This is a retrospective study of patients receiving ALPPS at Kaohsiung Chang Gung Memorial Hospital and Chiayi Chang Gung Memorial Hospital between January 2013 and December 2017. Patients with acceptable cardiopulmonary function and acceptable anesthesiological risks were considered for liver resection.

Patients were divided into two groups according to their pathological findings based on Ishak modified staging: [11] non-fibrotic liver group (NFL, stage 0) and fibrotic/cirrhotic liver group (FL, stage 1–5/6). Patients with liver tumors underwent relevant investigations and radiologic evaluation, including a fibroscan, contrast-enhanced computed tomographic angiography, and measurement of liver volumes.

According to the preoperative image-based volumetric planning, an FLR volume ≤ 35% of standard liver volume (SLV) in a non-cirrhotic liver patient or ≤ 40–45% of SLV in patients with Child–Pugh A liver cirrhosis or steatosis was considered a potential indication to perform the ALPPS procedure. Indocyanine green retention rate (ICG) should be ≤ 25% at 15 min. The indications of major resection for liver tumors, which result in insufficient FLR included a huge liver mass with size ≥ 8 cm, multiple tumors, and a central location with inevitable major vascular resection. Patients with untreatable liver tumors in the FLR, extrahepatic metastases, and medical contraindications to major hepatectomy were excluded from ALPPS surgery.

The severity of perioperative morbidity and mortality was graded according to the Clavien-Dindo classification [12]. Major complications were defined as Clavien grade ≥ 3B. Grading of post-hepatectomy liver failure (PHLF) was defined according to the International Study Group of Liver Surgery (ISGLS) classification [3].

FLR and total liver volume were gauged before (baseline FLR0) and weekly after ALPPS-I from computed tomography (CT) scans. SLV was calculated by the Urata formula [13]. FLR volumetric changes before and after ALPPS-I were expressed in the form of absolute gain in volume increment and changes in the FLR/SLV ratio. Absolute kinetic growth rate (AKGR) was assessed by the change in volume divided by the number of waiting days to CT volumetry (ml/day). The relative kinetic growth rate was expressed as ratio of AKGR/FLR0.

Liver resections and anatomy were defined according to the Brisbane 2000 Terminology and COUINAUD segmentation of the liver [14, 15].

The meticulous techniques of hepatectomy have been described in a previous paper [16]. The hanging maneuver was carried out through dissecting the retrohepatic area at the caudal end, coursing in a cephalad fashion toward the space between the right and middle hepatic veins. An angled long aortic clamp was gently inserted and advanced in a cephalad direction at the 10–11 o’clock position toward the dissected space between the right and middle hepatic veins. When the tip of the clamp reached the space between the veins, a Penrose tube tape was clenched in the tip of the clamp [17]. The hilar structures were explored, and a vascular loop was passed around the right hepatic artery (HA). The right portal vein (PV) was identified and clamped followed by ligation with silk after the liver partition. A nearly complete or complete parenchymal split [18] with a Cavitron Ultrasonic Surgical Aspirator (CUSA) to the level of the hilar plate just above the inferior vena cava was decided not to induce uncontrollable bleeding to avoid a premature resection of the tumor-bearing liver and possible postoperative bleeding complicated with infection, which would result in a premature second laparotomy. A perioperative Doppler ultrasound check was performed to ensure patency of the intrahepatic vessels except the right PV. A sterile plastic bag (IV bag) was applied in the subhepatic space between the liver and gastroduodenal tracts and transection space to decrease adhesion. At the end of the procedure, the right HA, RPV, main PV, and right intrahepatic duct were encircled with vascular loops to facilitate their identification during ALPPS-II. A cholecystectomy was performed, and a transcystic tube was inserted to check biliary anatomy, followed by marking the transection line of the biliary tract for cutting the bile duct in ALPPS-II and leakage at the end of the liver partition. Two drain tubes were placed in the subphrenic and raw surfaces of the liver. We avoided excessive dissection of the surrounding tissues to prevent adhesions.

During the second stage, the abdomen was entered through the previous incision. Intra-abdominal fluid was collected for routine microbiological examinations. The tumor-bearing liver was resected after transecting the right hepatic duct, right HA, right PV, and right hepatic vein. Possible clots from the stump of the right PV were flushed followed by appropriate suturing. A large amount of saline water was irrigated through the entire abdominal cavity with drains at the right subphrenic and raw surfaces of the transection.

Hematoxylin–eosin-stained slides and immunohistochemical analysis of specimens were reviewed and fibrosis was scored according to the Ishak modified staging. A score of 0–6 was given according to the degree of fibrosis: 0 (no fibrosis), 1–5 (mild–severe fibrosis), and 6 (cirrhosis) [11].

Once discharged home, the patients were followed at our outpatients’ clinic once every month for the first 3 months and thereafter, once every 3 months.

All continuous variables were expressed as medians with interquartile ranges and compared using the Mann–Whitney U test. Categorical variables were compared with Fisher's exact test. The Kaplan–Meier curve was calculated for survival analysis and comparison between groups was tested by the Log-rank test. Spearman’s correlation was used to measure association between the fibroscans and liver volumes. The change in liver volume, including pre-ALPPS-I and -II FLR, pre-ALPPS-I and -II FLR/SLV, increment in FLR/SLV, absolute increment in FLR volume, absolute increment in FLR volume/FLV0, increment in FLR volume /FLV0 per day, rate of hypertrophy, and increment in FLR/SLV per day during the follow-up time were calculated using generalized estimating equations (GEE). GEE were also used to compare these changes between the fibrosis/cirrhosis and normal liver groups. The 2-sided p values < 0.05 were considered statistically significant. Statistical analysis was performed using SPSS Statistics for Windows 25.0 (IBM Corp., Armonk, NY, USA).

The institutional review board gave approval for this study (No. 201902147B0). Informed consent was obtained from all patients before the procedures. All cases were reviewed and discussed by the HCC multidisciplinary committee.