Journal of Indian Association of Pediatric Surgeons
Journal of Indian Association of Pediatric Surgeons
                                                   Official journal of the Indian Association of Pediatric Surgeons                           
Year : 2023  |  Volume : 28  |  Issue : 3  |  Page : 233--241

Preoperative triple-phase three-dimensional-multi-detector computed tomography imaging of the hepatic vascular tree: An accurate road map prior to tumor resection in hepatoblastoma

Aparajita Mitra1, Veereshwar Bhatnagar1, Sandeep Agarwala1, Arun Kumar Gupta2, Manisha Jana2, M Srinivas1, Anjan Kumar Dhua1,  
1 Department of Paediatric Surgery, All India Institute of Medical Sciences, New Delhi, India
2 Department of Radiodiagnosis, All India Institute of Medical Sciences, New Delhi, India

Correspondence Address:
Veereshwar Bhatnagar
ESIC-PGIMSR and ESIC Medical College and Hospital, NH-3, KL Mehta Rd, Behind BK Hospital, New Industrial Town, Faridabad - 121 012, Haryana


Objective: To evaluate the hepatic vasculature/tumor relations in hepatoblastoma patients with three-dimensional (3D) reformatted images after triple-phase multi-detector computed tomography (MDCT) and to compare these with the surgical findings to judge the accuracy of this investigation. Materials and Methods: The study was carried out in hepatoblastoma patients after appropriate neo-adjuvant chemotherapy, prior to resection. Images were postprocessed at a dedicated workstation for multi-planar reformations, maximum intensity projection, curved planar reformations, and volume-rendered technique reconstructions. The reporting was done as per a specific protocol by both the radiologist and surgeon (per-operative findings) and the accuracy of MDCT ascertained as per concordance between the surgical and imaging findings. Results: Fourteen children (13 boys, 1 girl) underwent surgery. Clinically, relevant information regarding vascular, tumor involvement, and interface with vessels was provided by the study in all cases. Although all tumors were deemed resectable on preoperative imaging, one procedure was abandoned due to an unanticipated portal cavernoma. While a few anatomical variations were unexpectedly encountered during surgery, there was good concordance overall between findings on imaging and surgical exploration. Conclusions: MDCT with 3D reformatting provides accurate virtual representations of the hepatic tumor. This allows simulation of surgical resection with decreased risk of vascular injury and postoperative liver failure.

How to cite this article:
Mitra A, Bhatnagar V, Agarwala S, Gupta AK, Jana M, Srinivas M, Dhua AK. Preoperative triple-phase three-dimensional-multi-detector computed tomography imaging of the hepatic vascular tree: An accurate road map prior to tumor resection in hepatoblastoma.J Indian Assoc Pediatr Surg 2023;28:233-241

How to cite this URL:
Mitra A, Bhatnagar V, Agarwala S, Gupta AK, Jana M, Srinivas M, Dhua AK. Preoperative triple-phase three-dimensional-multi-detector computed tomography imaging of the hepatic vascular tree: An accurate road map prior to tumor resection in hepatoblastoma. J Indian Assoc Pediatr Surg [serial online] 2023 [cited 2023 Sep 23 ];28:233-241
Available from:

Full Text


Pediatric oncosurgeons have always sought tools which simplify the surgery of solid tumors and ensure patient safety. The surgical decision-making process has to be supported by the appropriate kind of imaging which provides a clear idea of the spatial anatomy of the tumor, especially in relation to the major vessels.

The advent of multi-detector computed tomography (MDCT) in 1998 ushered in a new era in diagnostic imaging as it allowed the acquisition of volume data without the misregistration of anatomic detail. The triple-phase technique with MDCT was proposed by Foley et al.[1] This rapidly became a substitute for conventional angiography when combined with computer-aided imaging.

Long-term survival in hepatoblastoma depends on complete surgical resection. Approximately 50% tumors are deemed unresectable and particularly complex in terms of hepatic involvement at presentation, especially in a developing country such as ours. With the use of neoadjuvant chemotherapy, up to 85% of these deemed unresectable tumors become resectable with a significant decrease in eventual loss of functioning parenchyma.[2] The overall survival rate for children with hepatoblastoma is reported to be 65%–70%.[3]

There is a need for safe surgical resection of hepatoblastoma and this requires that the hepatic vein confluence, the portal vein bifurcation and the retrohepatic inferior vena cava (IVC) be free from tumor. Recent protocols of ongoing large scale trials in hepatoblastoma such as the AHEP0731 protocol have even defined when to intervene surgically. It is advised that the radiographic margins from each of the aforementioned key vascular structures should be at least 1 cm.[4]

There have been very few studies in which the capability and accuracy of MDCT in delineating vascular territories before hepatic resection for hepatoblastoma have been explored. While magnetic resonance imaging (MRI) has largely supplanted MDCT in the evaluation of pediatric liver tumors, it is a fairly costly investigation with the added need for general anesthesia which precludes its widespread application in resource-challenged hospitals catering to huge populations. Thus, computed tomography (while adhering to the As Low AS Reasonably Achievable [ALARA] principle) still remains the investigation of choice in centers like ours. This study attempts to compare the information gleaned from the preoperative imaging of the hepatic vascular tree and its relation to a tumor, with the anatomy observed during surgery.

 Materials and Methods

This cross-sectional study was conducted at a single institution over a period of 2 years (August 2013 to July 2015) after the requisite clearance from the Institutional Ethics Board. It aimed to evaluate the hepatic vasculature and vessel-tumor relations in hepatoblastoma using three-dimensional reformatted images after triple-phase MDCT. The information thus obtained was applied during surgical tumor resection. The potential of this investigation was assessed in terms of predefined anatomical parameters.

Fourteen children with hepatoblastoma planned for resection were included in the study. They had undergone prior tumor localization and characterization by fine-needle aspiration cytology and had received appropriate neo-adjuvant chemotherapy. The risk categorization was done as per the International Childhood Liver Tumor Strategy Group Société Internationale d'Oncologie Pédiatrique - Epithelial Liver Tumor Study Group (SIOPEL) guidelines [Table 1]. Those with inoperable tumors were excluded.{Table 1}

Multi-detector computed tomography imaging technique

The children were admitted for the sole purpose of the imaging study as per the hospital protocol. After positioning the child on the gantry, sedation was achieved with a combination of Injection Midazolam (dose - 0.1–0.2 mg/kg) and Injection Fentanyl (dose 1–2 μg/kg) intravenous bolus with full preparedness for general anesthesia if required.

The SOMATOM AS Plus 64-detector row computed tomography (CT) machine (Siemens, Erlangen, Germany) was used for all the studies. A pediatric protocol with the proprietary CareDose4D technique was used to keep the radiation dose to the minimum while adhering to the As Low as Reasonably Allowable (ALARA) principle.

Mechanical injection of the calculated amount (2 ml/kg) of nonionic iodinated contrast (Iohexol, Iodine concentration of 300 mg/ml) was given at a rate of 5 ml/s through a preplaced 20 G cannula using a mechanical power injector. Each study began with an unenhanced acquisition of the liver with the region of interest (ROI) kept to a minimum. The liver including the celiac axis and the mesenteric artery were imaged in a single breath hold per phase with a predetermined time of acquisition. The timing of each of the phases was determined by bolus tracking technology (SmartPrep, GE Healthcare). This necessitates placing the ROI around the descending aorta and measuring the change in aortic enhancement over time.

The same team of radiologists (AKG, MJ) with an expertise in pediatric imaging used a workstation to generate three-dimensional (3D) images of the tumor and hepatic vascular anatomy. The software used was the syngoInSpace4D (Siemens, Erlangen, Germany) with the Advanced Bone Removal Tool which allowed rapid segmentation and removal of bony structures along with fast 3D Multi-Planar Reformations, Maximum Intensity Projection, Curved Planar Reformations, and Volume-Rendered Technique overview of vascular structures.

The images were uploaded to a picture archiving and communication system and evaluated by the radiologist and the operating surgeon in tandem with emphasis on the following parameters:

Tumor location, relation with the surrounding vessels, segments occupied, and impingement on vasculatureOrigin and course of the coeliac axis, the superior mesenteric artery (SMA)The course of the hepatic arteries, point of bifurcation, and existence of any accessory vesselsThe branching of the portal vein and relation with the hepatic segmentsThe hepatic veins, their confluence with each other, and the IVC.

The extent of resection was planned based on the extent of the tumor. The intraoperative findings were recorded and verified by direct visualization of the lesion, its dimensions, the segments involved, and the hepatic vasculature. Both the radiological and intra-operative findings were recorded on multiple pro forma and the images archived.

Statistical analysis

The collated data were analyzed using STATA software version 11 (Stata Corp LP, Texas, USA). The findings on MDCT were compared with the actual in vivo visualization of both the tumor and the tumor-vasculature relationship. The findings were analyzed and tests of marginal homogeneity (Stuart-Maxwell test) were applied. P <0.05 was considered statistically significant.


A total of 14 children were included in this study (13 boys and one girl). The mean age at presentation was 15.3 (9–72 months). There were 8 infants (57%). One patient was classified as Pretreatment extent 2 (PRETEXT 2), Standard Risk (SR), three patients (21.4%) were PRETEXT 3, SR while the other 10 (71.4%) were PRETEXT 3, High Risk (HR).


Details of pretreatment imaging were available for both groups of patients-those presenting to our institution for the first time as well as those referred from elsewhere (having received some or no therapy). The involvement of segments was noted [Figure 1]a. It is interesting that none of the seven patients referred from other hospitals had undergone a triple-phase contrast-enhanced computed tomography (CECT) as an initial investigation.{Figure 1}

Prior to any treatment

Segment 1, 2, and 3 in 14.3% each, segment 4A in 11 (78.5%), segment 4B in 10 (71.4%), segment 5 in 13 (92.8%), segment 6 in 10 (71.4%), segment 7 in 9 (64.3%), and segment 8 in 12 (85.7%) patients.

Prior to resection

Segment 1 in 2 (14.3%), segment 2 in 3 (21.4%), segment 3 in 2 (14.3%), segment 4A in 7 (50%), segment 4B in 7 (50%), segment 5 in 10 (71.4%), segment 6 in 7 (50%), segment 7 in 6 (42.8%), and segment 8 in 8 (57.1%) patients.

Pretherapy vasculature

The information regarding vasculature on pretherapy imaging is presented in [Table 2].{Table 2}


One patient (7.2%) had pulmonary metastasis at presentation and underwent pulmonary metastasectomy twice (left lower lobe and right upper lobe) after right hepatectomy.

Neoadjuvant chemotherapy

The children were stratified into the risk categorization advocated by the SIOPEL group and received appropriate neo-adjuvant chemotherapy, the details of which are detailed in [Table 3].{Table 3}

Comparison of multi-detector computed tomography and intra-operative findings

All tumours were deemed resectable on the basis of CT findings after chemotherapy.


Coeliac axis, superior mesenteric artery, and accessory hepatic artery

These vessels were visualized from their point of origin and traced proximally on the 3D reconstruction. Neither was reported to be involved or displaced in any of the images.

The intra-operative findings were concordant for a lack of involvement or displacement by tumor bulk in 13 of 14 children (92.8%). These were not evaluated in one child in whom the procedure was abandoned because of an unresectable tumor with portal cavernoma and excessive bleeding peroperatively.

None of the preoperative images had identified any accessory hepatic arteries. This was at odds with Case 1, in whom an accessory artery was seen arising from the SMA during right hepatectomy.

Common hepatic artery

None of the children had an involvement of the common hepatic artery (CHA) on pretherapy or preoperative imaging which was concordant with the intra-operative findings in all 14 (100%) cases.

Right hepatic artery

Two children (14.3%) were found to have significant findings concerning the right hepatic artery (RHA) on preoperative imaging-

 Case 4

A 9-month-old male with a tumor in the right lobe, without any vascular involvement on initial imaging or evidence of the SIOPEL risk factors. He received four cycles of neoadjuvant cisplatin monotherapy but was shifted to the PLADO regimen in view of rising AFP levels. Surgery was done after 1 cycle of PLADO chemotherapy.

 Case 9

A 10-month-old male with a right lobe tumor, no vascular involvement or SIOPEL risk factors. Classified as PRETEXT 2, SR in view of involvement of segment 5, 6, and 7 on initial imaging.

However, only one of these children (case 9) showed evidence of RHA involvement during surgery. A test of marginal inhomogeneity (Stuart-Maxwell) did not show any significance of this discordance.

Anatomical variant: Left hepatic artery

None of the 14 children ever had an involvement of the left hepatic artery (LHA) reported on imaging. Hence, there was 100% concordance with the peroperative findings. An anatomical variant was noted during surgery, although the LHA was found to be originating from the left gastric artery in one child (Case 6).

Portal vein

Main portal vein

 Preoperative imaging

Main portal vein (MPV) was reported to be involved in or closely related to the tumor in two cases (14.2%), Cases 6 and 7. In one of these, a prechemotherapy CT had reported splaying of the hepatic veins and an IVC thrombus additionally.

 Surgical finding

Case 6 showed encasement of the MPV during surgery, while in Case 7, the MPV was clear of infiltration. MPV involvement was an unanticipated finding in Case 9 (previously mentioned in the context of RHA involvement). The MPV was displaced anteriorly and to the right with the tumor infiltrating its posterior aspect.

Thus, one case (7.1%) each was truly positive, falsely positive, and falsely negative according to the preoperative MDCT.

Left portal vein

The left portal vein (LPV) was found to be involved in two children (14.3%), Case 6 and Case 7 (who also had MPV involvement reported), on the preoperative MDCT. Both children were found to have an encased LPV at surgery.

Thus, there was 100% concordance between the findings on imaging and surgery.

Right portal vein

Four children (28.5%) had significant findings concerning the right portal vein (RPV) on preoperative imaging:

 Case 1

A 12-month-old boy, classified as PRETEXT 3/SR and involvement of segments 4b, 5 and 6. Preoperative CT picked up a significant mass effect on the posterior segmental branch of the RPV.

 Case 6

Previously described, with involvement of MPV also reported.

 Case 9

Previously described, with involvement of the RHA also reported.

 Case 13

A 72-month-old boy with no vascular findings and involvement of segments 4, 5, 6 and 7 reported on pretherapy imaging. He was labeled PRETEXT3/HR in view of portal lymphadenopathy.

Of these children, only two children, cases 6 and 9 were found to have a RPV encased by the tumor during laparotomy. Another child with no previously reported RPV finding had a vein completely encased by the tumor. Thus two cases were falsely positive, two were truly positive, and one was a false negative for RPV involvement on MDCT.

Inferior vena cava and hepatic veins

Inferior vena cava

None of the 14 children ever had an involvement of the IVC reported on imaging and there was 100% concordance with the peroperative findings.

An IVC thrombus was reported in the prechemotherapy imaging in one child (Case 7) but had resolved after four cycles of PLADO chemotherapy and preoperative CT, corroborated by ultrasonography (USG) Doppler showed a normal IVC.

Right hepatic vein

The right hepatic vein (RHV) was involved on preoperative MDCT in one case (7.1%). This boy, Case 9, also had evidence of middle hepatic venous involvement and there was 100% concordance with the per-operative findings.

Middle Hepatic Vein

The middle hepatic vein (MHV) was found to be affected by the tumor in two children (14.3%):

 Case 4

Previously described with RHA involvement reported on preoperative MDCT. The tumor was lying between the MHV and the LHV with a focal narrowing and 180° abutment of the vein (1.7 cm from the ostium into the IVC).

 Case 9

Also had RHV involvement reported on MDCT.

Of these children, only Case 9 had similar findings during surgery. However, a test of marginal inhomogeneity (Stuart-Maxwell) did not show any significance of this discordance.

Left hepatic vein

None of the 14 children ever had an involvement of the LHA reported on imaging and there was 100% concordance with the peroperative findings.

Anatomical variant: Left and middle hepatic vein confluence

A common trunk for the left and MHV was reported in 8 children (57.1%) on MDCT with per-operative visualization corroborating 7 of these (87.5%) findings.

Another child was seen to have a common confluence of the left hepatic vein (LHV) and MHV during surgery despite the MDCT not having highlighted this anatomical finding.

Anatomical variant

Accessory hepatic veins: An accessory hepatic vein was reported in Case 4 on preoperative MDCT. It was an inferior hepatic vein with an ostium into the IVC which was caudal to the RHV-IVC ostium. This was verified during surgery.

Case 2 was found to have a large accessory hepatic vein during surgery which had not been reported on imaging.

Test of marginal inhomogeneity (Stuart-Maxwell) did not show any significance of the discordance in venous anatomy.

Segments involved

The preoperative MDCT was evaluated as per the aforementioned vascular landmarks for segmental involvement and compared with the operative findings [Figure 1]b. The differences between the absolute numbers reported were not significant.


The shorter scan times (hence obviating the need for general anesthesia), extended scan range, and the improved longitudinal resolution of MDCT are especially beneficial for the pediatric age group. The development of software which can virtually reconstruct and simulate surgical resection planes allows surgeons to deconstruct tumor-organ models at will to accurately predict postoperative consequences and eventually, survival.

Hepatoblastoma is the most common malignant liver tumor in children and accounts for 0.5%–2% of all pediatric tumor masses. Current treatment recommendations from the International Childhood Liver Tumor Strategy Group (SIOPEL) stratify patients according to preoperative staging and the PRETEXT (pretreatment EXT-ent of disease) staging system.[5]

The spatial anatomy of the tumor

Reconstruction using special software provided high-quality images of the tumor location. The added advantage of color which virtually simulated the in-vivo appearance, made visualization and projection much simpler. Furthermore, the images could be rotated in both coronal and axial planes to get maximum information [Figure 2]a and [Figure 2]b. The slightly more labor intensive 3D-MDCT and computerized reformatting of images was reserved till surgical resection was actually being planned, i.e. after at least 3 or 4 cycles of chemotherapy.{Figure 2}

Variations in vascular anatomy in our patient cohort

Arterial visualization

It is important to know the variants of the hepatic artery supply which can occur in up to 50% of patients and may unnecessarily complicate surgery if their existence is not known beforehand. Adult series of hepatic metastases from colorectal cancers has shown that this is especially true in the case of those planned for intra-arterial chemotherapy.[6]

Stemmler's group[7] compared both dual-phase 4-and 8-MDCT and the 2D and 3D images for the quality of arterial branch visualization. The concordance between the MDCT angiographic and surgical findings was 93%. Surprisingly, the mean scores for branch order visualization were significantly higher with the 2D than with the 3D studies (P < 0.0001). This could be a result of the postprocessing required to reconstruct an MDCT angiogram. An appropriate threshold such that arteries are vivid, but surrounding parenchyma and early enhanced veins are excluded [Figure 3]. This implies that the small-diameter or faintly opacified arteries do not show up in the final reconstruction.{Figure 3}

Accurate postprocessing is essential to identify anatomical variants. Michels[8] described the variations of hepatic blood supply and showed that the classic hepatic arterial anatomy with the right and left branches arising from the main hepatic artery, is seen in approximately 55% of patients. In our study, a variation in anatomy of the LHA, i.e. an origin from the left gastric artery, a Michel type II anomaly, was missed on the MDCT. This only serves to emphasize that it is the digital reconstruction which is paramount, the information is always available.

Despite the findings of the CT angiography, it is essential to be mindful of the origin of the hepatic artery before ligation. Failure to do so has resulted in hepatic ischemia and hepatic failure. One of the possible scenarios in which the surgeon may have difficulty in ascertaining the origin of a vessel is when a large tumour encroaches upon or distorts the hilum. An early take-off from the CHA may be the saving grace for the LHA when there is a large hilar tumour which has already involved the RHA.

While some groups[9],[10] report accuracy rates of 97%–98% for this modality, others have been less enthusiastic.[11] Sahani et al.[10] noted excellent arterial opacification on CT visualizing tertiary order branches as small as 1 mm in diameter as well as the artery to segment 4 (which represents a mental watershed in hepatic surgeries). We found arterial anomalies on imaging in 2 of 14 patients. A replaced LHA from the left gastric artery was seen in one of our patients. This anomaly was not detected on CT but was appreciated during surgery.

Portal vein visualization

The MPV and its tributaries, the right and LPVs were visualized with clarity although there were times when a closely abutting tumour-vein interface could not be conclusively labelled as “invasion”-possibly one of the few failures of this technique. A child with an extensive tumour involving the caudate lobe and segments 4, 5, and 8 and infiltrating segment 3 and 4 from behind the ligamentum venosum, had an anteriorly displaced portal vein. Peroperatively, the posterior aspect of the vein was found to be infiltrated by the neoplasm leading to significantly difficult dissection. Another child had a suspicion of main portal venous infiltration but eventually the left branch was found to be incarcerated within the mass while the main vein was free. The LHA was also involved.

The imaging and operative findings concerning the LPV were in agreement. The RPV though, had a number of false positives and false negatives. This was not significant statistically and considering that these children eventually underwent right or extended right hepatectomy, not significant in terms of long-term outcomes either. It is intuitive that the vein to be preserved be pronounced free of involvement if one is to be confident of good flow to the future liver remnant without the development of portal hypertension in the long run. Unanticipated postoperative liver failure due to an ischemic or congested remnant is one of the most dreaded complications and this is where virtual reconstructions and simulations of resections score.

Hepatic veins and the inferior vena cava

There was excellent visualization of the outflow tract into the IVC overall. The imaging findings were congruent with the operative findings regarding the IVC, the right and LHVs. There was one false positive MHV report in a child with a tumour wrapped halfway around the circumference of the vein. The focal narrowing picked up on CT was not actually a site of invasion and the vein could be dissected with ease.

The left and MHVs often join to form a common trunk before draining into the IVC. Our series showed this variant on CT in 8 children (57.1%). This approximated the anatomy in 7 (50%) patients. Other groups have reported this in 69% patients and a 100% agreement between CT and intraoperative sonography in the evaluation of hepatic veins (95% confidence interval, 91%–99.9%).[10]

Accessory hepatic veins were seen in 2 (14.2%) cases, one of which was unanticipated. It is important to characterize these veins prior to surgery as they may represent sole drainage of a segment such as the inferior accessory hepatic veins, which typically drain segments V and VI directly into the IVC., At times, a large (>5 mm) tributary vein may drain segment VIII into the MHV, and resection of the MHV during a left hepatectomy may compromise venous drainage of segment VIII, resulting in congestive ischemia and atrophy. These aberrant vessels are also important in certain surgical settings such as liver transplantation and chemoembolization, especially the accessory inferior RHV, which must be dissected separately during surgery.[12]

The detailed images of the vascular supply along with a bird's eye view of the whole abdomen obtained by this technique has obviated the need for conventional angiography, a sentiment echoed by other groups working in this field. Kinoshita et al., used this technique for a group of ten neo-infantile tumors along with volumetry analyses to give an idea about the functional liver remnant. The clarity of the vascular territory and the location of the feeding vessels allowed them to predict the consequences of various types of resections and change their management as appropriate.[13]

The drawbacks

One of the few conceivable disadvantages of this technique is the exposure to radiation when compared to the other imaging modalities such as USG and MRI. The adoption of dose modulation as per body weight[8] and the availability of the requisite software for 3D reconstruction has brought exposures below those for conventional CT studies in keeping with the concept of ALARA (As Low As Reasonably Achievable).[14]

This is not the best technique for demonstrating the biliary tract and this is a significant disadvantage when one considers the implications of prolonged bile leak from missed ducts or obstructive jaundice following iatrogenic injury. Accessory ducts are quite prone for injury unless the surgeon is aware of their existence as was the case in a 9-month-old boy (Case No. 4) with a mass in segment 4 and 5 who underwent extended right hepatectomy. A right accessory duct which was inadvertently transected and subsequently ligated. Unfortunately, a prolonged bile leak necessitated the placement of a Roux loop of jejunum over this duct.

Another drawback is the large data sets, typically between 500 and 1000 images. This of course makes mandatory the use of workstations rather than film to analyze abdominal multisection CT data. Eventually, with much higher speed, the timing of the enhancement is crucial.[15]

Finally, it is not just the hardware or the software which has a bearing on outcomes but human resource as well. So another essential prerequisite, but also a limiting factor, is the need for a trained radiologist who can re-format the data appropriately and work in concert with the surgeon to interpret the images.

Overall, this imaging technique is an excellent way of ascertaining the spatial relations of the tumor with the vessels, but it is not infallible. There are issues with image acquisition and reconstruction and an unavoidable dependence on the right kind of proprietary software. The lack of information about the biliary drainage system is an unfortunate consequence of the technique itself. A larger cohort would provide a greater variety of tumours and clinical situations and allow one to judge the sensitivity and specificity of this technique with greater accuracy.


Triple-phase CECT studies of children with hepatoblastoma can be reconstructed using appropriate software on workstations by trained radiologists to provide the surgeon with a vascular road map prior to complex hepatic resections. Accurate knowledge of segmental involvement and inflow/outflow tracts allows the simulation of surgical resection and the prediction of postoperative outcomes.


All authors have contributed to the concept and design, acquisition of data, or analysis and interpretation of data.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Foley WD, Mallisee TA, Hohenwalter MD, Wilson CR, Quiroz FA, Taylor AJ. Multiphase hepatic CT with a multirow detector CT scanner. AJR Am J Roentgenol 2000;175:679-85.
2Stocker JT. Hepatic tumors in children. Clin Liver Dis 2001;5:259-81, viii-ix.
3Chung EM, Lattin GE Jr., Cube R, Lewis RB, Marichal-Hernández C, Shawhan R, et al. From the archives of the AFIP: Pediatric liver masses: Radiologic-pathologic correlation. Part 2. Malignant tumors. Radiographics 2011;31:483-507.
4Murphy AJ, Ayers GD, Hilmes MA, Mukherjee K, Wilson KJ, Allen WM, et al. Imaging analysis of hepatoblastoma resectability across neoadjuvant chemotherapy. J Pediatr Surg 2013;48:1239-48.
5Roebuck DJ, Aronson D, Clapuyt P, Czauderna P, de Ville de Goyet J, Gauthier F, et al. 2005 PRETEXT: A revised staging system for primary malignant liver tumours of childhood developed by the SIOPEL group. Pediatr Radiol 2007;37:123-32.
6Fasel JH, Muster M, Gailloud P, Mentha G, Terrier F. Duplicated hepatic artery: Radiologic and surgical implications. Acta Anat (Basel) 1996;157:164-8.
7Stemmler BJ, Paulson EK, Thornton FJ, Winters SR, Nelson RC, Clary BM. Dual-phase 3D MDCT angiography for evaluation of the liver before hepatic resection. AJR Am J Roentgenol 2004;183:1551-7.
8Michels NA. Newer anatomy of the liver and its variant blood supply and collateral circulation. Am J Surg 1966;112:337-47.
9Takahashi S, Murakami T, Takamura M, Kim T, Hori M, Narumi Y, et al. Multi-detector row helical CT angiography of hepatic vessels: Depiction with dual-arterial phase acquisition during single breath hold. Radiology 2002;222:81-8.
10Sahani D, Saini S, Pena C, Nichols S, Prasad SR, Hahn PF, et al. Using multidetector CT for preoperative vascular evaluation of liver neoplasms: Technique and results. AJR Am J Roentgenol 2002;179:53-9.
11Bogetti JD, Herts BR, Sands MJ, Carroll JF, Vogt DP, Henderson JM. Accuracy and utility of 3-dimensional computed tomography in evaluating donors for adult living related liver transplants. Liver Transpl 2001;7:687-92.
12Sahani D, Mehta A, Blake M, Prasad S, Harris G, Saini S. Preoperative hepatic vascular evaluation with CT and MR angiography: Implications for surgery. Radiographics 2004;24:1367-80.
13Kinoshita Y, Souzaki R, Tajiri T, Ieiri S, Hashizume M, Taguchi T. A preoperative evaluation for neo-infantile liver tumors using a three-dimensional reconstruction of multidetector row CT. Oncol Rep 2009;21:881-6.
14Recommandations Dosimétriques TDM SFIPP/IRSN. Available from: [Last accessed on 2015 Jul 26].
15Ros PR, Ji H. Special focus session: Multisection (multidetector) CT: Applications in the abdomen. Radiographics 2002;22:697-700.