Laparoscopic Versus Open Liver Resection for Colorectal Liver Metastases—Which Is a More Suitable Standard Practice?

Two international consensus conferences for laparoscopic liver resection (LLR) were held separately in Louisville, USA (2009) and Morioka, Japan (2014).1,2 The Morioka consensus conference included independent jury members according to the Zurich-Danish model of consensus conference, and used GRADE to assess secure evidence.3–5 According to the jury recommendations, minor LLR was positioned to be a standard practice but was still in the assessment phase (IDEAL stage 3) as it is adopted by more surgeons, whereas major LLR was positioned to be an innovative procedure in the exploratory phase (IDEAL stage 2b).4 To our knowledge, a completed randomized controlled trial (RCT) comparing the efficacy of LLR and open liver resection (OLR) for patients with colorectal liver metastases (CRLM) has not yet been published. To date, only 3 studies have assessed the utility of LLR versus OLR using propensity score matching (PSM) analysis involving more than 100 CRLM patients treated by LLR.6–8 A well-designed PSM analysis can reduce selection bias of 2 procedures and can provide a similar reliable outcome of RCT;9 however, the results cannot be applied to all patients but are limited to selected patients. More recently, a meta-analysis of these PSM studies has been published.10 Large-scale PSM studies and this recent meta-analysis have demonstrated almost similar results: longer or similar operative time, reduced intraoperative blood loss or less transfusion requirement, shorter hospital stay, lower or comparable morbidity, and equivalent mortality in LLR patients compared with OLR patients.6–8,10 We congratulate Dr Fretland et al11 for publishing the first elegant the Oslo laparoscopic versus open liver resection for colorectal metastases (OSLO-COMET) RCT in this issue of Annals of Surgery. Their RCT was registered in ClinicalTrials.gov (NCT01516710) in January 2012.12 Another RCT comparing open versus laparoscopic left lateral liver sectionectomy (Orange II study; NCT00874224) was discontinued because of a delay in patient recruitment.13 The OSLO-COMET RCT is a single-center study in Oslo University Hospital with data collection performed from February 2012 to January 2016. Enrollment was limited to CRLM patients undergoing a parenchyma sparing minor liver resection, defined as a resection of less than three Couinaud segments. The authors who operated have an experience of more than 400 LLRs at the start of this study; therefore, they have sufficiently overcome the learning curve for minor LLR.14 Mortality, morbidity, and resection margins were the three principal comparators considered by the jury in the Morioka conference2 and were clearly resolved in this RCT. Mortality was low (0/133 in LLR and 1/147 in OLR) and comparable in the 2 groups. Morbidity (complication of Accordion grade 2 or higher)15 was a primary endpoint and was significantly lower in LLR than in OLR. Accordion grade 2 included complications that required pharmacological treatment, transfusion, or total parenteral nutrition. Major morbidity is usually defined as Accordion grade 3 or higher (Clavien-Dindo ≥ grade III) and is an indicator of poor survival in CRLM patients undergoing liver resection.15–17 Using this criterion of Clavien–Dindo ≥ grade III, morbidity was judged to be comparable (11% in LLR; 14% in OLR) in the 2 groups. It is essential to keep adequate resection margins for malignant disease; there have comparable percentages of involved resection margins on different surgical approaches (6% in LLR; 7% in OLR). Although the amount of blood loss or transfusion rate was less in LLR than in OLR in all PSM studies,6–8,10 the difference in the amount was small but comparable to that in these RCT groups. LLR is thought to be superior to OLR because laparoscopy allows a better magnified view, and pneumoperitoneum reduces bleeding from low pressure vessels during liver transection.2 Recently, it has been reported that intraoperative conversion due to an unfavorable intraoperative event is associated with worse outcomes, namely, longer operative time with substantial blood loss, longer hospital stay, increased morbidity, and increased mortality.18 High-risk factors for conversion included neoadjuvant chemotherapy, prior liver resection, malignant tumor resection, posterosuperior location, and major resections.18 Although the conversion rates were low in this RCT (2% for conversion to laparotomy and 5% to hand-assisted LLR), this specific disadvantage of LLR should be continuously investigated. Furthermore, recurrence and long-term prognosis are extremely important issues for CRLM patients undergoing LLR. In the future, secondary endpoints (ie, overall survival, disease-free survival, and recurrence-free survival) will be evaluated. Meta-analysis of a PSM study10 reported that a better 3-year overall survival can be achieved by LLR compared with OLR. When comparing LLR and OLR, it is important to evaluate the correct complexity of liver resection before surgery. This RCT used 2 complexity examination systems and found no difference between the groups. Some key factors involved include tumor location and size, operative method, tumor proximity to the major vessels, and liver function. However, these systems do not include a quantification of adhesions after previous operations. Adhesion is one of the most frequent unfavorable intraoperative findings.18 In fact, this RCT included somewhat repeated liver resections (18% in LLR; 9% in OLR), and LLR has been generally thought to be more convenient for repeated resections than OLR.19 The extent of adhesion after LLR is less than that after OLR, and in the second operation, the area that requires adhesiolysis is limited in LLR. Although RCT can provide high-quality evidence, it is difficult to perform in the modern era because more patients choose LLR rather than OLR. For some liver surgeons, their operative skills for LLR might not be as good as those for OLR because of their lack of familiarity with the laparoscopic procedure or insufficient LLR experience. In contrast, some skilled liver surgeons for LLR believe that LLR is theoretically better than OLR.20 The Morioka consensus conference has strongly recommended more educational activities and a prospective registry for further promotion.2 In Japan, the Japanese Endoscopic Liver Surgery Study Group has already started organizing hands-on educational workshops using an animal model followed by the establishment of a nationwide prospective registry.21 More than 600 participants have been targeted for these workshops and much effort has been exerted for the safe promoting of LLR. In October 2015, preoperative online registration was required for all candidates who intended to perform LLR. “Pretreatment,” “surgical findings,” “discharge,” and “readmission” data are presently being collected. In June 2017, covering 21 months of observation, 4257 patients which included 22% of CRLM patients and 291 facilities were registered. Of late, about 300 patients have been registered monthly and the 30-day and 90-day postoperative mortalities were 0.07% and 0.21%, respectively. Recently, the International Laparoscopic Liver Society has been established.22 International Laparoscopic Liver Society has several important missions, namely, facilitate the promotion and educational activities of LLR for meaningful improvements in patient care and coordinate international registries. In conclusion, this OSLO-COMET RCT article by Dr Fretland et al in this issue of Annals of Surgery has provided evidence of the safety and less invasiveness of LLR. However, further evidence is required from additional RCTs or multicenter prospective registries. Importantly, the future of LLR also strongly depends on education initiatives which should be carefully planned and regularly implemented.