Non-grasping en bloc mediastinal lymph node dissection for video-assisted thoracoscopic lung cancer surgery

Accurate lymph node (LN) staging is an important component of the assessment and management of patients with resectable non-small cell lung cancer (NSCLC). Although the role for mediastinal LN sampling (MLNS) versus mediastinal LN dissection (MLND) with regard to improvement in survival of early-stage NSCLC is still controversial, the necessity for LN assessment for staging purpose is undeniable [1-5]. Therefore, current guidelines recommend that systemic mediastinal LN sampling or dissection with at least three N2 stations assessed is mandatory for all patients with resectable NSCLC [3,4,6].

Video-assisted thoracoscopic surgery (VATS) became a viable treatment option for NSCLC with an equivalent oncological resection (lobectomy plus MLND or MLNS) [5,7-11] and equivalent or even superior oncological outcomes [12]. However, MLND through VATS has always been an aporia. Except for technical difficulties caused by limited access ports and narrow operative space, surgeons always face a dilemma that on one hand they should accomplish “completeness” with stronger grip or traction of the target LNs, while on the other hand they should avoid cutting or crushing the target LNs. Recently, some experts depicted their techniques in performing VATS MLND, and it’s not unique that all of them had used strategies to grasp the target LNs [13-15]. The present study has developed different and novel strategies to avoid direct grasping of the target LNs. We proposed the concept of “single-direction VATS lobectomy” [16], accompanied with which we gradually developed a stylized method of MLND named as “non-grasping en bloc MLND”. The aim of this study is to depict our stylized technical tricks and to evaluate its quality and safety.

Although MLND as associated with improved survival of early-stage NSCLC remains controversial, we still advocate the method in our practice. We think that with complete removal of all resectable LNs, detection of micrometastasis or skip lesions, and the proportion of complete R0 resections are increased, which lead to more accurate tumor staging and reduced local recurrence [17-20]. Although video-assisted thoracoscopic MLND is commonly practiced, there may be concerns regarding oncological effectiveness based on the completeness of MLND [21]. Several previous studies demonstrated the feasibility and efficacy of MLND through VATS [5,7-11]. Some experts also depicted their techniques in performing thoracoscopic MLND using traditional “grasping” technique [13-15]. We designed this retrospective study to share our experience in optimizing surgical techniques and to present our outcomes of non-grasping VATS MLND.

The advantages of VATS, such as magnified view and modified visualization, improve the ability to recognize normal structures and to identify anomalous conditions, minimizing the risk of complications during nodal dissection. On the other hand, the disadvantages include two-dimensional view only and limited mobility of surgical devices because of small access ports. Therefore, alternating introduction and removal of different instruments should be avoided as possible. In addition, fragile as LN is, direct grip should be avoided as much as possible in terms of avoiding cutting or crushing the target LN, which may cause dissemination of cancer cells to the pleural or mediastinal space if the node is involved [9]. In our experience, we gradually developed the method of “non-grasping en bloc MLND” that incorporates several important features compared with traditional grasping technique. These features include, first, diversified use of the MES saves trouble of alternating instruments and provides more convenience for the surgeon. Moreover, the combination of the suction and hemostatic device (EcH or US) provides a clear operating field. Second, the non-grasping strategy avoids damage to LNs, and this meets the principles of surgical oncology. Third, en bloc dissection of the bounded fat block makes sure that there is no LN missed. Fourth, three-dimensional anatomization following specific orders according to different anatomic features makes MLND more vivid and concise. Fifth, even for cases with “fixated” mediastinal LNs, the technique of “non-grasping” does well during dissection, and we think that the usage of grasping technique will not help much in dissection but causing troublesome bleeding or damages to the LNs.

When evaluating the feasibility of the video-assisted thoracoscopic MLND, two significant concerns arise, namely, efficacy and safety. The best way to elucidate the completeness is that there would be neither a node nor fat tissue around each station to be found after dissection. We think that skeletonization of the anatomic region is mandatory to each station with clearly exhibited anatomic landmarks. The cutting or crushing of LN may lead to a problem in counting the number of the retrieved LNs as mentioned by Watanabe [9]. In our practice, we think that the non-grasping and en bloc strategy may decrease the probability of cutting or crushing LNs and facilitate to precisely count the number of them. We anatomized the dissected block of each station to count the number of LNs immediately after the operation. The anatomized block would be checked independently by pathologists after all specimens were sent to them.

In this study, we obtained comparable results with other studies [5,11]. The number of N2 LNs and total LNs (N1 + N2) were 9.5 ± 4.0 and 16.0 ± 5.9, respectively. In addition to the extent of MLND, the incidence of postoperative upstaging is also a significant parameter for evaluation. Previous studies found comparable results for VATS vs. open MLND. In these studies, the incidence of upstaging from N0 to N2 disease was found to range from 1.3% to 2.3% for MLND through VATS [7,11,22]. And Amer et al. reported much higher rate of nodal upstaging (16.6%) even though rigorous preoperative staging protocol including PET/CT was performed to each patient [15]. In our study, the incidence of upstaging from N0 to N2 disease was 12.2%. We believe that our persistence in en bloc dissection might have contributed to find more unexpected N2 disease.

Studies so far have demonstrated comparable postoperative mortality and morbidity of MLND by VATS vs. open lobectomy, which indicates that MLND by VATS is a safe procedure [23]. However, when pursuing oncological completeness, we have encountered three cases of hoarseness with unilateral vocal cord palsy (1 case: right side; 2 cases: left side), which might be attributed to thermal or mechanical injury to the recurrent laryngeal nerve caused by the EcH or US. Therefore, we suggest that there is no need to anatomize the recurrent laryngeal nerve desperately when performing MLND. Moreover, there were three cases of postoperative chylothorax, which had been treated faultlessly, in the early days of our practice. We had not realized that this problem might be due to the thoracic duct injury other than small lymphatic leakage until we encountered two cases of chyle leakage when dissecting station 4 R LNs. For these two cases, we ligated the thoracic duct just above the diaphragm resulting in no postoperative chylothorax. Therefore, when dissecting station 4 (R or L) LNs, we should keep in mind that the vulnerable thoracic duct is crossing to the left just behind the ascending aorta. In short, we argue that only with clear anatomic map in mind and meticulous operation could we decrease complications.

During the early stages of our practice in video-assisted thoracoscopic surgery for lung cancer, we had explored different techniques trying to make MLND more convenient and easier. At last, we developed this non-grasping en bloc technique and found it to be more comfortable for us to better perform VATS MLND. However, the techniques used during the exploring stages were diversified. It was hard for us to compare the non-grasping technique with the former techniques. And this is definitely a limitation of our study. Most thoracic surgeons think that the cutting and crushing of LN may cause dissemination of cancer cells to the pleural or mediastinal space if the node is involved [9]. We performed the non-grasping technique abiding by principles of surgical oncology. There were seldom cut or crushed LNs in our practice, and this was also the reason that the total number seemed to be less than that reported in other studies. Although we failed to evaluate the incidence of cut or crushed LNs, which was also a limitation of our study, we did believe that the non-grasping and en bloc strategy might contribute to reduce the incidence of cut or crushed LNs. Moreover, our preoperative staging protocol was not so rigorous due to economic issues when surgical indication was deemed. Preoperative invasive staging protocol (i.e., mediastinoscopy) was used mainly in patients with multiple or bulky mediastinal LNs. For cases with resectable mediastinal LNs, preoperative invasive staging protocol was not critically applied, and this was a very important limitation of our study. Therefore, we routinely performed complete MLND rather than MLNS for every patient in order to accomplish adequate evaluation of mediastinal LNs. And this might be the main reason for a high rate of postoperative nodal upstaging. We’d like to improve our preoperative staging protocol in the future.

In conclusion, video-assisted thoracoscopic MLND was equivalent to a comparable dissection through a thoracotomy in the hands of an experienced thoracoscopic surgeon with regard to both morbidity and oncological efficacy [24]. The method of “non-grasping en bloc MLND”, which enables en bloc dissection of mediastinal LNs while saving trouble of excessive interference of instruments and potential damage to the target LN, is safe, concise, and effective with a promising application prospect.