Outcomes after Vascular Resection During Curative-Intent Resection for Hilar Cholangiocarcinoma: A Multi-Institution Study from the US Extrahepatic Biliary Malignancy Consortium

Abstract

Introduction

Surgical resection is the cornerstone of curative-intent therapy for patients with hilar cholangiocarcinoma (HC). The role of vascular resection (VR) in the treatment of HC in western centres is not well defined.

Methods

Utilizing data from the U.S. Extrahepatic Biliary Malignancy Consortium, patients were grouped into those who underwent resection for HC based on VR status: no VR, portal vein resection (PVR), or hepatic artery resection (HAR). Perioperative and long-term survival outcomes were analyzed.

Results

Between 1998–2015, 201 patients underwent resection for HC, of which 31 (15%) underwent VR: 19 patients (9%) underwent PVR alone and 12 patients (6%) underwent HAR either with (n = 2) or without PVR (n = 10). Patients selected for VR tended to be younger with higher stage disease. Rates of postoperative complications and 30-day mortality were similar when stratified by vascular resection status. On multivariate analysis, receipt of PVR or HAR did not significantly affect OS or RFS.

Conclusion

In a modern, multi-institutional cohort of patients undergoing curative-intent resection for HC, VR appears to be a safe procedure in a highly selected subset, although long-term survival outcomes appear equivalent. VR should be considered only in select patients based on tumor and patient characteristics.

Introduction

In the United States, cholangiocarcinoma has a reported autopsy prevalence of 0.01 to 0.46% and an incidence of 1–2/100,000 population, although the incidence is much higher in Asia(Hemming et al. 2011; Khan et al. 2012; Brito et al. 2015). Surgical resection or transplant is the only potentially curative treatment of hilar cholangiocarcinoma(Rocha et al. 2010), however, the role of en bloc vascular resection remains controversial(Shaib and El-Serag 2004; Patel and Singh 2007). Therefore, the aim of the current study was to explore the safety and long-term outcomes of vascular resection during surgical treatment of hilar cholangiocarcinoma using data from a large cohort of patients from the U.S. Extrahepatic Biliary Malignancy Consortium.

Methods

The U.S. Extrahepatic Biliary Malignancy Consortium (USEBMC) is a collaboration of 10 high-volume, academic institutions, and includes Emory University, Johns Hopkins University, New York University, The Ohio State University, Stanford University, University of Louisville, University of Wisconsin, Vanderbilt University, Wake Forest University, and Washington University in St. Louis. All patients at participating institutions with HC who underwent resection with curative intent from 1998 to 2015 and had long-term follow-up data, including vital status and recurrence status, were included. Exclusion criteria included planned palliative or noncurative resections (including bile duct resection only, missing data on vascular resection status, or incomplete follow-up data.

Standard demographic and clinicopathologic data were collected including age, sex, body mass index (BMI), American Society of Anesthesiologists (ASA) score, comorbidities (including history of hypertension, diabetes, COPD, congestive heart failure, renal failure, chronic steroid use, and tobacco use), tumor-related signs and symptoms, type of surgery and tumor-specific characteristics. In particular, data were collected on presence of jaundice or ascites and preoperative biliary drainage. Data on treatment-related variables, such as type of surgery and receipt of lymphadenectomy and intraoperative estimated blood loss (EBL), were also collected. Pathology review was performed by experienced GI pathologists at each institution, and staging was assigned as per American Joint Committee on Cancer (AJCC) 7th edition guidelines(Edge and Compton 2010). Institutional Review Board approval was obtained at each institution prior to data collection.

Resection margin and nodal status were ascertained based on final pathologic assessment. Tumor-specific characteristics included median tumor size, location, Bismuth type, and presence of vascular or perineural invasion. Length of hospital stay (LOS), reoperation, 30-day perioperative complications, and mortality were obtained. Complications were categorized based on the Clavien-Dindo classification system(Clavien et al. 2009). Date of last follow-up and vital status were collected for all patients through December 31, 2015. Survival information was verified with the Social Security Death Index when necessary.

The primary endpoints were overall survival (OS) and recurrence-free survival (RFS). Categorical variables were compared using the Fisher exact test or χ² test as appropriate, and continuous variables were compared using ANOVA. OS was measured from the time of resection to death or last follow-up. RFS was measured from time of resection to recurrence, death or last follow-up. Survival probabilities were estimated using the method of Kaplan and Meier and compared using the log-rank test. Prognostic factors for survival were evaluated using multivariate Cox proportional hazards regression. A p value <0.05 was considered significant. All statistical analyses were performed using STATA version 13.1 (StataCorp LP, College Station, TX, USA).

Results

A total of 201 patients in the U.S. Extrahepatic Biliary Malignancy Consortium database who underwent potentially curative resection for HC and met inclusion criteria were identified. Seventy patients were excluded for undergoing palliative, non-curative intent procedures, including 51 who underwent bile resection alone, while 23 patients were excluded for missing vascular resection status and 11 were excluded for incomplete follow-up data. In assessing the cohort, 170 patients (85%) did not undergo vascular resection, while 19 patients (9%) underwent PVR alone, and 12 patients (6%) underwent HAR, including 2 patients who had both PVR and HAR. Clinicopathologic characteristics by VR status are presented in Table 1. Patients who underwent HAR were significantly more likely to have received neoadjuvant chemotherapy than those who did not undergo VR and those who underwent PVR (3/12 versus 6/170 and 1/19, respectively, p=0.002).

TABLE 1.

Clinicopathologic Variables Associated with Vascular Resection in 201 Patients After Curative Resection for Hilar Cholangiocarcinoma¹

	No Vascular Resection n=170	PV Resection n=19	HA Resection n=12	p
Female (n, %)	101 (60)	9	6	0.515
Age (x, sd)	66 (10)	62	52	p<0.001
Race				0.804
White	129 (78)	15	10
Black	12 (7)	2	1
Asian	13 (8)	0	0
BMI (x, sd)2	26 (6)	2	0
Functional Status3		26	26	0.108
Independent	152 (99)			0.367
Partially Dependent	2 (1)	17	11
ASA class		1	0
1	2 (1)			.928
2	40 (28)	0	0
3	94 (67)	7	3
4	5 (4)	10	5
Hypertension	90 (55)	1	0
Diabetes Mellitus		12	10	0.060
Managed with Oral Medications	14 (9)			0.549
Insulin-Dependent	8 (5)	1	2
Prior Cardiac Event	23 (14)	2	0
Congestive Heart Failure	2 (0)	5	0	0.136
Dyspnea	4 (3)	0	0	0.830
Tobacco Use	42 (26)	0	0	0.687
COPD	9 (6)	3	5	0.203
Acute Kidney Injury	2 (1)	2	1	0.639
Preoperative sepsis	11 (7)	0	0	0.830
Preoperative jaundice	133 (80)	1	0	0.648
Primary Sclerosing Cholangitis	4 (3)	15	11	0.592
Ascites	6 (4)	0	1	0.310
Preoperative stent placement		0	1	0.435
Endoscopic stent only	68 (41)			0.142
Percutaneous stent only	35 (21)	5	4
Both	36 (21)	7	5
Neoadjuvant chemotherapy	6 (4)	1	3	0.002

¹Fields with missing data may not add to 100 percent

²BMI indicates body mass index; ASA class, American Society of Anesthesiologists classification.

³Functional status: Independent, the patient does not require assistance from another person for activities of daily living. This includes a person who is able to function independent with prosthetics, equipment or devices. Partially dependent, the patient requires some assistance from another person for activities of daily living. This includes a person who utilizes prosthetics, equipment or devices but still requires some assistance from another person for activities of daily living.

Operative and pathologic are presented in Table 2A. Post-operative characteristics are presented in Table 2B. The overall 30-day mortality in this cohort was 8%, The overall incidence of experiencing any complication was 69%. There was no difference in the incidence of Clavien-Dindo Grade III or higher complications among the no VR, PVR, and HAR groups (109/179, 9/19, and 8/12, respectively, p = 0.524). Postoperative liver failure showed a trend towards higher rates in patients who received portal vein resection, although this did not reach statistical significance (7/170, 3/19 and 0/12, for no VR, PVR and HAR group respectively, p=0.119). Patients received adjuvant therapy at comparable rates amongst all three groups.

TABLE 2A.

Operative Characteristics Associated with Vascular Resection in 201 Patients After Curative Resection for Hilar Cholangiocarcinoma¹

	No Vascular Resection n=170	PV Resection n=19	HA Resection n=12	p
Operative characteristics
Diagnostic laparoscopy	30 (18)	3	2	0.974
Resection type				0.380
Radical cholecystectomy (segments IVb+V) & portal lymph node dissection	4 (2)	0	0
Right hepatectomy & Bile duct resection	27 (16)	2	1
Left hepatectomy & Bile duct resection	50 (30)	4	6
Extended right hepatectomy & Bile duct resection	35 (21)	4	0
Extended left hepatectomy + Bile duct resection	14 (8)	2	3
Right trisectorectomy + Bile duct resection	19 (11)	5	1
Left Trisectorectomy + Bile Duct resection	16 (10)	2	0
Pancreaticoduodenectomy	3 (0)	0	0
Caudate resection	69 (41)	13	8	0.022
MIS Technique2				0.968
Open	167 (98)	19	12
Laparoscopic	1 (1)	0	0
Laparoscopic-converted-to-open	2 (1)	0	0
Intraoperative Blood transfusion	67 (43)	9	5	0.787
EBL (mL) (x, sd)	1011 (1880)	1020 (635)	2100 (1890)	0.349
Extent of portal vein resection
Partial		14	2
Complete		5	0
No-touch technique		3
Type of portal vein reconstruction
Primary		15	1
Venous patch		1	0
Prosthetic patch		1	0
Prosthetic conduit		2	0
Extent of hepatic artery resection
Right hepatic artery resection			3
Left hepatic artery resection			4
Hepatic artery reconstruction
End to end anastao			0
Vein graft			3
R0 resection status				0.912
R0	119 (70)	14	8
R1	51 (30)	5	4
Intraoperative frozen section margin collection	148 (88)	18	10	0.579
Frozen section margin positive	48 (33)	6	2	0.709
Intraoperative drain placement	157 (92)	17	12	0.539

¹Fields with missing data may not add to 100 percent

²MIS indicates minimally invasive surgery; EBL, estimated blood loss; LVI, lymphovascular invasion; PNI, perineural invasion

TABLE 2B.

Pathologic and Post-Operative Characteristics Associated with Vascular Resection in 201 Patients After Curative Resection for Hilar Cholangiocarcinoma¹

	No Vascular Resection n=170	PV Resection n=19	HA Resection n=12	p
Pathologic Characterisctics
Pathologic grade				0.607
1	30 (19)	3	4
2	92 (58)	12	5
3	35 (11)	1	3
4	1 (1)	0	0
Tumor stage				0.006
T1	15 (12)	0	2
T2a	36 (28)	1	2
T2b	57 (44)	5	3
T3	20 (15)	7	4
T4	2 (2)	2	0
N2 nodes sampled	27 (16)	7	5	0.016
Positive lymph nodes	63 (41)	5	2	0.176
LVI	53 (39)	7	4	0.640
PNI	110 (77)	15	7	0.664
Postoperative Characteristics
30-day mortality	12 (7)	3	0	0.153
Length of stay (days) (x, sd)	15 (10)	20	11	p<0.001
Readmission	42 (36)	4	4	0.738
Adjuvant therapy	79 (53)	8	6	0.569
Any complication	114 (69)	13	6	0.623
Recurrence	55 (35)	9	6	0.278
Recurrence Location				0.203
Local	13 (8)	4	2
Distant	24 (14)	5	4
Both	15 (9)	0	0

¹Fields with missing data may not add to 100 percent

The median follow-up time in this cohort was 22 months, during which time 70 patients (41%) developed recurrence of disease. Median OS and RFS by potential prognostic factors are shown in Table 3A. Independent factors associated with OS and RFS are shown in Table 3B. Median OS was comparable among the no VR, PVR, and HAR groups (21, 24, and 33 months; p = 0.818) (Figure 1A). On univariate analysis, RFS was comparable between VR groups (16.2, 17.1 and 30.2 months for no VR, PVR and HAR respectively; p=0.199) (Figure 1B).

TABLE 3A.

Univariate Analysis of Variables Associated with OS and RFS After Curative Resection for Hilar Cholangiocarcinoma

	N	Overall Survival		Recurrence Free Survival
		Median Survival (mo)	p value	Median Survival (mo)	p value
No vascular resection	170	21	0.818	16	0.199
Portal vein resection	19	24		17
Hepatic artery resection	12	33		30
Lymph node status
Negative	110	25	0.002	1935	0.004
Positive	70	16		11
Age
<60 years	58	28	0.050	20	0.231
>60 years	142	18		16
ASA class ≥3	52	24	0.954	15	0.623
ASA class <3	148	20		17
T stage
T stage <3	211	22	0.002	19	<0.001
T stage ≥3	44	17		11
Grade
I	37	32	0.136	22	0.05
II	108	19		16
III	39	23		13
R0 resection status
R0	92	23	0.291	17	0.518
R1	39	19		15

ASA class, American Society of Anesthesiologists classification.

TABLE 3B.

Multivariate Cox Regression of Factors Associated with OS and RFS After Curative Resection for Hilar Cholangiocarcinoma

	Overall Survival			Recurrence Free Survival
Variable	HR	95% CI	p value	HR	95% CI	p value
No vascular resection	Reference			Reference
Portal vein resection	0.9	0.5–2.2	0.993	1.7	0.8–3.3	0.138
Hepatic artery resection	1.0	0.5–2.2	0.964	0.6	0.3–1.3	0.176
Age >60	1.3	0.8–2.2	0.273	1.1	0.7–1.7	0.777
ASA class ≥3	0.8	0.6–1.2	0.328	0.7	0.4–1.1	0.099
T stage ≥3	1.7	1.1–2.7	0.034	2.1	1.3–3.5	0.003
Grade
I	Reference			Reference
II	1.1	0.7–1.8	0.676	1.1	0.7–1.7	0.735
III	1.2	0.8–2.1	0.590	1.7	1.1–2.9	0.030
Margin status
R0	Reference			Reference
R1	1.2	0.7–1.9	0.468	1.0	0.7–1.5	0.974
Lymph node positive	2.1	1.4–3.4	0.002	2.0	1.4–3.0	<0.001

ASA class, American Society of Anesthesiologists classification.

Figure 1.

Figure 1A. Kaplan-Meier estimates of OS for 201 patients who received no vascular resection (noVR), portal vein resection (PVR), or hepatic artery resection (HAR) during curative resection for HC.

Figure 1B. Kaplan-Meier estimates of RFS for 201 patients who received no vascular resection (noVR), portal vein resection (PVR), or hepatic artery resection (HAR) during curative resection for HC.

Discussion

In a large, modern, multi-institutional North American cohort of patients undergoing resection for hilar cholangiocarcinoma, vascular resection, including portal vein resection and/or hepatic artery resection, was performed in a select minority of younger patients with higher T stage disease. Both procedures were relatively safe in the perioperative period, and were associated with equivalent long-term survival outcomes when compared with patients who received no vascular resection, even when controlling for relevant clinical factors such as T stage and lymph node status. Vascular invasion remains a major obstacle in the treatment of patients with hilar cholangiocarcinoma, and traditionally invasion of vascular structure by tumor meant unresectable disease. However, with improvements in surgical technique and perioperative management, vascular resection and reconstruction in select patients who can tolerate the procedure offers an option for achieving R0 resection status and equivalent long-term survival(Matsuyama et al. 2016).

A major concern around vascular resection for hilar cholangiocarcinoma is safety and feasibility of the procedure. In this cohort, patients who underwent vascular resection had combined characteristics of more advanced disease and younger age at operation, likely reflecting surgeon selection of patients most likely to both tolerate and benefit from the procedure. Patients who received vascular resection also tended to have lower ASA class, although this did not reach statistical significance. The overall complication rate was comparable to previously published series in hilar cholangiocarcinoma (Nuzzo et al. 2012; Loehrer et al. 2013; Shubert et al. 2014), and neither complication rates nor readmission rates differed significantly by VR group. Patients with PVR did show a higher average length of stay, and had a trend towards higher rates of postoperative liver failure. The overall mortality rate in this cohort was 9%, which is similar to the rate recently reported for mortality after resection for HC in the ACS NSQIP database (Loehrer et al. 2013), and differences in mortality rates between VR groups did not attain statistical significance. However, it should be noted that no perioperative deaths were seen in the HAR group, while 3 patients died in the PVR group, 2 of which experienced postoperative liver failure. While some prior series have argued that PVR is a safer operation while HAR should be regarded with caution (Miyazaki et al. 2007), the present data show that PVR yielded slightly worse perioperative outcomes, although the number of patients analyzed may be too small to detect true differences.

Previous reports have suggested that PVR may provide some survival benefit to patients undergoing curative resection for HC, while the impact of HAR on long-term survival remains largely unclear(Ebata et al. 2003; Lee et al. 2006; Miyazaki et al. 2007; Hemming et al. 2011; de Jong et al. 2012). Not surprisingly, patients who received VR with PVR or HAR had higher T stage tumors representing more advanced disease. Despite this, the rate of R0 resection amongst all three groups is uniform. This suggests that local control is achieved with vascular resection to at least the same extent as in patients who do not undergo vascular resection, a result similar to what has been previously reported in Asian cohorts (Matsuyama et al. 2016). This fact may be a strong driving factor behind equivalency in survival outcomes between all three groups, despite more advanced T stage disease in PVR and HAR groups.

This analysis suggests, at the least, that PVR in the treatment of HC is a safe option in an effort to achieve R0 margins during resection for HC, both with respect to perioperative morbidity and mortality, and long-term survival. However, it remains to be seen if PVR confers survival benefit to patients with locally invasive HC. Many previous studies assessing the role of vascular resection in HC report non-significant differences survival outcomes, often due to the relatively low number of patients available in the studies. In a 2007 retrospective, single-institution study, Miyazaki, et al. evaluated 228 patients with HC, of which 161 underwent curative-intent resection(Miyazaki et al. 2007). While a trend was seen towards decreasing survival with VR, there were no statistically significant differences between groups. Several subsequent single- or two-institution studies in Japan, China, and the USA have failed to discern any statistically significant differences in survival in HC patients receiving no VR, PVR, or HAR(Igami et al. 2010; Hemming et al. 2011; Wang et al. 2015; Matsuyama et al. 2016) (Table 4 for more detail). Despite the multi-institutional nature of the current dataset, hilar cholangiocarcinoma remains a rare disease entity, especially in the western hemisphere. It is interesting to note that despite the large centers included in this dataset, there were relatively small volumes of HC patients relative to Asian series and the rate of vascular resection was in general lower. Larger sample sizes may be necessary to fully understand if vascular resection confers a survival benefit in this patient population.

TABLE 4.

Studies evaluating vascular resection in curative-intent resection for hilar cholangiocarcinoma.

Investigator/ Year	Country	Type Study	Years Data	# HC/ # Resected	5-year %OS no VR	5-year %OS PVR	5-Year %OS HAR	5-Year %RFS no VR	5-Year %RFS PVR	5-Year %RFS HAR	Conclusions: Improved Survival with Vasc Resection?
Miyazaki 2007	Japan	Retrospective, single center	1981–2004	228/161	30	16	0	NA	NA	NA	No
Igami 2010	Japan	Retrospective, single center	2001–2008	428/298	51	23	33	NA	NA	NA	No
Hemming 2011	USA	Retrospective, single center	1999–2010	NA/95	45	43	NA	NA	NA	NA	No
Wang 2015	China	Prospective, single center	2005–2012	277/154	36	25	25	NA	NA	NA	No
Matsuyama 2016	Japan	Retrospective, single center	1992–2014	249/172	NA	NA	NA	46	51	22	No

This is the first multicenter study investigating the impact of vascular resection on outcomes for patients with HC in a North American population. However, as a retrospective analysis, it has several inherent limitations. Notably, the likely presence of confounding variables that were either not noted or not known at the time of data collection, despite the large number of variables collected, limits conclusions of causality. Furthermore, there is inherent preoperative selection bias based on patients and disease factors in comparing those that underwent VR to those who did not. Finally, as a retrospective database, some amount of data in each field collected was not able to be confirmed. However, it is expected that such missing data was random in nature and therefore would not be expected to affect the main findings of the current study. Additionally, despite the multi-institutional nature of this database and the large number of total patients reviewed, only a small number underwent vascular resection, making it difficult to make firm statistically significant conclusions. Despite these limitations, this study represents the largest, multi-institutional North American analysis of vascular resection in hilar cholangiocarcinoma, and as such provides valuable insights into variations in practice and outcomes after vascular resection.

In conclusion, in this multi-institutional analysis of 201 patients that underwent surgical resection for HC, vascular resection was performed in a small number of patients and was a safe procedure that yielded equivalent survival outcomes. This suggests that for select patients, in the hands of experienced surgeons at high-volume centers, vascular resection for HC is a safe tool for achieving R0 margins in an effort to improve survival. Surgeons should always consider individual patient and tumor characteristics when selecting patients for appropriate treatment for HC.