Introduction
Introduction
Accurate staging of the mediastinum is a critical part of the management for newly
diagnosed cases of lung cancer. It is well recognised in non small cell lung cancer
(NSCLC) that stage at presentation is the main determinant of survival. Furthermore,
the degree of spread will determine the most appropriate and optimal treatment strategy.
NSCLC usually metastasises initially via the lymphatic system to locoregional lymph
nodes (hilar and mediastinal) and subsequently to distant organs by haematogenous
spread.
Surgical approaches to staging
Surgical approaches to staging
The traditional surgical approach to mediastinal staging is to perform a mediastinosocopy.
A standard cervical mediastinoscopy, when performed by an experienced thoracic surgeon
should allow access to the right and left paratracheal lymph node stations as well
as to the uppermost part of the subcarinal lymph node group (Fig. [1]
a). However, it is well recognised that mediastinoscopy is not always a straightforward
procedure and optimal yield is related to the experience of the operator. Less experienced
surgeons may not feel confident about fully exploring the mediastinum for fear of
damaging mediastinal structures. Pooled data showed that mediastinoscopy had a sensitivity
of 81 % and specificity of 100 % [1]. Although many case series report a higher sensitivity than this, it is probably
an accurate reflection of average clinical practice. A recent study showed that only
40 % of mediastinoscopies are performed adequately [2].
Although a mediastinoscopy has been long regarded as the gold-standard for mediastinal
staging it is important to realise that no one surgical procedure provides access
to all mediastinal lymph node stations. When staging NSCLC it is important to appreciate
the relative likelihood and distribution of mediastinal lymph node metastases. In
general the smaller and more peripherally situated a lung cancer, the less likely
it is to have metastasised to mediastinal lymph nodes. Histological type also influences
the risk of metastasis. In general, squamous and bronchoalveolar subtypes are less
prone to metastasise than adenocarcinoma. Because of the anatomy of the lymphatic
drainage system, the lobar location of a lung cancer will influence the likely site
of nodal metastases and this should be considered during staging investigations [3]. Right upper lobe tumours usually spread to the right paratracheal lymph nodes (levels
2R and 4R). Right middle lobe and lower lobe tumours from either side commonly spread
to the subcarinal nodes (level 7) initially. Left lower lobe tumours are unusual in
that they have a high risk of metastasising to contralateral right paratracheal lymph
nodes. On the contrary, left upper lobe tumours which have a propensity to metastasise
to aortopulmonary window (level 5) and para-aortic (level 6) lymph nodes initially,
may still be fully resectable despite being stage IIIA. If nodal spread is confined
to levels 5 and 6 they have a relatively better prognosis than other N2 nodal subsets
of stage IIIA.
These well recognised patterns of disease have historically influenced the surgical
approach to mediastinal staging. A standard cervical mediastinoscopy allows access
to levels 2R, 2L, 4R, 4L and the upper anterior part of level 7. Lymph node groups
5 and 6 lying in the aortopulmonary window and para-aortic area respectively may be
accessed by a left anterior mediastinotomy. With the advent of video assisted thoracoscopic
(VATS) procedures, a left VATS permits access to the whole of the subcarinal area
and to the posterior-inferior mediastinum including lymph node levels 8 and 9.
During the last two decades, developments in imaging technology, initially computed
tomography (CT), then positron emission tomography (PET) and most recently PET-CT
have shaped the staging algorithm for non-small cell lung cancer (NSCLC). However,
informative as these techniques are, there strengths and weaknesses must be fully
appreciated in order to avoid over-reliance on their results. Pooled data on CT and
PET scanning showed that the sensitivity and specificity for predicting malignant
involvement of mediastinal lymph nodes were 60 % and 81 % and 84 % and 89 % respectively
[4]. Recent data from Pozo-Rodriguez et al have shown that the negative predictive value
of combined CT and PET can be as high as 98 %. While these data suggest that a negative
PET-CT is sufficiently accurate to obviate the need for other pre-thoracotomy staging
investigations, the same can not be said for patients with a positive scan. The false-positive
rate of CT and PET is sufficiently high to demand pathological confirmation in order
to prevent patients being over-staged on radiological investigations alone and being
denied potentially curative treatment [6].
Fig. 1 Figure 1a shows the lymph node stations that can normally be accessed by a standard cervical
mediastinosocopy. Figure 1b shows the lymph node stations that can usually be accessed by endobronchial ultrasound
(EBUS-TBNA) and endoscopic ultrasound fine needle aspiration (EUS-FNA). The diagrams
of the lymph node stations are after Mountain CF and Dresler CM. Regional lymph node
classification for lung cancer staging. Chest 1997; 111 : 1718 - 1723.
Endoscopic staging procedures
Endoscopic staging procedures
Endoscopic ultrasound
Endoscopic ultrasound (EUS), which has been used by gastroenterologists for a number
of years for oesophageal staging, is increasingly being used to examine the mediastinum
in lung cancer. With the advent of curvilinear echoendoscopes which allow real-time
imaging of fine needle aspiration (FNA), it is possible to sample mediastinal lymph
nodes. The route of the oesophagus dictates the lymph node stations that are readily
accessible. Lymph nodes located in the aortopulmonary window (medial to the ligamentum
arteriosum - level 4L), subcarinal space (level 7), paraoesophageal region (level
8) and inferior pulmonary ligament (level 9) should be amenable to biopsy (Fig. [1]
b). Furthermore, EUS can be used to examine (and biopsy) lesions in the left adrenal
gland and the left lobe of the liver. The air-filled trachea lying anterior to the
oesophagus usually prevents imaging of the hilar and paratracheal lymph nodes although
occasionally left paratracheal lymph nodes can be seen. In expert hands the sensitivity
of EUS-FNA for detection of malignant lymph nodes is around 90 - 95 % [1]
[7]
[8]. However, the majority of these studies are retrospective and report on lymph nodes
> 1 cm in short-axis diameter as reported by CT scanning. This suggests that there
may be a selection bias present in these studies. Annema et al., have recently reported
the results of a large prospective study on the use of routine EUS-FNA in patients
with known or suspected lung cancer scheduled to undergo surgical staging procedures
[9]. They found that EUS-FNA prevented 70 % of surgical procedures on account of upstaging.
These findings support those of two other recent studies which both showed that EUS
can obviate the need for surgical staging procedures [10]
[11]. However, it should be noted that in all of these studies the inclusion criteria
stipulated lymph node size greater than 10 mm on CT and in Annema’s study the size
range was 13 - 77 mm with a mean of 24 mm [9]. This therefore raises the question as to the role of EUS in lung cancer patients
with a ‘CT-node negative’ mediastinum. Two groups have investigated this scenario.
Wallace et al., [12] showed that 17 of 69 (25 %) patients with lung cancer had unsuspected disease on
CT which was detected by EUS-FNA. More recently, LeBlanc et al., [13] found that EUS precluded inappropriate surgery in 12 % of patients without mediastinal
lymphadenopathy, based on the presence of N2 or N3 node positive disease. Although
neither of these studies incorporated PET scanning their results do indicate that
EUS probably has a significant role to play in staging of a ‘normal’ mediastinum.
Interestingly, LeBlanc et al. noted that EUS detected more occult metastases in cases
of hilar and lower lobe tumours than upper lobes [13]. Based on the known patterns of lymphatic drainage described above, this is not
surprising.
Overall, the emerging data on EUS-FNA shows that it has a major role to play in the
diagnosis and staging of lung cancer. Furthermore, it is complimentary to, rather
than competitive with, mediastinoscopy. Confirmation of this has come in two recent
studies which have clearly shown that EUS-FNA can add to the findings of mediastinoscopy
[14]
[15].
Transbronchial needle aspiration
Transbronchial needle aspiration
Transbronchial needle aspiration (TBNA) of lymph nodes and parabronchial lesions was
first described by Wang in 1983 [16] and although it is an established bronchoscopic technique it remains underutilised
[17]
[18]. The main reasons for this are its variable biopsy yield and long learning curve.
Until recently TBNA has been limited to ‘blind’ needle puncture guided by static CT
scans. This technique is highly operator dependent and reported sensitivity varies
between 20 % and 80 % [16]
[19]
[20]. Understandably the larger the target lymph node the higher the chance of success
and higher yields are obtained from certain lymph node stations such as level 4R (right
paratracheal) and level 7 (subcarina) [21]. Yield is also improved with at least seven passes of the TBNA needle and on site
cytopathology [22]. More recently Herth et al. [23] have demonstrated that a mechanical radial ultrasound mini-probe introduced through
the working channel of a flexible bronchoscope, can be used to localise target lymph
nodes prior to sequential blind biopsy using a standard bronchoscope and TBNA needle.
By this method, sampling sensitivity rose to 86 %. While this technique has improved
TBNA yields, it does not allow ultrasound-guided control of needle puncture and inadvertent
puncture of vascular structures remains a risk, albeit a small one.
Most recently endobronchial ultrasound guided transbronchial needle aspiration (EBUS-TBNA)
has been described. This utilises a curvilinear echoendoscope similar to those used
for EUS-FNA which allows real-time visualisation of needle puncture. Several groups
have now reported their initial results using this novel device with sensitivities
in the range of 85 - 95 % [24]
[25]
[26]
[27]
[28]. Using this device it has been shown that it is possible to successfully biopsy
lymph nodes down to 7 mm in diameter from the upper mediastinum lymph node stations
2, 3, and 4, subcarina (station 7) and hilar stations (10 and 11) (Fig. [1]
b). An endobronchial approach allows wider access to upper mediastinal and hilar nodal
stations than does a mediastinoscopy. The deeper regions of the subcarinal space are
easily examined from the right and left main bronchi and the hilar lymph nodes (station
10 and 11) can routinely be identified. Like EUS, EBUS-TBNA can be easily performed
under conscious sedation as a day case procedure and has a negligible complication
rate.
Combined endobronchial and endoscopic ultrasound for staging the mediastinum
Combined endobronchial and endoscopic ultrasound for staging the mediastinum
The recent advances in endobronchial ultrasound guided TBNA now make it possible to
consider using minimally invasive techniques to stage almost the whole mediastinum.
In conjunction with endoscopic ultrasound it should be possible to examine the upper
mediastinal, hilar and posterior-inferior lymph node stations (Fig. [1]
b). The only lymph node groups that are not accessible to either approach are station
5 (aortopulmonary window lateral to the ligamentum arteriosum and station 6 (para-aortic).
Two groups have reported their preliminary experience of this approach. In their paper
reporting their initial experience of the EBUS-TBNA bronchoscope, Rintoul et al.,
[27] reported on a subset of 7 patients on whom they had performed both EUS-FNA and EBUS-TBNA.
They found that the addition of EUS added extra information to the examinations in
all cases. The procedures were well tolerated in all cases and the mean time for the
combined procedures was 45 minutes. Subsequently Vilmann et al., [29] compared findings by EUS-FNA and EBUS-TBNA in 33 patients. Again they observed that
EUS and EBUS gave complimentary information - in a number of cases EBUS-TBNA gave
specific cancer diagnoses in lymph node groups that had not been obtained by EUS and
vice versa. This most likely relates to the fact that the approach to certain nodal
stations such as the subcarina is different by each method and resultant needle biopsies
sample different nodes and parts of nodes. In this series the accuracy of EUS-FNA
and EBUS-TBNA, in combination, for the diagnosis of mediastinal cancer was 100 %.
Mediastinal re-staging after induction chemotherapy for NSCLC
Mediastinal re-staging after induction chemotherapy for NSCLC
The use of neo-adjuvant chemotherapy to downstage NSCLC prior to surgery has focussed
attention upon the problems inherent in re-staging of the mediastinum. It is recognised
that post-chemotherapy imaging modalities such as CT and PET loose sensitivity and
specificity. Re-mediastinoscopy is also more technically challenging due to adhesions
and fibrotic change consequent upon the initial surgery and chemotherapy [30]
[31]. As a result diagnostic accuracy is lower with more incomplete procedures and complications
[32]. The minimally invasive approaches of EBUS-TBNA and EUS-FNA offer new approaches
to staging and re-staging the mediastinum around neo-adjuvant chemotherapy. To date
there is only one small study using EUS that has examined its role in re-staging the
mediastinum. Annema et al., [33] reported on 18 patients who underwent EUS-FNA following chemotherapy for stage IIIA-N2
disease. They reported a diagnostic accuracy of 83 % for correct prediction of N2
nodal status. EUS-FNA negative cases were confirmed by surgical exploration. This
result is similar to published results using re-mediastinoscopy as a staging tool.
It is likely that if EBUS-TBNA and EUS-FNA were combined in this situation even better
results might be possible. It is interesting to note that in 2 of the 3 false negative
cases in Annema’s series, disease positive lymph nodes were found in the paratracheal
regions - an area difficult to routinely sample by EUS.
Pros and cons of minimally invasive mediastinal staging
versus surgical staging
Pros and cons of minimally invasive mediastinal staging
versus surgical staging
EBUS and EUS clearly offer great potential for staging the mediastinum in a minimally
invasive manner. From the patients point of view they are far less invasive than surgery
and can be performed as a day case under conscious sedation and local anaesthetic
without the need for an in-patient stay and general anaesthetic. From the clinicians
point of view they are quick to undertake and in combination offer a wider access
to different mediastinal stations than mediastinoscopy or thoracoscopy alone. However,
while it is likely that these techniques will replace the need for surgical staging
procedures to a large extent, it is unlikely that they will replace them entirely.
At present there have been no controlled trials comparing the accuracy of EBUS and
EUS with mediastinoscopy or thoracoscopy. Future studies are required to define the
best use and combination of these techniques. There will be times when it is important
to obtain large biopsies rather than cytological FNA samples. While cytological specimens
are usually sufficient to diagnose carcinoma, the same can not be said of lymphoma
or mesothelioma. In our practice if mediastinal biopsy is required to diagnose or
stage either of these conditions we proceed directly to surgery. We have also found
on occasion that cytological samples from large necrotic or very hard scirrhous lymph
nodes are inadequate for diagnosis and surgical biopsy is necessary. Larger 19-gauge
‘histology’ EUS needles are under development which may help in certain situations
but it is unlikely that needles of this size will be possible for EBUS given size
restrictions imposed by anatomy.
The reduction in the number of mediastinoscopies being performed will also have implications
on surgical training and experience [34]. As stated earlier, a mediastinoscopy should not be regarded as a straightforward
procedure and its accuracy like any invasive procedure is related to the expertise
of the clinician performing it. There is an understandable concern in surgical circles
that if trainee surgeons are not exposed to sufficient numbers of the ‘easier’ cases
they will not be in a position to manage a challenging one.
NSCLC staging algorithm of the future (Fig. [2])
NSCLC staging algorithm of the future (Fig. [2])
In summary, the standard of care for mediastinal staging is changing. As techniques
such as PET-CT and EBUS-TBNA and EUS-FNA become more widespread the diagnostic and
staging algorithm for lung cancer will change. In future, the staging algorithm for
NSCLC might start with a contrast CT of the neck, chest and abdomen and then for patients
who appear to have radically treatable disease progress to a PET-CT scan with brain
imaging as clinically indicated. Based on this, diagnostic and staging tissue biopsy
procedures would be performed with the ethos being to diagnose and stage to the highest
stage in as few steps as possible. This might involve an initial combination of bronchoscopy
and/or EBUS and/or EUS. Surgical staging procedures might be held in reserve when
other staging studies are equivocal or when the need for larger amounts of tissue
is clinically indicated. In the situation where re-staging following neo-adjuvant
therapy is indicated initial staging using a combination of EBUS and EUS would allow
surgical staging to be reserved for re-staging if necessary although it is likely
that EBUS and EUS will have a major role to play here too.
Fig. 2 Potential NSCLC staging algorithm of the future for patients with potentially radically
treatable disease incorporating early use of PET-CT and targeted biopsy of disease
using minimally invasive approaches whenever possible.
