Hepatic Arterial Supply in 1297 CT-Angiographies

Key words

hepatic arteries - CT-angiography - anatomy

Introduction

The hepatic arterial supply and its variability play an important role, particularly in liver transplantation and living donor liver transplantation. In the Eurotransplant area, 18 151 postmortal liver transplantations were performed between 2002 and 2013 [1]. Of course, precise knowledge of variants of the hepatic arterial supply is essential. Vascular complications can, for example, lead to insufficient blood flow to the liver with subsequent transplant loss [2] [3] [4]. Moreover, in hepatic tumor embolizations [5] [6] as well as in extrahepatic abdominal surgeries involving the stomach, pancreas or gall bladder [7] [8] [9], variants of the hepatic vascular supply can present anatomical problems.

Currently, CT-angiography is the diagnostic standard for pre-operative and pre-interventional assessment of the hepatic arterial anatomy. Commonly Michels’s classification is used for a precise description [10]. N. A. Michels performed 200 autopsies for his study and categorized variants of the hepatic arterial supply into 10 different types. His classification was modified by Hiatt in 1994 [11], who analyzed 1000 medical records and classified variants into 6 categories. Other important publications came from Saba et al with 1600 patients [12], Covey et al. with 600 patients [13] and Koops et al. with 604 patients [14], who all largely followed Michels’s classification. Further studies in recent years have come from Sureka et al. with 600 patients [15], Song et al. with 5002 patients [16], Abdullah et al. with 932 patients [17], Gruttadauria et al. with 701 patients [18] and Soin et al. with 527 patients [2]. All of these studies, however, developed their own classifications without derivation from Michels’s classification. Therefore, the comparability is limited.

The purpose of this study was the analysis, evaluation and classification of hepatic arterial supply variants and the determination of their frequency distribution in CT-angiographies.

Materials and Methods

At a university hospital, all CT-angiographies covering the abdomen were analyzed retrospectively between January 1, 2010 and August 30, 2012. The hospital comprises a liver transplant center and radiologists are trained on the evaluation of liver anatomy. Therefore, deviations of the hepatic arterial anatomy are usually reported regardless of the indication for the examination. S.N and S. K. (> 3 years and > 1 year of experience, respectively) performed most of the included examinations themselves and wrote the later considered reports, paying special attention to the liver anatomy. The reports were verified by U. T. (> 15 years of experience), again with a special focus on the hepatic arterial supply.

C. L. conducted a blinded re-evaluation of all CT examinations and compared his results to the radiological reports. Additionally, C. L. and S. N. together re-evaluated all variants other than type I. Any discrepancies were resolved by consensus in consultation with U. T. The assessment of the hepatic arterial supply was thus redundant.

Out of 1568 CT-angiographies, 1297 could be included in this study ([Table 1]). Only one CT-angiography per patient was used. In patients with multiple CT-angiographies, only pre-operative and pre-interventional examinations were considered and the scan with the best quality was chosen. 271 patients were excluded, mainly because of iatrogenic or disease-related changes in the vascular supply, i. e. status post total or split liver transplantation, hemihepatectomy, Whipple or Billroth I & II procedure, vascular interventions (e. g. mesenteric or celiac artery bypass, transcatheter arterial chemoembolization (TACE)) or aortic dissection (Stanford A/B). Patients with a transjugular intrahepatic portosystemic shunt (TIPS) were excluded as a precaution because of potential artifacts that could impair the detection especially of smaller vessels. 9 CT-angiographies were excluded because of poor image quality.

Table 1 Patient demographics.

sex	number	percentage	∅-age in years (min. – max.)
male	797	61.45	61.64 (20 – 102)
female	500	38.55	61.63 (20 – 92)
total	1,297	100	61.74

Written informed consent to the CT scan and contrast application also included optional consent to anonymous use of imaging data and was given for each of the included examinations. The local ethical committee approved the study.

CT-Angiography

All images were acquired with a 64-slice CT scanner (2005 GE LightSpeed VCT 64 Slice CT, GE Healthcare, Milwaukee, WI, USA). A standardized scanning protocol was used: arterial phases were timed using bolus tracking with a monitoring delay of 7 s for scans starting in the abdomen, 3 s for scans starting in the chest. A monitor interscan delay of 3 s and a diagnostic delay of 3 s were used. The scan triggered when a threshold of Δ100-HU in the thoracic or abdominal aorta was reached. The dose of the contrast agent (Ultravist 300; Bayer Schering Pharma, Berlin, Germany) was adjusted according to the type of CT-angiography (usually 100 ml for abdominal scans, 120 ml for combined thoracoabdominal scans); flow speed was typically 4 ml/s. Contrast injection was followed by an NaCl flush (40 ml). Images were acquired with a primary slice thickness of 0.625 mm, a table speed of 39.37 mm/s and a pitch of 0.984:1. Multiplanar reconstructions (axial, coronal and sagittal) with a standardized slice thickness of 3 mm were sent to the picture archiving system and used for the evaluation. Image review was done using our standard workstation (PACS: Cerner ProVision, Cerner Corporation, Kansas City, MO, USA; RIS: Lorenzo RadCentre, iSoft Health GmbH, Mannheim, Germany) with high-resolution displays (Model MDCC 2121, Barco N. V., Kortrijk, Belgium). Patients never received oral contrast prior to the scan.

Classification

The classification of the hepatic arterial supply was based on Michels’s classification [10] ([Fig. 1]). Not yet classified variants were assigned to the most suitable type and recorded in an extended classification. We also used a) and b) subtypes for these variants as Michels did with type VIII. The arterial supply of the left hepatic lobe from the left hepatic artery (LHA) and of the right hepatic lobe from the right hepatic artery (RHA) was considered the textbook type (Michels type I). Both arteries derive from the proper hepatic artery, which is the continuation of the common hepatic artery (CHA) after the gastroduodenal artery (GDA) branches off ([Fig. 2]). Variants of the arterial supply result from arteries that originate from the superior mesenteric artery (SMA), from the left gastric artery (LGA) or directly from the aorta ([Fig. 3]). They can either exist in addition to the vessels of the textbook type or replace them completely. If a double arterial supply of a hepatic lobe was identified, the vessel was classified as accessory, whereas if the supply occurred exclusively through one vessel, it was marked as aberrant (replaced). This declaration, however, can be misleading, since all arteries are end arteries with a selective distribution to a definite area of the liver [10].

Fig. 1 Schematic illustrations of Michels’s classification of hepatic arterial supply variants with types I-X. Legend: LHA = left hepatic artery, RHA = right hepatic artery, GDA = gastroduodenal artery, LGA = left gastric artery, SA = splenic artery, aLHA = accessory left hepatic artery, rLHA = replaced left hepatic artery, aRHA = accessory right hepatic artery, rRHA = replaced right hepatic artery, MHA = middle hepatic artery. Note: only type VIIIa shown, type VIIIb corresponds to the inverted case with an aberrant LHA from the LGA with an accessory RHA from the SMA.

Fig. 2 3D-reconstruction showing Michels type I (textbook type). Legend: LHA = left hepatic artery, RHA = right hepartic artery, CHA = common hepatic artery, CT = celiac trunk, LGA = left gastric artery, SA = splenic artery, SMA = superior mesenteric artery. Note: accessory renal arteries bilaterally

Fig. 3 Series of axial reconstruction slices showing Michels type 2. Legend: Top row shows tracking of the right hepatic artery, bottom row shows tracking of the replaced left hepatic artery. Open arrow = celiac trunk, black arrowhead = common hepatic artery continuing as proper hepatic artery when the gastroduodenal artery branches off and finally as the right hepatic artery, thick arrow = splenic artery, black and white arrow head = left gastric artery continuing as replaced left hepatic artery

Results

According to Michels’s classification, 937 cases (72.2 %) showed the textbook type (type I). In 360 patients (27.8 %), deviations from the normal supply were identified, including 26 (2.0 %) patients with not yet classified variants. The overall second most frequent variant was an accessory LHA from the LGA (type V, n = 114, 8.8 %), followed by an aberrant RHA from the SMA (type III, n = 83, 6.4 %) and an aberrant LHA from the LGA (type II, n = 55, 4.2 %). A complete overview of all variants is provided in [Table 2].

Table 2 Distribution of the found variants

Michels type		n	%
I		937	72.2
II		55	4.2
III		83	6.4
IV		20	1.5
V		114	8.8
VI		20	1.5
VII		6	0.5
VIII		11	0.8
IX		25	2.0
X		0	0
not yet classified types		n	%
Ia:	SMA+CT = 1 truncus	4	0.3
IIa:	CHA from aorta	1	0.1
IIb:	LGA from aorta	1	0.1
IIIa:	CHA from aorta	2	0.2
IVa:	no MHA	10	0.8
Va:	CHA from aorta	4	0.3
VIIa:	aRHA from aorta	1	0.1
XI:	RHA+LHA separated from CT	3	0.2

SMA = superior mesenteric artery, CT = celiac trunk, CHA = common hepatic artery, LGA = left gastric artery, MHA = middle hepatic artery, aRHA = accessory right hepatic artery, RHA = right hepatic artery, LHA = left hepatic artery

The complete outflow of the CHA from the LGA (type X) could not be found in any of the examined patients. Out of the 26 patients with non-classified supply types, the most frequent variant was an aberrant RHA originating from the SMA, with an aberrant LHA originating from the LGA as in type IV, but without the middle hepatic artery (MHA) (new as type IVa) (n = 10, 0.8 %). In four cases (0.3 %) we found a modification showing the SMA and the celiac trunk (CT) forming a common trunk, a so-called celiacomesenteric trunk, with a subsequent normal blood supply of the liver from the CHA (new as type Ia). In another four cases (0.3 %) we found a modification of type V with the CHA originating directly from the aorta (new as type Va). Three patients (0.2 %) showed a separate origin of the RHA and the LHA from the CT (new as type XI). Two patients (0.2 %) had an aberrant RHA from the SMA as in type III with the CHA originating directly from the aorta (new as type IIIa). Three variants could be detected only once (0.1 %): type VII with an accessory RHA arising directly from the aorta (new as type VIIa), type II with the CHA arising directly from the aorta (new as type IIa) and type II with the LGA arising directly from the aorta (new as type IIb) ([Fig. 4]).

Fig. 4 Schematic illustrations of newly discovered types of hepatic arterial supply. Legend: LHA = left hepatic artery, RHA = right hepatic artery, GDA = gastroduodenal artery, LGA = left gastric artery, SA = splenic artery, aLHA = accessory left hepatic artery, rLHA = replaced left hepatic artery, aRHA = accessory right hepatic artery, rRHA = replaced right hepatic artery

Discussion

In the past few years, CT-angiography has increasingly been used to get an overview of the vascular situation prior to surgeries and angiographic interventions. Today it is the diagnostic standard for pre-operative and pre-interventional evaluation of the hepatic vascular supply. Advantages are, for example, the possibilities of 2 D and 3 D image reconstructions. Varying image qualities due to artifacts as well as patient and scanner characteristics can be a disadvantage for a standardized evaluation. Furthermore, the slice thickness of reconstructions and anatomical cross-sections of the vessels play a role during the assessment. For example, several studies have shown the benefit of conventional angiography in the detection of smaller intrahepatic vessels, aberrant arteries or the LGA [19] [20] [21]. However, these studies compared conventional angiographies to CT-angiographies with rather large slice reconstructions between 8 and 10 mm. When compared to CT-angiographies with thinner slice thicknesses, a good correlation of both imaging modalities was seen in newer studies and CT-angiography was found to be sufficient to evaluate the hepatic arterial anatomy [22] [23] [24] [25] [26] [27].

The importance of precise knowledge of the hepatic arterial anatomy is evident when vascular complications are considered. In liver transplantations these typically include stenosis and hepatic artery occlusions, steal-syndromes and aneurysms [28]. Currently, the incidence of vascular complications is documented with a rate of 0.7 % to 12.9 % [4] [28] [29]. For the TACE procedure Maeda et al. reported that the incidence for arterial injuries in this connection is up to 16 % for each artery and up to 48 % for each patient [30]. Moreover, in extrahepatic surgeries or interventions involving the stomach, esophagus or pancreas [8] [31] [32], vascular variants can lead to unexpected bleeding or an impairment of the hepatic arterial supply.

One explanation for the unequal gender distribution in the study population could be that many diseases of civilization preferentially affect men (e. g. cardiovascular, hepatobiliary or malignant diseases) [33] [34] [35].

In our study the textbook type (type I) was found in 72.2 % of cases, which is in line with the results of other studies using CT-angiography or conventional angiography for the evaluation [2] [14]. In the literature, the data vary across autoptic and radiological studies from 55 % to 79 % ([Table 3]). Possible reasons for these fluctuations could be the inclusion and exclusion criteria, the case number or the evaluation method (autopsy/CT-angiography/conventional angiography; CT slice thickness).

The middle hepatic artery (MHA) described by Michels was not further considered, because it originates from the LHA or RHA except for type IV [10]. Wang et al. reported that the MHA exists in 71 % of the patients, regardless of the hepatic arterial supply and derives directly or indirectly from the normal or a replaced CHA [36].

Table 3 Comparison with results of other studies.

variants	own results n = 1297 CT-angio	Michels [9] n = 200 autopsy	Koops et al. [13] n = 604 angio	Covey et al. [12] n = 600 angio	Saba et al. [11] N = 1629 CT-angio
type I	72.2 %	55.0 %	79.1 %	61.3 %	61.4
type II	4.2 %	10.0	2.5	3.8	7.5
type III	6.4 %	11.0	8.6	8.7	10.6
type IV	1.5 %	1.0	1.0	0.5	1.3
type V	8.8 %	8.0	0.5	10.7	6.7
type VI	1.5 %	7.0	3.3	1.5	6.9
type VII	0.5 %	1.0	0.2	1.0	0.7
type VIII	0.8 %	2.0	0.2	3.0	1.9
type IX	2.0 %	4.5	2.8	2.0	1.6
type X	0	0.5	0	0	0.3
other	2.0 %	0	1.8	7.5	1.1

Our results were close to those documented for conventional angiographies, the actual gold standard for vascular diagnostics, and furthermore widely match the results concerning the distribution of accessory and replaced vessels. Depending on the study, variations other than the textbook type were between 12 % and 49 % [7] [11] [31] [37] [38]. In our study the proportion was 27 %. The two most frequent variants were an accessory LHA from LGA (type V) with 8.8 % and a replaced RHA from the SMA (type III) with 6.4 %. In Michels’s study, type III ranked second (11 %), type II ranked third (10 %) and type V (8 %) came in fourth. It was conspicuous that the percentage of replaced vessels found by Michel was higher (type II and III). Two divergent pictures emerged concerning the distribution of the accessory vessels (type V and VI): while our results for type V corresponded approximately to those of Michels (8.8 % vs. 8.0 %), the results for type VI differed clearly (1.5 % vs. 7.0 %) [10]. The evaluation of other studies [39] [40] also revealed that the number of accessory vessels was less than that found by Michels. According to Koops et al., accessory arteries might be underrepresented in studies with CT-angiographies because of their partly very small size, or because the differentiation between replaced and accessory vessels might be limited [14]. The complete origin of the CHA from the LGA (type X) could not be found in any patient, which corresponds to the studies of Koops et al. and Covey et al. [13] [14].

In 26 patients not yet classified variants were detected. This result did not meet the expectations, since their incidence in other studies was higher. These unclassified variants mostly affected the origin of the vessels from the celiac trunk. In order to take them into account, they were categorized into subgroups of the existing classification ([Table 4]).

Tab. 4 Modified classification including so far unclassified variants.

type I: LHA + RHA from CHA from CT

type Ia: LHA + RHA from CHA, CT + SMA from 1 common trunk

type II: rLHA from LGA + RHA from CHA

type IIa: rLHA from LGA + RHA from CHA, CHA from aorta

type IIb: rLHA from LGA from aorta, RHA from CHA

type III: LHA from CHA + rRHA from SMA

type IIIa: LHA from CHA from aorta + rRHA from SMA

type IV: rLHA from LGA + rRHA from SMA + MHA from GDA

type IVa: rLHA from LGA + rRHA from SMA without MHA

type V: LHA+RHA from CHA + aLHA from LGA

type Va: LHA + RHA from CHA from aorta + aLHA from LGA

type VI: LHA+RHA from CHA+ aRHA from SMA

type VII: LHA+RHA from CHA + aLHA from LGA + aRHA from SMA

type VIIa: LHA + RHA from CHA + aLHA from LGA + aRHA from aorta

type VIIIa: LHA from CHA + aLHA from LGA + rRHA from SMA type VIIIb: rLHA from LGA + RHA from CHA + aRHA from SMA

type IX: LHA + RHA from CHA from SMA

type X: LHA + RHA from CHA from LGA

type XI: LHA + RHA divided from CT

LHA = left hepatic artery, RHA = right hepatic artery, CHA = common hepatic artery, CT = celiac trunk, SMA = superior mesenteric artery, rLHA = replaced left hepatic artery, rRHA = replaced right hepatic artery, GDA = gastroduodenal artery, LGA = left gastric artery, aLHA = accessory left hepatic artery, aRHA = accessory right hepatic artery, MHA = middle hepatic artery. Note: type VIII as already subdivided by Michels into a/b subgroups.

• References

• 1 Eurotransplant International Foundation. Yearly statistics, Year 2002–2013. Accessed 18 August, 2014. Available from: http://eurotransplant.org/cms/index.php?page=yearlystats
• 2 Soin AS, Friend PJ, Rasmussen A et al. Donor arterial variations in liver transplantation: management and outcome of 527 consecutive grafts. Br J Surg 1996; 83: 637-641
• 3 Merion RM, Burtch GD, Ham JM et al. The hepatic artery in liver transplantation. Transplantation 1989; 48: 438-443
• 4 Duffy JP, Hong JC, Farmer DG et al. Vascular complications of orthotopic liver transplantation: experience in more than 4,200 patients. J Am Coll Surg 2009; 208: 896-903 ; discussion 903–895
• 5 Clark TW. Complications of hepatic chemoembolization. Semin Intervent Radiol 2006; 23: 119-125
• 6 Sueyoshi E, Hayashida T, Sakamoto I et al. Vascular complications of hepatic artery after transcatheter arterial chemoembolization in patients with hepatocellular carcinoma. Am J Roentgenol 2010; 195: 245-251
• 7 Suzuki T, Nakayasu A, Kawabe K et al. Surgical significance of anatomic variations of the hepatic artery. Am J Surg 1971; 122: 505-512
• 8 Klingler PJ, Seelig MH, Floch NR et al. Aberrant left hepatic artery in laparoscopic antireflux procedures. Surg Endosc 2004; 18: 807-811
• 9 Gadzijev EM. Surgical anatomy of hepatoduodenal ligament and hepatic hilus. J Hepatobiliary Pancreat Surg 2002; 9: 531-533
• 10 Michels NA. Blood supply and anatomy of the upper abdominal organs with a descriptive atlas. Philadelphia, Pa: Lippincott; 1955
• 11 Hiatt JR, Gabbay J, Busuttil RW. Surgical anatomy of the hepatic arteries in 1000 cases. Ann Surg 1994; 220: 50-52
• 12 Saba L, Mallarini G. Anatomic variations of arterial liver vascularization: an analysis by using MDCTA. Surg Radiol Anat 2011; 33: 559-568
• 13 Covey AM, Brody LA, Maluccio MA et al. Variant hepatic arterial anatomy revisited: digital subtraction angiography performed in 600 patients. Radiology 2002; 224: 542-547
• 14 Koops A, Wojciechowski B, Broering DC et al. Anatomic variations of the hepatic arteries in 604 selective celiac and superior mesenteric angiographies. Surg Radiol Anat 2004; 26: 239-244
• 15 Sureka B, Mittal MK, Mittal A et al. Variations of celiac axis, common hepatic artery and its branches in 600 patients. Indian J Radiol Imaging 2013; 23: 223-233
• 16 Song SY, Chung JW, Yin YH et al. Celiac axis and common hepatic artery variations in 5002 patients: systematic analysis with spiral CT and DSA. Radiology 2010; 255: 278-288
• 17 Abdullah SS, Mabrut JY, Garbit V et al. Anatomical variations of the hepatic artery: study of 932 cases in liver transplantation. Surg Radiol Anat 2006; 28: 468-473
• 18 Gruttadauria S, Foglieni CS, Doria C et al. The hepatic artery in liver transplantation and surgery: vascular anomalies in 701 cases. Clin Transplant 2001; 15: 359-363
• 19 Aspestrand F, Kolmannskog F. CT and angiography in chronic liver disease. Acta Radiol 1992; 33: 251-254
• 20 Chambers TP, Fishman EK, Bluemke DA et al. Identification of the aberrant hepatic artery with axial spiral CT. J Vasc Interv Radiol 1995; 6: 959-964
• 21 Berchtenbreiter KM. Vergleich der Identifikation hepatischer, perihepatischer sowie abdomineller Gefäße: Computertomographie versus Digitale Subtraktionsangiographie. Ludwig-Maximilians-Universität zu München; 2002
• 22 Freund M, Wesner F, Reibe F et al. Spiral CT angiography for preoperative planning in patients with epigastric tumors: comparison with arteriography. J Comput Assist Tomogr 1996; 20: 786-791
• 23 Coskun M, Kayahan EM, Ozbek O et al. Imaging of hepatic arterial anatomy for depicting vascular variations in living related liver transplant donor candidates with multidetector computed tomography: comparison with conventional angiography. Transplant Proc 2005; 37: 1070-1073
• 24 Kamel IR, Raptopoulos V, Pomfret EA et al. Living adult right lobe liver transplantation: imaging before surgery with multidetector multiphase CT. Am J Roentgenol 2000; 175: 1141-1143
• 25 Guiney MJ, Kruskal JB, Sosna J et al. Multi-detector row CT of relevant vascular anatomy of the surgical plane in split-liver transplantation. Radiology 2003; 229: 401-407
• 26 Takahashi S, Murakami T, Takamura M et al. Multi-detector row helical CT angiography of hepatic vessels: depiction with dual-arterial phase acquisition during single breath hold. Radiology 2002; 222: 81-88
• 27 Saylisoy S, Atasoy C, Ersoz S et al. Multislice CT angiography in the evaluation of hepatic vascular anatomy in potential right lobe donors. Diagn Interv Radiol 2005; 11: 51-59
• 28 Settmacher U, Stange B, Haase R et al. Arterial complications after liver transplantation. Transpl Int 2000; 13: 372-378
• 29 Harms J, Chavan A, Ringe B et al. Vascular complications in adult patients after orthotopic liver transplantation: role of color duplex sonography in the diagnosis and management of vascular complications. Bildgebung 1994; 61: 14-19
• 30 Maeda N, Osuga K, Mikami K et al. Angiographic evaluation of hepatic arterial damage after transarterial chemoembolization for hepatocellular carcinoma. Radiat Med 2008; 26: 206-212
• 31 Rong GH, Sindelar WF. Aberrant peripancreatic arterial anatomy. Considerations in performing pancreatectomy for malignant neoplasms. Am Surg 1987; 53: 726-729
• 32 Hemming AW, Finley RJ, Evans KG et al. Esophagogastrectomy and the variant left hepatic artery. Ann Thorac Surg 1992; 54: 166-168
• 33 Grau M, Subirana I, Vila J et al. Validation of a population coronary disease predictive system: the CASSANDRA model. J Epidemiol Community Health 2014; Epub 2014 Mar 11
• 34 Ratib S, West J, Crooks CJ et al. Diagnosis of liver cirrhosis in England, a cohort study, 1998–2009: a comparison with cancer. Am J Gastroenterol 2014; 109: 190-198
• 35 Siegel R, Ma J, Zou Z et al. Cancer statistics, 2014. CA Cancer J Clin 2014; 64: 9-29
• 36 Wang S, He X, Li Z et al. Characterization of the middle hepatic artery and its relevance to living donor liver transplantation. Liver Transpl 2010; 16: 736-741
• 37 Decurtins M, Friend PJ, Calne RY. Incidence and outcome of donor arterial anomalies in liver allografts. Transplant Proc 1987; 19: 2394-2395
• 38 Makisalo H, Chaib E, Krokos N et al. Hepatic arterial variations and liver-related diseases of 100 consecutive donors. Transpl Int 1993; 6: 325-329
• 39 Rygaard H, Forrest M, Mygind T et al. Anatomic variants of the hepatic arteries. Acta Radiol Diagn (Stockh) 1986; 27: 425-427
• 40 Chen CY, Lee RC, Tseng HS et al. Normal and variant anatomy of hepatic arteries: angiographic experience. Zhonghua Yi Xue Za Zhi (Taipei) 1998; 61: 17-23