Without going into great detail, it is useful to remember their usefulness in staging. Detailed anamnesis and physical examination can provide significant data about the degree of extension of the disease, which would allow for a substantial simplification of the tests to be done later. Thus, dysphonia, superior vena cava syndrome, Horner's syndrome or thoracic pain often reflects the invasion of anatomical structures adjacent to the lung which usually contraindicate surgical treatment. Equally, neurological symptoms or intense, persistent bone pain should lead us to suspect the existence of distant metastasis, has relevant therapeutic and prognostic implications. Chest radiography is usually the first test to give a high-probability suspicion of the existence of LC. In addition to its diagnostic value, the detection of pleural effusion, destruction of vertebra or ribs, mediastinal invasion, etc., can be decisive for establishing the degree of extension and notably simplify the process of staging.

After chest radiography, CT is usually the next imaging test to provide relevant information in the staging process. Its performance and limitations in the diagnosis of LC have been extensively studied and are well-known9,10 (Table 2). Regarding the primary tumor (T), CT is still the best method for the overall anatomical study of the thorax. It obtains detailed information about the size, location, and anatomical relationships with neighboring structures of tumors and can detect very small nodules that would normally not be detectable with chest radiography. As for the invasion of the chest wall, there is a reported sensitivity (SEN) of 83% and specificity (SPC) of 80%, but the only truly reliable sign of invasion is bone destruction.9 With regards to the invasion of the mediastinum by the tumor, some criteria predict resectability with a distance between the mass and the mediastinum ≤3cm, the visualization of a fatty layer between the structures, or a contact angle between the mass and the aorta of less than 90°. Contrarily, the radiological signs suggesting invasion of the mediastinal structures that would implicate unresectability are not very reliable and it is not acceptable to rule out surgery based on such findings.9

In general, a diameter of 1cm at the shortest point is accepted as the upper limit of normal, although this criterion is not useful to discern between malignant and benign lymph nodes.9,10 Around 40% of mediastinal lymph nodes suggestive of malignancy according to CT are benign, and 20% of those that are apparently benign are not in the end. Even among patients in clinical stage 1A, 5%–15% will show lymph node affectation in the surgical-pathological examination.10 These limitations in diagnostic performance (Table 2) mean that the CT findings need to be confirmed with other more reliable tests.

Liver and suprarenal glands. Isolated hepatic metastases are not frequent in NSCLC, but are in SCLC. The suprarenal glands are a frequent location for metastases, although their differentiation with benign adenomas often requires obtaining cytohistological samples. Therefore, during the same exploration, both chest and upper abdominal CTs are usually performed. As for the search for possible extrathoracic metastases, the recommendations are commented further ahead.

PET, a diagnostic modality based on the greater metabolic activity of the neoplastic cells, provides information of interest about tumor biology, but its capacity for spatial resolution is less than that of CT. As for the threshold of normality for the so-called standard uptake value (SUV), each center should establish its own cut-point. The development of PET/CT, which integrates the images of both procedures into one single exploration, improves diagnostic accuracy.11,12 For the evaluation of the mediastinum, said efficacy is higher than that of CT (Table 2), although it varies depending on the type: thus, for adenocarcinoma, the PPV of PET/CT is 50%, and the NPV is 77.8%; meanwhile, for squamous carcinoma said levels are 23.1% and 96.3%, respectively.13

Nevertheless, the high captation of glucose in benign processes such as granulomas and infections causes a rate of false positive results (FP) of 20%–25%. It is therefore recommended to confirm such findings by obtaining cytohistological samples before rejecting the option of surgery in a patient who is a potential candidate. Contrarily, given a negative PET result in the mediastinal evaluation, it is considered acceptable to proceed with the intervention without previous invasive tests, with the following exceptions: (a) centrally located tumors, usually in contact with the mediastinum; (b) tumors with low metabolic activity; (c) apparent N1 affectation; or (d) when CT detects lymph nodes with tumors whose smallest diameter is >15mm; in this latter situation, a meta-analysis revealed a post-test probability for tumor affectation of 21%.14,15

The PET or PET/CT results, given their high SEN for detecting distance metastasis, can be relevant for modifying the therapeutic plan, especially to avoid unnecessary thoracotomies.16 On the other hand, in patients who are candidates for radiotherapy, PET/CT can better outline the area to be radiated.17 Thus, its use is recommended for patients with clinical stages 1A–IIIA, provisionally subsidiaries for radical treatment, even though its usefulness in stage IA is less evident.10

A careful clinical evaluation is still the best method for the prediction of metastases. Non-specific symptoms, such as weight loss, asthenia, and osteomuscular pain or other more specific symptoms, such as subtle changes in mood or mild loss of strength in a limb, as well as biochemical or hematological alterations that are unexplained by any other reason (hypercalcemia, hypoalbuminemia, high LDH, anemia, etc.), are associated with the presence of metastasis. The diagnostic imaging tests are guided by the location of the present symptoms or signs. Thus bone pain justifies performing a gammagraphy, which has a SEN of 87% and an SPC of 67% for detecting bone metastasis.10 Given the frequent existence of degenerative or posttraumatic lesions, it is not rare to find doubtful images, in which case PET can be useful, as it is highly exact with SEN, SPC, PPV and NPV above 90%.9 PET, or PET/CT, are more resolute than CT for discerning hepatic or suprarenal lesions, especially when they are large, in which case the exactness is close to 100%. In lesions with a diameter of <15mm, the performance is much lower. In addition to its greater diagnostic efficacy in the usual locations, PET can detect metastases or other primary tumors in unsuspected places, such as the gastrointestinal tract, bladder, soft tissue, etc. Recent publications demonstrate that PET/CT could improve the performance of isolated PET or CT, although 15% of the abnormal images are classified as indeterminate.12,18

Table 2 summarizes some results from said studies, although the levels of NPV could be overestimated due to limitations in the follow-up derived from the short time that has passed since the introduction of PET/CT. The systematic practice of this test improves staging by detecting distance metastases left hidden after conventional staging (from 8% in patients in provisional stage I until 24% in stage III, using PET/CT), reducing in this manner the number of useless thoracotomies.16

But it is not free of disadvantages, as the existence of false positive (FP) results often requires carrying out risky procedures in order to obtain cytohistological confirmation of a lesion that finally turns out to be benign. In general, except for overwhelming clinical and radiographic evidence, the findings of images suggestive of metastasis, especially when it is a single, focal hypercaptation, should not lead to the exclusion of the patients from potentially curative treatments if histological confirmation of malignancy is not obtained.18 On the other hand, it must be remembered that, in some cases, the confirmation of a single metastasis (suprarenal or cerebral) does not necessarily imply renouncing surgical treatment of the primary tumor, as it could also be removed.8

With regards to the diagnosis of cerebral metastases, cranial CT or magnetic resonance imaging (MRI) should be carried out in cases with any suspicious neurological symptoms or signs, as well as in asymptomatic stage III patients in whom the possibility of aggressive treatment (surgery or thoracic radiotherapy) is considered.19 In earlier stages, although the indication is more arguable, given the greater incidence in the non-squamous types (even up to 20%), it seems recommendable.

Its use is recommended only under special circumstances, such as in tumors of the sulcus superior due to its superiority over CT for evaluating the invasion of the brachial plexus, the mediastinal vessels or the vertebral body. In such cases the diagnostic exactness for evaluating the tumor extension can reach 94%, compared to 63% with CT.9

In a systematic review of 17 studies, it has been established that TBNA has a SEN of 78%, a SPC of 100% and an FN rate of 28%.21 The diagnostic accuracy only comes close to ultrasound-guided needle aspiration in the subcarinal station; in the rest, its performance is poorer (58% vs 84%).22 When cost-efficient criteria are introduced, TBNA is less recommendable than endoscopic ultrasound techniques and mediastinoscopy (MED) when the prevalence of mediastinal lymph node affectation is higher than 25%.23

In a meta-analysis centered on the staging of LC by EBUS, a total of 11 studies were reviewed, including 1299 patients.24 The total SEN was 93% and the total SPC was 100%, with a mean prevalence of 53.2%. The analysis by subgroups shows that the selection of patients with abnormal lymph nodes on CT or PET and the availability of immediate cytopathologic diagnosis are independent factors that increase the total SEN to 94% and 97%, respectively. In the absence of immediate cytohistological diagnosis, the diagnostic performance depends on the number of needle aspirations done per lymph node. SEN is 69.8%, 83.7%, and 95.3% for the first, second and third aspirations, while the NPV is 86.5, 92.2, and 97.6%, respectively.25

The high diagnostic precision of EBUS for mediastinal lymph node staging even holds true for lymph nodes with a diameter <1cm. In a study of 100 patients with normal lymph nodes on CT (diameter between 5 and 10mm; mean, 8.1mm) a SEN was obtained of 92.3% and a SPC of 100%, with a NPV of 96.3%.26 In all, surgical controls were carried out using thoracotomy (85%) or MED (15%). And in a second study with 100 patients with LC and PET with a morpho-metabolically normal mediastinum, EBUS-TBNA had a SEN of 89%, a SPC of 100% and a NPV of 98.9% in the detection of lymph node metastases.27

Data have not been published for the cost-efficiency analysis, although it can save between 30% and 56% of mediastinoscopies.28

In a meta-analysis grouping 18 studies, it was concluded that the total SEN of the technique is 83% and total SPC 97%. When the examination is done in patients with pathologic CT, the SEN increases to 90%, with a SPC of 97%.29 It is possible to aspirate lymph nodes in patients with normal mediastinum on CT (diameter<1cm), with a SEN between 50% and 61% and a SPC between 98% and 100%.30

EUS-TBNA can also detect sub-diaphragmatic metastases (left suprarenal gland, lymph nodes of the celiac trunk and liver)31 and evaluate the presence of mediastinal invasion by the tumor (T4) with much more precision than with radiological techniques (SEN, 98%; SPC, 98%; FN, 1%, and FP, 30%).32

In a comparative randomized study, between MED and systematic EUS-TBNA in 104 patients candidates for surgery, it was demonstrated that the latter alternative reduced the number of unnecessary thoracotomies from 25% to 9%.33

The combination of EUS-TBNA and EBUS-TBNA enables complementary access to all the mediastinal lymph node stations, except the 6th. The SEN of this combination is 93%, with a NPV of 97%. In a cost analysis model from the year 2007 and according to the standards of American health insurance providers, it was estimated that the combination of EBUS and EUS is more cost-efficient than MED with the prevalence of affected lymph nodes over 32.9%. Under this percentage, the most cost-efficient approach is EUS.23

The two most recent guidelines for clinical practice13,21 about the clinical staging of LC coincide in recommending, with a degree of recommendation 1B,21 the cytohistological confirmation of the radiological or metabolic alterations that suggest mediastinal lymph node affectation. Both guidelines, completed before recent publications about the performance of EBUS27 and EUS,29 also recommend a minimally invasive surgical technique if this confirmation was done with an endoscopic technique and the results of the cytohistological study of the sample were either negative for malignancy or inconclusive.

In these cases, the most widely used surgical technique is MED. It can explore the right and left, upper and lower paratracheal lymph node stations as well as the subcarinal station. The minimal acceptable requirement for MED is the biopsy of at least one lymph node of the lower right and left paratracheal and the subcarinal stations.13 If the tumor is on the left, it is necessary to extend the exploration to the subaortic and para-aortic lymph node stations, especially if the tumor is hilar or of the upper lobe. Both the left parasternal mediastinotomy and the ECM give access to these two lymph node stations. The experience with 739 systematic mediastinotomies for staging left LC demonstrated the increase in positive explorations when combined with MED (28% of the explorations). In addition, the systematic sampling of the mediastinal fat when no lymphadenopathies are found at these levels reveals neoplastic infiltration in 6% of the cases.34 Systematically carried out in combination in patients who are candidates for lung resection, left parasternal MED and mediastinotomy or ECM reach the following results: SEN 86%; SPC 100%; diagnostic accuracy 94%; PPV 100%, and NPV 90%.35 When these explorations are negative, the possibility for finding metastasized lymph nodes in the lower mediastinal lymph node stations is very low (1.2% of cases)35; therefore its systematic exploration with thoracoscopic or ultrasound-guided transesophageal needle aspiration, when the imaging techniques are normal, would not be indicated. The rate of complications of these explorations is around 3%, and the mortality rate is around 0.1%.

There are few publications about ECM as a substitution for left parasternal mediastinotomy for staging left LC, but the results are very homogeneos.36–39 It is a valid alternative to left parasternal mediastinotomy with no specific associated complications. Its high NPV is relevant for indicating lung resection without the need for induction treatment.

Thoracoscopy, with or without video-assistance, provides a complete exploration of the pleural cavity and the ipsilateral mediastinum, if there are no pleural adherences that impede it. Under ideal conditions, it gives access to the lower right paratracheal, hilar, subcarinal and paraesophageal lymph node stations and the lower lung ligament on both sides, as well as the subaortic and para-aortic stations on the left side. Compared with MED, it has the disadvantage of being a unilateral exploration, and therefore it cannot rule out N3 affectation, except when both hemithoraces are explored. It is useful in the study of accompanying pleural effusions in order to confirm or rule out pleural affectation, and in the diagnosis of lung nodules if their peripheral location allows them to be biopsied or removed by means of atypical lung resection. Its use before thoracotomy, associated with pericardioscopy, can identify causes of unresectability and avoid exploration thoracotomies in 10% of patients.40

The adaptation of a video camera to the mediastinoscope has made the technique evolve in an unexpected way. Two methods of mediastinal lymphadenectomy by video-assisted MED have been described, which are equivalent to mediastinal lymph node dissection performed by thoracotomy. These are video-assisted mediastinoscopic lymphadenectomy (VAMLA)41,42 and transcervical extended mediastinal lymphadenectomy (TEMLA)43 (Table 2). It is interesting to note that, in the last 150 TEMLA, all the diagnostic parameters were 100%.

All these techniques are also useful for re-staging and are indicated in the pre-operative evaluation of second primary tumors, recidivated tumors and, currently more frequently, in the evaluation of tumor response after induction treatment in cases of locally advanced LC. For this latter indication, Table 2 summarizes the results of MED for re-staging without previous MED,44 re-mediastinoscopy (REMED),45–48 TEMLA49 and video-assisted thoracoscopy.50 In the case of MED for re-staging without previous MED, it is logical that the levels are higher than those of REMED, as the upper mediastinum is not altered by the adherences of the first surgical procedure. The results of REMED are very homogeneous when done systematically. Only one series stands out due to its low SEN and low NPV. However, in that series, in more than 60% of patients, in REMED the subcarinal lymph nodes were not biopsied, which are the most affected lymph nodes together with the lower right paratracheal nodes. TEMLA has the best values for re-staging, but it must be kept in mind that, in these cases, the staging was done by means of EBUS. Therefore, there were no mediastinal adherences, which are what complicate REMED. Last of all, video-assisted thoracoscopy has been used very little in re-staging, but the results of this unique multi-centered study50 indicate that they may be comparable to those of REMED.

For patients with tumors in stage I or II who can tolerate lung resection, the minimum procedure recommended is lobectomy.60 For patients with stage I who can tolerate a surgical intervention but not a lobectomy, sublobar resection (regulated or atypical segmentectomy) is preferable to non-surgical treatment.60 In stage I, the approach by means of minimally invasive surgery is as recommendable as standard thoracotomy.61 In patients with central or locally advanced tumors, as well as in those with N1, complete resection by means of bronchoplastic techniques offers better results than pneumonectomy.62

Definition of complete tumor resection. The IASLC defines complete tumor resection63 as that which meets the following criteria:

• (a)

  Free resection margins, as demonstrated microscopically. These margins should include the stumps of arteries, veins and bronchi, peribronchial soft tissue and any peripheral margin close to the tumor or to the rest of the resected tissue.

• (b)

  Systematic lymph node dissection as extensive as possible, or rather lobe-specific systematic lymph node dissection (as previously described). The samples should include at least six lymph nodes: three of the intrapulmonary and/or hilar stations and three of the mediastinal stations, one of which should be subcarinal.

• (c)

  There should not be isolated extracapsular extension of the tumor in the resected lymph nodes or in those located in the margin of the main lung tumor.

• (d)

  The highest removed mediastinal lymph node should be negative.

Processing and macroscopic study. The piece should optimally be set by inflating the airway with formaldehyde as a routine fixative. The measurements of the surgical piece and the tumor are taken once the piece is fixed.

The evaluation of the tumor should include: (a) parenchymatous or endobronquial location (if the bronchus is main, lobar or segmental) and distance from the pleura; (b) largest diameter; (c) description (shape, color, delimitation, cavitation); (d) anatomical extension: bronchial affectation, vascular invasion, visceral pleural invasion or the shortest distance from the visceral pleura (the IASLC recommends the use of staining for the elastic layer of the visceral pleura, whose affectation defines the invasion of the visceral pleura, in cases in which normal staining with hematoxylin and eosin are doubtful),64,65 extension to interlobar fissures and adhered tissues; (e) distance from margins, specifying the shortest distance (bronchial, vascular, surface of the resected parenchyma, adhered tissues); and (f) presence of other tumor nodules. Synchronic primary carcinomas are independently staged.

Evaluation of the surrounding parenchyma. Normal or abnormal, specifying the pathology.

Lymph nodes of the surgical piece. Location, number and size.

Lymph nodes remitted separately from the piece. With specifications of each lymph node station, number and size.