction therapy, we did not achieve completeness of resection in 2 patients, including 1 benign tumor. Nevertheless, we observed a high rate of R0 resections (93.5%) at pathologic evaluation compared with previous reports. This could explain the relatively high survival rates that we obtained, even though we failed to reveal a statistically significant association between R classification and OS [10, 13]. In the present series, 76.9% of patients with a malignant disease received an induction therapy. As outlined in previous studies, induction therapy is likely to favorably influence prognosis by improving patient selection, increasing the rate of R0 resections, and driving postoperative multimodal management with assessment of the pathologic response [15, 18, 19]. Independently from
Comments Mainly CT as neoadjuvant treatment 83.3% of major pathologic response rate 61% of 2-stage surgery (1 or 2 weeks interval) 72% of 3-year survival in R0 patients Upfront surgery in patients with “imminent” septic or neurologic complication 45% of sublobar lung resections Adjuvant RT or RCT in all cases Mainly CT as neoadjuvant treatment 35% of subclavian artery involvement 31% of pNþ cases Neoadjuvant CT alone in all cases Only 8/30 (27%) with vertebral involvement RCT as only induction treatment 34% of NSCLC, 28% of sarcoma pN0 in all cases Systematic mediastinoscopy to exclude cN2þ Mainly RCT as neoadjuvant treatment 48% of 2-stage surgery
RCT ¼ radiochemotherapy;
RT ¼ radiotherapy.
induction therapy, analysis of DFS among patients with nonlung malignancies suggests that local resection and “control” of the primary could allow prolonged survival despite recurrences. Indeed, patients harboring malignant tumors exhibited relatively low DFS compared with OS, whereas DFS was almost similar to OS estimates in the NSCLC group. Thoracic spine procedures, especially when combined with anatomic pulmonary resection, are usually associated with significant perioperative morbidity, probably owing to patients’ comorbidities, complexity of the surgical approach and dissection (especially after induction therapy), extent of resection, and operation time. Indeed, other authors have reported early morbidity rates ranging from 28% to 56% with associated mortality of approximately 6% [20–23]. Despite high rates of en bloc resections and hemivertebrectomies or total vertebrectomies, procedures theoretically associated with higher DFS but likely requiring longer intervention time, we currently report a 32% rate of postoperative major adverse events and no in-hospital deaths. Notably, we recorded only 1 patient requiring early surgical reoperation for hemothorax, whereas some authors reported a rate of required postoperative surgical procedure of up to 21% for treatment of complications such as chylothorax, bronchus fistula, and cerebrospinal fluid leakage [22]. Similar to this latest study, we failed to identify any factor associated with reoperation or other major adverse events in the early postoperative period.
GENERAL THORACIC
Patients (No.)
First Author [Reference]
GENERAL THORACIC
234
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
The observed 16.1% neurologic deficit rate, excluding “necessary” resections of nerve roots involved by the tumor, also corroborates results from similar cohorts reporting up to 25% of deficits in the early postoperative period and suggests the high functional impact of such a surgery, above all in the light of long-term follow-up showing that only 1 patient recovered from his neurologic dysfunction [20, 21, 23]. The outcome of late complications of thoracic spinal resections and their management is poorly described in the current literature. Here, we reported an acceptable rate of orthopedic functional impairment mostly using a single posterior incision. A surgical approach to perform spinal stabilization is still debated, with some surgeons advocating double incisions and a prone position for performing the vertebral portion [16, 20].
Limitations As suggested before, this analysis is retrospective and based on a relatively small subset of patients owing to the rarity of such conditions. It is therefore subject to several biases, in particular, the selection of cases or relative heterogeneity of the cohort, likely to have influenced and induced “counterintuitive” results, such as the absence of a significant prognostic impact of resection completeness, percentage of viable tumor cells after induction, or lymph node involvement. In addition, the few events that occurred did not allow us to perform multivariate analyses or identification of independent factors influencing survival or early and late morbidity. To understand more clearly such satisfactory survival, above all in NSCLC patients, our results probably have to be put in perspective with a broad spectrum of surgical series published in the literature.
Conclusion En bloc resections of thoracic tumors invading the spine allow the achievement of long-term survival with a low recurrence rate in highly selected patients often included in a multimodal strategy of care. However, these surgical procedures are associated with significant early-onset morbidity and sometimes the occurrence of very lateonset complications requiring intensive treatment and a long follow-up.
References 1. Dartevelle PG, Chapelier AR, Macchiarini P, et al. Anterior transcervical-thoracic approach for radical resection of lung tumors invading the thoracic inlet. J Thorac Cardiovasc Surg 1993;105:1025–34. 2. Maggi G, Casadio C, Pischedda F, et al. Combined radiosurgical treatment of Pancoast tumor. Ann Thorac Surg 1994;57:198–202. 3. Muscolino G, Valente M, Andreani S. Pancoast tumors: clinical assessment and long-term results of combined radiosurgical treatment. Thorax 1997;52:284–6. 4. Chadeyras JB, Mazel C, Grunenwald D. Vertebral en bloc resection for lung cancer: twelve years’ experience. Ann Chir 2006;131:616–22.
Ann Thorac Surg 2019;108:227–34
5. Sanli A, Onen A, Y€ ucesoy K, et al. Surgical treatment in nonsmall cell lung cancer invading to the chest wall (T3) and vertebra (T4). Tuberk Toraks 2007;55:383–9. 6. Yokomise H, Gotoh M, Okamoto T, et al. En bloc partial vertebrectomy for lung cancer invading the spine after induction chemoradiotherapy. Eur J Cardiothorac Surg 2007;31:788–90. 7. Anraku M, Waddell TK, de Perrot M, et al. Induction chemoradiotherapy facilitates radical resection of T4 non-small cell lung cancer invading the spine. J Thorac Cardiovasc Surg 2009;137:441–7. 8. Schirren J, D€ onges T, Melzer M, Sch€ onmayr R, Eberlein M, B€ ol€ ukbas S. En bloc resection of non-small-cell lung cancer invading the spine. Eur J Cardiothorac Surg 2011;40: 647–54. 9. Deslauriers J, Tronc F, Fortin D. Management of tumors involving the chest wall including Pancoast tumors and tumors invading the spine. Thorac Surg Clin 2013;23: 313–25. 10. Yokoi K, Taniguchi T, Usami N, Kawaguchi K, Fukui T, Ishiguro F. Surgical management of locally advanced lung cancer. Gen Thorac Cardiovasc Surg 2014;62:522–30. 11. Sakanoue I, Hamakawa H, Onishi E, Imai Y, Takahashi Y. Giant cell tumor of the rib with direct invasion into the thoracic spine. Gen Thorac Cardiovasc Surg 2017;65:293–6. 12. Koizumi K, Haraguchi S, Hirata T, et al. Surgical treatment of lung cancer with vertebral invasion. Ann Thorac Cardiovasc Surg 2004;10:229–34. 13. Fadel E, Missenard G, Court C, et al. Long-term outcomes of en bloc resection of non-small cell lung cancer invading the thoracic inlet and spine. Ann Thorac Surg 2011;92: 1024–30. 14. Maneenil K, Xue Z, Liu M, Boland J. Sarcomatoid carcinoma of the lung: the Mayo Clinic experience in 127 patients. Clin Lung Cancer 2018;19:e323–33. 15. Provost B, Missenard G, Pricopi C. Ewing sarcoma of the chest wall: prognostic factors of multimodal therapy including en bloc resection. Ann Thorac Surg 2018;106: 207–13. 16. Grunenwald DH, Mazel C, Girard P. Radical en bloc resection for lung cancer invading the spine. J Thorac Cardiovasc Surg 2002;123:271–9. 17. Fadel E, Missenard G, Chapelier A, et al. En bloc resection of non-small cell lung cancer invading the thoracic inlet and intervertebral foramina. J Thorac Cardiovasc Surg 2002;123: 676–85. 18. Watanabe SI, Nakagawa K, Suzuki K. Neoadjuvant and adjuvant therapy for stage III non-small cell lung cancer. Lung Cancer Surgical Study Group (LCSSG) of the Japan Clinical Oncology Group (JCOG). Jpn J Clin Oncol 2017;47: 1112–8. 19. Chen Y, Peng X, Zhou Y. Comparing the benefits of chemoradiotherapy and chemotherapy for resectable stage III A/ N2 non-small cell lung cancer: a meta-analysis. World J Surg Oncol 2018;16:8. 20. Mody GN, Bravo I~ niguez C, Armstrong K, et al. Early surgical outcomes of en bloc resection requiring vertebrectomy for malignancy invading the thoracic spine. Ann Thorac Surg 2016;101:231–6. 21. Dartevelle PG, Mitilian D, Fadel E. Extended surgery for T4 lung cancer: a 30 years’ experience. Gen Thorac Cardiovasc Surg 2017;65:321–8. 22. Collaud S, Waddell TK, Yasufuku K, et al. Long-term outcome after en bloc resection of non-small-cell lung cancer invading the pulmonary sulcus and spine. J Thorac Oncol 2013;8:1538–44. 23. Collaud S, Fadel E, Schirren J, et al. En bloc resection of pulmonary sulcus non-small cell lung cancer invading the spine: a systematic literature review and pooled data analysis. Ann Surg 2015;262:184.
Background. Vertebral involvement by a thoracic tumor has long been considered as a limit to surgical treatment, and despite advances, such an invasive operation remains controversial. The aim of this study was to characterize a single-center cohort and to evaluate the outcome, focusing on survival and complications. Methods. We retrospectively reviewed the data of all patients operated on for tumors involving the thoracic spine in an 8-year period. En bloc resection was generally performed by a double team involving thoracic and orthopedic surgeons. Distant follow-up was recorded for oncologic and functional analysis. Results. There were 31 patients operated on. An induction therapy was administered in 20 patients. Spinal resection (mostly including ‡2 vertebral levels) was combined with lobectomy in 48.3% of the patients, and osteosynthesis was required in 22 patients. We observed no in-hospital death and a major complications rate of 32.3%, including 5 patients with early neurologic com-
plications. There were 61.3% primary lung carcinomas, 12.9% extrapulmonary primaries, 9.7% metastases, and 16.1% benign tumors. Mean follow-up was 32.1 months. The 5-year overall survival rate was 81.3% in the entire cohort and 75.0% in patients with a malignant tumor. Occurrence of an early postoperative major complication was the only factor significantly associated with shorter overall survival (p [ 0.03). The 5-year disease-free survival rate was 37.0% in malignancies. Delayed complications occurred in 35.5% of patients, including persistent neurologic deficit in 12.9%, instrumentation migration in 19.4%, and local infection in 12.9%. Conclusions. En bloc resection of spinal thoracic tumors offers long-term survival and few recurrences in highly selected patients but is associated with significant delayed mechanical or infectious complications.
V
operative period [4–11]. However, the selection and eligibility criteria, as well as the type of multimodal treatment used in these different cohorts, are very heterogeneous, and a real consensus has not yet been reached. The aim of this study was to characterize a singlecenter cohort and to evaluate the short-term and longterm outcomes of resected thoracic tumors involving the spine in the setting of a multidisciplinary approach including thoracic surgeons, spine surgeons, medical oncologists, and pathologists. A secondary objective was to analyze the occurrence of adverse events related to such heavy surgical procedures, focusing on delayed complications and their management.
ertebral involvement by a thoracic tumor has long been considered as a limit to surgical treatment, and rightly so, because of the complexity of the posterior mediastinum anatomy and the poor short-term and longterm prognosis that was usually associated with the resection of such tumors [1–3]. With the progress of surgical techniques, the improvement of preoperative workups, and especially, the evolution of multidisciplinary care, including systemic treatments and radiotherapy for malignant lesions, some previously disqualified patients are now considered eligible for curative-intent surgical procedure. In this regard, some centers reported their experience and favorable oncologic results, although mostly restricted to the early post-
(Ann Thorac Surg 2019;108:227–34) Ó 2019 by The Society of Thoracic Surgeons
Accepted for publication Feb 8, 2019. *Drs Zhang and Fournel contributed equally to this work. Address correspondence to Dr Alifano, Department of Thoracic Surgery, Paris Center University Hospital, 27, rue de Faubourg Saint Jacques, 75014 Paris, France; email: [email protected].
Ó 2019 by The Society of Thoracic Surgeons Published by Elsevier Inc.
The Video can be viewed in the online version of this article [https://doi.org/10.1016/j.athoracsur.2019.02.019] on http://www.annalsthoracicsurgery.org.
0003-4975/$36.00 https://doi.org/10.1016/j.athoracsur.2019.02.019
GENERAL THORACIC
Departments of Thoracic Surgery, Pathology, and Orthopedic Surgery, Paris-Center University-Hospital, Assistance PubliqueHopitaux de Paris, Paris-Descartes University, Paris, France; and Department of Thoracic Surgery, Shanghai General Hospital affiliate to Shanghai Jiao Tong University, Shanghai, China
228
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Ann Thorac Surg 2019;108:227–34
Table 1. Preoperative Patients’ Characteristics
GENERAL THORACIC
Variables Sex Male Female Smoking Current Former >3 months No Major comorbidity Chronic obstructive pulmonary disease Ischemic heart disease Previous cancer Neoadjuvant treatment Radiotherapy Chemotherapy Radiochemotherapy None Response to neoadjuvant therapy Partial response Stable disease
Table 2. Perioperative Course and Pathologic Analysis No. (%) (n ¼ 31) 23 (74.2) 8 (25.8) 14 (45.2) 5 (16.1) 12 (38.7) 4 (12.9) 4 (12.9) 3 (9.7) 1 (3.2) 5 (16.1) 14 (45.2) 11 (35.5) (n ¼ 20) 11 (55) 9 (45)
No. ¼ number.
Material and Methods Ethics We retrospectively reviewed clinical records of all patients operated on in a curative intent from 2011 to 2018 in our Department of Thoracic Surgery at Cochin Hospital for a tumor with vertebral extension or development (at least at imaging and macroscopic intraoperative assessment). This study was performed in accordance with principles outlined in the Declaration of Helsinki and in agreement with French laws on biomedical research. The Comit e d’Ethique de la Recherche Clinique–Soci et e Franc¸aise de Chirurgie Thoracique et Cardio-Vasculaire approved the study (CERC-SFCTCV-2018-3-6-11-33-4-folu).
Preoperative Workup and Multimodal Management In addition to a physical examination, the preoperative workup generally included thoracoabdominal and cerebral computed tomography (CT), bronchoscopy, 18fluodeoxyglucose positron emission tomography, and evaluation of respiratory function by spirometry. The use of magnetic resonance imaging was systematic in the presence of close contact with the intervertebral foramina or medullary canal. A preoperative biopsy to obtain the histologic type of the tumor was most often requested, especially in patients who were likely to receive neoadjuvant therapy. All patients were discussed beforehand case-by-case at a multidisciplinary thoracic oncology meeting. The eligibility for surgical resection mostly relied on the patient’s general condition, the histologic type of the tumor, and its
Variables Type of pulmonary resection Lobectomy Segmentectomy Wedge Type of vertebral resection (largest resection considered) Total vertebrectomy Hemivertebrectomy Partial (transverse process and vertebral body wedge resection) Number of operated-on vertebral levels 1 2 3 Number of operated-on rib levels 0 1 2 3 4 Early complications Pneumonia Neurologic deficit Hemorrhage Early reintervention Mechanical ventilation Other Lung carcinoma Squamous cell carcinoma Adenocarcinoma Large cell neuroendocrine carcinoma Non-small cell carcinoma Sarcoma Osteosarcoma Ewing sarcoma Synovial sarcoma Other malignant tumor B3 thymoma Benign tumor Neurinoma Hemangioma pT status T2a,b T3 T4 pN status N0 N1 N2
No. (%) (n ¼ 31) 15 (48.3) 4 (12.9) 6 (19.4)
3 (9.7) 19 (61.3) 9 (29.0)
16 (51.6) 14 (45.2) 1 (3.2) 9 4 6 9 3
(29) (12.9) (19.4) (29) (9.7)
5 (16.1) 5 (16.1) 1 (3.2) 1 (3.2) 6 (19.4) 7 (22.6) 19 (61.3) 8 (25.8) 6 (19.4) 1 (3.2) 4 (12.9) 6 (22.6) 4 (12.9) 1 (3.2) 1 (3.2) 1 (3.2) 1 (3.2) 5 (16.1) 4 (12.9) 1 (3.2) 19 (61.3) 5 (26.4) 7 (36.8) 7 (36.8) 19 (61.3) 16 (84.2) 1 (5.3) 2 (10.5) (Continued)
Ann Thorac Surg 2019;108:227–34
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Table 2. Continued Variables
No. (%) (n ¼ 31)
R status R0 R1 Presence of vascular or lymphatic emboli Yes No
31 29 2 19 6 13
(100) (90.6) (9.4) (61.3) (31.6) (68.4)
No. ¼ number.
tumor dose of 45 Gy, avoiding, if possible, the lung hilum and uninvolved mediastinum. Response to induction therapy was evaluated according to Response Evaluation Criteria In Solid Tumors (RECIST) criteria, and a stable or responsive disease was required to confirm surgical indication, whereas a progressive disease was considered as a contraindication.
Technical Considerations The rules reported below were generally followed (Video): En bloc resections were performed by a multidisciplinary team including orthopedic and thoracic surgeons. In case of primary lung cancer, lobectomy associated with systemic lymph node dissection was the operation of choice, with sublobar resection and pneumonectomy being performed in case of necessity (poor cardiorespiratory function, or conversely, lobectomy not allowing a satisfactory oncological resection). Sublobar resections were usually Fig 1. Vertebral mapping of resected tumors. Each thoracic (T) spinal level is labeled with the number (N) and mean size of tumors involving it. In the entire series, some tumors involved more than 1 level.
GENERAL THORACIC
anatomical borders, which should anticipate the possibility of complete resection. The general policy of our team is to prescribe induction treatments for primary lung cancer involving the spine, especially in the case of superior sulcus tumors. For other primaries, decision for induction was based on anticipated sensitivity of the known malignancy to radiotherapy and chemotherapy. Induction in non-small cell lung cancer (NSCLC) consisted of platinum-based chemoradiotherapy with a
229
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Ann Thorac Surg 2019;108:227–34
GENERAL THORACIC
230
Fig 2. Overall survival (OS) in (A) all study patients, (B) non-small cell cancer (NSCLC) primary lung tumors, and (C) benign (curve 1) versus malignant tumors (curve 2). Disease-free survival (DFS) in (D) all study patients, (E) primary lung tumors, and (F) benign (curve 1) versus malignant tumors (curve 2). The vertical line indicates the 95% confidence interval at 60 months.
performed in extrapulmonary primaries invading both spine and lung. Posterolateral thoracotomy was performed with the patient positioned in Paulson lateral decubitus and lying anteriorly at 45 degrees on the operating table. The vertebral resection method depended on the degree of tumor invasion. In cases of infiltration of the posterior arch of the rib with contact of the transverse process (even if minimally infiltrated), the transverse process was resected; in case of larger infiltration of transverse process, intervertebral foramen (or foramina), vertebral lamina, or vertebral body or disk at maximum 1 cm within the vertebral midline, an hemivertebrectomy was performed. In case of more extensive involvement of vertebral body, a total vertebrectomy was performed. Hemivertebrectomy and total vertebrectomy could both be performed in more than 1 spinal level, 3 being the maximum number accepted. Spinal resection and stabilization was performed with a posterior midline incision without changing the patient’s position. After the pedicle screws were positioned, a laminectomy of the involved side was
performed. In case of total vertebrectomy, bilateral instrumentation was associated with a titanium cage filled with fragments of free bone graft. Nerve roots and accompanying segmental vessels were exposed and ligated inside the spinal canal to prevent cerebrospinal fluid leakage. These maneuvers allowed gentle traction of the dural sac and en bloc extraction of the specimen through the thoracotomy. Spinal rods were then implanted and secured on both sides.
Statistical Analysis Demographic data and major comorbidities of the patients were collected and analyzed as well as the radiologic evaluation of the tumor, the type of surgical intervention and associated multimodal treatment, the immediate operative follow-up, and pathologic features. Descriptive statistics are expressed as mean SD or median (range) for quantitative variables and as frequencies (percentages) for categorical variables. The follow-up was performed by interrogation of referring or family physicians and city registers. Survival rates,
Ann Thorac Surg 2019;108:227–34
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Table 3. Risk Factors Analysis Variable
p Value 0.4
60.7 (21.1–100.3) 73.4 (62.1–84.8) 0.9 70.8 (54.1–85.7) 74.9 (61.3–88.6) 0.6 74.8 (55.7–94.1) 46.2 (28.7–63.7) 0.69 74.5 (57.0–91.9) 72.7 (59.4–86.0) 0.39 64.8 (43.6–86.1) 39.2 (9.9–68.5) 0.09 74.9 (63.6–86.1) 16.9 (0.0–36.8) 0.11
14 (45.2%), and 3 levels in 1 (3.2%) and required spinal osteosynthesis in 22 patients (71.0%; Fig 1). The median number of resected ribs was 2 (range, 0 to 4). There were no in-hospital deaths. In-hospital stay was uneventful in 15 patients (48.4%), with a mean postoperative duration of stay of 13.9 9.5 days. Major complications occurred in 10 patients (32.3%), with 5 cases of early postoperative neurologic disorders (including phrenic nerve palsy), in particular. Notably, 1 patient who underwent repeat resection of a cervicothoracic compressive schwannoma (after initial medullary decompression performed urgently in another institution some weeks previously) experienced early postoperative spinal shock with persistent paraplegia. We also recorded 1 patient with a dural wound, followed by symptomatic cerebrospinal fluid leakage into the right pleura and pneumocephalus, with recovery at postoperative day 20 after medical therapy, including intravenous acetazolamide. One patient underwent a repeat operation at postoperative day 1 for diffuse bleeding that led to a major hemothorax. Other complications were minor and included medically treated events such as a urinary tract infection or an intestinal subocclusive syndrome. Univariate analysis did not reveal any factors significantly associated with occurrence of major complication.
Pathologic Analysis
70.5 (45.2–95.8) 14.0 (10.1–17.9) 0.03 38.9 (20.2–57.6) 81.3 (73.4–89.2)
CI ¼ confidence interval.
including non–cancer-related deaths, were calculated by the Kaplan-Meier method and compared by log-rank tests.
Results Study Population Characteristics Thirty-one patients who met the inclusion criteria were identified in the study period. Main patients’ characteristics are reported in Table 1. At the preoperative evaluation, 26 patients exhibited a malignant tumor, including 19 NSCLCs, whereas the 5 remaining patients had a benign disease. An induction therapy was performed in 20 patients, allowing partial response in 11 (55%) or stable disease in 9 (45%) at reevaluation. In patients who received radiotherapy, the median irradiation dose was 45 Gy (range, 30 to 56 Gy).
Operative Procedure and In-Hospital Course As summarized in Table 2, a pulmonary resection was performed in 80.6% of patients and was associated with spinal resection for en bloc tumor removal. Vertebral resection involved 1 level in 16 patients (51.6%), 2 levels in
Pathologic analysis of resected specimens revealed a majority of primary lung carcinomas (61.3%) and 5 benign tumors (16.1%; Table 2). Focusing on the T classification of the 19 NSCLCs, we observed a postinduction downstaging in 63.2% of patients. There were 2 patients with mediastinal lymph nodes involvement. In 1 squamous cell carcinoma of the lung and 1 benign tumor, resection was classified R1 because of positive margins at the foraminal level. Analyzing the response to neoadjuvant treatment, we observed a mean percentage of viable tumor cells as 17.1% 23.1% of the whole tumor area among all postinduction specimens, with complete response (no residual disease) in 4 patients (20%).
Prognosis and Delayed Complications Mean follow-up was 32.1 23.7 months. During the study time frame, 4 patients died, including 1 non–cancerrelated death 2 months after hospital discharge. All patients who died were smokers and had NSCLC. A Kaplan-Meier analysis showed that 5-year and 7-year survival rates were similar in each subgroup and equal to 81.3% (95% confidence interval [CI], 59.9% to 92.7%) in the entire population, 75.0% (95% CI, 48.9% to 90.4%) in those with malignant tumors, and 61.3% (95% CI, 30.6% to 85.4%) in those with NSCLC (Figs 2A–C). All patients with a benign disease were alive during follow-up (range, 24 to 83 months). Log-rank tests revealed that occurrence of major complications in the early postoperative period (p ¼ 0.03), analyzed as a binary categorical variable, was the only factor significantly associated with shorter overall survival (OS). There was no significant impact on OS of such factors as age, chronic obstructive pulmonary
GENERAL THORACIC
Chronic obstructive pulmonary disease Yes No Age 60 years >60 years Radiologic response to induction therapy Partial response Stable disease Number of resected vertebral levels 1 >1 Pathologic N status N0 Nþ Pathologic R status R0 R1 Viable tumor cells after induction therapy 10% >10% Occurrence of early major complication Yes No
Mean Survival Estimates Months (95% CI)
231
232
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
Ann Thorac Surg 2019;108:227–34
Table 4. Case Series of Patients With Long-Term Complications
Patient 1
GENERAL THORACIC
2
3
Type of Vertebral Resection
Late/Chronic Complication
Postoperative Time to Diagnosis of Complication (months)
Management
24
Instrumentation replacement
10
Extension of osteosynthesis
11
Local debridement and coverage with latissimus dorsi muscle flap Rehabilitation
T1-T2 Pedicle screw loosening and hemivertebrectomy þ 3 displacement ribs T1-T2 total Traumatic migration of spinal vertebrectomy þ 3 ribs instrumentation and neurologic deficit Secondary acute deep infection C7-T4 laminectomy þ T2 Persistent T6 paraplegia hemivertebrectomy þ 2 ribs Chronic restrictive respiratory failure
0
0
Chronic noninvasive ventilation Local debridement and coverage with paraspinal muscle flap Local debridement and coverage with trapezius muscle flap Rehabilitation Superficial debridement and vacuum therapy Rehabilitation
4
Instrument replacement
4
T2-T3 hemivertebrectomy Superficial and deep infection
57
5
T2-T3 Wound disunion and deep infection hemivertebrectomy þ 1 rib T5-T6 hemivertebrectomy Lower-back sensitive neurologic deficit T2 hemivertebrectomy Asymptomatic pedicle screw loosening superficial wound infection T3-T4 partial Mixed chronic respiratory failure with vertebrectomy phrenic nerve palsy T3-T4 hemivertebrectomy Asymptomatic pedicle screw migration out of the vertebra with major increase of kyphosis T1–3 hemivertebrectomy Asymptomatic dural screw migration without cord compression T1-T2 partial C8 sensitive deficit vertebrectomy
19
6 7 8 9
10 11
disease status, completeness of resection, or pN status (Table 3). Relapse of the disease occurred in 7 patients (22.6%) during the follow-up period, with 3 (42.9%) thoracic and 4 (57.1%) extrathoracic first recurrences. The 5-year disease-free survival (DFS) rates were 50.5% (95% CI, 30.0% to 74.8%) in the whole population, 37.0% (95% CI, 19.5% to 67.2%) in those with malignant tumors, and 61.1% (95% CI, 30.8% to 87.7%) in those with NSCLC (Figs 2D–F). At late follow-up, 11 patients (35.5%) experienced single or multiple distant complications. Permanent neurologic dysfunction occurred in 4 patients (12.9%) despite rehabilitation, and a displacement of the osteosynthesis hardware occurred in 6 patients (19.4%). Among these patients, half required a hospital readmission and reoperation for instrumentation replacement and spine stabilization. The occurrence of this late complication was significantly higher in patients with chronic obstructive pulmonary disease (p ¼ 0.02) and tumors with nerve root involvement (p ¼ 0.002). One of the 5 patients who experienced early postoperative neurologic complications completely recovered from a C8-T1 sensitive deficit. Four patients (12.9%) required readmission for late-onset
0 5
10 0
Surveillance Rehabilitation
infections, consisting of 1 superficial and 3 organ-space infections contiguous to the spinal instrumentation. At the study end point, all of the patients, except for those with late-onset infections, were successfully and conservatively treated, without implant removal, particularly. The history and management of late or persistent complications is detailed in Table 4.
Comment Despite advances in preoperative work-up and multimodal approach, surgical selection of patients with tumors invading the thoracic spine remains challenging because of the delicate regional anatomy, potential functional impact, or prognostic considerations. In the present study, we aimed to address these issues and showed that careful preoperative workup and doubleteam (thoracic and orthopedic) en bloc resection could allow safe and reliable surgical management, with no inhospital mortality and high long-term survival rates, even when excluding benign tumors. Such satisfactory oncologic results underline the paramount role of patient selection, neoadjuvant/adjuvant treatments, and surgical approach.
Ann Thorac Surg 2019;108:227–34
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
233
Table 5. Prognosis of Patients After Vertebral Procedures for Resection of Thoracic Malignancies in Relevant Studies Induction Treatment (%)
Chadeyras [4] Yokomise [6] Anraku [7]
32 6 23
66 100 100
NSCLC NSCLC NSCLC
31 33 48
13 0 17
65/24 68/68 58/40
Schirren [8]
20
30
NSCLC
40
20
z75/47
Fadel [13]
54
50
NSCLC
55
9
47/31
Provost [15]
30
97
Ewing sarcoma
27
7
z82/61
Mody [20]
32
78
Malignancy
56
6
z60/40
Collaud [23]
48
98
NSCLC
12
z73/61
CT ¼ chemotherapy;
No. ¼ number;
Major Histologic Complication Rate Rþ Type (%) (%)
38 (prolonged intensive care unit stay)
2-Year/5-Year Survival Rates (%)
NSCLC ¼ non-small cell lung carcinoma;
As shown in Table 5, the heterogeneity of published surgical cohorts of chest malignancies invading the spine led to a wide range of long-term survivals, suggesting that significant biases could have influenced our results, which slightly exceed the upper limit of the range [4, 6, 12, 13]. Like most authors in this research field, we acknowledge the bias of analyzing a small number of patients with few events occurring; thus, a part of these good survival results could be attributed to hazard. With respect to patient selection, we performed an extensive preoperative workup, similar to other teams, to avoid intraoperative discovery of unresectable disease and decrease the rate of incomplete resections [9,12–17]. Despite accurate preoperative imaging, the en bloc pattern of all resections, and a large subset of patients who received an induction therapy, we did not achieve completeness of resection in 2 patients, including 1 benign tumor. Nevertheless, we observed a high rate of R0 resections (93.5%) at pathologic evaluation compared with previous reports. This could explain the relatively high survival rates that we obtained, even though we failed to reveal a statistically significant association between R classification and OS [10, 13]. In the present series, 76.9% of patients with a malignant disease received an induction therapy. As outlined in previous studies, induction therapy is likely to favorably influence prognosis by improving patient selection, increasing the rate of R0 resections, and driving postoperative multimodal management with assessment of the pathologic response [15, 18, 19]. Independently from
Comments Mainly CT as neoadjuvant treatment 83.3% of major pathologic response rate 61% of 2-stage surgery (1 or 2 weeks interval) 72% of 3-year survival in R0 patients Upfront surgery in patients with “imminent” septic or neurologic complication 45% of sublobar lung resections Adjuvant RT or RCT in all cases Mainly CT as neoadjuvant treatment 35% of subclavian artery involvement 31% of pNþ cases Neoadjuvant CT alone in all cases Only 8/30 (27%) with vertebral involvement RCT as only induction treatment 34% of NSCLC, 28% of sarcoma pN0 in all cases Systematic mediastinoscopy to exclude cN2þ Mainly RCT as neoadjuvant treatment 48% of 2-stage surgery
RCT ¼ radiochemotherapy;
RT ¼ radiotherapy.
induction therapy, analysis of DFS among patients with nonlung malignancies suggests that local resection and “control” of the primary could allow prolonged survival despite recurrences. Indeed, patients harboring malignant tumors exhibited relatively low DFS compared with OS, whereas DFS was almost similar to OS estimates in the NSCLC group. Thoracic spine procedures, especially when combined with anatomic pulmonary resection, are usually associated with significant perioperative morbidity, probably owing to patients’ comorbidities, complexity of the surgical approach and dissection (especially after induction therapy), extent of resection, and operation time. Indeed, other authors have reported early morbidity rates ranging from 28% to 56% with associated mortality of approximately 6% [20–23]. Despite high rates of en bloc resections and hemivertebrectomies or total vertebrectomies, procedures theoretically associated with higher DFS but likely requiring longer intervention time, we currently report a 32% rate of postoperative major adverse events and no in-hospital deaths. Notably, we recorded only 1 patient requiring early surgical reoperation for hemothorax, whereas some authors reported a rate of required postoperative surgical procedure of up to 21% for treatment of complications such as chylothorax, bronchus fistula, and cerebrospinal fluid leakage [22]. Similar to this latest study, we failed to identify any factor associated with reoperation or other major adverse events in the early postoperative period.
GENERAL THORACIC
Patients (No.)
First Author [Reference]
GENERAL THORACIC
234
ZHANG ET AL THORACIC TUMORS INVADING THE SPINE
The observed 16.1% neurologic deficit rate, excluding “necessary” resections of nerve roots involved by the tumor, also corroborates results from similar cohorts reporting up to 25% of deficits in the early postoperative period and suggests the high functional impact of such a surgery, above all in the light of long-term follow-up showing that only 1 patient recovered from his neurologic dysfunction [20, 21, 23]. The outcome of late complications of thoracic spinal resections and their management is poorly described in the current literature. Here, we reported an acceptable rate of orthopedic functional impairment mostly using a single posterior incision. A surgical approach to perform spinal stabilization is still debated, with some surgeons advocating double incisions and a prone position for performing the vertebral portion [16, 20].
Limitations As suggested before, this analysis is retrospective and based on a relatively small subset of patients owing to the rarity of such conditions. It is therefore subject to several biases, in particular, the selection of cases or relative heterogeneity of the cohort, likely to have influenced and induced “counterintuitive” results, such as the absence of a significant prognostic impact of resection completeness, percentage of viable tumor cells after induction, or lymph node involvement. In addition, the few events that occurred did not allow us to perform multivariate analyses or identification of independent factors influencing survival or early and late morbidity. To understand more clearly such satisfactory survival, above all in NSCLC patients, our results probably have to be put in perspective with a broad spectrum of surgical series published in the literature.
Conclusion En bloc resections of thoracic tumors invading the spine allow the achievement of long-term survival with a low recurrence rate in highly selected patients often included in a multimodal strategy of care. However, these surgical procedures are associated with significant early-onset morbidity and sometimes the occurrence of very lateonset complications requiring intensive treatment and a long follow-up.
References 1. Dartevelle PG, Chapelier AR, Macchiarini P, et al. Anterior transcervical-thoracic approach for radical resection of lung tumors invading the thoracic inlet. J Thorac Cardiovasc Surg 1993;105:1025–34. 2. Maggi G, Casadio C, Pischedda F, et al. Combined radiosurgical treatment of Pancoast tumor. Ann Thorac Surg 1994;57:198–202. 3. Muscolino G, Valente M, Andreani S. Pancoast tumors: clinical assessment and long-term results of combined radiosurgical treatment. Thorax 1997;52:284–6. 4. Chadeyras JB, Mazel C, Grunenwald D. Vertebral en bloc resection for lung cancer: twelve years’ experience. Ann Chir 2006;131:616–22.
Ann Thorac Surg 2019;108:227–34
5. Sanli A, Onen A, Y€ ucesoy K, et al. Surgical treatment in nonsmall cell lung cancer invading to the chest wall (T3) and vertebra (T4). Tuberk Toraks 2007;55:383–9. 6. Yokomise H, Gotoh M, Okamoto T, et al. En bloc partial vertebrectomy for lung cancer invading the spine after induction chemoradiotherapy. Eur J Cardiothorac Surg 2007;31:788–90. 7. Anraku M, Waddell TK, de Perrot M, et al. Induction chemoradiotherapy facilitates radical resection of T4 non-small cell lung cancer invading the spine. J Thorac Cardiovasc Surg 2009;137:441–7. 8. Schirren J, D€ onges T, Melzer M, Sch€ onmayr R, Eberlein M, B€ ol€ ukbas S. En bloc resection of non-small-cell lung cancer invading the spine. Eur J Cardiothorac Surg 2011;40: 647–54. 9. Deslauriers J, Tronc F, Fortin D. Management of tumors involving the chest wall including Pancoast tumors and tumors invading the spine. Thorac Surg Clin 2013;23: 313–25. 10. Yokoi K, Taniguchi T, Usami N, Kawaguchi K, Fukui T, Ishiguro F. Surgical management of locally advanced lung cancer. Gen Thorac Cardiovasc Surg 2014;62:522–30. 11. Sakanoue I, Hamakawa H, Onishi E, Imai Y, Takahashi Y. Giant cell tumor of the rib with direct invasion into the thoracic spine. Gen Thorac Cardiovasc Surg 2017;65:293–6. 12. Koizumi K, Haraguchi S, Hirata T, et al. Surgical treatment of lung cancer with vertebral invasion. Ann Thorac Cardiovasc Surg 2004;10:229–34. 13. Fadel E, Missenard G, Court C, et al. Long-term outcomes of en bloc resection of non-small cell lung cancer invading the thoracic inlet and spine. Ann Thorac Surg 2011;92: 1024–30. 14. Maneenil K, Xue Z, Liu M, Boland J. Sarcomatoid carcinoma of the lung: the Mayo Clinic experience in 127 patients. Clin Lung Cancer 2018;19:e323–33. 15. Provost B, Missenard G, Pricopi C. Ewing sarcoma of the chest wall: prognostic factors of multimodal therapy including en bloc resection. Ann Thorac Surg 2018;106: 207–13. 16. Grunenwald DH, Mazel C, Girard P. Radical en bloc resection for lung cancer invading the spine. J Thorac Cardiovasc Surg 2002;123:271–9. 17. Fadel E, Missenard G, Chapelier A, et al. En bloc resection of non-small cell lung cancer invading the thoracic inlet and intervertebral foramina. J Thorac Cardiovasc Surg 2002;123: 676–85. 18. Watanabe SI, Nakagawa K, Suzuki K. Neoadjuvant and adjuvant therapy for stage III non-small cell lung cancer. Lung Cancer Surgical Study Group (LCSSG) of the Japan Clinical Oncology Group (JCOG). Jpn J Clin Oncol 2017;47: 1112–8. 19. Chen Y, Peng X, Zhou Y. Comparing the benefits of chemoradiotherapy and chemotherapy for resectable stage III A/ N2 non-small cell lung cancer: a meta-analysis. World J Surg Oncol 2018;16:8. 20. Mody GN, Bravo I~ niguez C, Armstrong K, et al. Early surgical outcomes of en bloc resection requiring vertebrectomy for malignancy invading the thoracic spine. Ann Thorac Surg 2016;101:231–6. 21. Dartevelle PG, Mitilian D, Fadel E. Extended surgery for T4 lung cancer: a 30 years’ experience. Gen Thorac Cardiovasc Surg 2017;65:321–8. 22. Collaud S, Waddell TK, Yasufuku K, et al. Long-term outcome after en bloc resection of non-small-cell lung cancer invading the pulmonary sulcus and spine. J Thorac Oncol 2013;8:1538–44. 23. Collaud S, Fadel E, Schirren J, et al. En bloc resection of pulmonary sulcus non-small cell lung cancer invading the spine: a systematic literature review and pooled data analysis. Ann Surg 2015;262:184.