Indications of Laparoscopic Repeat Liver Resection for Recurrent Hepatocellular Carcinoma

Abstract Aim This study aimed to evaluate the indications of laparoscopic repeat liver resection (LRLR) for recurrent hepatocellular carcinoma from the viewpoint of its difficulty. Methods One hundred and one patients who underwent LRLR and 59 patients who underwent open repeat liver resection (ORLR) were included. The difficulty was classified according to the preoperative predictive factors for difficult LRLR, including an open approach during previous liver resection, history of two or more previous liver resections, history of previous major liver resection, tumor near the resected site of the previous liver resection, and intermediate or high difficulty with the difficulty scoring system. We compared the surgical outcomes between the LRLR and ORLR groups based on the difficulty class (low‐ or intermediate difficiulty class, 0 to 3 predictive factors; high difficiulty class, 4 or 5 factors). Results In the low‐ or intermediate difficiulty class, intraoperative blood loss and the proportion of patients with postoperative complications were significantly lower in LRLR than in ORLR, and the duration of the postoperative hospital stay was significantly shorter in LRLR than in ORLR. In the high difficiulty class, total operative time and operative time before starting hepatic parenchymal resection were significantly longer in LRLR than in ORLR, and there were no significant differences in other surgical outcomes between the two groups. Conclusion LRLR is recommended for patients in the low or intermediate difficulty class. However, LRLR does not have an advantage with longer operative time for patients in the high difficulty class compared with ORLR.


| INTRODUC TI ON
Liver resection is commonly accepted as a curative treatment for hepatocellular carcinoma (HCC), which is usually featured in treatment guidelines. 1,2 Although repeat liver resection has also been accepted as an effective treatment for recurrent HCC (HCCR) in patients with preserved liver function, [3][4][5][6] it is a challenging clinical procedure because the adhesion around the previous liver resection site can often make operative procedures difficult and sometimes results in incidental complications. 7,8 Laparoscopic liver resection (LLR) is a minimally invasive treatment modality and has demonstrated feasible short-and long-term outcomes comparable to open liver resection. [9][10][11][12] Recently, the difficulty scoring system (DSS) for initial LLR has been proposed based on the experience at three high-volume centers in Japan. 13 A multicenter validation study demonstrated that the DSS could predict the difficulty of surgical procedures and short-term outcomes in patients who underwent initial LLR. 14 Although most previous studies have demonstrated the safety of laparoscopic repeat liver resection (LRLR) for patients comparable to initial LLR or open repeat liver resection (ORLR), [15][16][17][18][19][20][21] it is difficult to compare the advantages and disadvantages of LRLR because the difficulty of LRLR is different from those of initial LLR and ORLR. Therefore, the feasible indication of LRLR remains unclear.
Some previous studies have shown that unfavorable surgical outcomes, such as a long operative time and large intraoperative blood loss, after ORLR or LRLR were associated with a history of previous open liver resection, two or more previous liver resections, the relationship between current and previous locations of liver tumor, and a high DSS score. 15,[22][23][24][25] Our previous study revealed five preoperative predictive factors for difficult LRLR: a history of an open approach during previous liver resection, history of two or more previous liver resections, history of previous major liver resection, tumor near the resected site of the previous liver resection, and intermediate or high difficulty in the DSS. 26 These are similar to the risk factors for unfavorable surgical outcomes as described above.
This study aimed to evaluate the indications of LRLR by comparing intra-and postoperative outcomes of LRLR with those of ORLR, based on risk factors for unfavorable surgical outcomes, including the difficulty classification of LRLR reported in our previous study.

| Preoperative risk (predictive) factors for unfavorable surgical outcomes and difficulty classification for LRLR
The preoperative risk factors for unfavorable surgical outcomes after LRLR were selected based on previous studies 15, [22][23][24][25] and our own experiences, including our previous study. 26 The risk factors in the previous studies 15 (Table 1). Prolonged operative time had been defined as >321 min (equivalent to the 75th percentile for the study population; n = 25). Severe adhesion had been defined by the occurrence of one or more of three situations: the patient required >120 min before the start of liver dissection; injury occurred to other organs due to the dissection procedure; or the patient required conversion to open surgery because of the adhesion. Accordingly, 27 patients were classified as having severe adhesion. 26 In our previous study, 26

| Statistical analysis
The Mann-Whitney U-test was used to compare continuous variables. Categorical variables were summarized as numbers and percentages and compared between groups using Fisher's exact test or TA B L E 1 Associations between five predictive factors for LRLR and surgical outcomes in patients who underwent LRLR

| Patients' backgrounds between patients who underwent LRLR and ORLR
Patients' backgrounds are described in Table 3. There were no differences in age, sex, body mass index, comorbid liver diseases, or tumor diameter between the LRLR and ORLR groups. The proportion of patients with a history of previous open liver resection was significantly lower in the LRLR group than in the ORLR group. The proportion of patients with a history of previous major liver resection tended to be lower in the LRLR group than in the ORLR group.
Although there was no difference in the proportion of patients in the

| Surgical outcomes between patients who underwent LRLR and ORLR
A significantly longer total operative time, less intraoperative blood loss, and shorter postoperative hospital stay were observed in the LRLR group than in the ORLR group. There was no difference in the operative time before starting liver parenchymal resection and the proportion of patients with postoperative complications between the groups (Table 4).

| Comparisons of surgical outcomes between LRLR and ORLR in the low or intermediate difficulty class
In the low or intermediate difficulty class (patients with 0 to 3 preoperative predictive factors), there were no significant

TA B L E 2
Surgical outcomes in patients who underwent LRLR according to difficulty classification differences in the total operative time and operative time before starting hepatic parenchymal resection (Table 5). Intraoperative blood loss was significantly less in the LRLR group than in the ORLR group (P < .0001), and the duration of postoperative hospital stay was significantly shorter in the LRLR group than in the ORLR group (P < .0001). Moreover, the proportion of patients with postoperative complications was also significantly lower in the LRLR group than in the ORLR group (P = .014). In the LRLR group, one patient had bile leakage. In the ORLR group, two patients had bile leakage, one had intractable ascites, and two had pleural effusion.

| Comparisons of surgical outcomes between LRLR and ORLR in the high difficulty class
In the high difficulty class (patients with 4 or 5 preoperative predictive factors), total operative time and operative time before starting hepatic parenchymal resection were significantly longer in the LRLR group than in the ORLR group (P = .0009 and P = .0017, respectively; Table 6). There were no significant differences in intraoperative blood loss and duration of postoperative hospital stay between the two groups.

| D ISCUSS I ON
Some previous studies reported that LRLR was not inferior to ORLR in terms of short-and/or long-term outcomes. [15][16][17][18][19][20][21] They reported less blood loss and a shorter hospital stay in LRLR than in ORLR due to minimal damage to structures surrounding the liver, reduction of adhesion formation, and the need for adhesiolysis in LRLR. [15][16][17][18] Although these studies concluded that LRLR can be performed safely in selected patients, the definition of such "selected patients" remains unclear. LRLR can often be difficult, and a safe laparoscopic procedure is not always performed for all patients with HCCR.
Therefore, in this study we evaluated the surgical outcomes of LRLR vancing; therefore, our "current" difficulty classification will not be suitable in the future. However, presently the indication of a surgical approach for repeat liver resection is unclear, and our present classifications may be useful to decide the surgical approach in patients who require repeat liver resection. Third, our present study classified the difficulty simply based on the number of predictive factors, with consideration that the five predictive factors were indicated as the independent risk factors in our present and previous studies. 26 The importance and weight may be different in each factor. A scoring system should be established by a large-number study, considering the importance and weight of each factor, based on the present results.
In conclusion, the difficulty classification evaluated by five pre-