Oncological outcomes of major liver resectionfollowing portal vein embolization: a systematic review and a meta-analysis
Mariano Cesare Giglio,MD; Alexandros Giakoustidis, PhD; Ahmed Draz, MD; Zaynab A.R. Jawad MRCS MSc; Madhava Pai, MSc, FRCS; Nagy A Habib ChM FRCS; Paul Tait MA FRCR and Long R. Jiao, MD FRCS
Hepatopancreatobiliary Surgical Unit, Department of Surgery and Cancer,Hammersmith Hospital, Imperial College London
Correspondence:
Professor Long R. Jiao, MD, FRCS
Professor of Surgery, Chair and Consultant in HPB Surgery
Hammersmith Hospital , Imperial College London
Du Cane Road, London W12 0HS
Tel: +44(0) 20 8383 3937
Fax: +44(0) 20 8383 3212
Email:
We declare that there is no financial interest in or financial conflict to disclose with the subject matter or materials discussed in the paper
Outcomes after portal vein embolization
SYNOPSIS
Preoperative portal vein embolization has no adverse effect on hepatic recurrence or overall survivalin patients with CRLM undergoing major liver resection.
ABSTRACT
Background Preoperative portal vein occlusion with either percutaneous portal vein embolization (PVE) or portal vein ligation (PVL) is routinely used to induce liver hypertrophy prior to major liver resection in patients with hepatic malignancy. While this increases the future liver remnant (FLR) and hence the number of patients suitable for resection, recent evidence suggests that induction of liver hypertrophy pre-operatively may promote tumour growth and increase recurrence rates. Aim of the current study is to evaluatethe impact of PVE on hepatic recurrence rate and survival in patients with colorectal liver metastases (CRLM).
MethodsMedline, Embase and Web of Science databases were searched to identify studies assessing the oncological outcomesof patients undergoing major liver resection for CRLM following PVE. Studies comparing patients undergoing one stage liver resection with or without pre-operative PVE were included. The primary outcomewas hepatic recurrence (HR). Secondary outcomes were 3- and 5-year overall survival (OS).
ResultsOf the 2131 studies identified, sixnonrandomized studies (n=668) met the eligibility criteriacomparing outcomes of patients undergoing major liver resection withor without PVE (n=182 vs.n=486 respectively). The median follow-up time ranged from 23.5 to 46 months. There was no significant difference in HR (OR, 0.78; 95% CI, 0.42 to 1.44, p=0.41), 3-year OS (OR, 0.80; 95% CI, 0.56 to 1.14, p=0.22) or 5-year OS (OR, 1.12; 95% CI, 0.40 to 3.11, p=0.82).
ConclusionPVE in patients with CRLMhas no adverse effect on hepatic recurrence or overall survival following major liver resection.
INTRODUCTION
Liver resection remains the only treatment offering bothlong-term survival1–3 and potential cure4 in patients with colorectal liver metastases (CRLM). The aim of surgery is to offer curative resection with preservation of a sufficient future liver remnant (FLR)5to minimizepost-operative morbidity and mortality6. Unfortunately, at the time of diagnosis less than 25% of patients are suitable for surgery7 with insufficient FLRbeinga principle limiting factor8.Portal vein occlusion (PVO) with either percutaneous portal vein embolization (PVE) or operative portal veinligation (PVL) has been widely usedto induce pre-operative hypertrophy of the FLR in order to increase the resection rate9–11.Although PVE can increase the liver volume on average by 11.9%12, recent studies have shown that it could also promote tumour growth13–154.Furthermore, the rate of tumour growth was more rapid than liver regeneration15. As a result, the impact of PVE on hepatic recurrence (HR) and long-term oncological outcomes after resection of CLRM remains debatable with some studies reporting a poor overall survival in the PVE group15,16.
Over the last 5 years, there has been an increasing interest in surgical approaches to induce liver hypertrophy such as Associating Liver Partition and Portal vein ligation for Staged hepatectomy(ALPPS)17, Radio-frequency-assisted Liver Partition with Portal vein ligation (RALPP)18 and Associating Liver Tourniquet and right Portal occlusion for Staged hepatectomy technique (ALTPS)19. The advantages of these procedures over PVE are still being evaluated. The aim of this meta-analysis is to evaluate the impact of PVEon hepatic recurrence and survival following major liver resectionin patients with CRLM.
METHODS
Search strategy
A systematic literature search of Medline, Embase and Web of Science databases was carried out. The following search terms were used: “portal”, “embolization”, “embolisation”, “occlusion”, “ligation”, “metasta*” (Table, Supplemental Digital Content 1 containing the full electronic search strategy for Medline)with no language or publication status restrictions. The reference list of retrieved articles was used to identify additional eligible studies. The last search was carried out on 10thOctober 2015.
Study selection
To be eligible for the analysis studies had to: (1) include patients over 18 years old undergoing one-stage liver resection for colorectal liver metastases with curative intent; (2) compare patients undergoing major liver resection withpre-operative PVE (PVE group) with those not undergoing pre-operative PVE (noPVE group); (3) report the rate of hepatic recurrence in both groups; (4) follow-up patients for hepatic recurrence for at least 12 months. Definition of major liver resection includes: 1) left or right hepatectomy; 2) right trisegmentectomy and 3) right or left extended hepatectomy. Two reviewers (M.C.G. and A.G.) independently assessed the retrieved references at title and abstract level to identify potential eligible studies. Any conflict was resolved by a third reviewer (A.D.)until a consensus was reached. The full-text of these studies was then retrieved for further analysis.
Data extraction, quality and risk of bias assessment
Two authors independently extracted or calculated relevant data from each included study by completing an electronic database with the following information: first author, year of publication, country of origin, indication for PVE,time between PVO and surgery, duration of the follow-up, characteristics of patients (Table, Supplemental Digital Content 2), rate of hepatic recurrence, 3 year and 5 year overall survival. When the rates of overall survival at 3 and 5 years were not available in the text, thesewere retrieved by the analysis of the published Kaplan-Meier survival curve.
The Newcastle-Ottawa Scale (NOS)20 was used to appraise the quality of the included studies. This tool assessesthe risk of bias in three domains: Comparability of patients,Selection of patients and evaluation of theOutcome. We used the NOS to appraise the last two domains and evaluatedthe former with a tailored assessment, thus being able to considerthe comparability of the two groups for more than the twofactors allowed by the NOS. We achieved this as follows.Patients requiring PVEbefore surgery usually have a higher tumour burden than patients not requiring PVE and consequently the first groupmaybe at higher risk of recurrence (baseline RecRISK) independently from the embolization.Thus we considered if the groups (PVE and no PVE) werecomparable for factors known to influence the risk of hepatic recurrence after resection of CRLM21: type (synchronous vs metachronous), number,size, distribution (unilobar vs bilobar) of metastases,perioperative chemotherapy, rate of R0 resections, stage of primary tumour.Publication bias was assessed by graphical exploration using funnel plots of study resultsagainst precision of the study.
Outcome analysis
The primary outcome of this meta-analysis was hepatic recurrence (HR). Secondary outcomes were 3-and 5-year overall survival (OS).Odds ratio (OR) and 95 % confidence intervals (95% CI) were used as summary measures. The point estimate of the OR was considered statistically significant at p0.05 if the 95% CI did not include 1. They were calculated with the fixed effects Mantel-Hænzel model unless there was significant heterogeneity across studies in which case the random effects DerSimonian and Laird model was used. The heterogeneity among studies was tested by the Q statistic and quantified by the I2statistic22. As a guide, I2 values of less than 25% indicates low heterogeneity, 25–50 % moderate heterogeneity, and greater than 50 % indicates high heterogeneity. For dichotomous analyses with zero count cells, 0.5 was added to each cell for the analysis.Sensitivity analyseswereperformed to test if the results were robust to our assumptions andmethodological decisions.The influence of each individual study included in the meta-analysis was investigated by omitting each study in turn and re-estimating the summary effect.When a fixed effect model was used on the first, the analysis was repeated using a random effect model. Statistical analysis was performed using STATA 12 statistical software (STATA Corp, College Station, Texas, USA). The study was reported in accordance with the MOOSE guidelines for reporting meta-analysis of observational studies23.
RESULTS
Study selection
The literature search identified a total of 3322 studies. 1191 were found to be duplicates and were removed. Of the 2131 remaining studies, 2104 were discarded because they failed to meet the inclusion criteria.The full text of the remaining 27 articles was examined in more detail. Six studies were included in the meta-analysis (Figure 1)8,14,16,24–26.
Study characteristics
The included studies consisted of 668 patients undergoing liver resection for CRLM with pre-operativePVE (n=182) and without pre-operative PVE (n=486, Table, Supplemental Digital Content 2). In all studies, PVE was the chosen method for induction of liver hypertrophy before major resection. However, in one study8, nine patients in the PVE group(28%)eventually underwent PVL as the FLR was considered insufficient at the time of laparotomy. In two studies16,25 a few patients had two-stage hepatectomy(n=25 and n=11 respectively). The mean time between PVE and surgery was 52.8 days. The median follow-up time ranged from 23.5 to 46 months. Other characteristics of the included studies are shown in Table 1.
Quality assessment and risk of bias
Results of the assessment of the quality of the studies and the risk of bias are shown in Table 2.
The NOS evaluation revealed no concerns about the selection of patients (all included studies reported 4 / 4) and the quality of outcome evaluation ranged between 2 and 3 (out of a maximum score of 3). The two group of patients were comparable in terms of R0 resections and use of perioperative chemotherapy in all the included studies. Inthree studies 8,14,16the PVE group had a higher baseline RecRISK compared with the no PVE group, while the opposite occurred in one study24.
Outcomes analyses
There was no difference in HR between patients in the PVE and no PVE group (OR, 0.78; 95% CI, 0.42 to 1.44, p=0.41;Figure2). OSat 3 years was similar between the two groups (OR, 0.80; 95% CI, 0.56 to 1.14; p=0.22; Figure 3a) with no significant heterogeneity (Q, 7.37, p = 0.19). At 5 years follow-up, there remained no significant difference in survival (OR, 1.12; 95% CI, 0.41 to 3.11, p=0.82, Fig. 3b) but heterogeneity between the studies was detected (I2 = 79.5%, p = 0.001). The robustness of the results was confirmed by sensitivity analyses. When a fixed effect model was used on the first, analyses using a random effect model confirmed the results (3 years OS: OR, 0.9; 95% CI 0.65 to 1.95). Influence analysis showed that results (HR and OS) were no dependent on the inclusion of any study. Furthermore, this analysis identified the study by Oussoultzoglou E et al 24 as the source ofthe inter-study heterogeneity detected about the HR rate(I2, 54.2%; Q, 10.9; p = 0.05). In fact after excluding this studythere remained no significant heterogeneity (I2 = 0%; p = 0.79) nor significant difference in HR (OR, 1; 95% CI, 0.68 to 1.47). No difference in HR was found when the analysis was restricted to studies comparing patients undergoing right and extended right hepatectomy (OR, 0.54; 95% CI, 0.13 to 2.30; p=0.4). Funnel plot for the primary outcome showed symmetry (Figure, Supplemental Digital Content 3) which was confirmed using Peters’ (p = 0.91) and Harbord’s modified (p = 0.70) tests.
DISCUSSION
The effects of PVE on tumour growth in patients with CRLM have been previously reported showing in the embolized lobe an increase intumour volume14 (20.8% at 3 weeks) and tumour growth rate27,28. Similar findings have been reported in the non-embolized lobe15 with the growth rate of metastases having been shown to be more rapid than that of the regenerating liver parenchyma15. Different factors have been considered responsible for these changes28. As a consequence of portal flow reduction in the embolized lobe, arterial compensation occurs with decreased hepatic artery resistance29 and increased arterial blood flow30 which could explain the progression of intrahepatic metastases whose blood supply is exclusively arterial31. As the tumour in the embolized lobe will be resected, these effects are not necessarily clinically relevant, unless this progression involves the hilum or the planned resection plane28. Conversely in the non-embolized lobe, factors involved in liver regeneration may also stimulate tumour growth and the reactivation of dormant micrometastases28,32to increasethe risk of liver recurrence.
The current study revealed that preoperative PVE does not adversely affect oncological outcomes after major liver resection in patients with CRLM. No difference in the incidence of hepatic recurrence was found when comparing PVE and no PVE groups.Although inter-study heterogeneity was detected, influence analysis identified the study by Oussoultzoglou E et al 24as the source of this heterogeneity. Indeed these authors found that PVE reduced HR after right hepatectomy for CRLM. However in this study the PVE group had more patients with single metastasis (Table 2) than no PVE group.Due to the small number of patients in PVE group (n=23) and the strong degree of correlation between presence of single lesion and PVE, the multivariate analysis might have not been able to discriminatethe independent contribution of these two variablesto the reduced recurrence rateidentified in this group. In this meta-analysis, after excluding this study with the result from pooled analysis being more consistent, there remained no difference in HR.
Our analysis also showed no significant difference in 3- or 5-year OSbetween the two groups. These results are consistent with most of the evidence published in literature 8,11,24,26,33–37. Several studies have analysed the effect of PVE on disease-free survival (DFS), with controversial results, showing both unaffected 24,33 and compromised 14,16,25 DFS. In most of them, theDFS analysed was not liver specific, affected by colorectal recurrence at the site the primary tumour, peritoneal and other extra-hepatic sites.Ardito et al.26analysed the liver specific DFS curves for the two groups.At 5-years there was no significant difference, confirmed by multivariate analysis. Theliver specific DFS curves, however, revealedthat patients in the PVE group had experienced recurrence of liver metastases much earlier than those without PVE. Pamecha et al 25reported that the median time to hepatic recurrence was shorter in the PVE group compared with those not undergoing PVE (12 vs. 24 months). Hoekstra et al.24 reported a significantly higher rate of recurrence in the PVE group compared with those in the no PVE group (42% vs 4%) at the first postoperative scan at 3 monthsalthough some argued that this findingmight be due to lesions present at the time of resection. These reports suggest that patients undergoing PVE may develop earlier recurrence.
There are some limitations to thismeta-analysis. It is based on nonrandomized studies as no RCT has been conducted which may be considered unethical. Consequently an inherent risk of selection bias is present, as patients requiring PVE before surgery usually have a higher tumour burden than patients not requiring PVE. As shown, we assessed this bias considering the comparability of the two groups of patients for many variables known to influence the risk of recurrence after resection 21; information about other variables (CEA, Disease free-interval time) were not available in most of the studies and we could not take them into account. The included studies reported exclusively survival data of patients who successfully underwent liver resection following PVE; thereforean intention-to-treat survival analysis, including patients who became unresectable after PVE, was precluded. The small number of available studies limits the statistical power of this meta-analysis. Nevertheless,as one can seeby looking at the confidence intervals in the forest plot, the statistical power of this analysis was greater than the one of any individual study published so far. As such, this meta-analysis adds to literature additional new information 38.
In conclusion, the current analyses and the literature review suggest that pre-operative PVE should not affect oncologic outcomes of patients undergoing liver resection for CRLM.
.
REFERENCES
1. Rees M, Tekkis PP, Welsh FKS, O’Rourke T, John TG. Evaluation of long-term survival after hepatic resection for metastatic colorectal cancer: a multifactorial model of 929 patients. Ann Surg. 2008;247(1):125-135. doi:10.1097/SLA.0b013e31815aa2c2.
2. Choti MA, Sitzmann J V, Tiburi MF, et al. Trends in long-term survival following liver resection for hepatic colorectal metastases. Ann Surg. 2002;235(6):759-766.
3. Capussotti L, Muratore A, Mulas MM, Massucco P, Aglietta M. Neoadjuvant chemotherapy and resection for initially irresectable colorectal liver metastases. Br J Surg. 2006;93(8):1001-1006. doi:10.1002/bjs.5386.
4. Jaeck D, Bachellier P, Guiguet M, et al. Long-term survival following resection of colorectal hepatic metastases. Association Française de Chirurgie. Br J Surg. 1997;84(7):977-980.
5. Abdalla EK, Adam R, Bilchik AJ, Jaeck D, Vauthey J-N, Mahvi D. Improving resectability of hepatic colorectal metastases: expert consensus statement. Ann Surg Oncol. 2006;13(10):1271-1280. doi:10.1245/s10434-006-9045-5.
6. Schindl MJ, Redhead DN, Fearon KCH, Garden OJ, Wigmore SJ. The value of residual liver volume as a predictor of hepatic dysfunction and infection after major liver resection. Gut. 2005;54(2):289-296. doi:10.1136/gut.2004.046524.
7. Khatri VP, Petrelli NJ, Belghiti J. Extending the frontiers of surgical therapy for hepatic colorectal metastases: is there a limit? J Clin Oncol. 2005;23(33):8490-8499. doi:10.1200/JCO.2004.00.6155.
8. Mueller L, Hillert C, Möller L, Krupski-Berdien G, Rogiers X, Broering DC. Major hepatectomy for colorectal metastases: is preoperative portal occlusion an oncological risk factor? Ann Surg Oncol. 2008;15(7):1908-1917. doi:10.1245/s10434-008-9925-y.
9. Kinoshita H, Sakai K, Hirohashi K, Igawa S, Yamasaki O, Kubo S. Preoperative portal vein embolization for hepatocellular carcinoma. World J Surg. 1986;10(5):803-808.
10. Makuuchi M, Thai BL, Takayasu K, et al. Preoperative portal embolization to increase safety of major hepatectomy for hilar bile duct carcinoma: a preliminary report. Surgery. 1990;107(5):521-527.
11. Azoulay D, Castaing D, Smail A, et al. Resection of nonresectable liver metastases from colorectal cancer after percutaneous portal vein embolization. Ann Surg. 2000;231(4):480-486.
12. Abulkhir A, Limongelli P, Healey AJ, et al. Preoperative portal vein embolization for major liver resection: a meta-analysis. Ann Surg. 2008;247(1):49-57. doi:10.1097/SLA.0b013e31815f6e5b.