Technique of vessel-skeletonized parenchyma-sparing hepatectomy for the oncological treatment of bilobar colorectal liver metastases

This study retrospectively reviewed a cohort of 152 CRC patients with CRLMs who had achieved pathologically confirmed curative resection after having undergone hepatectomy between August 2000 and January 2019 at Okayama University Hospital (Figure 1).

There was no case treated by ablative procedure. The following demographic and clinical data were reviewed through medical records: gender, age at first hepatectomy, lymph node metastasis, genetic profiles, and location of the primary lesion. Clinical data of CRLMs was summarized in size, number, localization, contact with major hepatic veins, extrahepatic spread, information of pre- and postoperative chemotherapy, and preoperative risk assessed with a nomogram by Beppu T et al. (Beppu’s nomogram) and the Genetic And Morphological Evaluation (GAME) score [18, 19]. Beppu’s nomogram and GAME score are associated with postoperative recurrence risk and survival. Postoperative bile leakage and morbidity were assessed based on the definitions of the International Study Group of Liver Surgery [20] and Clavien-Dindo classification [21]. In the pathological examination of the resected specimens, a margin width ≥ 1 mm was defined as R0 resection, and a margin width < 1 mm was defined as R1 resection [22]. R1 status was categorized as R1-Par for tumor exposed to parenchyma or detached from main Glissonean pedicles, and R1-Vas for tumor detached from hepatic vein trunk.

In this study, we focused on patients with 8 or more bilobar CRLMs because of definitely considered to be unresectable. Among our initial cohort of 152 CRC patients with CRLMs, a total of 33 patients possessed 8 or more bilobar CRLMs. The 33 patients were classified into either the VESPAH or Major Hx group by their treated surgical procedure clearly indicated in the medical record; 13 patients were classified into the Major Hx group, and the remaining 20 patients were into the VESPAH group.

The liver anatomy and resection terminology in this paper is based on the Brisbane classification [23]. Liver resections consisting of at least three adjacent liver segments of resection, such as left or right hemihepatectomy (Hr2) and trisectionectomy (Hr3), were defined as major hepatectomy.

Oncologically, it is possible to achieve tumor detachment from just the outer layer of the hepatic vessels. VEssel-Skeletonized PArenchyma-sparing Hepatectomy (VESPAH) permits tumor detachment from the outer layer of the hepatic vessels, even in cases where space between a CRLM and a hepatic vessel is extremely narrow. On the other hand, tumor detachment is not recommended in cases where a metastatic tumor is widely adhered to a hepatic vein or if the radiological findings demonstrate deformed and/or stenotic hepatic vein. In contrast to hepatic veins, Glissonean pedicles can be exposed at the cut surface of the liver and preserved as a liver-skeleton (Figure 2a). However, in cases where metastatic lesions invade the Glissonean pedicles, the lesions should not be detached from the Glissonean pedicles (Figure 2b). In particular, in cases of intrahepatic peripheral bile duct dilatation adjacent to liver metastases, the minimum anatomical area perfused by the Glissonean pedicle should be resected along with the pedicle division on the proximal side of its tumor invasion. In other words, extensive invasion to the first/second-order of Glissonean pedicles requires hemi-hepatectomy or sectionectomy, while Glissonean pedicles beyond the third-order require segmentectomy/sub-segmentectomy. And in other aspect, Vessel-Skeletonization can contribute to navigation to small multiple lesions located in the deep liver parenchyma and to gaining surgical margin. Parenchyma-sparing hepatectomy allowing R1 resection is firstly reported as R1 vascular resection by Italian group [24, 25]. In this analysis, detached from main Glissonean pedicles was defined as R1-par. And, in other aspect, exposure of the liver skeleton can guide liver resection to ensure resection of deep small lesions. These are the ways in which VESPAH is both similar and different from their R1 vascular resection.

Regarding the technical details of VESPAH, precise identification by intraoperative ultrasonography (IOUS) was indispensable for the removal of multiple lesions located in the superficial or deep liver parenchyma and for confirmation of the hepatic vasculature. Parenchymal transection was performed using the cavitron ultrasonic surgical aspirator (CUSA) under intermittent pedicle clamping as Pringle maneuver. In order to shorten the time of Pringle maneuver and to reduce bleeding, if possible, some parts of liver transection could be undergone without the Pringle maneuver. To ensure efficient and certain removal of the lesions, a winding transection line could be set, resulting in the formation of a polymorphic curved cut-surface instead of a flat plane. And intrahepatic vessels thicker than 1–2 mm were ligated or clipped. Operative management for cases where a hepatic vein was adhered metastatic lesions was based on the IOUS findings and preoperative images. The hepatic vein could then be detached from the adjacent tumor at just the outer layer of the vessel by meticulous dissection with Metzenbaum scissors or CUSA. However, in cases where the tumor was adhered a large area of the hepatic vein or if deformities and/or stenotic changes were present, complete resection of the hepatic vein was required. The hepatic vein was not reconstructed in cases where other collateral veins, such as an inferior right hepatic vein or an adequate intrahepatic communicating vein, were present as documented in the preoperative imaging and/or IOUS [26].

The degree of liver transection depended on the type of VESPAH procedure performed (Figure 3). First, in the basic procedure that involved wedge or partial resection of the liver (Hr0), multiple resections, similar to cherry-picking, were performed (Figure 3a) [27]. An example of such case was presented in Figure 4a. In the transected area of the Hr0 procedure, Glissonean pedicles and/or hepatic veins were sometimes widely exposed as a result of detaching the tumor from these structures or of vessel navigation to reach lesions in the deep parenchyma.

Second, in the basic procedure that involved anatomical segmentectomy of the liver (HrS), the Glissonean pedicles invaded by the tumors were sacrificed to minimize the degree of anatomical resection. Each Glissonean pedicle was transected via an extrafascial and transfissural approach with major liver transection [28]. This transection area was determined by staining an anatomical segment via ultrasound-guided puncture and the injection of dye [29]. Thus, in the HrS procedure, several lesions were combined adjacently or separately (Figure 3b). In some special cases, resection of the entire caudate lobe (S1) extending to the ventral liver was performed (An example of such case was presented in Figure 4b).

Finally, in the basic procedure of sectionectomy of the liver (Hr1), two adjacent liver segments were removed. The Glissonean pedicle was ligated and transected via an extrahepatic approach without liver transection [28]. After the demarcation line was confirmed, a transaction area that included several colorectal liver metastases close to the border was determined (Figure 3c). In cases where the colorectal liver metastatic lesions were in contact with an umbilical plate and a middle hepatic vein, anatomical left lateral section of via an intrafascial approach with extension to upper segment 4 and detachment from the middle hepatic vein were included in the VESPAH procedure. Additional Hr0 and/or HrS procedures were added as necessary (An example of such case was presented in Figure 4c). Additionally, for numerous lesions scattered around the anterior sector, anterior sectionectomy was extended at the polymorphic transection line to the middle and posterior sectors with tumor detachment from all three hepatic vein trunks (An example of such case was presented in Figure 4d and Supplementary VIDEO 1. If necessary, Hr0 procedures and/or minimized HrS were combined with these basic VASPAH procedures of HrS and Hr1.

Concerning about perioperative management, quality of liver parenchyma is assessed by liver function test, Indocyanine green (ICG) clearance test, and 99mTc-GSA scintigraphy. And surgical plan is simulated by volumetric software, prior to surgery. It should also be noted that VESPAH increases the risk of bile leak due to the atypical extensive liver cut surface and the skeletonized Glissonean pedicle. Therefore, the Glissonean branch near the hilar region is treated thoroughly and carefully. In cases where the Glisson sheath near the hilum is extensively exposed on the liver cut surface, and leak test are performed routinely. The leak-point is repaired by suturing or ligation, and if there is still a tendency to leak, a C-tube is placed to decompress the biliary tract. As for bile leak, our policy is to wait for spontaneous healing with Drain replacement, but Endoscopic retrograde biliary drainage (ERBD) and reoperation are also considered for Grade B/C.

Overall survival (OS) was calculated from the date of the first hepatectomy for colorectal liver metastases to the date of death or last follow-up (for censored patients). Recurrence-free survival (RFS) was calculated from the date of the first hepatectomy to the date of the first documentation of local, regional, or distant relapse or the appearance of a second primary lesion, as determined by routine computed tomography scans, magnetic resonance imaging scans, or both, performed every 2–3 months. In principle, adjuvant chemotherapy was introduced as soon as possible, following the recovery of liver function. Not only hepatic recurrence, but also extra-hepatic metastatic recurrence like lung and lymph node metastasis were treated with surgical resection whenever possible. And early recurrences are treated with repeat resection after a period of 6–12 months, including chemotherapy.

Statistical analyses were performed using JMP Pro software version 14.3.0 (SAS Institute, Inc., Cary, NC, USA). Continuous variables are expressed as the medians with ranges, and the Wilcoxon/Kruskal-Wallis test was used to compare continuous variables. Categorical variables were expressed as numbers and percentages and compared using the Chi-square test. OS and PFS were estimated by univariate analyses using the Kaplan–Meier method, expressed as medians, and compared with log-rank test. All reported P-values were calculated using two-sided tests, and P-values <0.05 were considered statistically significant.

This study conformed to the Declaration of Helsinki on Human Research Ethics standards and was approved by the Okayama University Hospital Institutional Ethics Board (number 2008-024). Since this study was retrospective in nature, there was no written informed consent from the investigated patients.

Of the total 152 patients, 108 (71.1%) had multiple CRLMs. Among them, 86 patients (56.6%) had multiple bilobar CRLMs, and among these patients, 33 (21.7%) had multiple bilobar CRLMs (≥8 liver metastases). Of the 33 patients with 8 or more bilobar CRLMs, 20 patients (13.2%) were treated with the VESPAH procedure (VESPAH group), and the remaining 13 patients (8.6%) were treated with major hepatectomy (Major Hx group). The demographics of both groups are summarized in Table 1.

All patients in both groups were treated with preoperative chemotherapy before hepatectomy. During preoperative chemotherapy treatment, targeted agents were more frequently used in the VESPAH group (n=20 [100%]) than in the Major Hx group (n=8 [72%]; P=0.011).

CRLMs in the VESPAH group were removed by partial resection (Hr0, n=3), minimal anatomical segmentectomy (HrS, n=8), or minimal anatomical sectionectomy (Hr1, n=9). In contrast, metastases in the Major Hx group were removed by Major Hx, i.e., right/left hemihepatectomy (Hr2, n=10) or right/left trisectionectomy (Hr3, n=3). In the VESPAH group, the patients who underwent the Hr0 procedure had a median of 8 resection areas (range, 6–14), and a median of 9 CRLMs (range, 8–16) were removed. Of the patients who were treated with the HrS or Hr1 procedure, additional Hr0 procedures were combined with basic HrS or Hr1. The patients treated with the HrS procedure underwent a median number of 4 Hr0 resections (range, 1–7), and a median of 9 CRLMs (range, 8–18) were removed. The Hr1 patients underwent a median of 4 Hr0 resections (range, 3–5), and a median of 13 CRLMs (range, 8–53) were removed. In the Major Hx group, the patients treated with the Hr2 procedure underwent a median of 1 Hr0 resection (range, 1–5), and a median of 11 CRLMs (range, 8–41) were removed. The patients who underwent the Hr3 procedure were not treated in combination with Hr0 resection, and a median of 10 CRLMs (range, 9–10) were removed.

As for intermittent vascular occlusion during hepatectomy, the VESPAH group required the Pringle maneuver more frequently and for longer time than the Major Hx group (the median time was 84 min in the VESPAH and 38 min in the Major Hx group, P=0.001). Concerning intra- and perioperative outcomes, there was no operative mortality, and no difference was observed in the mean volume of blood loss, occurrence of posthepatectomy hemorrhage or bile leakage between the two groups; however, the incidence of post hepatectomy liver failure (PHLF) grade B/C was less frequent in the VESPAH group than in the Major Hx group (incidence of PHLF Grade B/C was in 5% in the VESPAH group and 38% in the Major Hx group, P=0.048). After the patients physically recovered and regained remnant liver function, 18 patients (90%) in the VESPAH group received adjuvant chemotherapy within 2 months after surgery, whereas 5 patients (38%) in the Major Hx group received adjuvant chemotherapy within the same time period (P=0.002). The VESPAH group resulted in significantly higher rate of R1 surgery than the Major Hx group (rates of R1-Par/R1-Vasc was 15%/40% in the VESPAH group, and 8%/0 in the Major Hx group, P=0.015). However, the recurrence rate and type after hepatectomy were not different between the two groups. Local hepatic recurrence was comparable between two groups (rate of local recurrence was 7.1% in the VESPAH group, and 14.3% in the Major Hx group, P=0.599). According to surgical margin status, hepatic local recurrence was seen in 8.3% of R0, in 33% of R1-parc, and none of R1-vasc (P=0.269; Supp. Figure 3). Although there was no significant difference, the median number of repeat hepatectomy for intrahepatic recurrence during the patients’ lifetimes tended to favor the VESPAH group (P=0.06).

The median RFS after hepatectomy in the VESPAH group was almost the same as that in the Major Hx group (8 months in the VESPAH and 8 months in the Major Hx group, P= 0.257; Figure 5a). Furthermore, there was no significant difference in median OS after hepatectomy between the two groups; the median OS was 47 months in the VESPAH group and 33 months in the Major Hx group (P=0.481; Figure 5b). Moreover, the 5-year OS rate was 43.8% in the VESPAH group and 39.1% in the Major Hx group.