The results of the qualitative variable are expressed as percentages and absolute frequencies. For the qualitative variable the media and standard deviation are expressed. The comparison of discrete variables was carried out by means of the chi-square test, and that of independent means of quantitative variables using Student's t-test. P<.05 was considered statistically significant. For multivariate analysis a model of binary logistical regression with the Hosmer–Lemeshow goodness of fit test. In the multivariate analysis all those factors were included which produced a P<.1 result in the bivariate analysis. For the statistical analysis the software used was SPSS version 15.0 for Windows. The prevalence, sensitivity and specificity, predictive positive value and predictive negative value were determined using standard definitions.

In 54 patients (28.1%) the TBNA was the only endoscopic technique that enabled BC to be diagnosed. Of the 192 patients, 37 presented non-microcytic BC, T1–T3, with no widespread metastasis, and surgery was viable, which suggested the appropriateness of staging mediastinoscopy. In 25 (67.6%) (19 N2, 6 N3) of these cases TBNA revealed the presence of a mediastinal tumour, as a result of which mediastinoscopy was unnecessary. Of the remaining 12 patients, one opted against mediastinoscopy, and of the other 11, tumours were shown to be absent in 7 cases (true cases of negative TBNA) and present in 4 (3 adenocarcinomas, 1 epidermoid) (false TBNA negatives). The flow of patients is shown in Fig. 1. The prevalence of mediastinal tumours in this sub-group of patients was 74.4%, and TBNA, and the sensitivity of the TBNA was 86.2%, its specificity 100% and its negative value 63.6%.

Fig. 1.

Flow of operable patients with potentially non-microcytic carcinoma. MED: medianoscopy; negative MED: the medianoscopy revealed no lymph node tumours; in the mediastinoscopy mediastinal lymph not tumours are observed; TBNA: conventional transbronchial aspiration.

(0.09MB).

The added cost of the performance of TBNA is the sum of the cast of needle aspiration and the cytological study carried out on the sample. That is to say: 80EUR+38EUR per patient, the total cost for the whole study population therefore being (192×118) 22656EUR. The saving represented by the TBNA is the result of two factors: (a) the avoidance of 25 staging mediastinoscopies: which, multiplied by 3000EUR means a saving of 75000EUR. (b) In 54 additional cases, TBNA was the only endoscopic technique that enabled BC to be diagnosed. In 10 of these patients the diagnosis had already been made by other means (in 7 pleural liquid, in 2 by supraclavicular adenopathy aspiration and in one case by aspiration of the lesion occupying the hepatic space). But in the case of the 44 remaining patients, it would have been necessary, in the absence of TBNA, to use some other diagnostic technique. In 31 of these patients, the chosen technique would have been the CT-guided FNAC (a saving of 23777EUR), in 10 a bronchoscopy would have been performed (a saving of 1580EUR), and in 3 the chosen method would have been the performance of a diagnostic mediastinoscopy (a saving of 9000EUR). The total amount saved in this section comes to 34357EUR. The total economic impact has therefore been a saving of 86701EUR during the period of the study, which represents a saving of 451.57EUR per patient.

This aspect has been analysed on numerous occasions. In previous studies6,7,19,20 the main factors associated with the obtaining of a valid sample have been the size of the adenopathy, the location of the adenopathy in the right or central lymph nodes and the fact of it being a microcytic carcinoma. The best study in this respect is that of Fernández-Villar et al.,21 as it is based on a large number of patients and is one of the few studies in which a multivariate analysis is carried out. In this series the main factor associated independently with the obtaining of diagnostic samples was the diameter of the adenopathy, with the presence of indirect signs in the location of the aspiration also reaching statistical significance. In our study, like in the other published studies, as the diameter of the adenopathy increases, so too does the probability of obtaining a diagnostic aspiration.

While the cut-off point at which a TBNA is considered advisable has conventionally been less than 10mm at its minor axis, of the 10 patients who presented adenopathies that were smaller than this 4 produced valid samples, even though only one resulted in cancer diagnosis (one case of microcytic carcinoma).

In the multivariate analysis for obtaining a valid aspiration, the diameter of the adenopathy is of particular relevance, in line with studies published hitherto. What is also of statistical significance, and this is an aspect which is analysed here for the first time, is the number of lymph nodes biopsied during the bronchoscopy (in 34% of the bronchoscopies carried out more than one lymph node, as opposed to 17.4% in the study conducted by Fernández-Villar, but it must be borne in mid that a cytologist was present in the bronchoscopy room), without significant differences being observed in the diameter (18.7±6.4mm in the patients in which more than one adenopathy was biopsied. as opposed to 17.5±6.4mm in patients in whom only one was; P=.22). The percentage of microcytic carcinoma was higher in whom only one adenopathy was biopsied (36 out of 125 [38.8%] as opposed to 9 out of 67 [13.4%]; P=.02).

In the multivariate analysis for obtaining a diagnostic aspiration it is possible to observe a trend, which is close to achieving statistical significance in favour of the presence of microcytic carcinoma, which could be explained by the well-known greater aggressiveness of this histological type.6,7,22 No significant reduction has been observed in the diagnostic reliability of TBNA between bronchoscopies in which other additional diagnostic techniques (brushing, bronchioalveolar lavage, bronchial biopsy, transbronchial biopsy) and those in which only one endoscopic technique was used during diagnostic bronchoscopy.

With regard to the laterality of the adenopathies, there is no difference in the probability of either valid or diagnostic aspiration whether the left nodes, right nodes, and nodes situated in the centre are biopsied or bilateral nodes, a result which is reinforced by the fact that the size is similar (17.0±5.2mm in the left nodes and 17.4±6.3mm in the right or middle nodes, P=.72). There was no significant difference in the percentage of microcytics (23.2% in the left; 20.4% in the right or central; P=.71). Nor were any differences observed between the right paratracheal, the left paratracheal, the subcarinal or the hilar adenopathies, although there was a tendency, albeit not a significant one, towards a greater diagnostic performance in the hilar adenopathies, without any significant difference in size being observed (18.3±5.7mm in the hiliar adenopathies; 17.2±6.2 in the non-hilar adenopathies; P=.39), probably on account of the fact that the there are more reliable anatomical references for aspiration of hilar adenopathies.

Although there is no canonical definition for the concept of the clinical usefulness23,24 of a diagnostic test such as TBNA, it could be said to be clinically useful when the result permits us to confirm (or rule out) a given diagnosis, or when it has resulted in the diagnostic or therapeutic approach to the patient being modified.

In our study, TBNA is the only endoscopic technique that has resulted in a BC diagnosis in 28.1% of the patients. This result is slightly lower than that obtained by Fernández-Villar,22 and similar to results reported by other authors.6,16,19 Of the 54 patients in whom TBNA was the only endoscopic technique which produced a diagnosis, in 44 cases (22.9%) additional diagnostic tests could be avoided, so that money is saved, diagnostic delays are avoided and mortality might well be reduced.

On the other hand, the group in which TBNA has a direct impact on the diagnostic and therapeutic approach to be adopted is that made up of operable patients with potentially resectable non-microcytic carcinoma. Of the 37 patients in our series belonging to this group, TBNA produced a positive N2 or N3 diagnosis in 25 (67.6%), with the resulting avoidance of a mediastinoscopy. In this group TBNA showed a sensitivity of 86.2%, and a negative predictive value of 63.6%, with a mediastinal tumour prevalence of 74.4%.

These results are similar to those recorded by other authors.9,10,16,25 There is a small difference between the percentage of patients with non-microcytic carcinoma in whom a diagnostic aspiration was obtained with respect to the whole group (61.2%) and the operable and potentially resectable group (67.6%). This difference could be explained by the fact that in the total group of patients more than one adenopathy was biopsied in 46 out of 147 (31%), while in the operable group this was done in 19 out of 35 (55%) (the most frequent combination was aspiration in adenopathies 4R and 7). Undoubtedly, the performance of a more sophisticated technique, such as PTB-EBUS or PET-EUS would have improved on those results,26,27 but it must be borne in mind that, whether or not it is available, this method requires a second examination and its economic cost is higher. TBNA can be carried out during diagnostic bronchoscopy, which improves diagnosis and staging in a single endoscopic action. Our data call into question the performance of the above-mentioned examinations, in terms of both effectiveness and of patient wellbeing, unless the TBNA has been carried out beforehand, when pathological mediastinal adenopathies have been observed in the thoracic CT scan.

Bearing in mind the added cost which the performance of an aspiration represents, and the saving resulting from the avoidance of additional diagnostic techniques and staging mediastinoscopies, the introduction of TBNA has, in our clinical setting, led to a net saving of 451.57EUR per patient. This result is comparable to the results recorded by other groups.9,10,16

Recently two articles were published offering an economic analysis of different diagnostic strategies in the evaluation of mediastinal tumours in BC; in both cases specialised software was used based on a decision tree.

Harewood et al.28 examined the cost-effectiveness of 7 variants: mediastinoscopy, TBA-USEB, conventional TBNA, TBA-EBUS+PET-EUS, TBNA+PET-EUS, and TBNA+TBNA-EBUS, in the mediastinal staging of patients with non-microcytic carcinoma. They conclude that the combination of TBA-EBUS and PET-EUS is the most economical strategy when there is a high probability of lymph node metastasis, while if it is low PET-EUS is preferable. However, in order to arrive at these results the authors assume 36% sensitivity for the TBNA, a figure which is much lower than that given by other authors, and they also begin with the assumption of 30% prevalence of mediastinal tumour, which differs greatly from what is normally regarded to be the case.29 Furthermore, they assign a cost to TBNA which is only slightly inferior to TBA-EBUS (1430 as opposed to 1711EUR). Steinfort et al.30 compare 4 diagnostic strategies in mediastinal staging of patients with non-microcytic carcinoma and pathological adenopathies in thoracic CT scans or in the PET: PTB-EBUS with histological confirmation of the negative results, TBA-EBUS without histological confirmation of the negative results, conventional TBNA and mediastinoscopy. The final result is that the most cost-effective strategy is TBA-EBUS without histological confirmation of the negative results. However, this work is based on the premise that TBNA is performed during a bronchoscopy intended exclusively for mediastinal staging, and a similar cost is assigned to it as to TBA-EBUS. The result might have been different if it had been assumed that TBNA would be carried out during the diagnostic bronchoscopy, which must necessarily have already been carried out on all patients. Besides, as the same authors show, the cost of the acquisition of the ecobronchoscope has not been factored into the analysis.

The main limitation of this paper is that it is a based on a retrospective study. Nevertheless, the fact that the patients suspected of having BC have access to a thoracic CT scan prior to the diagnostic bronchoscope, and that in these patients TBNA was systematically performed, minimises this drawback. Another limitation of the paper, which is one that it has in common with all studies of this kind, consists in the fact that, in view of the high degree of specificity described when the recommended technical principles are applied, the diagnostic results of the TBNA are not confirmed by means of a surgical technique, and are assumed to be true positives. Furthermore, as it is also common to studies of this kind, most of the diagnostic results have not been confirmed by means of mediastinoscopy, either on account of the irresectability of the tumour or the inoperability of the patient. Only operable patients with potentially resectable tumours have been subjected to a mediastinoscopy for confirmation of non-diagnostic results.

The main strength of the study is that it has included a homogeneous, complete, and non-selected group of students, as this is the only centre in our area where patients suspected of having BC are admitted and as we take full responsibility for the entire process with regard to the diagnosis and follow-up of these patients. Furthermore, the fact that most of the bronchoscopies are carried out by the same endoscopist minimises the variations attributable to technical personnel and individual skill.