INTRODUCTION

The liver is the largest organ and gland in the body, located in the right upper quadrant of the abdominal cavity. The liver receives its blood supply from 70 % hepatic portal system and from 30 % hepatic arterial system. The hepatic arterial system anatomy has a variable structure and variation rate varies between 25 % and 75 % (^{Standring et al., 2008}). These vessels may be accessory, occurring in addition to the normal arterial supply, or replaced representing the primary arterial supply to the lobe. These arteries arise from superior mesenteric artery (SMA), left gastric artery (LGA), abdominal aorta (AA) or other visceral branches (^{Hiatt et al., 1994}).

Liver transplantation, laparoscopic cholecystectomy and other liver surgeries have caused increased interest in the liver arterial system anatomy (^{Chen et al., 2009}). Arterial variations are important in the planning and performance of all surgical and radiological procedures in the upper abdomen. However, surgical errors resulting from the inability to understand the anatomy of the liver arterial system cause serious consequences and even mortality for patients (^{Singh et al., 2014}). For this reason, many studies examining the anatomy and variations of the liver arterial system have been conducted and the Michels Classification has been defined as the most accepted classification in the literature (Michels, 1951; ^{Gruttadauria et al., 2001}; ^{Brandhagen et al., 2003}; ^{Koops et al., 2004}; ^{Prabhasavat & Homgade, 2008}; Pérez- Saborido et al., 2012).

Knowing the diameter of hepatic arteries is the great importance for vascular and biliary reconstruction, hepatic artery thrombosis and especially liver transplantation (^{Douard et al., 2002}). In addition, arterial diameters must be known for the correct selection of the catheter in the treatment of angioplasty, angiography and hepatic metastatic tumors (^{Watanabe et al., 2005}). However, information about artery diameters is still insufficient (^{Da Silveira et al., 2009}).

Although sufficient information is available in the literature on the anatomy and variations of the liver arterial system, there is no comprehensive study on the relationship between anatomical variations and artery diameters. Therefore, in this study, it was aimed to examine the anatomy and variations of the hepatic arterial system and measure the diameters of the related arteries with Multidetector Computed Tomography (MDCT) images.

MATERIAL AND METHOD

The study was carried out on the images obtained from the MDCT device (Sensation 64, Siemens, Erlangen, Germany) serving in the Selçuk University Department of Radiology, after permission of Non-Interventional Clinical Research Ethics Committee was obtained (Selçuk University, 218/312-12.09.2018). A total of 500 patients were included in the study, including 216 women (16-90) and 284 men (14- 90). 13 patients with circulatory and vascular diseases or who could not been obtained optimal images were excluded from the study because they fell outside the study criteria.

While the patients were in supine position, the measurements were performed on the 0.3 mm sectional ab- dominal images (from diaphragm to pubic region) taken in the arterial phase using MDCT. Using the Enlil Pacs imaging program (Enlil PACS Viewer, Eroglu Yazılım, Eskisehir, Turkey) at the workstation, hepatic arterial system (AA, celiac trunk (CT), common hepatic artery (CHA), proper hepatic artery (PHA), right hepatic artery (RHA), left hepatic artery (LHA) variations, localization, initial diameter and relationships with other arteries were recorded. While the diameters of hepatic arterial system vessels were measured on the axial plan, the variations of the related arteries were determined on the coronal plan. The Michels classification was taken as reference in the examination of variation situations (Michels).

Statistical analysis of the data was done with SPSS version 19.0 package program (SPSS Inc., Chicago, IL, USA). Statistical analysis included means, standard deviations and percentages of the data. Kolmogorov-Smirnov test was used to suitability of the data for normal distribution and it was determined that datas were not homogeneous. Mann-Whitney U test was used for binary group comparisons in which the significant difference between the sexes and the variant/normal measurements were evaluated. The relationship between variation incidence and sex was analyzed by Chi-square test. The results were evaluated in the 95 % confidence interval and the data with p value less than p<0.05 were considered statistically significant.

RESULTS

The mean age of the 500 patients in the study was 60.4 ± 17, the mean age of 216 women was 61.1 ± 17.1, and the mean age of 284 men was 59.9 ± 16.9.

In the study, the data were analyzed and evaluated statistically in two main groups; variations of the hepatic arterial system and diameter measurements of these arteries.

Variations of the hepatic arterial system. Classic anatomy (Type I, Fig. 1) was observed in 428 (85.6 %) of 500 cases, and also hepatic arterial system variation was found in 72 (14.4%) of them. In cases with variation, Type III (Fig. 2) was detected in 63 cases (87.5 %), Type II (Fig. 3) in five cases (6.9 %) and Type IV (Fig. 4) in one case (1.4 %). In addition to these, we found three rare variations that not defined in the Michels classification. Two of these were the replaced LHA originated from CHA (Fig. 5), and the other one was the the replaced RHA originated from CT (Fig. 6) was recorded. The distribution of variations by sex was shown in Table I.

The frequency of hepatic arterial variations and the relationship between sexes were analyzed. Depending on this result, no statistically significant relationship was found between the frequency of variation and sex (p=0.068).

Diameter measurements of the hepatic arterial system. In this study, the transverse diameter of AA, the initial diameters of CT, CHA, PHA, RHA, LHA were examined according to sex and overall distribution. Arterial diameter measurements; classical anatomy, variational anatomy and all cases were evaluated separately (Table II).

When the diameter measurements of cases compatible with classical anatomy and variational anatomy were examined, it was determined that CHA and PHA diameters were larger in cases compatible with classical anatomy and LHA diameter was larger in cases with variations. And these differences were founded statistically significant (p<0.05).

When the arterial diameter measurements were compared by sex, it was found that all means were greater in men than in women. Except for the PHA diameter, all these measurement differences were statistically significant (p<0.05).

Fig. 1 A. Arterial system compatible with Michels Type I on coronal section, B. Arterial system compatible with Michels Type I on inspace coronal section. 

Fig. 2 A. Arterial system compatible with Michels Type III on coronal section, B. Arterial system compatible with Michels Type III on inspace coronal section. SMA; superior mesenteric artery, RHA; right hepatic artery. 

Fig. 3 A. Arterial system compatible with Michels Type II on coronal section, B. Arterial system compatible with Michels Type II on inspace coronal section. LGA; left gastric artery; LHA; left hepatic artery. 

Fig. 4 A. Arterial system compatible with Michels Type IV on coronal section, B. Arterial system compatible with Michels Type IV on inspace coronal section. RHA; right hepatic artery; LHA; left hepatic artery. 

Fig. 5 A. The replaced LHA originated from CHA on coronal section, B. The replaced LHA originated from CHA on inspace coronal section. CHA; a. hepatica communis, LHA; left hepatic artery. 

Fig. 6 A. The replaced RHA originated from CT on coronal section, B. The replaced RHA originated from CT on inspace coronal section. CT; celiac trunk, RHA; right hepatic artery. 

Table I Common and rare types of hepatic arterial system variations and distribution by sex according to Michaels Classification. 

Table II Diameter measurements of cases compatible with classical anatomy, variational anatomy and all cases by sex. 

^{Data expressed as mm ± standard deviation of mean. * A statistical difference was determined according to sex groups}.

DISCUSSION

In today's conditions, progress in liver surgery and microvascular reconstruction techniques, and improvements in imaging methods both increase surgical success and reduce complications (^{Orguc et al., 2004}). For example, liver and hepatic arterial system surgeries are usually planned according to the classical anatomy pattern. However, arterial system variations are also encountered during surgical dissection. Sometimes, the variations are detected by the leakage that occurs during liver harvesting operation or graft preparation on the "bench". These situations increase arterial complications risk and reveal that variations may be overlooked during surgical dissection (^{Abdullah et al., 2006}). With these examples, importance of the arterial variations detected in the donor and recipient before the operation in reducing the risk of complications was demonstrated (Orguc et al.; Abdullah et al.).

Therefore, images obtained from MDCT angiographies of 500 cases randomly selected from the normal population were used in this study. Hepatic arterial system variations and diameter measurements of the related arteries were evaluated on these images.

There are many studies in the literature on the anatomy and variations of the hepatic arterial system. However, Michels' dissection study performed on 200 cadavers in 1952 is accepted as the basis for the classification of hepatic arterial system anatomy and variations (Michels). In the literature, it was determined that many researchers reference the Michels classification in their studies. The values in these studies are shown in Table III.

As seen in the Table III, it was determined that the most common variation in most studies in the literature was Type III (Hiatt et al.; Gruttadauria et al.; Brandhagen et al.; Koops et al.; Prabhasavat & Homgade; ^{Kim et al., 2012}; Pérez- Saborido et al.; ^{Taha Ali et al., 2012}). In other studies in the literature, the most common variation type was Type II (^{Todo et al., 1987}; Prabhasavat & Homgade) and Type V (^{Özbek, 2005}). In our study, 500 cases were evaluated according to Michels classification. Classical anatomy (Type I) was found in 85.6 % of the cases and variational anatomy was found in 14.4 %. In cases with variation, the most common variation was Type III with a rate of 87.5 % and followed by Type II variation with a rate of 6.9 %. In addition, Type IV was found with a rate of 1.4 %, while three cases that did not fit the Michels classification were also determined. Among these three cases, two of them were replacement LHA originating from CHA (2.8 %) and one of them replacement RHA (1.4 %) originating from CT. Thus, our study was determined as the study with the highest rate of classical anatomy in the literature after Prabhasavat et al. These differences between the literature data are explained by the variability of geographical, ethnic and genetic factors or the embryological development process (as a result of the partial or complete persistence of fetal development) (Özbek).

Table III Comparison of hepatic arterial system variations and frequency with literature studies, It was shown according to the Michels classification, values were shown as percentages (%). 

Knowledge of normal arterial diameters plays an important role in the accurate radiological diagnosis of arterial aneurysms, moreover it reduces complications and increase surgical success (^{Nghiem et al., 1999}). These parameters are great essentials in the determination of anastomosis and reconstruction techniques, choice of the appropriate grafts and stends, vascular modelings, revealing the conditions that may prevent the operation, and the follow-up of liver transplantations (^{Anton et al., 2018}). For example, prior knowledge of normal and expected values for specific arteries can assist in the early diagnosis of an arterial stenosis even before clinical signs of low arterial flow. Another example is the need for anastomosis of the accessory arteries as well as the normal arteries in case of a variation that requires protection from ischemia and additional anastomosis. Therefore, it is also necessary to change the surgical plan (Anton et al.). Otherwise, organ transplantation may result in failure due to graft failure or insufficient regeneration (Winter et al., 1995). For this reason, a detailed evulation of hepatic arterial system morphometry in the preoperative period is essential for a successful transplantation (^{Vandamme et al., 1969}).

There are very few studies in the literature to determine the diameter measurements of the hepatic arterial system. These studies are also limited in terms of their scope, parameters and methodology. The method, sample numbers and measurement values of these studies are shown in Table IV.

Table IV Comparison of arterial diameter measurements with literature studies, Data expressed as mm ± standard deviation of mean, Multidedector Computed Tomography; MDCT, Magnetic Resonance Imaging; MRI, Doppler ultrasonography; DU. 

Da Silveira et al. cadaver study, ^{Noda et al. (2018}) MR and CT angiography and Chen et al. angiography studies were closest to our study in terms of their scope and methodologies.

However, differences were observed in the measurement data of the studies mentioned with our study data. For example, the artery diameters in the study of Da Silveira et al. were larger than our study. This case could be explained by measuring larger vessel diameters due to the non- functioning of smooth muscles in cadavers. The artery diameters in the studies of Noda et al. and Chen et al. were smaller than in our study. The reasons for these differences were that Chen et al. measured the inner diameter of arteries with angiography method, and Noda et al. measured the arterial diameters on images of individuals with clinical symptoms using different contrast material and different imaging device. As the general reasons for these discrepancies in measurement values could be explained due to differences in imaging methods and devices, ethnics origin, sex, cadaveric or alive cases, variation types.

Another important point about our results was that the PHA diameter in cases compatible with classical anatomy was greater than 3 mm, and the PHA diameter in cases with variations was less than 3 mm. These values were consistent with the studies of ^{Ishigami et al. (2004}) and Da Silveira et al. These results draw attention to the knowledge that variational arterial system anatomy increases the risk of complications in the recipient after transplantation and creates a higher risk in hepatic arteries smaller than 3 mm (Nghiem et al.; Douard et al.; Da Silveira et al.).

In conclusion, hepatic arterial system of 500 cases was evaluated retrospectively on MDCT images in our study. Classical anatomy was found in 85.6 % of the cases and the most common variation was determined as Type III. When arterial diameter measurements were examined, it was found statistically significant that CHA and PHA diameter was larger in cases compatible with classical anatomy, and LHA diameter was larger in cases showing variation. Also, it was determined that the diameter measurements of men were larger than women when these groups were evaluated by sex, and statistically significant difference in all diameters except PHA.

Our study, which examines the frequency of hepatic arterial system variations and the relationship of arterial diameters with sex and variations, is an original study on its scope and sample. We believe that this study will further advance with some modifications and additions and will make a great contribution to the branches of surgery and radiology.