Abstract
Objective: Remediastinoscopy is a valuable tool in restaging non-small cell lung cancer after induction therapy for mediastinal nodal involvement as it provides pathological evidence of response and may select patients for subsequent thoracotomy. However, long-term survival data after remediastinoscopy are scarce. Methods: From November 1994 to April 2003, a remediastinoscopy was performed in 32 patients (29 men, 3 women) after induction therapy for locally advanced non-small cell lung cancer. Mean age was 67.8 years (range, 47–83). Neoadjuvant chemotherapy was given in 26 patients and chemoradiotherapy in 6. Follow-up data were completed in January 2005. Results: Remediastinoscopy was technically feasible in all patients. There were five false-negative remediastinoscopies, resulting in a sensitivity of 71%, specificity of 100% and accuracy of 84%. Follow-up was complete in all patients. Median survival time for the whole group was 21 months (95% confidence interval [CI] 9–33). Median survival time in patients with a positive remediastinoscopy was 7 months (95% CI 5–9), with a negative remediastinoscopy 41 months (95% CI 13–69), and with a false-negative remediastinoscopy 24 months (95% CI 5–43). The difference between positive and negative remediastinoscopies was highly significant (p = 0.003). In the combined group of patients with positive and false-negative remediastinoscopies (n = 17), median survival time was 8 months (95% CI 3–13). The difference with negative remediastinoscopy remained significant (p = 0.012). In a multivariate analysis, including sex, age, histology and nodal status at repeat mediastinoscopy, only nodal status was a significant independent prognostic factor (p = 0.015). Conclusions: Remediastinoscopy is a valuable restaging procedure after induction therapy. Prognosis is poor in patients with persisting mediastinal nodal involvement, proven at repeat mediastinoscopy.
1 Introduction
Mediastinoscopy remains the gold standard for invasive mediastinal nodal staging in non-small cell lung cancer (NSCLC) [1]. Although technically more difficult than the initial procedure, remediastinoscopy has shown to be feasible [2–5]. After neoadjuvant therapy, remediastinoscopy is a valuable tool in restaging locally advanced NSCLC [6,7]. It not only offers pathological information on mediastinal invasion and tumour response but also can select patients who will most likely benefit from a subsequent thoracotomy. Since only one preliminary survival analysis was reported in literature, our aim was to determine survival in a series of patients undergoing remediastinoscopy, and to detect independent prognostic factors for survival [8]. Accuracy of the procedure was also determined.
2 Materials and methods
From November 1994 to April 2003, a repeat mediastinoscopy was performed in 32 patients (29 men, 3 women) after induction therapy for locally advanced non-small cell lung cancer. Mean age was 67.8 years (range, 47–83). All patients had proven mediastinal involvement (N2 or N3 disease) at first mediastinoscopy and were given induction or neoadjuvant therapy. The sampled lymph nodes were classified according to the original map described by Naruke et al. [9]. Regarding histology, 16 patients had adenocarcinoma, 14 patients squamous cell carcinoma and 2 patients large cell carcinoma. Neoadjuvant cisplatin-based chemotherapy was given in 26 patients and chemoradiotherapy in 6. Regarding radiotherapy, 45 Gy were given to the primary tumour and involved mediastinal nodes. Follow-up data were completed in January 2005; so, minimum follow-up for surviving patients was 21 months. Survival analysis was performed by the Kaplan–Meier method. By univariate and multivariate analyses, significant prognostic factors in relation to survival were determined. A p value less than 0.05 was considered significant.
3 Results
Remediastinoscopy was technically feasible in all patients. There was no mortality; we encountered one haemorrhage from a bronchial artery during repeat mediastinoscopy, which was controlled by packing.
Remediastinoscopy was positive in 12 patients and negative in 20. Thus, the latter group was clinically downstaged to N0 or N1 and these patients underwent lung resection. The former group received radiotherapy. During thoracotomy, a systematic nodal dissection was performed. There were five false-negative remediastinoscopies, giving a diagnostic sensitivity, specificity and accuracy of remediastinoscopy of 71%, 100% and 84%, respectively. In the false-negative group, micrometastases were discovered after definitive pathological examination in the subcarinal (station 7) and tracheobronchial (station 4) nodes. Three patients in our study initially had N3 status. One patient had a positive remediastinoscopy and was further treated by radiotherapy. Two patients were downstaged to N0 and underwent thoracotomy with extensive mediastinal lymph node dissection.
Follow-up was complete in all patients. During follow-up, 23 patients died, mostly of distant metastases. Median survival time for the whole group was 21 months (95% confidence interval [CI] 9–33). Median survival time in patients with a positive remediastinoscopy was 7 months (95% CI 5–9), with a negative remediastinoscopy 41 months (95% CI 13–69), and with a false-negative remediastinoscopy 24 months (95% CI 5–43) [Table 1, Fig. 1]. In univariate analysis, the difference between positive and negative remediastinoscopies was highly significant (p = 0.0026). In the combined group of patients with positive and false-negative remediastinoscopies (n = 17), median survival time was 8 months (95% CI 3–13) [Table 1, Fig. 2]. In univariate analysis, the difference with negative remediastinoscopy remained significant (p = 0.012). In a forward stepwise multivariate analysis, including sex, age, histology and nodal status at repeat mediastinoscopy, only nodal status was a significant independent prognostic factor (p = 0.015). Relative risk of death in patients with positive remediastinoscopy was 2.88.
After induction therapy, 12 patients had a PET scan with fluorodeoxyglucose. There were four true-positive, four true-negative, two false-positive and two false-negative outcomes. This number is too small to permit a valid statistical comparison with remediastinoscopy.
4 Discussion
Although the accuracy of repeat mediastinoscopy is lower than that of a first mediastinoscopy, it is a valuable restaging tool as it provides pathological proof of downstaging or persisting mediastinal involvement. Our findings are in accordance with the previously published results of the diagnostic value of remediastinoscopy [3,4,7]. Invasive restaging remains necessary to correctly determine the clinical stage after induction therapy since the accuracy of computed tomography and magnetic resonance imaging is low, not only in predicting response after induction therapy but also in primary staging of lung cancer [5,10,11]. On the other hand, PET scanning has shown to be a promising tool in detecting residual viable disease in the primary tumour but not in the mediastinal lymph nodes, so that a positive result still needs histological confirmation [12–16].
In the present study, median survival time of patients with a positive and/or false-negative remediastinoscopy was significantly lower than those with a negative repeat mediastinoscopy. With a positive remediastinoscopy, the relative risk of dying was 2.88, compared to those patients who had a negative repeat mediastinoscopy. Thus, only those patients with a tumour downstaged to N0 or N1 after induction therapy will benefit from a lung resection. A clear association between survival and mediastinal lymph node clearance after chemotherapy in locally advanced NSCLC has been previously reported [17–21]. Betticher et al. [21] showed that persisting N2 after induction therapy is associated with a poor prognosis. Albain et al. [20] reported a 5-year overall survival after induction chemoradiotherapy and surgery of 41% if pN0 at surgery, 24% if pN1–3 and 8% when no surgical resection was performed.
There is only one other report in literature specifically studying survival after remediastinoscopy [8]. In this preliminary study, 5-year survival rate for patients with pN0 (n = 12) was 20% in contrast to 0% for patients with cN2 or pN2 after remediastinoscopy (n = 24). This difference was not significant, probably due to the less number of patients with long-term follow-up. In our series with a minimum follow-up of 21 months for surviving patients, a highly significant difference between positive and negative remediastinoscopies was found. Considering age, sex, histology and nodal status at repeat mediastinoscopy, only nodal status was a significant independent prognostic factor for survival in our study. Pathological findings at repeat mediastinoscopy not only determine prognosis but also select those patients who will benefit from a subsequent thoracotomy.
In conclusion, although accuracy is lower than after first mediastinoscopy, remediastinoscopy is a valuable restaging procedure after induction therapy in patients with non-small cell lung cancer. Prognosis is poor in patients with persisting mediastinal nodal involvement. In this way, futile thoracotomies are avoided.
Appendix A
Conference discussion
Dr G. Leschber (Berlin, Germany): Can you explain the rate of false-negative lymph nodes that you found? Were these lymph nodes from stations that you didn’t reach? How can you explain that there is a relatively high rate of false-negative lymph nodes?
Dr De Waele: Mostly the false-negative nodes were due to micrometastases surrounded by a large amount of fibrosis and maybe also to a lack of consistent biopsies, but biopsies were taken of all the lymph node stations that were positive during the first mediastinoscopy.
Dr Leschber: So you go for the same lymph node stations that you do in the first mediastinoscopy?
Dr De Waele: Yes, the positive ones.
Dr Leschber: So you look at all the same?
Dr De Waele: Yes.
Dr H. Abunasra (Leicester, UK): You talked about the technical feasibility of repeat mediastinoscopies. I’m interested to know if you had any problems, like any mortalities or any other problems that may have arisen from doing such a procedure. The other question I would like to ask you is, what do you think about PET scanning? Do you think that has a role in such a procedure?
Dr De Waele: Regarding the first question on mortality and morbidity, none of our patients died due to repeat mediastinoscopy. One patient had a bleeding from a bronchial artery which could be controlled by packing. Regarding the second question, I think the PET scan has a high negative predictive value for staging of the primary tumour, but it lacks accuracy for restaging of the mediastinum.
Dr Y. Kim (Seoul, Korea): It is a very good idea if you have true-positive lymph nodes, you can stop the procedure and avoid thoracotomy. However, to be clinically relevant, instead of combining the true-positive with the false-negative group, it would have been better to combine the false-negative group with the true-negative group. Did you try it? In this way you could use it in the clinical setting. For example, when you do a redo mediastinoscopy and if the lymph nodes are negative, you could go ahead and perform a thoracotomy. Do you have that kind of information?
Dr De Waele: No, I haven’t.
Dr R. Rami-Porta (Barcelona, Spain): My question is about inclusion of patients in the induction protocol. In our experience, which you quoted, our survival for pN0 patients was rather disappointing, only 20%, and I’ve seen that your median survival is higher than the one that we had, but we included patients with multiple N2 disease, multiple levels. What sort of patients did you include, only single N2 or multiple N2?
Dr De Waele: Most of them were single N2, but there were 3 patients who originally presented with N3 status.
Dr D. Branscheid (Grosshansdorf, Germany): What is your result? What did you do with the patients after you performed the second mediastinoscopy? What was your decision? I saw that you had a 20% 5-year survival rate. Would you exclude them from any operation?
Dr De Waele: When there was persisting N2 status, the patients were excluded from further thoracotomy. Only in patients with pathological downstaging a thoracotomy was performed.
Dr M. Tsuboi (Tokyo, Japan): My question is regarding the correlation between the false-negative result and the tumour location, primary tumour location. Have you ever examined a correlation between the false-negative result and the primary tumour location?
Dr De Waele: No, we have not examined that at the present time.
Dr T. Grodzki (Szczecin, Poland): It was an extremely interesting presentation, but I have some doubts, because the groups were in fact incomparable because 20 of them underwent thoracotomy, and survival in the positive group without thoracotomy, the median survival was 7 months. Am I right?
Dr De Waele: Yes.
Dr Grodzki: And with the false-negative, I mean finally positive, who underwent thoracotomy, it was 24 months. Am I right?
Dr De Waele: Yes.
Dr Grodzki: So how can you compare it, because there was another treatment. Did you refer the rest who did not undergo thoracotomy for any radiation or chemo or something like that, or did you just let them go?
Dr De Waele: Perhaps Professor Van Schil can answer that question.
Dr Van Schil: The patients who had a positive repeat mediastinoscopy were treated by chemoradiotherapy or radiotherapy alone. There is an ongoing discussion whether patients with positive N2 nodes discovered at repeat mediastinoscopy should undergo surgery, but in this study they went for radiotherapy or chemoradiotherapy.
Dr Grodzki: Yes, but you showed that those who by mistake went to thoracotomy presented much better results than those who were not treated surgically.
Dr Van Schil: Indeed, but the false-negative cases were mostly micrometastases discovered at thoracotomy, and the ones who were positive at remediastinoscopy were clearly positive nodes remaining after induction chemotherapy. So there is probably a different biological behaviour in lymph nodes that remain positive after the induction chemotherapy.
Dr Grodzki: So you just confirmed, that the word ‘positive’ does not necessarily mean the same in every condition?
Dr Van Schil: No. When you perform the remediastinoscopy and the results are negative, you don’t know beforehand that it will be false-negative, of course. This fact will only be discovered after the operation when you have the definite pathological examination of the lymph nodes and the primary tumour removed at thoracotomy.
Survival after remediastinoscopy divided into positive, negative and false-negative subgroups.
Survival after remediastinoscopy divided into positive, negative and false-negative subgroups.
Survival after remediastinoscopy divided into negative and combined positive and false-negative subgroups.
Survival after remediastinoscopy divided into negative and combined positive and false-negative subgroups.
Different subgroups, divided according to findings at remediastinoscopy
Different subgroups, divided according to findings at remediastinoscopy
