From conventional two-stage hepatectomy to ALPPS: Fifteen years of experience in a hepatobiliary surgery unit

Jvier Mupoey Iá?ez , , Ev Mrí Montlvá Orón , Andre Boscà Roledo ,Alonso Cmcho Rmírez , An Hernndo Snz , Plo Grnero Cstro , Alerto Alegre Delgdo , Rfel López-Andújr

a Hepatobiliary Surgery and Transplant Unit, La Fe University and Polytechnic Hospital, Valencia, Spain

b Department of Radiology, La Fe University and Polytechnic Hospital, Valencia, Spain

Keywords:ALPPS Two-stage hepatectomy Colorectal liver metastases Posthepatectomy liver failure

ABSTRACT

Introduction

Hepatectomy of large tumor load may result in postoperative liver failure and associated complications due to excessive liver parenchyma removal [1] . The definition of resectability has evolved through years until the Americas Hepato-Pancreato-Biliary Association Consensus Conference, held in 2006, stated that the indication of resection depends on the presence of sufficient future liver remnant (FLR) after complete (R0) resection.

In an attempt to resect patients with presumably insufficient FLR, two-stage hepatectomies have been developed including portal vein ligation or portal vein embolization in the first stage to induce hypertrophy of the FLR [2-4] . In 20 0 0, Adam et al. described the conventional two-stage hepatectomy (TSH) technique for patients with bilateral multinodular colorectal liver metastasis(CRLM) [5] . The biggest drawback of this approach, described in nearly one-third of patients is the risk of tumor progression during the 4-8 weeks that the FLR hypertrophy needed to avoid posthepatectomy liver failure (PHLF) [ 6 , 7 ].

In 2011, a new concept of two-staged hepatectomy was described [8] . The technique was termed with the acronym ALPPS(associating liver partition and portal vein ligation for staged hepatectomy). It combined portal vein ligation andinsituparenchymal transection during the first-stage hepatectomy, allowing a faster regeneration of the FLR and therefore reducing the possibility of tumor progression. A high resection rate was achieved with ALPPS,but also a high morbidity and mortality rate, representing the major drawback of this technique in the first experiences [9] . Numerous comparisons between TSH and ALPPS have been made since then, with heterogeneous results [10-14] .

Since the ALPPS technique was reported, considering its potential advantages, our surgical team decided to explore this option in patients with CRLM or primary liver tumors with bilateral involvement, performing ALPPS technique when possible. The primary endpoint of this study is overall survival (OS) and diseasefree survival (DFS) evaluation of both techniques: ALPPS vs. TSH during the last 15 years in CRLM patients. Secondary outcomes are resection rate, perioperative morbidity and mortality, and FLR growth of both surgical techniques.

Methods

Study design

A single-center retrospective study was designed to compare TSH with ALPPS in patients with initially unresectable bilateral liver tumors and CRLM. Data on patient demographics, clinical characteristics, tumor type, chemotherapy treatment, operative data, liver volumetric measures, rate of complete resection, postoperative complications, survival, recurrence and rescue treatments were collected.

Patients

All patients who underwent TSH or ALPPS technique in our institution between January 2005 and January 2020 were included in the study. The main criterion for performing these techniques in clinical practice was patients with bilateral affecting liver tumors who underwent hepatectomy with an FLR to total liver volume (TLV) ratio (FLR/TLV)＜25% and/or FLR to body weight ratio (FLR/BWR)＜0.5 [15] , and FLR/TLV＜35% and/or FLR/BWR＜0.7 when liver parenchyma damage: steatosis, fibrosis or chemotherapy-induced liver damage (sinusoidal obstructive syndrome or steatohepatitis) was suspected in the radiological and laboratory tests. Per protocol preoperative liver biopsy was not usually performed, and liver parenchyma damage was confirmed intraoperatively in stage 1 surgery when necessary. ALPPS technique was introduced from January 2012 onwards in all cases. The preoperative study of CRLM patients consisted of tumor markers(CEA, CA19-9), computed tomography (CT) scan, magnetic resonance imaging (MRI) and/or positron emission computed tomography. Patients were evaluated in a multidisciplinary committee, and all of them were informed about the risks of the procedure and signed informed consent.

Liver volumetry

Patients were evaluated before surgery with CT, using Liver Analysis IntelliSpace Portal software (Koninklijke Philips NV, Amsterdam, Netherlands). TLV and FLR excluding the volume of liver tumors were measured. FLR/TLV and FLR/BWR were calculated.In TSH patients, the postoperative control by CT was performed approximately four weeks after stage 1 and repeated every two weeks until a sufficient FLR was reached. ALPPS patients had a CT control on postoperative day 7.

Patients were scheduled for the stage 2 operation if after stage 1 the liver volume reached an FLR/TLV ≥30% and/or FLR/BWR ≥0.5 for patients with normal liver parenchyma, or an FLR/TLV of at least 35%-40% and/or FLR/BWR ≥0.7 when liver parenchyma damage was confirmed.

Surgical technique

All the procedures included in the study were open resections with curative intent. TSH patients in stage 1 underwent wedge resections in FLR associated to intraoperative portal vein ligation of the diseased hemi-liver. Stage 2 consisted in major hepatectomy with arterial, bile duct and hepatic vein section of diseased liver and parenchymal transection. ALPPS technique in our institution has been previously described by Montalvá Orón et al. [16] , and in all ALPPS procedures a complete parenchymal transection and portal vein ligation were performed.

Perioperative management and follow-up

For CRLM cases, neoadjuvant chemotherapy was administered individually, and re-evaluation with CT or MRI was done after 4-6 cycles usually. Disease response was checked according to Response Evaluation Criteria in Solid Tumors (RECIST) [17] . Surgery was performed if stable disease or response was checked; otherwise, chemotherapy was continued with a new line to try to achieve a response. Chemotherapy after stage 1 was administered in TSH patients in case of interstage disease progression or delay to make stage 2. In all cases adjuvant chemotherapy was administered after the surgical interventions if patient status allowed it.Follow-up was performed with CT or MRI and tumor markers every 4 months during 1st year and every 6 months onwards.

Outcome measurements

The primary outcome was the evaluation of OS and DFS at 5 years after TSH and ALPPS techniques in CRLM patients. Secondary outcomes were complete resection rate (failure of treatment was considered in patients who could not reach stage 2 due to insufficient FLR growth, disease progression or surgical complications that prevented the second operation from being performed), perioperative morbidity within 30 days after final intervention according to Clavien-Dindo classification [18] (PHLF was defined according to International Study Group of Liver Surgery criteria [19] ), perioperative mortality at 90-day and liver volumetric response after stage 1.

Statistical analysis

Categorical variables were represented as percentages and compared using the Chi-square test. Continuous variables were expressed as median with interquartile range and compared using the Mann-WhitneyUtest. OS was calculated from the date of stage 1 until death or the last follow-up, and DFS was calculated from the date of stage 2 until the first recurrence. Survival analysis was performed by intention-to-treat. The OS and DFS curves in both groups were generated using the Kaplan-Meier method and compared with the log-rank test. For all tests, the statistical significance was two-tailed, defined asP＜0.05. The analysis was performed using SPSS version 22.0 (IBM, Chicago, IL, USA).

Results

Patients’ characteristics

A total of 38 staged hepatectomies were performed by our institution: 17 TSH and 21 ALPPS during the study period. Baseline demographic and disease characteristics are represented in Table 1 .There were no significant differences in patients’ age, sex and comorbidities. In TSH patients, primary tumor was colorectal cancer in 88.2% (n= 15), 5.9% (n= 1) pancreatic neuroendocrine tumor and 5.9% (n= 1) gastrointestinal tumor. ALPPS patients had as primary tumor colorectal cancer in 81.0% (n= 17), fibrolamellar hepatocarcinoma in 4.8% (n= 1), peripheral cholangiocarcinoma in 4.8%(n= 1), and gastric and duodenal neuroendocrine tumor in 9.5%(n= 2).

Table 1 Baseline demographic and disease characteristics of study patients.

There were no significant differences in primary tumor characteristics in CRLM patients, and Fong risk score [20] for recurrence after hepatectomy was similar between both groups. Within the TSH group, 93.3% (n= 14) were treated with neoadjuvant chemotherapy (in one case, the patient refused the treatment),and 100% (n= 17) in the ALPPS group (P= 0.155), but secondline chemotherapy was administered in 0 patient in TSH vs. 29.4%(n= 5) (P= 0.024) in ALPPS because of no response to the first treatment. Biological agents were administered in 53.3% (n= 8)patients in the TSH group and 82.4% (n= 14) patients in the ALPPS group (P= 0.077), particularly panitumumab was significantly more frequently administered in the ALPPS group ( Table 1 ).Otherwise, interstage chemotherapy was administered in 23.5%(n= 4) in TSH and 0 patients in ALPPS group (P= 0.016) ( Table 2 ).

Table 2 Surgical procedures.

Procedure details

Surgical interventions are represented in Table 2 . There were no significant differences between serum bilirubin, creatinine and INR values before stage 1 and stage 2 in both groups. The median interstage duration was 65 (63-75) vs. 10 (9-11) days (P= 0.008)in the TSH vs. ALPPS groups. Stage 1 surgical time was longer in ALPPS patients: 330 (294-365) vs. 235 (180-285) min (P= 0.012),while stage 2 surgical time was longer in the TSH group: 290 (248-301) vs. 80 (60-115) min (P= 0.013). Stage 1 intraoperative blood transfusion requirement was higher in ALPPS: 0 vs. 2 (0-2.5) U red blood cell transfusions (P= 0.031), while in stage 2 it was higher in the TSH group: 2 (0-2.5) vs. 0 (0-0.3) U (P= 0.043).

Complete resection rate was 76.5% (n= 13) in the TSH group and 85.7% (n= 18) in the ALPPS group (P= 0.426). In TSH patients, stage 2 was not carried out in 23.5%: 4 patients with CRLM,due to interstage disease progression. While in the ALPPS group 3 patients (14.3%) did not reach complete resection: 1 with duodenal neuroendocrine tumor liver metastasis, and 2 with CRLM, all of whom were due to PHLF grade C. Complete R0 resection was achieved in 84.6% (n= 11) vs. 77.8% (n= 14) (P= 0.159) of the patients who fulfilled the two stages procedures in the TSH and ALPPS groups, respectively.

Volumetry study outcome

There were no significant differences in preoperative liver volumes between both groups. Volumetric studies are shown in Table 3 . In TSH patients preoperative FLR volume was 367.5 (224.5-486.7) mL while in ALPPS patients it was 371.0 (300.5-481.5) mL(P= 0.435), meaning an FLR/TLV of 27.3% (18.4%-36.7%) vs. 25.5%(20.1%-30.6%) (P= 0.389) and an FLR/BWR of 0.49 (0.31-0.60) vs.0.54 (0.43-0.66) (P= 0.564) in the TSH and ALPPS groups, respectively.

Kinetic growth rate was higher in the ALPPS group: 3.49(1.37-8.29) mL/d in TSH vs. 26.06 (15.75-36.55) mL/d in ALPPS(P= 0.014). Postoperative FLR growth was similar between the two groups: median postoperative FLR/TLV was 34.1% (26.4%-41.0%) in the TSH group vs. 40.5% (31.7%-51.7%) (P= 0.373) in the ALPPS group. FLR increase was 51.0% (17.7%-85.1%) vs. 45.5% (26.1%-68.2%)(P= 0.254) and FLR/BWR achieved was 0.68 (0.58-0.89) vs. 0.81(0.67-1.10) (P= 0.456) in TSH and ALPPS patients, respectively.

Postoperative morbidity and mortality

Overall major morbidity (Clavien-Dindo ≥3a) (stage 1 + stage 2) was 41.2% (n= 7) in THS and 33.3% (n= 7) (P= 0.389) in ALPPS patients; 90-day mortalities were 11.8% (n= 2) in THS vs. 19.0%(n= 4) (P= 0.654) in ALPPS ( Table 4 ).

There were no significant differences in major morbidity and mortality between the two groups. Pre-stage 1 and pre-stage 2 ALPPS risk scores for mortality proposed by Linecker et al. [21] were also similar between the two groups.However, some differences were observed. In the TSH group, major morbidity (Clavien-Dindo ≥3a) was 17.6% (n= 3) after stage 1, comprising one patient experiencing intra-abdominal bleeding requiring re-operation, and two patients who needed abdominal percutaneous drainage due to postoperative ascites and collection.After stage 2, 29.4% (n= 5) developed major morbidity: two patients with postoperative bile leak were treated with percutaneous drainage, and one patient with PHLF grade B. Two patients died after stage 2: one due to PHLF grade C and one with sepsis and multiple organ failure caused by intra-abdominal infection.

In the ALPPS group the major morbidity after stage 1 was 23.8% (n= 5). One patient developed pleural effusion and intraabdominal collection requiring thoracic and abdominal percutaneous drainages, and 4 patients developed PHLF grade C, 3 of whom dying before stage 2, which was the major difference in comparison with TSH (0% vs. 19.0% of PHLF grade C after stage 1). In stage 2, major morbidity was 19.0% (n= 4): one patient with pleural effusion and one patient with intra-abdominal collec-tion, both were treated with percutaneous drainage. One patient with intra-abdominal collection developing sepsis with kidney injury was managed with percutaneous drainage and renal replacement therapy, and one patient died after stage 2 as a consequence of PHLF grade C ( Table 5 ).

Table 3 Liver volume study outcomes.

Table 4 Surgical morbidity and mortality.

Table 5 Deceased patients.

Overall survival and disease-free survival in CRLM patients

In CRLM patients, the median follow-up was 28.3 (18.1-73.5) months in TSH vs. 28.2 (7.5-47.7) months in ALPPS series.Intention-to-treat OS rates at 1 and 5 years for TSH were 80% and 33%, and for ALPPS patients 76% and 35%. Median OS was 24.4(13.2-72.8) vs. 23.2 (12.1-48.3) months (P= 0.913), respectively( Fig. 1 A).

DFS rates at 1 and 5 years were 36% and 27% in TSH vs. 33%and 27% in ALPPS group, respectively. Median DFS was 11.7 (7.5-61.1) vs. 5.6 (2.1-13.9) months in TSH vs. ALPPS patients (P= 0.261)( Fig. 1 B).

Recurrence types are described in Table 6 . In resected TSH patients (n= 11), liver-only recurrence occurred in 1, extrahepatic recurrence in 5 and liver and peritoneal in 1, and two cases had a rescue treatment: one with liver radiofrequency ablation and the other with hepatectomy and cytoreductive surgery with hyperthermic intraperitoneal chemotherapy, both with recurrence disease in the follow-up at the end of the study.

Fig. 1. Overall survival ( A ) and disease-free survival ( B ) in colorectal liver metastases patients.

Table 6 Follow-up in CRLM patients.

In resected ALPPS patients (n= 15), liver-only recurrence occurred in 4, extrahepatic recurrence in 5 and liver and lung in 1.Four cases had rescue treatment: one re-hepatectomy, one liver radiofrequency ablation, one transarterial chemoembolization with irinotecan eluting beads (DEBIRI) and one case with hepatectomy followed by lung metastasectomy. Only one of them was diseasefree at the end of the study. The rest of disease recurrences were treated with chemotherapy in both groups.

Discussion

ALPPS technique, from its original description by Schnitzbauer et al. [9] , has generated a big controversy [22] . Different changes and variations of ALPPS technique [23-25] have been developed in an attempt to improve the high morbidity and mortality reported in the first experiences [ 10 , 12 , 26 ]. The ever-constant question is whether these problems could be offset by the rapid hypertrophy of the FLR [27] and the increased resectability rate [28] achieved with ALPPS technique in relation to TSH. This increased resectability, however, does not necessarily translate into improved oncologic outcomes in the setting of CRLM [22] . Therefore, our study’s primary outcome is to assess whether the advantages of ALPPS can achieve differences in long-term survival compared to those of TSH.

As expected, in this series liver hypertrophy was faster in ALPPS patients, with an interstage time interval to volumetric control of 7 vs. 38 days. Still, in contrast with previous findings [ 13 , 29-31 ],hypertrophy or FLR reached was similar between both procedures;there was no case in which stage 2 was not achieved due to insufficient FLR hypertrophy in TSH.

It has been proven that the resectability rate is higher with the ALPPS technique against TSH [ 14 , 28 , 32 ]. Our results support these findings. In TSH the only reason of dropout was interstage disease progression in CRLM patients. At the same time, in the ALPPS group, it was the postoperative complications and deaths,especially PHLF, which is one of the biggest liver surgeon concerns about the ALPPS technique. One considerable finding in our study is that major postoperative complications (Clavien ≥3a) rates were similar between both groups (41.2% in TSH vs 33.3% in ALPPS).In the ALPPS group, however, there was higher mortality. Therefore, complications were more lethal in the ALPPS group, related to PHLF grade C after stage 1.

In relation to morbidity and mortality, our series have some differences with other similar studies comparing TSH and ALPPS:Ratti et al. [33] , Adam et al. [34] , Chia et al. [27] and the first randomized trial (LIGRO trial) by Sandstr?m et al. [28] . In these studies, morbidity was quite similar between both techniques (14%-43%vs. 28%-43% in TSH vs. ALPPS) but mortality was lower than our series (2%-9% in TSH vs. 8%-12% in ALPPS) even though it is also higher in ALPPS patients in these studies, except Adam et al. It is particularly striking that major complications and mortality in our ALPPS patients occurred after stage 1, but not after stage 2, as usually reported in the studies previously mentioned. We do not have a clear explanation for that. We presume it is related to the learning curve of the procedure itself, probably a too high risk was assumed in some patients, as we can check with the ALPPS risk score reported by Linecker et al. [21] , where a higher risk of 90-day or in-hospital mortality was related to patients with more than 67 years, with biliary tumors after ALPPS-stage 1, and those after ALPPS-stage 2 with major interstage complications, raised serum creatinine or bilirubin. In accordance with these data, all the cases with postoperative mortalities were older than 67 years, except one case who underwent the second ALPPS procedure of our series: a CRLM relapse after a previous major hepatectomy, and considering the current experience gained from that time, we think it should not have been resected. On the other hand, none of the deceased patients had biliary tumors.

It is known that CRLM candidates to these types of surgeries have a high tumor burden. Recurrence is so frequent that some studies do not encourage to carry on with these techniques [ 35 , 36 ].Oldhafer et al. [30] reported 85.7% recurrence with a median time of 8 months after surgery, and Wanis et al. [37] reported a 3-year DFS of 13% in CRLM resected with ALPPS technique. Olthof et al. [22] compared 70 ALPPS resections with patients receiving systemic therapy with palliative intent, showing that early oncological results were not superior in resected patients. Despite this, there are some studies with better results [ 34 , 38 ]. Robles-Campos et al. [39] compared 21 matched cases with tourniquet-ALPPS and TSH, reporting 1-year OS of 81% vs. 76.2%, similar to the present series and 1-year DFS of 66.7% vs. 44.6% between TSH and ALPPS, higher than our results. Remarkably, and unlike most previous findings, in LIGRO trial [40] , ALPPS patients had a significantly longer median OS than those randomized to TSH (46 vs.26 months, respectively). On multivariable analysis, allocation to ALPPS, resection of liver metastases, ASA, size of the largest metastases, and postoperative complications (Clavien-Dindo ≥3a) had a significant impact on the outcome. Moreover, there may be some other factors likely to play a role for tumor recurrences in these patients, such as mutations in genes associated with negative oncologic outcome after surgical resection [41] . All these variables could explain the differences between such heterogeneous results in different studies mentioned above, including ours. It is striking in our series that a higher proportion of patients in the ALPPS group required the second-line preoperative chemotherapy than the TSH group. Possibly these factors are as important as achieving a complete resection surgery regarding long-term oncological results.

Based on our experience, ALPPS is the best option to get a complete resection in CRLM patients, as the dropout was higher in TSH(23.5% vs. 11.8% for CRLM cases), in which the only cause was disease progression. Perhaps this dropout acts as a selection of oncological favourable patients candidates to stage 2, as other authors have already proposed [ 42 , 43 ], but this does not translate in longterm OS nor DFS differences at 5 years in our series. On the other hand, sufficient FLR hypertrophy was achieved in all cases in TSH,even though now we know that rescue-ALPPS [28] is a good option for insufficient FLR after stage 1 in TSH surgery.

The main limitation of this study is that it is a retrospective evaluation of patients during a long period of time, in a singlecenter, and the majority of surgeries performed in each group belong to different periods. Moreover, the small sample size does not allow firm conclusions to be drawn. The greatest value of the present study is that it reflects in a real way the evolution and long-term follow-up of two-stage hepatectomies for high liver tumor burden patients over time, before and after the appearance of the ALPPS technique in a hepatobiliary surgery unit.

In conclusion, ALPPS could be an effective alternative to TSH in bilateral affecting liver tumors, allowing higher resection rate thanks to a rapid FLR hypertrophy. It also has higher mortality,mainly due to severe PHLF, so it is of utmost importance to select the candidates based on the risk factors. In our experience, in CRLM patients similar long-term OS and DFS can be achieved with both techniques.

Acknowledgments

We thank Mr. Arash Javadinejad, English Instructor and Publication Editor at the Instituto de Investigación Sanitaria La Fe (Valencia, Spain), for professional English language editing.

CRediT authorship contribution statement

Javier Maupoey Ibá?ez: Conceptualization, Data curation, Formal analysis, Methodology, Visualization, Writing - original draft.Eva María Montalvá Orón: Conceptualization, Formal analysis,Methodology, Visualization, Writing - original draft. Andrea Boscà Robledo: Formal analysis, Resources, Writing - review & editing.Alonso Camacho Ramírez: Conceptualization, Data curation, Writing - review & editing. Ana Hernando Sanz: Conceptualization,Data curation, Writing - review & editing. Pablo Granero Castro:Conceptualization, Methodology, Writing - review & editing. Alberto Alegre Delgado: Data curation, Resources, Writing - review& editing. Rafael López-Andújar: Conceptualization, Methodology,Project administration, Supervision, Writing - review & editing.

Funding

None.

Ethical approval

This study was approved by the Ethics Committee for Research of La Fe University and Polytechnic Hospital (No. 2020-086-1).

Competing interest

No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.