Liver transplant is the best treatment option for HCC. However, due to the scarcity of liver grafts, especially in Asia, liver resection remains the standard and the first-line treatment for patients with resectable HCC and reasonable liver function15. AR was first proposed by Makuuchi et al16. AR is to resect the entire hepatic parenchymal tissue supplied by the portal venous system raining the tumor tissue. The purpose is to remove all possible microscopic tumors and daughter nodules, as HCC tends to invade and spread via the portal vein and its intra-segmental branches, resulting in intrahepatic metastasis4, 5. AR is done by following the anatomical landmark of the vascular landmark, so the adequacy and completeness of resection can be assessed17. AR helps to reduce hepatic dysfunction and disturbance of hepatic regeneration since it leaves only minimal or no area of necrosis18, 19. It could also reduce the chance of liver diffusion and distant metastasis resulting from drop-off of cancer tissue or tumor cells induced by manipulation during operation20. In addition, it also allows precise measurement of the functional liver remnant in preoperative planning, especially in patients with marginal liver function reserve. With these advantages, AR based on the Couinaud classification of liver segments is advocated. In contrast, there is NAR, which is also known as parenchyma-sparing resection. In NAR, a liver lesion is resected with no regard to the anatomical segments. NAR is to preserve as much liver as possible so as to reduce the risk of postoperative liver failure, especially in a background of cirrhosis.
Patient selection for AR or NAR would depend on factors including tumor location, underlying liver status, and the amount of normal tissue sacrifice during liver resection. AR is usually offered to patients with deeply seated or centrally located tumors and relatively better liver function, whereas NAR is usually offered to patients with peripherally located tumors and relatively suboptimal liver function. In this study, more patients in the AR group underwent major resection (69.7% vs 0% before PSM and 45.8% vs 0% after PSM), suggesting deeper and more centrally located tumors, as well as close vicinity to the major pedicle.
The results before PSM suggested that despite patients’ better preoperative liver function with slightly better indocyanine green clearance, the AR group had more complicated operations and hence stormier intraoperative and postoperative outcomes. Furthermore, due to the more advanced tumor status in the AR group, it was not surprising that the AR group had inferior results. The two groups of patients were eventually similar in disease stage and underlying liver function after PSM that reduced preoperative assessment bias. After matching, the groups had similar OS but DFS was significantly better in the AR group. The similar OS could be explained by the higher re-resection rate for the more intrahepatic recurrences in the NAR group. When liver status and tumor and patient factors allowed, re-resection were offered. NAR could preserve a maximal amount of liver parenchyma, allowing more chances of re-resection. In the AR group, despite better DFS, the patients suffered from more eventful intraoperative and postoperative outcomes, resulting in more complications and hence prolonged hospital stay.
The resection line of AR was marked by vascular demarcation after pedicle control. The dissection allowed precise resection of a liver segment based on the vascular supply to the specific segment or section. In contrast, the 1-cm margin resection line in NAR was marked by intraoperative ultrasound, and the resection line did not follow the vascular demarcation, resulting in a narrower resection margin. The principle of HCC resection is complete removal of tumor with adequate margin to ensure tumor clearance. AR might be thought to produce better surgical margin, but one should also consider other factors such as tumor location and underlying liver function before pursuing AR. In fact, a recent study showed that margin might not be necessary in AR21, although further validation is needed.
Macrovascular involvement was found to be an independent factor for poor prognosis. Invasion of the major branch of the hepatic or portal vein is present in about 10–40% of patients at diagnosis22, 23. Surgical resection of tumor with macrovascular invasion remains controversial. Poor OS has been reported, ranging from 8 to 18 months. Some patients might survive longer, depending on the level of the major vessel involvement24, 25. The presence of macrovascular involvement would require resection of the major pedicle, that is, by means of AR. Systemic therapy (such as sorafenib) regardless of the location or extent of macrovascular invasion has been recommended26, with similar survival achieved.
On multivariate analysis, macrovascular involvement and NAR were independent risk factors for disease recurrence. Even that patients in the AR group had more macrovascular invasion, the use of AR stood out to be a factor in reduced disease recurrence. Reported results of AR versus NAR were far from concordant. Marubashi et al27 and Okamura et al28 found no difference in DFS or OS. Ishii et al6 showed that AR could be advantageous over NAR for patients with good liver function and a solitary tumor smaller than 5 cm, while Suh et al29 reported that AR did not achieve superior OS or DFS. Even meta-analyses showed discordant results; some found survival benefits30 but some did not20. However, although one should interpret the results with caution, the evidence was pointing to better DFS with the use of AR. Together with patient optimization and improvement of perioperative conditions (e.g. reduced blood loss), a better outcome could be achieved by AR.
Up to date, there has not been any published randomized controlled trial comparing AR and NAR for HCC, and it would be difficult to use a randomized controlled trial to address this subject since it would be unethical and unfeasible to randomize lesions irrespective of their locations. For this, PSM analysis is a better analysis modality. This study used PSM to minimize baseline differences between patient groups so as to find out the true benefits of the two surgical modes. After matching, the two groups of patients had similar baseline liver function and similar tumor characteristics.
This was a retrospective study and therefore the possibility of selection bias, missing data and treatment heterogeneity throughout the years could not be completely avoided. With a single-center cohort, inter-observer variability and treatment heterogeneity in terms of perioperative management and operative technique were also limited. However, it is hoped that this retrospective study with a reasonably large sample size can shed some light on surgical management of HCC.