to help further characterize hepatic lesions. Resolution of some lesions may be slightly better with MRI.

Additionally, certain lesions, such as hemangiomas and cysts, can easily be identified based on MRI

characteristics when the CT characteristics are indeterminate (Fig. 57-16). The disadvantages of MRI are

its increased cost, increased amount of time to perform, inability to quickly screen other organs and

body cavities within the same session, and wide variability in quality from one imaging center to

another.

7 Magnetic resonance cholangiopancreatography (MRCP) is becoming more widely used to

noninvasively view biliary anatomy (Fig. 57-17). The scans are heavily T2 weighted, which maximizes

the signal from the biliary tree. No injection of contrast agent is needed, and under optimal conditions

the resulting imaging can rival that of formal cholangiography. Three-dimensional reconstructions can

be performed to view the biliary tree from multiple angles and can be helpful in distinguishing between

stones, strictures, and neoplasms. Again, these reconstructions are vital to donor selection in live donor

liver transplantation. Rarely is it necessary to proceed with endoscopic retrograde

cholangiopancreatography (ERCP) to define the distal intrahepatic ductal anatomy.

Figure 57-15. Three-dimensional reconstruction of the hepatic vasculature.

Ultrasonography

8 Hepatic ultrasonography can be applied transcutaneously or intraoperatively via open or laparoscopic

surgery. It can be useful in identifying lesions within the hepatic parenchyma, to describe the

consistency (i.e., fatty or cirrhotic) and identify dilation of the biliary tree and any abnormalities or

stones within the gallbladder. In hepatobiliary surgical centers, intra operative ultrasonography is used

routinely to assess the anatomy of the pedicles (portal vein, hepatic artery, and bile duct), the hepatic

veins, and the hepatic parenchyma. It is useful both to further identify and characterize lesions within

the hepatic parenchyma and to delineate their relationships within the eight anatomic segments of the

liver. Additionally, it is often helpful to delineate proximity of lesions to major vascular structures and

to survey for abnormal anatomy in planning a resection.

With ablation therapies more commonly employed, ultrasound has become indispensable in directing

the use of radiofrequency ablation. This ablation therapy can be performed percutaneously,

laparoscopically, and during open surgery.

9 Typically, intraoperative assessment of the liver involves examining the portal pedicles. The main

portal pedicle is identified within the hepatoduodenal ligament. It is followed superiorly to the portal

bifurcation into the main right and left pedicles. The portal pedicles are invested with the Glisson

capsule and have a very echogenic covering to them in contrast to hepatic vein branches. The main right

portal pedicle is followed toward the right where it gives off an anterior branch and a posterior branch

(Fig. 57-18A). The right anterior branch gives off separate pedicles to segment V (caudad) and to

segment VIII (cephalad). The right posterior branch gives off separate pedicles to segment VI (caudad)

and to segment VII (cephalad). The main left pedicle is usually much longer and courses intact to the

base of the umbilical fissure before branching into various segmental pedicles (Fig. 57-18B). At the base

of the umbilical fissure, the main left pedicle courses anteriorly toward the round ligament and gives off

a pedicle to segment IV medially and pedicles to segments II and III laterally. Next, if the falciform

ligament has been divided, the hepatic veins can easily be visualized using intraoperative

ultrasonography (Fig. 57-18C). As described previously, usually a larger right hepatic vein can be

1462

delineated and smaller left and middle hepatic veins joining into a common trunk before emptying into

the IVC are seen. Commonly, an umbilical hepatic vein branch can be identified coursing between the

middle and left hepatic veins and running under the falciform ligament. Not uncommonly, significant

accessory right hepatic veins can be seen emptying from the posterior surface of the right liver directly

into the IVC as it courses posterior to the liver. The identification of these accessory right hepatic veins

is potentially important for both vascular control and preservation of outflow from the liver (in

occasional cases where outflow of the remnant right liver can be supported by a very large accessory

vein). Finally, the hepatic parenchyma is systematically scanned to identify lesions within the liver (Fig.

57-18). It is sometimes useful to adjust the ultrasound settings on a known lesion defined preoperatively

to maximize the echogenicity in the hopes of identifying other occult lesions not identified

preoperatively.

Figure 57-16. A,B: T1-weighted magnetic resonance imaging (MRI) with gadolinium from the same patient as in Figure 57-14

with history of colorectal cancer and three lesions in the liver. Lesion 1 in segment VIII is irregular and rim enhancing and was a

colorectal cancer metastasis. Lesion 2 straddling segments IV and VIII has smooth borders, is not rim enhancing, and was found to

be a cyst. Lesion 3 straddling segments IV and III across the umbilical fissure is irregular, rim enhancing, and was a colorectal

cancer metastasis. C,D: T2-weighted MRI from the same patient with history of colorectal cancer and three lesions in liver. Lesion

1 in segment VIII is irregular and mildly bright and was a colorectal cancer metastasis. Lesion 2 straddling segments IV and VIII

has smooth borders, is very bright, and was found to be a cyst. Lesion 3 straddling segments IV and III across the umbilical fissure

is mildly bright and was a colorectal cancer metastasis. Colorectal metastases and many tumors are mildly bright on T2-weighted

MRI, whereas cysts and hemangiomas are typically very bright.

Positron Emission Tomography

Positron emission tomography (PET), especially when combined with CT (PET-CT), has become a

valuable tool in helping to select patients who will most benefit from aggressive liver resection. This

technique is based on the increased metabolism of glucose in neoplastic tissues. A glucose analog,

fluorodeoxyglucose, that is tagged with fluorine-18 is injected intravenously before scanning and is

retained preferentially in metabolically active tumors over normal tissue. Sometimes PET scans will

identify areas of occult disease within the liver, but more importantly, they can identify areas of

extrahepatic occult disease previously unsuspected. When combined with a CT scanner within the same

machine and the ability to fuse images, the areas of increased activity can be more precisely

anatomically identified (Fig. 57-19).

1463

Figure 57-17. Magnetic resonance cholangiopancreatography (MRCP) of a patient after cholecystectomy with mild dilation of

common hepatic duct (CHD). The pancreatic duct (PD) is also visible.

Correlation of Computed Tomographic Images with Segmental Anatomy

Preoperative CT remains the primary imaging modality used by most surgeons before hepatic resection.

Figure 57-14 is provided to help correlate CT images to the segmental anatomy defined previously. The

segments of the liver are defined using identifiable structures on the CT (Fig. 57-20).

PREOPERATIVE EVALUATION OF HEPATIC RESERVE

Whenever a surgical resection is planned, an important consideration is whether the remnant liver will

be sufficient to regenerate and sustain the patient long term. In patients with relatively normal hepatic

parenchyma (without active hepatitis, cirrhosis, or metabolic defects), up to 75% of the hepatic volume

can be resected with good recovery as long as the remnant liver has adequate portal venous and hepatic

arterial inflow, adequate hepatic venous outflow, and adequate biliary drainage. Many groups around

the world have used various strategies to predict hepatic reserve (Tables 57-1 to 57-3). None of these

tests or strategies has been demonstrated to clearly better predict outcome than another. Many centers

in the United States rely simply on the Child–Pugh or MELD score and the prediction of adequate liver

remnant volume after resection. In select circumstances, it may be of benefit to perform portal vein

embolization to the right or left half (rare) of the liver in the hopes of obtaining compensatory

hypertrophy of the other side before resection. This is especially useful when the predicted liver

remnant after resection is small or if the patient has an underlying hepatic dysfunction that may not

allow the remnant to fully regenerate and sustain the patient long term. To gain maximal growth of the

left lateral section of the liver, some centers will also embolize the portal vein branches to segment IV

in addition to the main right portal vein. The disadvantages of portal vein embolization include the need

to wait 3 to 4 weeks before resection to allow the compensatory hypertrophy to occur and, for more

central lesions, the need to commit to taking out one or the other side with an extended hepatectomy

without the benefit of intraoperative evaluation.

1464