Safety and feasibility of laparoscopic liver resection for intrahepatic cholangiocarcinoma: a propensity score-matched study

Originating from intrahepatic bile duct epithelial cells, intrahepatic cholangiocarcinoma (ICC) accounts for 10 ~ 15% of primary liver cancer and is second only to hepatocellular carcinoma (HCC) [1]. Having a hidden onset, ICC is apt to invade perihepatic organs, tissues, nerves, and lymph nodes, and most patients are generally in advanced stages when diagnosed and lack effective treatment [2]. Recently, with the in-depth study of the molecular pathogenesis of ICC, the treatment of chemotherapy, local therapy, immunotherapy, and targeted therapy for ICC are being further improved.

Some patients with initially unresectable ICC have the opportunity to shrink tumors or even resect tumors radically after comprehensive treatment, and comprehensive treatment also assists ICC patients to obtain some progress during postoperative adjuvant treatment [3‐6]. Despite the rapid development of the abovementioned preoperative neoadjuvant and postoperative adjuvant therapy, only about 35% of patients could perform radical surgery, and the 5-year survival rate after surgery is only 25 ~ 40%, far lower than that of HCC [7, 8]. Of course, it should be pointed out that these studies are frequently based upon the practice of open liver resection (OLR) [9, 10]. Given the superiorities of minimal trauma, high-quality surgery, and fewer complications, laparoscopic liver resection (LLR) is controversial in treating ICC for the biological characteristics of diseases [9, 11]. Controversies about LLR for ICC concentrate on the insufficient quality of lymph node dissection [12], indistinct surgical margin due to the lack of tactile impression, and tumor dissemination caused by the vibration of surgical instruments with energy such as ultrasonic scalpel and pneumoperitoneum implantation. Restricted by various practical factors, such as patients’ subjective choices toward surgical methods and the discrepancy of pathological results, randomized controlled trials (RCTs), the gold standard for analyzing clinical problems, might be challenging to carry out in surgical fields. As an alternative method to reduce the impact of data bias and confounding variables, propensity score matching (PSM) analysis could not be used to replace RCTs with observational studies wholly. Still, in the case of data collection restriction in RCTs, data from large observational cohorts after PSM might help solve some clinical problems. There are few articles using PSM analysis to explore the safety and feasibility of LLR in treating ICC, and most of the existing articles are only case–control studies, or neglect the impact of the tumor, non-tumor, and surgical-related factors on the observation results, or lack exploration on the difference of short-term and long-term prognosis. Therefore, this study aimed to compare the short-term and long-term prognosis of LLR and OLR in treating ICC after balancing the corresponding confounding factors via PSM analysis and explore the independent factors affecting the long-term prognosis of ICC through the Cox proportional hazards regression model, hoping to provide certain ideas for the diagnosis and treatment of ICC.

From December 2010 to December 2021, excluding 4 patients with mixed liver cancer, 4 combined with other malignant diseases, and 10 not followed up for more than 1 year, a total of 170 ICC patients who met the inclusion criteria were ultimately included in this study (Fig. 1). The distribution of the admission time of patients ultimately screened in this study (n = 170) was shown in Fig. 2. Among the 170 patients, 97 ICC patients who underwent laparoscopic hepatectomy and 30 patients who were converted from laparoscopy to laparotomy formed the laparoscopic liver resection group (LLR group), and 43 ICC patients who underwent laparotomy formed the open liver resection group (OLR group). After a 2:1 PSM analysis, a total of 105 patients (70 patients in the LLR group and 35 patients in the OLR group) were included in further analysis.

As the second most frequent primary liver cancer, ICC possesses the characteristics of threatening invasiveness and terrible prognosis, and the morbidity and mortality of ICC have been increasing recently [17]. Unlike HCC, even today, when the concept of comprehensive systemic treatment is advocated extensively, an endless stream of targeted or immunotherapeutic drugs have not brought breakthrough survival benefits to ICC [18, 19], and hepatectomy is still considered the essential treatment [20]. Given the superiorities of minimal trauma, high-quality surgery, and fewer complications, laparoscopic liver resection (LLR) is controversial in treating ICC for the biological characteristics of diseases [21]. Controversies about LLR for ICC concentrate on the insufficient quality of lymph node dissection [12], indistinct surgical margin due to the lack of tactile impression, and tumor dissemination caused by the vibration of surgical instruments with energy such as ultrasonic scalpel and pneumoperitoneum implantation. Nowadays, the rising laparoscopic lymph node tracing technology with ICG fluorescence enables laparoscopic lymphadenectomy the advantage of visualization [22, 23]. Moreover, there exists insufficient evidence to exhibit that laparoscopic surgical instruments with energy and pneumoperitoneum increase the incidence of tumor dissemination and pneumoperitoneum implantation. And the intraoperative overturning and extrusion of the tumor-bearing liver segment by traditional open surgery might lead to iatrogenic tumor dissemination (including abdominal dissemination and incision implantation), determined by the high metastatic characteristics of ICC. From this point of view, LLR accords with the concept of “no touch.” Therefore, this study aimed to evaluate the safety and feasibility of LLR for the treatment of ICC and explored the independent factors affecting the long-term prognosis of ICC.

The distribution of the admission time of patients ultimately screened in this study was shown in Fig. 2. The reasons why there were more patients in the LLR group than in the OLR group in our center were as follows: First, our center is a hospital that takes minimally invasive surgery as the core development direction. It could be seen from the figure that we had tried to treat ICC via LLR since 2010. Second, 30 patients converted from laparoscopy to open surgery were finally included in the LLR group in this study, the reason of which was that we considered the conversion from laparoscopy to laparotomy as a remedial measure for the failure of LLR implementation and an evaluation index for the perioperative effect of the LLR group. If these patients converted to open surgery were excluded from the LLR group, the LLR operation might obtain a higher evaluation than the actual effect. The original intention of this study is to realistically evaluate the impact of LLR and promote the surgical methods that are conducive to the rehabilitation of patients.

To reduce the impact of data bias and confounding variables, we implemented PSM analysis. After PSM, the comparison results between the LLR and the OLR groups showed no difference in demographic characteristics and preoperative indexes (Table 1). In terms of intraoperative results, we found significant differences in some outcomes between the two groups after PSM. The intraoperative blood transfusion and blood loss in the LLR group were less than those in the OLR group. It should be pointed out that there was no statistical difference in lymphadenectomy rate, anatomical hepatectomy rate, and duration of surgery between the two groups, which broke the previous view that LLR took more extended time and could not implement adequate lymphadenectomy [12, 24]. Compared with the OLR group, the short-term prognosis of the LLR group was advantageous, and the morbidity of major postoperative complications (OLR vs LLR = 9 (25.7) vs 6 (8.5)) in the LLR group was lower than that of the OLR group. The long-term prognosis is an important aspect to evaluate the safety and feasibility of surgical technology. Before PSM, the OS of LLR was statistically better than that of OLR. Although the difference in OS of the LLR group compared with the OLR group after PSM was not that significant, LLR still maintained a certain advantage in OS. It should be noted that the 5-year survival rate of LLR for ICC in our center was higher than that of related published literature, which focused on LLR versus OLR for ICC after PSM [25‐27]. There was no statistical difference in RFS between the two groups before and after PSM. Consequently, LLR could enable ICC patients to receive an equivalent long-term prognosis in contrast with OLR.

In contrast with OLR, the superior short-term prognosis and comparable long-term prognosis of the LLR group might derive from the practice of the concept of enhanced recovery after surgery (ERAS) in this field and the application of high-tech equipment under laparoscopy. To enhance a high proportion of R0 resection and anatomical hepatectomy in the LLR group, besides conventional procedures, we also adopted three-dimensional reconstruction watershed analysis and laparoscopic fluorescence imaging systems using indocyanine green (FIS-ICG). A retrospective study published by our center, including 43 patients (with 6 ICC patients among them), pointed out that ICG-negative staining technology could help surgeons accomplish “hepatectomy along portal vein watershed,” that is, the complete resection of liver parenchyma supported by the tumor-bearing portal vein. And preoperative three-dimensional reconstruction was of great significance for the intraoperative identification of the target Glissonean pedicle and liver parenchyma cross section [28]. To further explore the safety and effectiveness of LLR with FIS-ICG for ICC patients, another study of our center retrospectively analyzed the clinical information of 11 ICC patients who underwent laparoscopic anatomical hepatectomy with FIS-ICG. All patients obtained R0 resection and experienced an excellent short-term prognosis. That study demonstrated that the ICG fluorescence of ICC tumors themselves was negative. Still, the violated liver could display two ICG fluorescence modes: marginal fluorescence (only the liver parenchyma surrounding tumors showed fluorescence, which was common in ICC with mass forming type), and segmental fluorescence (the liver segment suffering cholestasis due to ICC infiltration showed prominent fluorescence, which was common in ICC mixed with mass forming type and peribiliary infiltrating type) [29]. Of course, there exist limitations in the hepatectomy of tumor-bearing Glissonean pedicle or portal vein watershed under fluorescence navigation: for tumors in some complicated segments (such as S7 and S8), or cases that are difficult to obtain the fluorescent watershed accurately, the conventional regular resection instead reveals more practical significance. However, formulating laparoscopic anatomical resection schemes based on portal vein watershed is one of the directions of future research and exploration.

Furthermore, the Cox proportional hazards regression model exhibited that no matter before or after PSM, preoperative serum CA12-5 and postoperative hospital stay were independent factors affecting overall survival, and lymph node metastasis independently influenced recurrence-free survival. Still, surgical procedure (LLR or OLR) was not an independent factor affecting the long-term prognosis of ICC patients, which was consistent with the Kaplan–Meier analysis results of the two groups after PSM. Some studies had shown that CA12-5 was highly expressed in serum and neoplastic tissue of ICC and related to the long-term prognosis of ICC patients [30‐32]. However, the AJCC/TNM staging system (8th edition) did not list CA12-5 as one of the prognostic factors of ICC, and in clinical practice, clinicians may not put CA12-5 in the same important position as CA19-9. As an essential factor influencing the long-term prognosis of ICC patients after hepatectomy, the postoperative hospital stay could be regarded as a “composite variable.” Although collinearity diagnosis showed no highly correlated linear relationship between the postoperative hospital stay and other variables, the influencing factors of postoperative hospital stay of ICC patients deserve further exploration. Of course, the view of “postoperative hospital stay was an independent prognostic factor” proposed in this study is not advocating the pursuit of the reduction of necessary hospital stay and high bed turnover rate blindly but suggests that patients with extended postoperative hospital stay might suffer long-term prognostic hazard, and the doctors responsible for them should pay attention during follow-up. Compared with hepatocellular carcinoma, ICC is prone to regional lymph node metastasis. To accurately assess the lymph node status of ICC patients, many guidelines or expert consensuses suggest that the number of lymph nodes in the detected areas (around hepatoduodenal ligament, hepatic artery, and the pancreatic head) should be ≥ 6 [33, 34]. However, there is still no agreement on whether to routinely perform extended lymph node dissection, especially in the case of insufficient exposure under laparoscopy and difficulty in expanding the scope of lymph node dissection. Existing studies pointed out that when encountering the following conditions, we could merely perform routine lymph node dissection in LLR without expanding the scope of dissection: (1) tumor diameter < 3 cm, (2) preoperative images did not indicate vascular invasion, and (3) serum CA19-9 and CEA were not elevated before operation [24]. In addition, it should be pointed out that the actual lymphadenectomy rate (n ≥ 6) of the total population included in the study is not high (accounting for 39.4%), which is mainly due to the reason that without pathological examination, a large proportion of ICC could not be diagnosed by preoperative or intraoperative clinical data. When the preoperative imaging examination suggested non-ICC with no abnormal lymph nodes, we would followed the surgical method toward non-ICC (such as hepatocellular carcinoma) and did not implement lymphadenectomy in time. Therefore, exploring the techniques to improve the ability of preoperative differential diagnosis of ICC and other liver tumors and frequently using the intraoperative frozen-section examination of tumor and lymph nodes are conducive to the diagnosis of ICC and the evaluation of lymph node metastasis.

Based on the slightly large sample size of our center, this study demonstrated the short-term prognostic advantages of the LLR group compared with the OLR group and explored the prognostic factors of ICC patients after hepatectomy, but there still exist the following limitations. Although we collected relevant data from a prospective database, the bias caused by this retrospective study itself was the inevitable limitation. Thus, we used PSM analysis to reduce the impact of data bias and confounding variables. In addition, given that preoperative neoadjuvant and postoperative adjuvant therapy significantly affect the long-term prognosis of ICC patients after hepatectomy [3‐5], treatment strategies other than surgery are also worth being included as variables for further exploration.

In conclusion, compared with ICC treated by OLR, the LLR group obtained superior perioperative period outcomes, such as less intraoperative bleeding and fewer severe complications. In the long run, LLR could enable ICC patients to receive an equivalent long-term prognosis compared to OLR. In addition, ICC patients with preoperative abnormal CA12-5, lymph node metastasis, and more extended postoperative hospital stay might suffer from a worse long-term prognosis. However, these conclusions still need multicenter extensive sample prospective research to demonstrate.