-
PDF
- Split View
-
Views
-
Cite
Cite
Mathieu Lacquet, Johnny Moons, Laurens J Ceulemans, Paul De Leyn, Dirk Van Raemdonck, Surgery for mediastinal neurogenic tumours: a 25-year single-centre retrospective study, Interactive CardioVascular and Thoracic Surgery, Volume 32, Issue 5, May 2021, Pages 737–743, https://doi.org/10.1093/icvts/ivab002
- Share Icon Share
Abstract
Mediastinal neurogenic tumours are uncommon and often benign neoplasms mostly located in the posterior mediastinum and usually diagnosed incidentally. We reviewed our results after surgical resection. We compared patient characteristics and tumour nature between children and adults. Differences between thoracoscopic and open approach were analysed.
Departmental thoracic surgical database was queried for primary mediastinal neurogenic tumours resected between 1992 and 2017. Data included demographics, pathology, tumour nature, symptoms, surgical approach and postoperative morbidity/mortality.
Fifty-one patients (8 children and 43 adults) underwent tumour resection. Pathology revealed nerve sheath tumour in 1 child (12.5%) versus 36 adults (83.7%; P < 0.001) and ganglion cell tumour in 7 (87.5%) versus 5 (11.6%; P < 0.001). Two adults had a paraganglioma. Malignancy was present in 2 children (25%) versus 2 adults (4.6%; P = 0.049). All malignant tumours caused symptoms while most patients with benign tumours (38/47) were asymptomatic (P < 0.001). Surgical approach included thoracotomy, thoracoscopy and cervicotomy (n = 19/31/1) of which 2 were combined neurosurgical approach. All malignant tumours were approached via thoracotomy while the majority of patients with benign tumours (31/47) underwent thoracoscopy (P = 0.007). No significant difference was noted in overall morbidity between thoracoscopic versus open approach (45.2% vs 42.1%; P = 0.83). Hospital stay was significantly shorter following thoracoscopy (7.4 ± 3.3 vs 13.1 ± 9.8 days; P = 0.001).
Children carry a higher incidence to present with a malignant tumour originating from ganglion cells while most tumours in adults are benign, originating from the nerve sheath. The majority of patients with mediastinal neurogenic tumours are asymptomatic. Most tumours are amenable for thoracoscopic resection.
INTRODUCTION
Neurogenic tumours arise from tissues originating in the embryonic neural crest and can be further classified on whether these grow from nerve sheath, nerve cells (ganglia) or paraganglia. Among the group of nerve sheath tumours, schwannomas (neurilemmomas) and neurofibromas are most common. Malignant transition of these tumours is rare. Tumours of the autonomic ganglia exhibit a range of histological dedifferentiation, from completely benign (ganglioneuromas) to malignant (ganglioneuroblastomas and neuroblastomas). Many of these tumours can produce catecholamines. Paragangliomas can be either benign or malignant [1, 2]. Mediastinal neurogenic tumours (MNTs) may be associated with any neurogenic structure in the mediastinum. However, these are typically located in the posterior mediastinum near the costo-vertebral sulcus, in connection with the sympathetic chain, rami communicantes or intercostal nerves. Phrenic or vagal nerves are rarely involved. Paravertebral tumours are in close proximity to the neural foramina and can extend into the spinal canal, so-called ‘Dumbbell-sign’ tumours (Figs. 1–3) [3].
![Dumbbell-sign tumour in the left costo-vertebral sulcus [MRI (magnetic resonance imaging) coronal + transverse section] (adult case number 4).](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/icvts/32/5/10.1093_icvts_ivab002/2/m_ivab002f1.jpeg?Expires=1747872303&Signature=xleevsoHoOthiGd5sExnglfvKh33rHnXtTaWcGD2OhVZ37Y0X9agqZ9sAJlufS-2zJOcQd7towQLhMaIiI4WtLAND0KDR0ydXwhzQ~QrWgq0wuSsv55fj7AltzCrDhcJVaJD6Vs0DR1PpmRp3ChSxsAbb~q6s2OcSImWDl3RG~TJ6BWCGQ91Ot51XNjezEaTXra7y1kq9CzjTvkgQi-1hd8mWEK6OqoFwBge8XmV3BGLu~OOU8DFryhQCE1z2ygJfO48zinNTrPfDSobpj70QcD9t5~1zfXZnAaJatGAtZKSvl5003zCiYy-N5n1NZtRDl5TgE7tlz8o0Jwy9O6iGA__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Dumbbell-sign tumour in the left costo-vertebral sulcus [MRI (magnetic resonance imaging) coronal + transverse section] (adult case number 4).
![Dumbbell-sign tumour in the left costo-vertebral sulcus [CT (computed tomography) transverse section] (child case number 5).](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/icvts/32/5/10.1093_icvts_ivab002/2/m_ivab002f2.jpeg?Expires=1747872303&Signature=l4PMP6QMyIzcBwukrbxXhv7q0y9TQQkjpBbGqcdNi3akIDSRC1R0LwgBsjzL8rjfaR9v1GfCF4kdA9m6G2bJTW52-xs27cOf85loMXuxPAkFbqObqvdbujily7IAnIeEKQg-r09F5Cx6EDhh8rHbvOmGsIfa2IcSk2Xiu-pqtx931G5vKdvQwS8JcdFztIYmyRg7oA6zbD29TCzWY~bTWIvZjxh03e3RYk7CcKesLMYONfnW9yWVK24KkS6HanFK865II-kC39JWY5NqiU16QojDjbLSzCnzl0HZLuL-rqpESomCntlSUKCZ5Ww1goYwxPCVWp50J2PkN85bW3p2LQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Dumbbell-sign tumour in the left costo-vertebral sulcus [CT (computed tomography) transverse section] (child case number 5).

Dumbbell-sign tumour in the left costo-vertebral sulcus (CT coronal section) (child case number 5).
The actual incidence of these tumours is unknown but has been reported to account for 10–34% of all mediastinal tumours. Literature has shown that epidemiology of neurogenic tumours depends on whether the patient is an adult or a child [1].
Benign mediastinal neurogenic tumours tend to grow slowly. When the malignant forms predominate or in the presence of a large (benign) tumour, the symptoms can be quite marked [4]. Symptoms are mostly due to local growth and pressure on related structures and as such, these tumours present like chest wall pain, cough, dyspnoea, hoarseness, stridor, superior caval vein syndrome or muscle weakness.
Neurogenic tumours can be associated with neurofibromatosis type 1 (NF-1, ‘von Recklinghausen’s disease’) [5]. The presence of multiple neurofibromas is highly suggestive of this link. NF-1 is a genetic disorder causing tumours along the nervous system that can grow anywhere in the body. Common symptoms include Lisch noduli in the eye, cutaneous neurofibromas and ‘café au lait’ spots. The incidence of malignancy is higher compared with the general population.
The mainstay of treatment for MNTs, benign or malignant, is complete surgical resection. Hereby, symptoms due to local growth in the mediastinum can be prevented, diagnostic confirmation can be obtained and malignant degeneration can be avoided [2].
The aim of this study was to compare patient characteristics and tumour nature between children and adults with MNTs. Furthermore, we investigated possible differences in outcome between a thoracoscopic and open approach.
MATERIALS AND METHODS
Ethical statement
The study was approved by the KU Leuven Educational Supervisory Committee on Medical Ethics, delegated to the Ethics Committee for Research at the University Hospitals Leuven (MP001969) (6 November 2017).
The departmental thoracic surgical database of the University Hospitals Leuven, Belgium, was queried for patients with MNTs resected between 1992 and 2017.
Collected data included demographics, symptoms, size, location, surgical approach, postoperative morbidity and mortality and final pathology.
All patients who underwent resection of a primary MNT were included in this retrospective study. Patients with secondary neurogenic tumours (metastasis), tumours located outside the mediastinum (thoracic wall, axillary, subpleural, abdominal, brachial plexus, etc.), non-neurogenic mediastinal or pleural tumours (solitary fibromas, thymic tumours, bronchogenic cysts, Ewing sarcoma, etc.) and patients not referred for surgery were excluded.
Primary end-points were: (i) patient characteristics and tumour nature comparing children (age <18 years) versus adults and (ii) differences in outcome between surgical approach (thoracoscopy versus thoracotomy).
Follow-up was performed by attending surgeons or oncologists with a patient variable time interval. Follow-up data were collected for analysis up to July 2020. Clinical data and survival status were obtained by reviewing medical records. All patients had a registered follow-up of a least 1 month postoperative. In this period, we documented postoperative morbidity as already known or newly developed complications. Complications were graded according to the Clavien–Dindo classification [6]. Postoperative chest wall discomfort was evaluated in the first follow-up moment (1 month postoperative). It was documented as persistent disturbing (impaired daily life activities) chest wall pain and the need for painkillers.
Statistical analysis was performed using the Z-test, the Mann–Whitney U-test for continuous variables and the Chi-square test for categorical variables where appropriate. The difference was considered to be significant for any P-value <0.05.
RESULTS
Children versus adults
Between 1992 and 2017, 51 patients (8 children/43 adults; 27 females/24 males) underwent resection of a primary MNT. Two children (25%) versus 2 adults only (4.6%) presented with a malignant tumour (P = 0.049). Pathology revealed nerve sheath tumour in 1 child (12.5%) versus 36 adults (83.7%) (P < 0.001) and ganglion cell tumour in 7 (87.5%) versus 5 (11.6%) (P < 0.001). Two paragangliomas were seen in adults (1 benign and 1 malignant) (Fig. 4). Patient characteristics and tumour nature are compared between children and adults in Table 1. Postoperative complications are listed in Table 2. Nine patients (17.6%) were diagnosed with Horner’s syndrome postoperatively. All patients with a malignant tumour (4/4) presented with symptoms while the majority (38/47) of patients with benign tumours were asymptomatic (P < 0.001). In this study, 17/51 (33%) patients had a preoperative biopsy.

Tumour origin in children versus adults (children versus adults: P < 0.001 for nerve sheath tumour and for ganglion cell tumour; P = 0.54 for paraganglioma).
. | Children (<18 years), N = 8 . | % . | Adults, N = 43 . | % . | P-value . |
---|---|---|---|---|---|
Patient characteristics | |||||
Median age (years) | 6 ± 7 | – | 47 ± 13 | – | – |
Gender (♀/♂) | 5/3 | – | 22/21 | – | 0.56 |
Symptoms: yes/no | 3/5 | 37.5 | 10/33 | 23.3 | 0.40 |
Morbidity: number | 5 | 62.5 | 19 | 44.2 | 0.34 |
Perioperative mortality: yes/no | 0/8 | 0 | 2/41 | 4.7 | 0.54 |
Tumour nature | |||||
Malignancy: yes/no Median age (years) Pathology (number) | 2/6 0.17 ± 0.12 1 NB 1 GNB | 25 – – – | 2/41 38 ± 1.4 1 NS 1 PG | 4.6 – – – | 0.049 – – – |
Tumour cell origin (number) | |||||
Nerve sheath tumour Ganglion cell tumour Paraganglioma | 1 (1 SN) 7 (5 GN , 1 GNB , 1 NB) 0 | 12.5 87.5 0 | 36 (30 SN, 5 NF, 1 SN-NF) 5 (5 GN) 2 | 83.7 11.6 4.7 | <0.001 <0.001 0.54 |
. | Children (<18 years), N = 8 . | % . | Adults, N = 43 . | % . | P-value . |
---|---|---|---|---|---|
Patient characteristics | |||||
Median age (years) | 6 ± 7 | – | 47 ± 13 | – | – |
Gender (♀/♂) | 5/3 | – | 22/21 | – | 0.56 |
Symptoms: yes/no | 3/5 | 37.5 | 10/33 | 23.3 | 0.40 |
Morbidity: number | 5 | 62.5 | 19 | 44.2 | 0.34 |
Perioperative mortality: yes/no | 0/8 | 0 | 2/41 | 4.7 | 0.54 |
Tumour nature | |||||
Malignancy: yes/no Median age (years) Pathology (number) | 2/6 0.17 ± 0.12 1 NB 1 GNB | 25 – – – | 2/41 38 ± 1.4 1 NS 1 PG | 4.6 – – – | 0.049 – – – |
Tumour cell origin (number) | |||||
Nerve sheath tumour Ganglion cell tumour Paraganglioma | 1 (1 SN) 7 (5 GN , 1 GNB , 1 NB) 0 | 12.5 87.5 0 | 36 (30 SN, 5 NF, 1 SN-NF) 5 (5 GN) 2 | 83.7 11.6 4.7 | <0.001 <0.001 0.54 |
GN: ganglioneuroma; GNB: ganglioneuroblastoma; NB: neuroblastoma; NF: neurofibroma; NS: neurogenic sarcoma; PG: paraganglioma; SN: schwannoma; SN-NF: hybrid schwannoma–neurofibroma.
. | Children (<18 years), N = 8 . | % . | Adults, N = 43 . | % . | P-value . |
---|---|---|---|---|---|
Patient characteristics | |||||
Median age (years) | 6 ± 7 | – | 47 ± 13 | – | – |
Gender (♀/♂) | 5/3 | – | 22/21 | – | 0.56 |
Symptoms: yes/no | 3/5 | 37.5 | 10/33 | 23.3 | 0.40 |
Morbidity: number | 5 | 62.5 | 19 | 44.2 | 0.34 |
Perioperative mortality: yes/no | 0/8 | 0 | 2/41 | 4.7 | 0.54 |
Tumour nature | |||||
Malignancy: yes/no Median age (years) Pathology (number) | 2/6 0.17 ± 0.12 1 NB 1 GNB | 25 – – – | 2/41 38 ± 1.4 1 NS 1 PG | 4.6 – – – | 0.049 – – – |
Tumour cell origin (number) | |||||
Nerve sheath tumour Ganglion cell tumour Paraganglioma | 1 (1 SN) 7 (5 GN , 1 GNB , 1 NB) 0 | 12.5 87.5 0 | 36 (30 SN, 5 NF, 1 SN-NF) 5 (5 GN) 2 | 83.7 11.6 4.7 | <0.001 <0.001 0.54 |
. | Children (<18 years), N = 8 . | % . | Adults, N = 43 . | % . | P-value . |
---|---|---|---|---|---|
Patient characteristics | |||||
Median age (years) | 6 ± 7 | – | 47 ± 13 | – | – |
Gender (♀/♂) | 5/3 | – | 22/21 | – | 0.56 |
Symptoms: yes/no | 3/5 | 37.5 | 10/33 | 23.3 | 0.40 |
Morbidity: number | 5 | 62.5 | 19 | 44.2 | 0.34 |
Perioperative mortality: yes/no | 0/8 | 0 | 2/41 | 4.7 | 0.54 |
Tumour nature | |||||
Malignancy: yes/no Median age (years) Pathology (number) | 2/6 0.17 ± 0.12 1 NB 1 GNB | 25 – – – | 2/41 38 ± 1.4 1 NS 1 PG | 4.6 – – – | 0.049 – – – |
Tumour cell origin (number) | |||||
Nerve sheath tumour Ganglion cell tumour Paraganglioma | 1 (1 SN) 7 (5 GN , 1 GNB , 1 NB) 0 | 12.5 87.5 0 | 36 (30 SN, 5 NF, 1 SN-NF) 5 (5 GN) 2 | 83.7 11.6 4.7 | <0.001 <0.001 0.54 |
GN: ganglioneuroma; GNB: ganglioneuroblastoma; NB: neuroblastoma; NF: neurofibroma; NS: neurogenic sarcoma; PG: paraganglioma; SN: schwannoma; SN-NF: hybrid schwannoma–neurofibroma.
Postoperative complications . | Children (0–18 years), N = 8 . | Adults, N = 43 . | P-value . |
---|---|---|---|
Total number (%) | 5 (62.5 %) | 19 (44.2%) | p = 0.34 |
Type of complication (number) [Clavien–Dindo classification] | - Horner syndrome (3) [1] - Subcutaneous emphysema and pneumothorax (2) [1] | - Horner syndrome (5) [1] - Horner syndrome + paresis (brachial plexus) (1) [1] - Warm dry hand (3) [1] - Warm dry hand + axillary paresthaesia (1) [1] - Hoarseness (2) [1 + 2] - Galactorrhea (1) [2] - Unilateral elevation of the diaphragm (unilateral phrenic injury) (1) [1] - Atelectasis (2) [1] - Respiratory insufficiency (2) [5] - Liquorrhea (1) [1] |
Postoperative complications . | Children (0–18 years), N = 8 . | Adults, N = 43 . | P-value . |
---|---|---|---|
Total number (%) | 5 (62.5 %) | 19 (44.2%) | p = 0.34 |
Type of complication (number) [Clavien–Dindo classification] | - Horner syndrome (3) [1] - Subcutaneous emphysema and pneumothorax (2) [1] | - Horner syndrome (5) [1] - Horner syndrome + paresis (brachial plexus) (1) [1] - Warm dry hand (3) [1] - Warm dry hand + axillary paresthaesia (1) [1] - Hoarseness (2) [1 + 2] - Galactorrhea (1) [2] - Unilateral elevation of the diaphragm (unilateral phrenic injury) (1) [1] - Atelectasis (2) [1] - Respiratory insufficiency (2) [5] - Liquorrhea (1) [1] |
Clavien–Dindo classification: Grade I: any deviation from the normal postoperative course without the need for ‘pharmacological treatment’ or surgical, endoscopic and radiological interventions. Grade II: requiring pharmacological treatment with IV drugs. Grade III: requiring surgical, endoscopic or radiological intervention. Grade IV: life-threatening complication requiring IC/ICU (intensive care/intensive care unit) management. Grade V: death of a patient.
Postoperative complications . | Children (0–18 years), N = 8 . | Adults, N = 43 . | P-value . |
---|---|---|---|
Total number (%) | 5 (62.5 %) | 19 (44.2%) | p = 0.34 |
Type of complication (number) [Clavien–Dindo classification] | - Horner syndrome (3) [1] - Subcutaneous emphysema and pneumothorax (2) [1] | - Horner syndrome (5) [1] - Horner syndrome + paresis (brachial plexus) (1) [1] - Warm dry hand (3) [1] - Warm dry hand + axillary paresthaesia (1) [1] - Hoarseness (2) [1 + 2] - Galactorrhea (1) [2] - Unilateral elevation of the diaphragm (unilateral phrenic injury) (1) [1] - Atelectasis (2) [1] - Respiratory insufficiency (2) [5] - Liquorrhea (1) [1] |
Postoperative complications . | Children (0–18 years), N = 8 . | Adults, N = 43 . | P-value . |
---|---|---|---|
Total number (%) | 5 (62.5 %) | 19 (44.2%) | p = 0.34 |
Type of complication (number) [Clavien–Dindo classification] | - Horner syndrome (3) [1] - Subcutaneous emphysema and pneumothorax (2) [1] | - Horner syndrome (5) [1] - Horner syndrome + paresis (brachial plexus) (1) [1] - Warm dry hand (3) [1] - Warm dry hand + axillary paresthaesia (1) [1] - Hoarseness (2) [1 + 2] - Galactorrhea (1) [2] - Unilateral elevation of the diaphragm (unilateral phrenic injury) (1) [1] - Atelectasis (2) [1] - Respiratory insufficiency (2) [5] - Liquorrhea (1) [1] |
Clavien–Dindo classification: Grade I: any deviation from the normal postoperative course without the need for ‘pharmacological treatment’ or surgical, endoscopic and radiological interventions. Grade II: requiring pharmacological treatment with IV drugs. Grade III: requiring surgical, endoscopic or radiological intervention. Grade IV: life-threatening complication requiring IC/ICU (intensive care/intensive care unit) management. Grade V: death of a patient.
The mass was located in the posterior mediastinum in 96.1% (49 patients). Two patients had an MNT located in the middle mediastinum. Five patients had associated NF-1 (9.8%). Dumbbell-sign tumours were noticed in 8 patients (15.7%).
The number of patients who are currently still in follow-up is listed in Table 3. Two patients died in hospital postoperatively in the context of respiratory distress (1 adult patient with a malignant paraganglioma with invasion of the lung after left pneumonectomy; 1 adult patient with a benign schwannoma after thoracoscopic excision and combined lung volume reduction surgery). Overall general mortality is 7.8% (4 adult patients; 1 malignant, 3 benign MNTs). Two patients (1 child, 1 adult) with a malignant MNT received adjuvant therapy.
. | Adult (N = 43) . | Child 0–18 years (N = 8) . |
---|---|---|
Follow-up | ||
1 month (n) | 41 | 8 |
3 months (n) | 36 | 8 |
6 months (n) | 19 | 7 |
1 year (n) | 16 | 7 |
2 years (n) | 13 | 7 |
5 years (n) | 7 | 6 |
>5 years (n) | 4 | 5 |
> 10 years (n) | 1 | 3 |
Mortality | ||
MNT-related mortality Postoperative in hospital (n) >3 months (n) | 2 0 | 0 0 |
General mortality (n) | 4 | 0 |
. | Adult (N = 43) . | Child 0–18 years (N = 8) . |
---|---|---|
Follow-up | ||
1 month (n) | 41 | 8 |
3 months (n) | 36 | 8 |
6 months (n) | 19 | 7 |
1 year (n) | 16 | 7 |
2 years (n) | 13 | 7 |
5 years (n) | 7 | 6 |
>5 years (n) | 4 | 5 |
> 10 years (n) | 1 | 3 |
Mortality | ||
MNT-related mortality Postoperative in hospital (n) >3 months (n) | 2 0 | 0 0 |
General mortality (n) | 4 | 0 |
MNT: mediastinal neurogenic tumour.
. | Adult (N = 43) . | Child 0–18 years (N = 8) . |
---|---|---|
Follow-up | ||
1 month (n) | 41 | 8 |
3 months (n) | 36 | 8 |
6 months (n) | 19 | 7 |
1 year (n) | 16 | 7 |
2 years (n) | 13 | 7 |
5 years (n) | 7 | 6 |
>5 years (n) | 4 | 5 |
> 10 years (n) | 1 | 3 |
Mortality | ||
MNT-related mortality Postoperative in hospital (n) >3 months (n) | 2 0 | 0 0 |
General mortality (n) | 4 | 0 |
. | Adult (N = 43) . | Child 0–18 years (N = 8) . |
---|---|---|
Follow-up | ||
1 month (n) | 41 | 8 |
3 months (n) | 36 | 8 |
6 months (n) | 19 | 7 |
1 year (n) | 16 | 7 |
2 years (n) | 13 | 7 |
5 years (n) | 7 | 6 |
>5 years (n) | 4 | 5 |
> 10 years (n) | 1 | 3 |
Mortality | ||
MNT-related mortality Postoperative in hospital (n) >3 months (n) | 2 0 | 0 0 |
General mortality (n) | 4 | 0 |
MNT: mediastinal neurogenic tumour.
Surgical approach
Surgical approach for tumour resection included thoracotomy (n = 19), thoracoscopy (n = 31) or cervicotomy (n = 1). The number of thoracotomies versus thoracoscopies per 5-year interval over the 25-year study period is shown in Fig. 5. No significant difference was seen in the proportion of thoracotomy versus thoracoscopy over the past 25 years. Two patients had a combined neurosurgical approach for a Dumbbell-sign tumour. All malignant tumours (4/4) were approached via thoracotomy while the majority of patients with benign tumours (31/47) underwent thoracoscopic resection (P = 0.007). A thoracoscopic approach was used in 2 children (25%) versus 29 adults (67.4%) (P = 0.024).

The number of thoracotomy versus thoracoscopy per quintile over the 25-year study period (1992–2017).
No significant difference was noted in overall morbidity (45.2% vs 42.1%; P = 0.83), neither in postoperative chest wall discomfort (27.6% vs 50.0%; P = 0.15) between thoracoscopic versus open approach. Tumour weight was significantly lower in thoracoscopic cases (41.7 ± 43 vs 282.8 ± 288.5 g; P < 0.001). Hospital stay was significantly shorter after thoracoscopy (7.4 ± 3.3 vs 13.1 ± 9.8 days; P = 0.001). Based on the pathology report, postoperative imaging and operative report, resection was considered complete (R0) in 82.4% (42/51); more specific, 29/31 (93.5%) after thoracoscopy and 13/19 (68.4%) after thoracotomy (P = 0.019).
DISCUSSION
MNTs are relatively uncommon tumours. This was confirmed in our study with a low incidence of resected tumours over a 25-year period in a large thoracic surgical department in Belgium. Specific prevalence remains unknown, but neurogenic tumours constitute 10–34% of all mediastinal tumours [3]. In our study, the mass was located in the posterior mediastinum in 96.1%. According to the literature, neurogenic tumours constitute 75% of all primary tumours in the posterior mediastinum [7].
In general, for cystic lesions and well-encapsulated solid tumours that can be resected primarily, no tissue biopsy is needed preoperatively. It is our policy to go for complete surgical resection if the patient is symptomatic or anxious about the tissue diagnosis. Therefore, a preoperative biopsy will not affect the surgical plan in most cases. However, for mediastinal tumours with signs of invasion on imaging and suspicion of malignancy, a biopsy was performed. Our pathologists often do not rely on fine-needle aspiration to make a diagnosis, as cytology alone may be misleading potentially. An incisional biopsy by minimally invasive surgery is preferred. In our study, 17/51 (33%) patients had a preoperative biopsy [4, 8].
Nerve sheath tumours dominated the adult population, while ganglion cell tumours were significantly more present in children. Children had significantly more malignant tumours (25% vs 4.6%). The low number of children (n = 8) included in our study probably underestimated the rate of malignancy in children as most studies mention higher percentages (>50%) [2, 9]. Probably, not all children diagnosed with an MNT were referred for surgical treatment. Children with malignant tumours had a lower median age when compared with all resected children (0.17 vs 6 years). Fraga et al. [10] reported 13 children <1 year old, all presenting with malignant tumours (12 neuroblastomas and 1 ganglioneuroblastoma).
Surgery for MNTs has a relatively high complication rate, as already mentioned in Table 2. This rate is inherent to the position of these tumours in the mediastinum and the anatomical structures these are involved with. As MNTs are most frequently originating from the sympathetic chain, Horner’s syndrome is the most feared complication after surgery. Nine patients (17.6%) were diagnosed with Horner’s syndrome postoperatively. The most feared complication for Dumbbell-sign tumours is spinal cord damage or a leak from the dural sac [11]. The complication rate is remarkable because of their presence in a group of patients with mostly benign tumours. On the other hand, without surgical resection, symptoms due to local growth are rather realistic to occur. As we stated before, complete resection is also important for diagnostic confirmation and avoidance of malignant degeneration. Most of the complications can be categorized as a grade 1 according to the Clavien–Dindo classification.
Because of the benign aspect of most MNTs, long-term follow-up of these tumours in our centre was not deemed necessary. Twenty-five adults with resection of a benign MNT and a good clinical and radiological evolution (25/41; 61%) had a loss of follow-up at our centre after 1–6 months.
Overall general mortality is 7.8% (4 adult patients; 1 malignant, 3 benign MNTs), of which 2 with unknown cause because of the loss of follow-up. Two patients (1 child, 1 adult) with a malignant MNT received adjuvant therapy.
Neurogenic benign tumours of the mediastinum are often incidental findings. About 80% of the patients were asymptomatic at the moment of diagnosis. Takeda et al. [8] and Ribet and Cardot [12] reported symptomatic neurogenic tumours in 16–37% of cases. On the contrary, malignant tumours are symptomatic more frequently as seen in our series (n = 4/4; 100%).
NF-1 was reported in 5 patients (2 schwannomas, 2 neurofibromas, 1 neurogenic sarcoma). As mentioned previously, the presence of multiple neurofibromas is suggestive of a link with NF-1. In 30–45% of patients with neurofibromas, the presence of NF-1 is found [13]. It may also be associated with other nerve sheath tumours, with a higher risk of malignancy compared with the general population.
Up to 10% of neurogenic tumours in the mediastinum extend into the spinal canal [14]. In our retrospective study, Dumbbell-sign tumours were noticed in 8 patients (15.7%). Due to the minor involvement of the spinal canal in most patients, only 2 patients needed a combined neurosurgical approach. Our centre only preferred a combined approach whenever deemed necessary to avoid iatrogenic neurologic injury. One patient had a two-stage tumour resection via laminectomy and thoracotomy, the other patient had a single-stage combined laminectomy and thoracoscopic resection. A single-stage combined resection should be safe with good long-term effects and a lower risk of cord injury [15–18]. Our patient with a single-stage combined resection had a 5-year follow-up with an uncomplicated course and a good clinical and radiological outcome.
Certainly small and (presumably) benign tumours are amenable for thoracoscopic resection [19]. In our study, thoracoscopic resection seems to be a feasible way to treat mediastinal neurogenic tumours for these smaller tumours as well, resulting in a significantly shorter hospital stay, as also described by Petty et al. [20] and Yamaguchi et al. [21]. As reported in the literature [7], we found less postoperative chest wall discomfort, though in our study, this was not significant. No difference was noticed in overall morbidity between thoracoscopy and thoracotomy. We would have expected to see an increase in the use of thoracoscopy versus thoracotomy during the 25-year study period due to technical improvements. However, as shown in Fig. 5, this was not the case indicating that thoracotomy is still needed to remove larger tumours.
Technological improvements in video-assisted surgery with smaller endoscopic instrumentation have made clear that thoracoscopy is also recommended for children. Moreover, scoliosis following thoracotomy in neonates is described in >30% [22]. In our study, only 25% of children had thoracoscopic resection compared with 67.4% of adults. A more recent approach with robot-assisted thoracoscopic surgery is a promising technique that has been described as a simple and safe procedure, certainly for barely reachable tumours [23]. So far, none of our patients had a robot-assisted thoracoscopic resection in our department.
Kaba et al. [24] stated that open surgery is indicated in large-sized tumours (>6 cm), after previous thoracic surgery, when invasion into the spinal canal is expected, or when tumour is located in the apex. In these cases, an open approach should lower the risk of perioperative complications.
With regards towards the surgical approach to resect these mediastinal tumours, we have learned with more experience over the years. In case the tumour looks well circumscribed and encapsulated, we nowadays always try to mobilize the tumour by VATS; larger tumours can be extracted via an additional subxyphoidal port in case the mediastinal mass cannot be extracted through a standard thoraco-port; in case the tumour looks invasive and thus more likely to be malignant, an open approach seems more likely. A supraclavicular approach or median sternotomy can be helpful to resect a posterior neurogenic tumour in the apex of the chest with signs of invasion into neurovascular structures on imaging.
Limitations
Our research was a single-centre retrospective study on a small patient cohort (n = 51) over a 25-year study period with only a limited number of children (n = 8) included. Furthermore, we do not know the denominator of patients presenting with MNTs as not all patients may have been referred for surgery, especially young children with malignant tumours. The strength of this study was the large number of adult patients referred to our tertiary centre for resection (n = 43).
CONCLUSIONS
Children carry a higher incidence to present with a malignant tumour, originating from ganglion cells, while most tumours in adults are benign, originating from the nerve sheath. The majority of patients with neurogenic tumours are asymptomatic. The mainstay of treatment for MNTs, benign or malignant, is complete surgical resection. Hereby, symptoms due to local growth in the mediastinum can be prevented, diagnostic confirmation can be obtained and malignant degeneration can be avoided. Horner’s syndrome is a frequent complication after surgery. Most tumours, in particular small and presumably benign tumours, are amenable for thoracoscopic resection resulting in a shorter hospital stay.
ACKNOWLEDGEMENTS
The authors would like to thank the other staff members in the Department of Thoracic Surgery at the University Hospitals Leuven, Belgium, for their commitment in the clinical care of the patients included in this study: Willy Coosemans, MD, PHD; Herbert Decaluwé, MD, PHD; Lieven Depypere, MD, PHD; Philippe Nafteux, MD; PHD; Hans Van Veer, MD.
Conflict of interest: none declared.
Author contributions
Mathieu Lacquet: Investigation; Project administration; Resources; Visualization; Writing—original draft. Johnny Moons: Data curation; Formal analysis; Investigation. Laurens J. Ceulemans: Writing—review & editing. Paul De Leyn: Writing—review & editing. Dirk Van Raemdonck: Conceptualization; Methodology; Supervision; Writing—review & editing.
Reviewer information
Interactive CardioVascular and Thoracic Surgery thanks Mohammad Behgam Shadmehr, Pascal Alexandre Thomas and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.
A previous version of this abstract was submitted but not accepted for presentation at the 26th European Conference on General Thoracic Surgery ESTS 2018.
REFERENCES
Sceusi E, Vaporciyan AA. Benign and malignant neurogenic tumors of the mediastinum in children and adults. In: LoCicero III J, Colson YL, Feins RH, Rocco G (eds). Shields’ General Thoracic Surgery 8th edition. Mediastinal Surgery. Philadelphia, PA-Baltimore, MD-New York, NY-London-Buenos Aires-Hong Kong-Sydney-Tokyo: Wolters Kluwer
ABBREVIATIONS
- cancer
- adult
- child
- ganglia
- nerve sheath tumors
- paraganglioma
- surgical procedures, operative
- thoracoscopy
- thoracotomy
- mediastinum
- morbidity
- mortality
- neoplasms
- neurosurgical procedures
- pathology
- benign neoplasms
- mediastinal neural neoplasm
- excision
- tumor excision
- nerve sheath
- posterior mediastinum
- neurogenic tumors