Dealing with an Insufficient Future Liver Remnant: PVE and Two-Stage Hepatectomy

Abstract

Colorectal liver metastases (CLM) are not always resectable at the time of diagnosis. An insufficient future liver remnant is a factor excluding patients from curative intent resection. To deal with this issue, two-stage hepatectomy was introduced approximately 20 years ago. It is a sequential treatment strategy for bilateral CLM, which consists of preoperative chemotherapy, portal vein embolization, and planned first and second liver resections. This article reviews current evidences supporting use of two-stage hepatectomy.

1. Introduction

Approximately 15% of patients presenting with colorectal cancer have synchronous colorectal liver metastases (CLM), and approximately 30% have metachronous CLM.¹ Liver resection has been proven to have a survival benefit over chemotherapy alone and provides 5-year overall survival (OS) rates that range from 40% to 58%.^2–4 However, due to extent of disease, patients with CLM are not always candidates for curative intent resection at the time of diagnosis.⁵ Fortunately, preoperative chemotherapy downsizes CLM and can therefore increase the number of patients eligible for curative resection.⁶ Similarly, CLM patients can be excluded from curative intent resection due to an insufficient future liver remnant (FLR), as the low hepatic functional reserve of small FLR can lead to post-hepatectomy liver failure.^7,8 Portal vein embolization (PVE) was first reported in the 1980s to deal with an insufficient FLR.^9–11 Subsequently, two-stage hepatectomy for bilateral CLMs was reported in the early 2000s as the next technical advancement for improving resectability.^12,13 This article reviews the historical background, safety, and oncological outcomes of two-stage hepatectomy to overcome an insufficient FLR.

2. Portal vein embolization

During PVE, the portal venous system draining to the affected liver planned for resection is embolized. The purpose of PVE is to induce hypertrophy of the non-embolized liver in order to reduce the risk of postoperative hepatic insufficiency. The rationale behind PVE is based on an animal experiment during which ligation of the portal vein caused atrophy of the PV-occluded liver and led to hypertrophy of the non-PV-occluded liver.¹⁴ In 1975, Honjo et al. reported the first clinical use of portal vein ligation for unresectable primary and metastatic liver cancers.¹⁵ In the 1980s, Makuuchi et al.,⁹ Matsuoka et al.,¹⁰ and Kinoshita et al.¹¹ reported the use of percutaneous trans-hepatic PVE before liver resection. In 1994, Kawasaki et al. reported the use of PVE prior to liver resection for bilateral CLMs. Since that time, PVE has been adopted more commonly as a safe, image-guided procedure that increases the volume of FLR.^16–20

3. Minimal requirement of future liver remnant volume

Minimal requirements of FLR volume are summarized in Table 1. However, interpretation of these minimal requirements must take various factors into account: (1) the definition of normal liver and injured liver vary, (2) total liver volume used to calculate percent minimal requirement of FLR vary. Generally, the minimal requirement of FLR in patients with normal liver ranges from 20 to 30% using CT-based non-tumor liver volume or formula-based liver volume.^7,21–23 Studies reported that the FLR/standardized liver volume < 20–25% ^7,23 or the FLR/total liver volume < 25–26.6%^21,22 were associated with higher rate of major complication or hepatic dysfunction.

Table 1.

Minimal requirement of future liver remnant volume

	Region	Minimal requirement	Total liver volume
Normal liver
Vauthey et al. 7	USA	25%	Formula-based liver volume*
Shoup et al. 21	USA	25%	CT-based non-tumor liver volume
Shindle et al. 22	Europe	26.6%	CT-based non-tumor liver volume
Kishi et al. 23	USA	20%	Formula-based liver volume†
Injured liver (chronic hepatitis/cirrhosis and chemotherapy)
Kubota et al. 8†	Asia	40% or 50%‡	CT-based non-tumor liver volume
Shirabe et al. 24	Asia	250mL/m2§	NA
Azoulay et al.17	Europe	40%¶	CT-based non-tumor liver volume
Shindoh et al. 25	USA	30%∥	Formula-based liver volume†

Abbreviations: CT, computed tomography; NA, not applicable.

^*(Standardized liver volume) = 706.2×(body surface area) + 2.4 by Urata K, et al. ⁹³

^†(Standardized liver volume) = −794 + 1267.28×(body surface area) by Vauthey JN, et al.⁹⁴

^‡Based on indocyanine green retention rate at 15 minutes.

^§Categorized as injured liver based on cohort of patients with hepatocellular carcinoma.

^¶Fibrosis or cirrhosis

^∥Chemotherapy > 3 months

For patients with injured liver (chronic hepatitis or chemotherapy), safe limits of FLR after liver resection are reported to be 30–50% of FLR/CT-based non-tumor liver volume or formula-based liver volume.^8,17,24,25 The safety of liver resection is also influenced by the degree of liver injury because FLR volume itself is not correlated with hepatic functional reserve. Therefore, it is difficult to generalize the safety limit of FLR in patients with injured liver. In primarily hepatocellular carcinoma patients, Kubota et al. stratified the minimal requirement of FLR/CT-based non-tumor liver volume using the indocyanine green retention rate at 15 minutes (ICG-R15). Patients with ICG-R15 < 10% tolerated 40% of FLR/CT-based non-tumor liver volume after liver resection and patients with ICG-R15 ≥ 10% tolerated 50% of FLR/CT-based non-tumor liver volume after liver resection. ⁸ Other groups have also used FLR volume and ICG-R15 for predicting liver dysfunction and/or postoperative mortality.^26,27 Liver scintigraphy is another approach to estimate hepatic functional reserve.²⁸ Dinant et al. reported that assessment of 99mTc-mebrofenin uptake in the FLR predicted better than measurement of liver volume. ²⁸ For patients with injured liver, integration of FLR volume and functional assessment is ideal to evaluate “functional FLR” on an individual basis.

FLR volume has been used as a practical decision-making metric for PVE. Kinetic growth rate, defined as the degree of hypertrophy divided by the number of weeks elapsed after PVE, has also been reported to be a sensitive predictor of hepatic functional reserve.²⁹ A kinetic growth rate of at least 2.0% per week is protective of hepatic insufficiency²⁹ defined as a peak bilirubin level > 7.0 mg/dL postoperatively ³⁰ (Figure 1).

Figure 1. Rates of hepatic insufficiency based on kinetic growth rate.

Kinetic growth rate < 2.0% is predictive of hepatic insufficiency irrespective of standardized future liver remnant. KGR, kinetic growth rate; sFLR, standardized future liver remnant. (J Am Coll Surg. 2013 Feb;216(2):201–9, with permission.)

4. Two-stage hepatectomy for patients with bilateral colorectal liver metastases

4.1. Development of two-stage hepatectomy

Although liver resection is the current standard of care for CLM, most patients are unresectable at the time of diagnosis.⁵ In most cases, the contraindication to curative intent resection is due to small FLR. In 1999, Lygidakis et al. first used the nomenclature, “two-stage hepatectomy” for sequential treatment strategy consisting of portal vein ligation in a first step, transarterial immunochemotherapy, and hemihepatectomy in a second step for treating advanced liver metastases. ³¹ This approach did not include liver resection in the first step. In 2000, Adam et al. proposed a strategy of two-stage hepatectomy in patients with bilateral CLM.¹² The concept of this new strategy is that overall intention is curative, but a first-stage liver resection is not intended to remove all CLMs. A second-stage liver resection is performed after a while to allow hypertrophy of the FLR and to decrease the risk of postoperative liver failure. This approach of two-stage hepatectomy did not include use of PVE in all cases. In 2003, this strategy was further refined by Jaeck et al.^13,32 Namely, the first stage hepatectomy aims to remove CLMs located in the hemi-liver with less tumor burden (most typically the left liver). PVE is then performed 2 to 5 weeks after the first stage, followed by a second stage hepatectomy (mostly right hepatectomy or extended right hepatectomy) performed 5 to 8 weeks after PVE. This treatment pathway is widely adopted as “two-stage hepatectomy” for bilateral CLMs to achieve safe and curative liver resection.

4.2. Portal vein embolization or ligation?

Two-stage hepatectomy for bilateral CLMs is a sequential treatment which consists of preoperative chemotherapy, clearance of one hemi-liver, PVE, and hemi-hepatectomy. For primary gastrointestinal tumors with bilateral liver metastases, a similar concept using a planned two-step approach was proposed as “two-step surgery” by Kianmanesh and Belgihiti et al. in 2003.^33,34 The first step of this strategy consisted of resection of the primary tumor, clearance of metastases in left liver, and right portal vein ligation. The second step was to perform a right or extended right hepatectomy. Compared to portal vein ligation, PVE is minimally invasive, therefore it is preferred in patients who do not need a staged approach. However, for a two-stage hepatectomy, portal vein ligation may be a viable option because it can be performed in the operating room during the first stage liver resection. Thus, it remains controversial whether PVE or portal vein ligation is used for two-stage hepatectomy for bilateral CLMs.

Table 2 compares the current literature regarding changes in liver volume based on the use of PVE vs. portal vein ligation. Although background characteristics and procedures were heterogeneous in these studies, PVE was associated with better or similar volume increases compared to portal vein ligation.^34–38 Specifically, van Lienden et al. demonstrated that collateral flow and reperfusion of the ligated portal venous system developed in patients who underwent portal vein ligation, causing smaller increases in FLR.³⁸ Because the success of two-stage hepatectomy relies on liver hypertrophy of FLR between the first stage and the second stage, PVE is frequently selected based on the higher reported increases in liver volume.

Table 2.

Portal vein embolization vs. portal vein ligation

	PVE			PVL			Hypertrophy
	N	Volume increase	Morbidity	N	Volume increase	Morbidity
Broering et al. 35	17	188 mL	5.8%	17	123 mL	5.8%	PVE > PVL
Aussilhou et al. 34	18	35%	NA	17	38%	NA	PVE = PVL
Capusotti et al. 95	31	53.4%	3.2%	17	43.1%	0%	PVE = PVL
Robles et al. 37	18	40%	NA	20	30%	NA	PVE > PVL
Van Lienden et al. 38	14	41.6%	NA	7	8.1%	NA	PVE > PVL

Abbreviations: PVE, portal vein embolization; PVL, portal vein ligation; NA, not available.

4.3. Outcomes of two-stage hepatectomy

Short- and long-term outcomes of two-stage hepatectomy are shown in Table 3. The completion rates for planned first and second liver resections ranged from 60% to 90%.^{12,32,39–48} Postoperative morbidity rates were 20–60% and mortality rate was 0–15%. The rates of 3-year and 5-year overall survival in the reported series were 30–80% and 30–60%, respectively. The relatively low rates of morbidity and mortality and acceptable survival outcomes support the efficacy of two-stage hepatectomy for selected patients with initially unresectable bilateral CLMs. Among 91 patients who did not complete planned two-stage hepatectomy from 9 studies,^39–47 the main reasons were tumor progression after first stage hepatectomy (80%), insufficient FLR (11%), poor physical status (5%), portal vein thrombosis (2%), and death after first stage hepatectomy (2%) (Figure 2). Clearly, it is crucial to protect against tumor progression after first-stage hepatectomy and PVE. Use of preoperative chemotherapy or targeted therapy with bevacizumab during liver hypertrophy was reported to be effective and did not affect growth of FLR.^49,50 Recently, a fast-track two-stage hepatectomy pathway was reported, combining the first stage hepatectomy and PVE in a hybrid interventional radiology/operating suite. ⁵¹ This streamlined approach can reduce the time between first- and second-stage hepatectomy and allow early return to intended oncologic treatment.

Table 3.

Outcomes of two-stage hepatectomy for bilateral colorectal liver metastases

	Region	Number	Preoperative chemotherapy, %	PVE, %	PVL, %	Completion rate, %	Postoperative morbidity, %	Postoperative mortality, %	3-year OS, %	5-year OS, %
Adam et al. 12	Europe*	16	75	44	0	81	38	15	35	NA
Jaeck et al. 13	Europe†	33	91	100	0	76	56	0	54	NA
Tanaka et al. 39	Asia	24	64	73	0	100	23	0	33	NA
Wicherts et al. 40	Europe*	59	97	78	0	69	59	7	60	42
Homayounfar et al. 41	Europe	24	75	0	100	63	58	5	NA	NA
Tsai et al. 42	USA/Europe	45	71	7	71	78	26	6	84	NA
Brouquet et al. 43	USA‡	65	100	70	0	72	49	6	84	64
Tsim et al. 44	Europe	38	91	95	0	87	33	0	50§	NA
Narita et al. 45	Europe†	80	84	86	4	76	54	0	59	32
Muratore et al. 46	Europe	47	79	58	23	77	44	0	65	NA
Turini et al. 47	Europe	48	100	100	0	71	20	6	59	35
Passot et al. 48	USA‡	109	100	73	0	82	27¶	6	68∥	49∥

Abbreviations: PVE, portal vein embolization; PVL, portal vein ligation; NA, not available.

^*Reports from the same institution.

^†Reports from the same institution.

^‡Reports from the same institution.

^§Patients who underwent curative treatment.

^¶Complications graded as the Clavien-Dindo classification ≥ 3

^∥In 89 patients who underwent second-stage resection

Figure 2. Reasons for failure of two-stage hepatectomy.

Summative 91 patients who did not complete planned second-stage hepatectomy in 9 studies^39–48.

4.4. Technical refinements for Increasing FLR

Insufficient FLR before the second-stage hepatectomy is the second most common reason for non-completion of the two-stage hepatectomy approach. A number of technical tips have been shown to increase FLR. First, embolization using small spherical particles (tris-acryl microspheres) (Figure 3A) may contribute to better hypertrophy than embolization using large nonspherical particles (polyvinyl alcohol) (Figure 3B): increase of FLR, 69.0% ± 30.7% vs. 45.5% ± 40.9%, P = 0.001.⁵² It should be noted that there are many embolic agents for PVE, including n-butyl cyanoacrylate, ethiodized oil, fibrin glue, ethanol, and microparticles. Each embolic agent has unique technical demands, and no randomized controlled trials have been performed to compare them. Thus, the selection of embolic agents is based on cost, availability, and institutional expertise. Second, for patients undergoing extended right hepatectomy, embolization of segment 4 portal vein together with the right portal vein has been reported to be safe and useful for increasing the liver volume of the left lateral section (Figure 4).^32,53,54 Kishi et al. demonstrated a 54% increase in left lateral section liver volume after combined PVE of the right portal vein and segment 4 portal vein compared to 26% increase after PVE of the right portal alone.⁵⁴ The complication rate after PVE for the right portal vein and segment 4 portal vein was approximately 0–10%.^53,54 Finally, sequential ipsilateral hepatic vein embolization after PVE has been reported to result in a greater FLR increase compared to PVE alone.^55–59 According to the largest series including 12 patients, the mean proportion of FLR/total liver volume was 39.7 ± 0.6% 1–2 weeks after PVE, and 44.2 ± 1.1% 2 weeks after hepatic vein embolization. There was no report of embolization or dislodgement of coil in the heart or lung.⁵⁵ However, more studies are needed to assess the benefit of sequential ipsilateral hepatic vein embolization after PVE.

Figure 3. Particles used for portal vein embolization.

(A) Tris-acryl microspheres (B) Polyvinyl alcohol

Figure 4. Embolization to both the right portal vein and segment 4 portal vein.

(A) Coils in portal vein 4 are marked with white arrows and the embolized portal vein 4 is outlined in purple. Black arrows indicate anterior and posterior branches of the right portal vein. (B) The right liver and segment 4 with portal vein embolization shows atrophy, whereas segments 2 and 3, the future liver remnant, results in hypertrophy.

4.5. Factors associated with failure of two-stage hepatectomy

Previous studies have reported that patients who undergo an incomplete two-stage hepatectomy have worse survival than patients able to successfully complete two-stage hepatectomy. ^40,43,45,60 Thus, the selection of patients who are expected to complete two-stage hepatectomy is important to improve the outcomes for this procedure. Table 4 shows factors associated with failure of two-stage hepatectomy. Number of CLM, largest CLM diameter, longer duration of preoperative chemotherapy, and tumor progression were commonly reported factors. For patients who possess these risk factors, alternative treatment options need to be further investigated.

Table 4.

Factors associated with failure of two-stage hepatectomy

	Regions	N	Failure number	Factors
Narita et al. 45	Europe	80	29	1. Number of CLM in FLR ≥ 3
				2. Age > 70 yr
Giuliante et al. 60	Europe	130	28	1. Tumor progression during chemotherapy*
Imai et al. 96	Europe	125	44	1. CEA > 30 ng/mL
				2. Largest CLM diameter > 4 cm
				3. Preoperative chemotherapy > 12 cycles
				4. Tumor progression during chemotherapy
Passot et al. 48	USA	109	20	1. Number of CLM > 5
				2. Largest CLM diameter > 5 cm
				3. Preoperative chemotherapy > 6 cycles
				4. PVE
				5. Major complication after the first stage

Abbreviations: CLM, colorectal liver metastasis; FLR, future liver remnant; CEA, carcinoembryonic antigen; PVE, portal vein embolization.

^*Analyzed in 126 patients after excluding

4.6. Chemotherapy-associated liver injury

Because perioperative chemotherapy is part of the treatment pathway for two-stage hepatectomy, consideration of chemotherapy-associated liver injury is important. Types of liver injury are specific to regimen of chemotherapy. Oxaliplatin-based regimens are associated with sinusoidal obstruction syndrome. ⁶¹ Irinotecan-based regimens are associated with steatohepatitis. ^62–64 Chemotherapy-associated liver injury and postoperative outcomes are summarized in Table 5. After preoperative chemotherapy, approximately 10% of patients developed sinusoidal obstruction syndrome and approximately 20% of patients developed steatosis, although chemotherapy regimens are different and patients with initially unresectable CLM often receive several regimens.^63,64 Specifically, in patients who received oxaliplatin-based regimens, approximately 30–60 % of patients develop sinusoidal obstruction syndrome.^65–67 Use of bevacizumab with oxaliplatin-based regimens was reported to have a protective effect on the incidence and severity of sinusoidal obstruction syndrome.^68–70 Some studies have shown that patients with chemotherapy-associated liver injury had worse postoperative outcomes,^63,67,71,72 and other studies demonstrate similar postoperative morbidity and mortality between patients who underwent chemotherapy and patients who did not. ^64–66,73 With respect to the relationship between duration of chemotherapy and postoperative outcomes, evidence suggests prolonged preoperative chemotherapy is associated with a higher risk of postoperative morbidity. It should be noted that the definition of cut-off value for ‘prolonged’ preoperative chemotherapy is different by studies. Increased postoperative morbidity was reported in patients who received preoperative chemotherapy > 6, 9, or 12 cycles. ^71,72,74,75

Table 5.

Chemotherapy-associated liver injury and postoperative outcomes

	n	Preoperative chemotherapy, %	Irinotecan, %	Oxaliplatin, %	SOS, %*†	Steatosis, % *‡	Findings
Vauthey et al. 63	406	61	23	20	9	15	Patients with steatohepatitis had an increased 90-day mortality.
Aloia et al. 71	92	82	0	57	NA	13	Overall morbidity rates tended to be higher in patients undergoing preoperative chemotherapy.
Pawlik et al. 64	212	72	26	15	5	18	Similar postoperative morbidity and mortality§
Sahajpal et al.	96	55	17	0	NA	NA	Similar postoperative morbidity and mortality§
Nakano et al. 72	90	100	22	69	41	38	Patients with SOS has higher morbidity rate. Use of oxaliplatin > 6 cycles was associated with SOS.
Kandutsch et al. 65	63	79	0	100	43	22	Similar postoperative morbidity and mortality§
Komori et al. 66	27	56	0	100	33	0	Similar postoperative morbidity and mortality§
Takamoto et al. 97	136	40	47	95	16	11	2–4 weeks after chemotherapy cessation improves liver function.
Soubrane et al. 67	78	100	0	100	59	12	High grade SOS was associated with more postoperative morbidity.

Abbreviations: SOS, sinusoidal obstruction syndrome; VE, portal vein embolization; PVL, portal vein ligation; NA, not available.

^*Of patients who underwent chemotherapy

^†Grade 2/3 sinusoidal dilatation⁶¹

^‡Steatosis ≥ 30%.

^§Between patients with and without preoperative chemotherapy.

4.7. The role of gene mutation in two-stage hepatectomy

Recently, molecular alterations in CLM have been a focus for identifying patients who may benefit from liver resection.^76–79 Previous studies have shown that mutations in BRAF and KRAS are associated with a poor outcome after CLM resection.^76,80–84 Passot et al. demonstrated the importance of RAS as a biologic marker to select patients with bilateral CLM for liver resection.⁴⁸ In this series, the 5-year OS rate was 67 % in patients with RAS wild-type, compared to only 12% in patients with RAS mutation.

4.8. Two-stage hepatectomy using laparoscopic approach

Laparoscopic liver resection has been increasingly performed due to its advantages over open liver resection, in terms of better surgical and postoperative outcomes in selected patients.^85,86 Some case series reported the feasibility of two-stage hepatectomy using laparoscopic approach.^87–89 Each group employed different procedures such as a laparoscopic first-stage liver resection followed by an open second-stage hepatectomy, laparoscopic first- and second-stage hepatectomy, and etc. These studies emphasized that minimal adhesion during a second-stage hepatectomy is the benefit of laparoscopic approach. However, laparoscopic right hepatectomy or laparoscopic extended-right hepatectomy should be considered as having high surgical complexity.⁹⁰ The use of laparoscopic two-stage hepatectomy should be limited to centers with advanced experiences in hepatobiliary open and laparoscopic surgery.

4.9. Repeat surgery for recurrence after two-stage hepatectomy

Recurrence after two-stage hepatectomy is frequent because this approach is typically employed for patients with multiple and bilateral CLMs.^{25, 27, 34–42} Two studies have focused on the prognosis of recurrence after two-stage hepatectomy.^91,92 Imai et al reported that 38 patients (61%) underwent repeat surgery for recurrence after two-stage hepatectomy and patients who underwent repeat surgery had better overall survival than patients who did not (46% vs. 26%, P = 0.004). Lillemoe et al. reported that 31 patients (37%) underwent resection for recurrence.⁹² RAS mutation and first recurrence in multiple sites were associated with worse survival. Specifically, the 5-year OS rate in patients with RAS mutation who underwent repeat surgery for recurrence after two-stage hepatectomy was 38%, whereas the 5-year OS rate in patients with RAS wild-type who underwent repeat surgery was 86%.

5. Conclusion

Two-stage hepatectomy is an established treatment pathway which consists of perioperative chemotherapy, PVE, and two planned liver resections to deal with an insufficient FLR. This approach was originally designed to improve resectability of patients with bilateral CLMs and has been refined both in terms of the method for occluding the portal vein and advancements in chemotherapy over the past 20 years. Studies demonstrate that two-stage hepatectomy is safe for preserving sufficient FLR and is associated with better OS in patients who complete both planned liver resections than in patients who did not. Additionally, repeat surgery for patients who developed recurrence after two-stage hepatectomy has a survival benefit. Tumor progression between the first and second liver resections is the primary reason for failure of two-stage hepatectomy and remains a major limitation. This issue needs to be further investigated to increase the rate of completion of two-stage hepatectomy and to effectively use the two-stage hepatectomy approach.

Synopsis.

Colorectal liver metastases are not always resectable at the time of diagnosis. An insufficient future liver remnant is an issue to exclude patients from curative intent resection. Portal embolization and two-stage hepatectomy developed as safe and oncologically-effective methods to deal with insufficient future liver remnant.

ABBREVIATIONS

CLM

colorectal liver metastases

OS

overall survival

FLR

future liver remnant

PVE

portal vein embolization

OS

overall survival

FLR

future liver remnant