Chapter 10 Ultrasound-Guided Liver Resection for Hepatocellular Carcinoma Guido Torzilli Keywords Hepatic surgery · Ultrasound-guided percutaneous therapies · Liver resection · Contrast-enhanced Ultrasonography performed intraoperatively Introduction Hepatic surgery performed without a parenchyma-sparing policy carries relevant risks for patients’ survival due to the not negligible occurrence of postoperative liver failure. In particular, the coexistence of liver cirrhosis in most cases of hepatocellu- lar carcinoma (HCC) has a considerable adverse effect on the surgical results. As a matter of fact, recent series are still associated with mortality rates above 5%, which is not negligible [1]. For this reason and for the broadening of ultrasound-guided percutaneous therapies [2], the role of surgical treatment of HCC as the first-choice treatment is now reserved only for patients with normal bilirubin level, no signs of portal hypertension, and moreover carriers of single small HCC [3]. Imaging tech- niques also have been introduced as aids for surgeons in performing liver resection. In fact, since the early 1980s, intraoperative ultrasonography (IOUS) has been used to guide hepatic surgery in patients with liver cirrhosis [4]. Now, liver resections can be carried out with no mortality, even if cirrhosis is associated, combining the needs for oncological radicality and liver parenchyma sparing. This goal is mainly achievable because of IOUS [5, 6]. Recently, the demonstration of the feasibility and efficacy of contrast-enhanced ultrasonography performed intraoperatively (CE- IOUS) has further stressed the relevance of IOUS guidance during liver surgery [7, 8]. In this chapter, technical aspects of IOUS and the impact of this tool during surgery for HCC for both staging and resection guidance are discussed. G. Torzilli (B) Department of Surgery, Istituto Clinico Humanitas IRCCS, University of Milan, School of Medicine, Milan, Italy 135 K.M. McMasters, J N. Vauthey (eds.), Hepatocellular Carcinoma, DOI 10.1007/978-1-60327-522-4_10, C Springer Science+Business Media, LLC 2011 136 G. Torzilli Technical Aspects For a proper IOUS, high-frequency echoprobes (7.5–10 MHz) are necessary and should have a flat shape to allow their management in deep and narrow spaces. For this purpose, T-shaped probes, interdigital probes, and microconvex probes are available. Main factors for probe selection are its volume, its stability, and the wide- ness of the ultrasonographic scanning window: the best probe should be small, thin in width, and short in transverse length, stable, and with a wide ultrasonographic scanning window. In this sense the microconvex probe represents the best compro- mise among all these requirements. Indeed, the T-shaped probe is more stable but has a lower ratio between lateral length and ultrasonographic scanning window than the microconvex one. Linear transducers with enlarged scanning windows are also available now: in the future this solution may combine stability with larger scanning windows (Fig. 10.1). For CEIOUS, we use a convex 3–6 MHz frequency and 1.88–3.76 MHz har- monic frequency transducer from Aloka (Aloka Co., Tokyo, Japan). Once CEIOUS is needed, 4.8 mL sulphur-hexafluoride microbubbles (SonoVue R , Bracco Imaging, Italy) is injected intravenously through a peripheral vein by the anesthesiologist. For HCC, CEIOUS is used for characterizing the new lesions eventually detected Fig. 10.1 The scanning area of this IOUS image is trapezoidal with a flat upper part that rep- resents the contact area between the probe and the liver and a scanning window which enlarges as it gets deeper. It is also evident, how at IOUS, the portal vein (PV) and the hepatic vein (HV) have different thicknesses of their walls: in particular the wall of the portal branch is thicker, as commonly happens 10 Ultrasound-Guided Liver Resection for Hepatocellular Carcinoma 137 at IOUS [7]: the rationale is to check the vascular pattern during contrast enhance- ment of each new l esion. Because in the case of HCC it is very important to identify the arterial vascularization, which lasts from 20–30 sec, each nodule has to be care- fully evaluated and this demands multiple injections in the presence of multiple nodules. This may no longer be a necessity once the new hepato-specific contrast agents become commercially available for clinical use (for the moment available only in Japan): indeed, behaving as the hepato-specific contrast medium used in magnetic resonance does, they could provide new criteria for nodule differentiation. Furthermore, the contrast enhancement remains visible from several minutes to even hours after injection, thus CEIOUS should gain that panoramicity and reduce the need for reinjections. These features make their use extremely promising and begin further scenarios for the application of CEIOUS in patients who undergo surgery for HCC. Ultrasound Liver Anatomy A background of perfect knowledge of the liver anatomy surgically and ultrasono- graphically is needed in order to perform IOUS properly. For surgical anatomy, Brisbane Terminology is considered here [9]. After entering into the abdominal cav- ity, liver mobilization dividing the round and falciform ligaments, and division of eventual adhesions to free the antero-superior and inferior surfaces of the liver are the steps that should precede the liver exploration with IOUS. Of course, adhesions with other organs or structures should not be divided in the event there is the possi- bility that they are expressions of tumor infiltration: in this eventuality, IOUS could be helpful for ruling out or confirming the tumor invasion and then changing the surgical strategy accordingly. By pulling the round ligament, the liver surface is widely exposed and following the portal branches and the hepatic veins, the liver can be studied in its entirety. The probe should be managed using enough pressure to ensure good contact with the liver surface but not to compress the intrahepatic vascular structures and in par- ticular the hepatic vein. The three main hepatic veins are readily identified at their junction with the inferior vena cava (IVC) positioning the probe at this level and tilting it upward once the confluence of the hepatic veins into the IVC is recog- nized. Then gently withdrawing the probe, the hepatic vein paths can be traced into the liver. Hepatic veins appear as echofree zones into the liver parenchyma with the vessel wall which appears as a thin hyperechogenic line (Fig. 10.1): hepatic vein wall thickness can be larger in the cirrhotic liver and its lumen thinner in function of the hard stiffness of the organ. The portal vein branches can be followed first positioning the probe horizontally above the segment 4 inferior to visualize the first-order bifurcation and then first-, second-, and third-order portal branches can be followed with the probe. Because of the existence of the Glisson’s capsule, the portal pedicles, which run together with the arteries and the bile ducts, have thicker vessel walls compared with the hepatic 138 G. Torzilli vein and for this reason they appear at IOUS as echofree zones surrounded by a thicker hyperechogenic layer (Fig. 10.1); furthermore, other parallel thinner vascu- lar structures are visible, namely the arteries and bile ducts of the Glissonian triad. However, in principle, distinction between hepatic veins and portal branches should be based not only on their appearance but mainly on their anatomy: indeed in the cir- rhotic liver, as already mentioned, the vessel wall of the hepatic vein could be thicker and not immediately differentiable from a peripheral portal branch. Following the portal pedicles at the sectional, segmental, and subsegmental levels and positioning it in relation to the hepatic vein it is possible to precisely define the location of the IOUS target in terms of sections and segments. The appearance of bile ducts at IOUS is worthwhile mentioning because of their peculiarity. Indeed, normally they result as thin echofree zones in the Glissonian triad. Once dilated they appear more evidently as echofree zones and with a serpig- inous path pattern. The element that is difficult to recognize in the IOUS study of the bile ducts is their segmental anatomy. Indeed, bifurcation of sectional and seg- mental ducts is closer to the hilum compared with the portal branches and for that it is possible with one scan to visualize more than a segmental bile duct. If this fact is not considered it could be more difficult to address which part of the liver is not well drained. Conversely, if recognized, IOUS could allow the exact definition of the bile duct anatomy both in normal and pathological conditions. Indications The use of IOUS in liver resections can be schematically divided into three princi- pal phases: the liver exploration for the staging of the disease, the planning of the surgical strategy, and the guidance of the surgical maneuvers. Liver Exploration The hard and irregular surface of a cirrhotic liver makes the detection of s mall nodules by palpation difficult; IOUS allows the detection of new lesions in around 30% of cases [10]. However, most of the nodules detected by IOUS in the cirrhotic liver are not really tumors: in this way, IOUS introduces the risk of overestimat- ing the tumor stage. Indeed, except for those nodules with mosaic ultrasonographic pattern (Fig. 10.2a) that are malignant in 84% of cases, only 24–30% of hypoe- chogenic (dark) nodules (Fig. 10.2b), and 0–18% of those hyperechogenic (bright) (Fig. 10.2c) are neoplasm [10, 11]. To overcome this problem even biopsy seems inadequate. The only nodule that can be easily differentiated intraoperatively from a HCC or liver metastases is the small hemangioma which is often discovered primarily at IOUS; it has a typical ultrasonographic pattern, and moreover when compressed changes its size and appearance. Therefore, the problem of the dif- ferentiation of the lesions depicted at IOUS exploration becomes crucial. Further 10 Ultrasound-Guided Liver Resection for Hepatocellular Carcinoma 139 Fig. 10.2 (a) A mosaic pattern lesion at IOUS (arrows); (b) a hypoechogenic lesion at IOUS (arrows); (c) a hyperechogenic lesion a t IOUS (arrows) improvement in differential diagnosis of liver nodules with IOUS may be expected with the introduction and diffusion of the intraoperative use of the last-generation contrast agents. Contrast-Enhanced Intraoperative Ultrasonography More recently the introduction of CE-IOUS has set the rate for modified operative decision making on 30–40% of cases [7, 8]. Tumor vascularity as a criterion for differentiating the regenerative or dysplastic nodules from the HCC correlates well with the histological evidence of a progressive increase in unpaired arteries from dysplastic to neoplastic nodules in a cirrhotic liver [12]. Certainly, the pattern of vascular enhancement is insufficient for differentiating malignant from nonmalig- nant nodules in a cirrhotic liver with 100% specificity. However, CE-US provides differential diagnosis of FLL with a 95% specificity rate [13]; of course, it must be considered that this last rate referred to another type of lesion when compared to the CEIOUS target. Indeed, the intraoperative exploration takes advantage of the higher resolution of the ultrasonography done in direct contact with the liver. Therefore, the need for differentiating nodules detected at IOUS is mostly focused on lesions smaller than 1 cm: for these nodules the vascularity as the criterion for differential diagnosis is less specific. However, some improvements compared with conventional IOUS could be expected. For this reason, in the early 1990s attempts were made to use CE-IOUS with carbon dioxide as the contrast mate- rial for IOUS, however, the need for arterial catheterization made this technique too invasive [14]. In our preliminary experience CE-IOUS provided remarkable findings, either by adding information on nodular vascularity in patients with HCC, or by detect- ing nodules that were not visible at IOUS, in patients with colorectal cancer liver 140 G. Torzilli metastases [8]. Focusing attention on patients operated on for HCC specificity of CEIOUS is around 69% [7]. This value is probably not that high espe- cially when compared with that reported for CE-US [13]. However, as we men- tioned before, the small size of the lesions targeted for CEIOUS study could explain this discrepancy: for these tiny nodules the neovascularity as criterion for differentiation between malignant and benign lesions has limits indepen- dent from the method we use for studying them. Therefore, CEIOUS can be helpful in a certain percentage of nodules but not in all: in this perspective the rate of 69% of specificity is encouraging as it means that we can pro- vide proper information with this new technique in seven out of ten lesions we detect at the time of laparotomy. For the remaining three, even histology may be lacking as we know that there is no common agreement among Western and Eastern pathologists on the definition of early HCC and dysplastic lesions [12, 15]. A new perspective in this sense will certainly be provided by a more exten- sive use of the new contrast agent, at the moment only clinically available in Japan [16]. This agent, having a Kupffer phase adds a further criterion for differentiat- ing those nodules detected at IOUS, and for disclosing others eventually missed at IOUS; in this sense this new contrast agent mimics the contrast agents used with magnetic resonance. In practice, at CEIOUS we can follow in real-time the enhancement of the liver parenchyma with vessels appearing hyperechogenic instead of the echofree pattern at the unenhanced US. Any lesion with a pathological behavior appearing as hypoe- chogenic with or without inner vessels and with or without an arterial phase in which it is enhanced prior to t he remaining liver parenchyma is removed (Fig. 10.3). Those lesions that disappear once the contrast enhances the liver are not considered neoplastic, and thus are not removed. Fig. 10.3 (a) At IOUS a small hyperechogenic nodule is found (arrows) close to the middle hep- atic vein (MHV) and the segment 8 main portal branch (P8); (b) at CE-IOUS, the black hole (arrows) of that nodule in the late phase is evident showing a pathological pattern 10 Ultrasound-Guided Liver Resection for Hepatocellular Carcinoma 141 Planning of the Surgical Strategy IOUS exploration of the liver could have a great impact on the surgical strategy; however, more recently the impact of IOUS on operative decision making, when compared with those of preoperative imaging techniques, is reported to be just around 4–7% [17, 18]. The problem of the impact of IOUS on the operative decision making depends on two main factors: the surgical policy of each specific team and the type of tumor. Indeed, the relatively low rates reported [17, 18] are also partially motivated by the surgeon’s surgical policy: in fact, because a considerable num- ber of patients undergo major hepatectomies, new nodules detected by IOUS in the same hemiliver would not have modified the surgical strategy. Recently, it has been shown how major hepatectomies are carried out in the minority of patients [5, 6] just because of the extensive use of the IOUS guidance for achieving parenchymal- sparing resections, so that detection of new nodules is more suitable for changing the surgical strategy. IOUS allows an accurate three-dimensional reconstruction of the relationship among the tumor, the portal branches, and hepatic veins; this is a fundamental step in the definition of the proper surgical strategy. Indeed, surgical decision making should be obtained having portal branches and hepatic veins as landmarks to reduce the risk of major morbidity and mortality. Definition of the tumor–vessels relationship is relevant for planning the type of resection, and based on that, specific and original operations can be performed [5, 6]. IOUS easily allows the surgeon to recognize if an HCC is separated by some normal parenchyma from the vessel, if it is in contact with the vessel without invading its wall (Fig. 10.4), or conversely if the HCC is invading the vessel wall, is determining the proximal bile duct dilation, or if it is associated with a tumor thrombus. Extension of the hepatectomy is always considered for the parenchyma fed by infiltrated portal branch at IOUS [5]: vein Glissonian triad invasion is considered in the presence of portal or biliary tumor thrombus, in the absence of vessel wall visu- alization, or in the case of HCC in contact with the Glissonian triad with proximal bile duct dilation. Inversely, in the case of infiltration of a hepatic vein, an extension of the resection to the whole liver parenchyma theoretically drained by this vein is considered only if one of the following ultrasonographic signs is missing: • Presence of accessory hepatic veins at IOUS • Color-Doppler IOUS showing hepatopetal blood flow in the feeding portal branch once the hepatic vein is clamped [7] by means of encirclement more simply by vein compression at its extrahepatic route using the fingertip as described later • Communicating veins connecting adjacent hepatic veins (Fig. 10.5) Adopting these criteria we have been able to minimize the rate of major hep- atectomy and to devise new procedures. One of them is the Systematic Extended Right Posterior Sectionectomy (SERPS) as an alternative to right hemihepatectomy [19]. Another new procedure that can be accomplished adopting the aforementioned criteria is the so-called minimesohepatectomy [20]. This last represents an alterna- tive to the conventional mesohepatectomy in the case of tumors invading the middle 142 G. Torzilli Fig. 10.4 This patient was a carrier of a HCC in contact (arrows) with the right hepatic vein (RHV), which maintains intact its wall represented by an hyperechogenic layer. P6–7 = portal branch to segments 6 and 7 Fig. 10.5 Arrows indicate a thin communicating vein connecting the right hepatic vein (RHV) (a) with the middle hepatic vein (MHV) (b). This vein has been disclosed compressing the RHV (a): the latter has no flow (color) inside whereas flow directed from RHV to MHV (red color means upward, towards the probe, direction of the flow) is shown in the communicating vein (a,b), and flow directed toward the inferior vena cava (blue color means downward, opposite to the probe, direction of the flow) is shown inside the MHV (b). P8 = portal branch to segment 8 10 Ultrasound-Guided Liver Resection for Hepatocellular Carcinoma 143 hepatic vein at its caval confluence, and consists in a limited resection including the tract of the invaded vein without its reconstruction. Resection Guidance Systematic Segmentectomy In a cirrhotic patient, the liver volume to be resected must be determined with partic- ular care with the purpose of associating surgical radicality and noncancerous liver parenchyma sparing. Liver function tests and liver volumetry on CT scans help in this decision. Tumor dissemination from the main lesion through the portal branches cannot be detected with certainty by the pre- and intraoperative imaging modalities [21]. Consequently, some authors consider that the resected specimen should com- prise at least the portal area, which includes the lesion [21]. This last is impossible to be correctly identified without the aid of IOUS, especially in a cirrhotic liver where there are generally wide variations and abnormalities in the distributions of the portal branches. For this purpose systematic segmentectomy was devised in the early 1980s [4], and we have recent alternatives to this approach. Compression of the Portal Branch Initially used for tumors located in the left hemiliver [22], recently we have success- fully extended the application of this technique to any segmental location [23], and even to a sectional portion of the liver [24]. Once the feeding portal branch has been identified at IOUS, it i s compressed using the IOUS probe at one side of the liver and the finger at the opposite side (Fig. 10.6a–c): in this way it is possible to induce a transient ischemia of the portion of the liver distal to the compression site. This portion can be marked with the electrocautery, the compression released, and the resection carried out. This technique is simple, fast, noninvasive, and reversible: the possibility to modify the site of compression and then the resection volume allows us Fig. 10.6 (a) The portal branch to segment 7 (P7) is visualized at IOUS on the left and the sur- geon’s finger is positioned (F), and P7 is compressed on the right (arrow); (b) the hepatic ischemic area generated by compression with the surgeon’s finger and probe (P), which corresponds to the area to be resected is well evident on the liver surface (arrows); (c) the cut surface at the end of the segment 7 segmentectomy 144 G. Torzilli to size the resection in the function of the tumor f eatures and the status of the back- ground liver. Furthermore, the compression can be used in a countercompression perspective [23, 24], borrowing the philosophy proposed by Takayama et al. of defining the adjacent segmental margins to disclose those of the targeted segment which was applied by the author for segment 1 resection using the portal branch puncture technique [25]. Indeed for segments such as 8 and 4 superior for which the direct compression of the feeding portal branch could be infeasible, compressing the adjacent segmental branch allows the definition of their segmental margins [23]. Similarly, this technique can be applied to disclose the margin of the right anterior section of the liver by compressing the Glissonian pedicles to the right posterior section and to the left hemiliver, respectively [24]. For this last it avoids the need of portal pedicle dissection or blunt encirclement resulting in a simpler, safer, and equally anatomical procedure [26–28]. Hooking of the Portal Branch The segmental portal branches to segment 4 are generally divided in two groups, those for the superior and those for the inferior portion, but the most common branching pattern can be recognized in just half the cases [29]. These branches rather than being punctured, under IOUS guidance can be approached dissecting the umbilical portion: once exposed the vessel can be encircled with a suture and pulled under IOUS control to verify if it is the branch to segment 4 inferior or not. Then the proper portal branch can be ligated and divided and the discolored area that will appear on the liver surface should correspond to the segment 4 inferior which can be marked with the electrocautery to proceed with the liver dissection; this is a peculiar application of the so-called hooking technique [30]. Furthermore, the subsegment 4 superior could be resected just clamping the portal branch to the subsegment 4 inferior, as it is identified with the just-described hooking technique; the discol- ored subsegment 4 inferior caudally, the plan at IOUS which includes the middle hepatic vein laterally, and that marked by the falciform ligament medially, delimit the area to be resected. Limited Resection The anatomical versus nonanatomical surgical approach for HCC is still a controver- sial issue [21, 31–36], inasmuch as there are no really randomized studies comparing the two different operations. However, recent reports seem to confirm adequacy in terms of oncological radicality of limited resections for HCC once IOUS is exten- sively used [5, 6, 31, 36]. The IOUS-guided limited resection is simpler than the systematic segmentectomy because there is no need for identifying the area of the liver fed by the portal branch to be ligated, although nowadays, with the compres- sion technique, there could be no more need for puncturing the portal branches or dissecting the pedicles. Indeed, for limited resection, once the tumor is identified, the surgeon under IOUS control can mark with the electrocautery the border of the lesion and that of the area to be removed on the surface of the liver. To carry out . veins as landmarks to reduce the risk of major morbidity and mortality. Definition of the tumor–vessels relationship is relevant for planning the type of resection, and based on that, specific and original. volume, its stability, and the wide- ness of the ultrasonographic scanning window: the best probe should be small, thin in width, and short in transverse length, stable, and with a wide ultrasonographic scanning. the contact area between the probe and the liver and a scanning window which enlarges as it gets deeper. It is also evident, how at IOUS, the portal vein (PV) and the hepatic vein (HV) have different