Introduction

Minimally invasive thoracic cancer surgery has undergone a difficult path since the beginning of development in the 90s. To date, it is a preferable option for stage I non-small cell lung cancer (NSCLC) [1–5]. However, the issue of safety of this method remains relevant. Intraoperative and early postoperative complications account for 4–8% [6, 7] and can cause death. In this regard, it seems important to consider the risk factors of these complications and develop the action algorithms for emergency.

Material and methods

The study included 479 patients who underwent thoracoscopic lobectomy between 2009 and 2019. All surgeries were performed using a four-port technique with separate treatment of bronchial and vascular structures, as well as mandatory lymphadenectomy. Five surgeons of the hospital performed operations. Data included age, gender, clinical and postoperative TNM [8], conversions, intraoperative complications, additional unplanned surgeries and in-hospital mortality. Characteristics of patients are presented in Table 1.

Table 1. Characteristics of patients

We established the following criteria for intraoperative and early postoperative complications:

— damage to pulmonary vessels followed by severe bleeding;

— damage to the bronchi;

— damage to other organs.

Results

Potentially life-threatening events requiring additional surgical procedures occurred in 7.9% (38/479) of cases (bleeding — 35, bronchial injury — 3). Conversion to thoracotomy was not required in all cases. Thus, conversions due to intraoperative or early postoperative bleeding, as well as bronchial injuries amounted to 32.2% (20/62).

Damage to the vessels with severe bleeding was registered in 7.3% of patients (n=35), including one patient diagnosed with bleeding in early postoperative period on the operating table. Bleeding occurred in 8% (22/275) and 6.4% (13/204) of cases during left- and right-sided surgery, respectively. We achieved hemostasis without thoracotomy in 48.5% (17/35) of patients while 51.5% (18/35) of patients required emergency conversion. Mean blood loss in case of thoracoscopic hemostasis was 610 ml (±12.3), in case of conversion — 785 ml (±10.2), maximum blood loss — 1.5 liters.

Injury of segmental branches the pulmonary artery were the most common (61.1%). Bleeding from segmental branches of pulmonary veins comprised 27.8%, from azygos vein and vena cava — 11.1%.

In majority of cases (74.3%, 26/35), bleeding followed mobilization of vessels from surrounding tissues with advanced fibrosis (perivascular fibrosis, lymphadenopathy). In other cases, incomplete lung collapse and incorrect surgical technique contributed to vascular damage.

Most often, bleeding occurred in 2016 (6; 17.1%), 2014 (5; 14.3%), 2011 and 2018 (4; 11.4%, respectively). The largest share was noted in 2011 (16% of all surgeries). We found no significant differences when analyzing the correlation between bleeding rate and year (p=0.364; Pearson's χ²). Mean incidence of bleeding was 7.31% and remained the same despite accumulation of experience and learning curve (Table 2).

Table 2. Incidence of intraoperative bleeding depending on the year of surgery

The rate of conversion for bleeding steadily decreased over time. In 2011, bleeding was stopped via thoracotomy in both cases. In 2019, conversion rate decreased to a record 33.3%. It was a result of surgical experience. The number of cases of successful thoracoscopic hemostasis exceeded conversion rate after 2014. However, there were no significant differences depending on the year despite overall trend towards a lower conversion rate (p=0.875; Pearson's χ²) (Fig. 1, Table 3).

Fig. 1. Incidence of intraoperative bleeding depending on the year of surgery.

Table 3. Surgical approach for bleeding

In some cases, we provided final hemostasis via suturing of clipping the vessel without conversion. In 70.6% (12/17) of patients, bleeding was stopped via suturing the vascular defect. One patient required proximal and distal occlusion of pulmonary artery. In 29.4% (5/17) of patients, hemostasis was achieved via clipping the vessel. Hemostatic material did not lead to positive results in our practice. Blood transfusion was required in 26 patients (74.3%, 26/35).

Three patients had bronchus injury (7.9% in overall structure of intraoperative complications). In 2 cases, injury occurred during mobilization of the bronchus from surrounding tissues (bronchial wall perforation by active electrode of monopolar cauter). In one case, we erroneously intersected of intermediate bronchus by a stapler during elective right-sided upper lobectomy at the stage of separation of parenchyma between the upper and lower lobes. Two patients underwent conversion (suturing of the bronchial wall defect in one case and bronchoplasty in the other one). In one case, endoscopic suturing was successful.

There were no intraoperative lethal outcomes. None patient died after intraoperative complications. One patient died in long-term postoperative period from pulmonary embolism. However, this patient had no intraoperative and postoperative complications.

Discussion

Thoracoscopic anatomical lung resection is a relatively new method, and it is under the close attention of experts. The issue of safety is the most important. Reaction of specialists to intraoperative complications is different in open surgery and thoracoscopy. Those situations proceeding normally and often not causing concern in the first case can be fatal in thoracoscopy. The concept of “major complications” is common in the literature regarding intraoperative adverse events. However, there is currently no clear and generally accepted definition of this term. Decaluwe H. et al. [6] analyzed 3076 patients in 6 European centers and gave the following criteria: blood loss > 2 liters, erroneous intersection of bronchial and vascular structures, damage to other organs, intraoperative situation followed by additional surgery immediately or within 30 days after surgery. Major intraoperative complications occurred in 46 patients (1.5%). There were 88 vascular injuries and only 15 ones were classified as "major complications". Injury of other structures occurred in 37 patients. Overall in-hospital mortality was 1.4% (n=43), and 23% (n=10) of deaths were associated with major intraoperative complications. We would like to emphasize threshold blood loss chosen by the authors (2 liters). We find no a rationale for this choice, since much less blood loss can be fatal for a patient with comorbidities. Moreover, time of bleeding and lost volume replenishment rate are unclear. Thus, if we interpret not included cases as serious and potentially dangerous situations during thoracoscopy, then we get 4.1% (88 + 37 = 125/3076) of complications. In our sample, this value was 7.9%.

Intraoperative injury of one or another anatomical structure is possible at any stage. Erroneous intersections of vessels or bronchi is of particular attention. Flores R.M. et al. [9] reported 3 pulmonary artery transections, 2 bronchial transections and 1 vein transection (0.9%), as well as one unplanned pneumonectomy out of every 200 (0.47%) thoracoscopic procedures. Decaluwe H. et al. [6] reported 9 erroneous intersections of anatomical structures among 3076 (0.3%) patients. We had one case of erroneous bronchus transection. Speaking about the causes of this complication, we can distinguish two main areas, i.e. surgical experience and features of surgical technique. Ten surgeons operated on patients starting with early stages of mastering thoracoscopic anatomical lung resections at MSKCC (Memorial Sloan Kettering Cancer Center, New York, USA) between 2002 and 2010. However, erroneous intersections occurred regardless of time. Amer K. et al. [10] focus on surgical technique and report 3 erroneous intersections of the bronchi referring to staged treatment of the vessels and bronchus. The authors emphasize that suturing and transection of parenchyma after all previous measures caused this surgical complication. Primary separation of parenchyma provides adequate overview before intersection of bronchial and vascular structures and prevents erroneous intersection [11, 12]. Certain techniques (“tunneling”) make it possible to visualize the artery and bronchus even in patients with unclear interlobar sulcus [13].

There are a few large controlled randomized trials. Thus, it is unclear whether the incidence of vascular complications in thoracoscopic surgery does not differ from those in thoracotomy. There is an opinion that specialists and hospitals analyzing the results of thoracoscopy can underestimate the incidence of vascular damage if bleeding does not lead to conversion and does not have a significant clinical impact [9, 11, 14]. Kawachi R. et al. [7] reported 8.2% of vascular lesions in 73 patients who underwent thoracoscopic lobectomy. The same value was 1.7% among 176 patients after thoracotomy. There were no significant differences in blood loss. Particular attention should be paid to surgical access in case of complication. Surgeons are faced with a choice: to make a conversion to thoracotomy or to continue thoracoscopic surgery. This applies to not only intraoperative situation, but also redo surgery. Decaluwe H. et al. [13] analyzed data from several centers and found a significant number of effective thoracoscopic correction of complications (56.5%, bleeding — 55.6%) (Fig. 2). In our study, thoracoscopic hemostasis was successful in 45.8% of cases.

Fig. 2. Surgical access for complications.

Bleeding during thoracoscopic surgery is recognized as potentially the most dangerous due to difficult hemostasis. The causes may be difficult work with new equipment, no familiar landmarks in narrow surgical field and poor quality of electronics. Correct reaction to bleeding is important, but even more significant is ability to recognize a dangerous moment in time before it occurs. If the risk of vascular injury increases, it is recommended to perform conversion and avoid uncontrolled bleeding [11, 15].

Serious possible consequences of intraoperative bleeding during thoracoscopy force us to focus on developing the algorithm of actions for hemostasis. Some algorithms present certain stages in hemostasis [16, 17] (Fig. 3). The first step is pressing the defect with a gauze ball or adjacent tissues (fatty tissue, lung parenchyma) followed by waiting for several minutes. This approach is often sufficient for hemostasis. Igai H. et al. [17] reported effective simple compression in 25% of cases, hemostatic sealant — 65%, other methods (suturing, clipping) – 10%. These reports emphasize active and effective use of hemostatic materials.

Fig. 3. Algorithm for hemostasis in patients with intraoperative bleeding during thoracoscopy.

Conclusion

Thus, prevention of intraoperative complications including selection of patients, adequate surgical experience and correct surgical technique, sufficient equipment, timely decision-making on conversion (elective conversion) and ability to correct emergency ensure safety of thoracoscopic surgery. It is very important to openly record complications for subsequent analysis. Knowledge of the causes of errors followed by emergencies and their open collegial discussion are essential for full development of surgical technique.

Author contribution:

Concept and design of the study — Ryabov A.B., Pikin O.V.

Collection and analysis of data — Bagrov V.A., Rudakov R.V.

Statistical analysis — Bagrov V.A., Aleksandrov O.A.

Writing the text — Bagrov V.A.

Editing — Ryabov A.B., Pikin O.V.

The authors declare no conflicts of interest.