The anniversary of the first publication of Cancer provides an opportunity to review the progress in liver cancer surgery in the past 60 years. Indeed, the past half century has been witness to remarkable advances in liver surgery that have provided benefit to so many patients afflicted with primary or secondary liver tumors. However, no matter now astounding these past achievements may be, it is clear that the future holds even greater promise. Cancer 2008;113(7 suppl):1888–96. © 2008 American Cancer Society.
At the time that the inaugural edition of Cancer was published in 1948, liver surgery was in its infancy. One may go back centuries to identify critical developments in the history of medicine that culminated in the birth of this early era of liver surgery. More recent developments in the late 19th century included the introduction of ether anesthesia by the dentist William Morton and the first demonstration of its use by Henry Jacob Bigelow and John Collins Warren at the Massachusetts General Hospital in 1846. In 1867, Joseph Lister brought to surgery the principles of antisepsis, which were based on Pasteur's theory that bacteria cause infection. And, on the heels of these advances, surgeons began to address more frequently liver injuries suffered in battle with surgical intervention. The first resections described involved excision of protruding liver resulting from knife wounds. In 1870, during the Franco-Prussian War, Bruns excised a section of liver injured by a gunshot wound, and other similar operations were summarized by Dalton in 1888.1 At the time, the gravity of these injuries was extreme. In 1887, Elder published a series of 543 patients with ‘subcutaneous’ (blunt) or open liver injuries, all managed without surgical intervention, and reported mortality rates of 78% for subcutaneous injuries and 58% for open injuries.2
Early attempts at liver resection were ambitious given the limited understanding of liver anatomy, physiology, hepatic surgery techniques, and perioperative management. Suture control of liver cut surface was described by Posternski in 1885,3 and that report was followed by Keen's description of ligation of individual bleeding vessels.4 Schroeder and Beck described through-and-through mattress sutures with guards to prevent tearing of the delicate capsule.5, 6 These techniques notwithstanding, hemorrhage remained the formidable challenge associated with liver surgery of the time.
Which surgeon performed the first elective liver resection is a matter of some controversy, but the accomplishment often is attributed to A. Lius in 1886. He removed a left lobe adenoma in a woman aged 76 years. The woman died postoperatively from hemorrhage.7 Thus, the first successful elective liver resection has been attributed to Carl von Langenbuch, who removed a left lobe tumor in 1888.8 Liver anatomy at that time was poorly understood, which lead to continued reports of left hepatic lobectomy for removal of what is defined today as Couinaud segments 2 and 3.4
Liver surgery involved significant hemorrhage. Medical blood transfusion already had been introduced by James Blundell, a British obstetrician, for postpartum hemorrhage in 1818. In the early 1900s, the discovery that addition of an anticoagulant permitted blood storage for many days gave rise to early banks in time for World War I. During the second world war, Charles Drew's discovery that blood could be separated into blood plasma and erythrocytes permitted the storage and transfusion of separate components and gave rise to modern blood banks. Liver surgery as a field grew on the backs of blood transfusion services.
The first descriptions of a segmental anatomy of the liver were written by Francis Glisson in 1654 but were forgotten for centuries. Perhaps the inability to conduct liver surgery safely during those years rendered this information somewhat superfluous. However, with the birth of elective liver surgery in the late 19th century, an understanding of liver anatomy became more relevant. In 1898, Cantlie described that the right and left lobes are not separated by the falciform ligament.9 Others noted that the right lobe is comprised of anterior and posterior lobes and that the left lobe is comprised of lateral and medial lobes.10 This refined understanding of anatomy ushered in a new era in which specific planes were described that defined suitable surgical approaches to a right lobectomy (hepatectomy), left lobectomy (hepatectomy), and left lateral segmentectomy.11, 12 Lortat-Jacob and Robert reported the first procedure involving extrahepatic ligation of right portal structures and the right hepatic vein preceding parenchymal transection.13 And, with a clear description of the functional anatomic delineation of liver segments, Claude Couinaud set the stage for the next half century of rapid advances in liver surgery.14
The safety and efficacy of liver surgery has benefited immensely from advances in imaging. For centuries, palpation was the primary method for the detection of liver masses until the advent of medical ultrasound in the 1940s. With experience using ultrasound for oceanic mapping, the Naval Medical Research Institute was the first to apply this technology to the human body. Ultrasound represented a significant advance over existing, limited methods for tumor detection. Ultrasound also enabled accurate, percutaneous needle biopsy of liver lesions. However, the poor resolution of these early systems left much to be desired. The computed tomography (CT) scanner was devised and introduced through the separate research efforts of Sir Godfrey Hounsefield in the United Kingdom and Allan Cormack in the United States. They shared the Nobel Prize in Medicine and Physiology in 1979 for their work. And, as it pertained to liver surgery, hepatic lesions could now be detected much earlier, thereby slowly reducing the frequency of operations on patients with massive tumors. Studies of the blood supply to metastases revealed that, whereas the normal liver receives much of its blood supply from the portal vein, liver metastases derive their blood supply from the hepatic artery. This observation lead to the development of CT angioportography, a technique combining visceral arteriography with dynamic CT imaging. When the liver is imaged by CT with contrast in the portal venous phase, the different primary blood supply of healthy liver and metastases reduces contrast enhancement of metastases, making them more clearly apparent on scans. CT angioportography became the study of choice in the 1980s and 1990s for the detection of liver metastases.
Continued improvements in CT scanner technology, including faster helical scanning and the development of multihead detectors, were critically important in maintaining the progress of liver surgery. These newer scanners produce high-quality thin-section images in a short period. High-quality CT images obtained during arterial contrast enhancement have aided in the detection of small hepatocellular carcinomas. Portal venous phase imaging after intravenous contrast bolus replaced cumbersome CT angioportography. These scanning techniques also laid the foundation for 3-dimensional rendering of livers with measurement of liver volumes. These tools allow for accurate prediction of future liver remnant volumes to assess the likelihood of postoperative liver insufficiency.15 Radiographic liver volume measurements aided in the development of living-donor liver transplantation. In the 1980s and 1990s, implantable hepatic arterial infusion pumps were in vogue and required visceral angiography to define hepatic arterial anatomy. However, CT arteriography gradually replaced visceral angiography as the reliable study for this purpose.
With the advent of CT, transcorporal ultrasound soon was relegated to a secondary role in preoperative detection and characterization of liver tumors. However, Makuuchi et al demonstrated the value of intraoperative ultrasound in the early 1980s.16 Whereas low-frequency energy was required to achieve penetration through the abdominal wall during transcutaneous ultrasound, placement of the probe directly on the liver during surgery permitted the use of higher frequency transducers with much better resolution. Laparoscopic versions of transducers were developed to allow the performance liver ultrasound without the need for laparotomy. Despite advances in CT and magnetic resonance imaging (MRI) studies, intraoperative ultrasound remains an important tool for liver surgeons today.
Similar to the invention of CT scanning, MRI was the product of independent efforts spanning the Atlantic. Paul Lauterbur in Stonybrook, New York and Peter Mansfield of Nottingham, England shared the Nobel Prize in Medicine and Physiology for their discoveries concerning MRI. The first unit was constructed in 1977; and, since then, MRI has played a significant role in liver surgery. Its use in lesion detection and characterization nicely complemented that of CT. MRI demonstrated its utility in the diagnosis of benign simple cysts, hemangiomas, and focal nodular hyperplasia. Consequently, with greater use of MRI, it became less common to remove liver lesions without a correct preoperative diagnoses. More recently, MRI has found a role in assessing the extent of hepatic steatosis (eg, in response to chemotherapy or obesity). The benefits offered by MRI in surgical planning also are demonstrated by its ability to assess the extent of tumor involvement of bile ducts noninvasively and accurately with magnetic resonance cholangiopancreatography.
Gordon Brownell at Massachusetts Institute of Technology developed the first device for the detection of positron emission in 1953, but its more widespread use in oncology awaited the introduction of a scanner for humans and the development of (18F)fluorodeoxyglucose in the 1970s.17 The primary role in positron emission tomography (PET) in liver surgery has been to identify occult metastases before liver resection. Today, most patients undergo PET imaging as part of their staging evaluation before metastasectomy, leading to better patient selection and improved long-term survival statistics after liver resection.18
The history of liver surgery is a history of hemorrhage control. In the early 16th century,19 Ambroise Pare noted that, when the liver is wounded, much blood commeth out. Consequently, much effort has focused on the development of liver transection techniques that defy this colorful description. Liver surgeons have experimented with various devices for liver transection, methods to reduce hepatic blood flow, liver compression clamps, and, of course, blood replacement.
Devices for liver transection
Given the liver's propensity to bleed when injured, surgeons always have been enamored with devices, strategies, and tactics to ‘bloodlessly’ dissect through the liver. Finger fracture was described first by Keen at the turn of the 20th century4 but perhaps perfected in its implementation by Couinaud.20, 21 Dissection through the liver with the handle of a scalpel was advocated by Quattlebaum.22 The first electrocautery device, the surgical Bovie knife, was developed commercially in 1928 by 2 Americans, Cushing and Bovie; and, to this day, electrocautery serves as the primary means of achieving hemostasis in the liver. The use of ultrasound energy for dissection was introduced in 1984 by Hodgson and Delguerico as a way to disrupt hepatic parenchyma without injury to vascular structures.23 Liver dissection by pressurized, focused water jet also was introduced in the 1980s24 and remains popular in Europe. Lasers and microwaves also were introduced around the same time25, 26 but have not gained favor. Cryosurgery was described first in dog model as a method for achieving hemostasis.27 More recently, radiofrequency-conducting devices have been used in a strategy to precoagulate tissue along a transection plane before cutting through the liver.28 Differences in opinion and clinical trials comparing the merits of each of these devices notwithstanding,29 they all appear to be functional and helpful by facilitating control of blood vessels before cutting through them.
Control and modulation of hepatic blood flow
Clamps have been devised that compress the liver to achieve hemostasis. Stucke is credited with the first description of a liver clamp in 1961,30 followed by Storm and Longmire in 1971.31 However, surgeons are most familiar with the clamp devised by Lin in 1973.32 These clamps resemble a pair of windshield wipers opposing each other with a scissor-like handle used to apply pressure and lock the clamp in place. These clamps are relatively cumbersome to apply; Li and Mok subsequently described use of a circumferential tourniquet to achieve compression.33
Techniques to control hepatic vasculature to reduce blood loss played an important role in the progress of liver surgery. James Hogarth Pringle at the Royal Glasgow Infirmary described use of intermittent portal occlusion (between his finger and thumb) for the management of bleeding secondary to trauma.34 This maneuver—which now bears his name—remains in wide use today. However, the past 60 years have been witness to additional concepts and refinements in vascular control. With an objective of allowing liver surgeons to operate in a ‘bloodless’ field, a technique of total vascular isolation was reported in 1966 by Heaney et al and involved simultaneous clamping of the cava and the supraceliac aorta.35 A variation of this original technique without supraceliac occlusion was described subsequently by Bismuth et al.36 Schrock et al described use of a caval-atrial shunt consisting of a long catheter inserted through the atrial appendage into the inferior vena cava, with occluding tapes placed around the suprarenal and suprahepatic cava.37 Although simple in concept, these maneuvers all have been associated with significant morbidity. Even in elective liver surgery, the use of vascular clamps on the vena cava and porta hepatis to achieve hepatic vascular exclusion often is associated with hemodyanimc instability. Veno-veno bypass subsequently was introduced to eliminate this hemodynamic instability and simultaneously facilitate repair and/or replacement of the retrohepatic inferior vena cava.38, 39
Surgical approaches were refined further to provide selective vascular control to a single hepatic lobe to minimize contralateral hepatic warm ischemia time.40, 41 Surgical control of hepatic pedicle structures contained within the Glissonian sheaths before liver transection was championed by Launois and Jameison.42 McEntee and Nagorney43 and Voyles and Vogel44 first reported the use of stapling devices during liver surgery for transecting a portal vein and a hepatic vein, respectively. Use of these stapling devices has now become common in liver surgery.
During the era of liver resections associated with massive blood loss, surgeons and anesthesiologists had become accustomed to a practice of maintaining high central venous pressures for liver surgery to keep ahead of the anticipated blood losses. However, with portal pedical occlusion, the main source of potential hemorrhage is the hepatic veins and their tributaries. Hence, reduction of the central venous pressure was explored as a tactic to reduce intraoperative blood loss and was demonstrated to be both safe and effective.45 The use of low central venous pressure during liver surgery, combined with hemodilution before liver transection by blood removal (and reinfusion of the blood back into the patient after liver transection), also was devised as a strategy to reduce the need for banked blood transfusions during liver surgery.46
Liver cooling is another technique that was developed as a strategy to simultaneously reduce blood loss and improve the liver's tolerance for ischemia. The technique of hypothermia originally was introduced by Longmire and Marable for liver surgery involved cooling of the entire body.47 This cumbersome strategy was improved upon by Fortner and colleagues, who, in the early 1970s, reported a technique for in situ hypothermic liver perfusion with cold Ringer solution for major liver resection.48, 49 Further refinements to this technique included the use of protective agents (eg, prostaglandin E1)50 and in situ liver perfusion with cold University of Wisconsin solution.51
With improvements in liver surgery instruments, imaging, and tactics, liver surgeons have continued to push the frontiers, devising more complex approaches to improve safety, reduce morbidity, and expand the criteria for resectability.
Ex vivo resection
Surgical techniques used for liver transplantation have been applied to partial hepatecomies. One example involves the technique of removing the entire liver from the patient, resection of the hepatic tumors with the liver positioned on the back bench (ex vivo), followed by replacement of the liver into its original anatomic position. Although this operation is technically complex, the approach provides optimal exposure, minimal blood loss, and excellent visibility in a bloodless field. Pichlmayr et al described ex vivo tumor resections in a series that included 3 patients with in situ hypothermic perfusion.52 This initial experience was met with a 33% operative mortality rate. Subsequently, modifications, such as ‘ex situ-in vivo’ liver surgery, have been described.53 This approach preserves an intact inflow hepatic pedicle (‘in vivo’) after complete mobilization of the liver and inferior vena cava, inferior mesenteric and femoral to axillary vein bypass, complete vascular exclusion of the liver, and cold hepatic perfusion. The hepatic veins are then divided for exteriorization of the liver (‘ex situ’).
Staged, sequential hepatectomy
Patients with bilateral extensive tumor distribution were categorized historically as ‘unresectable’ on the basis of both poor tumor biology and the complexity and morbidity of the necessary operation. Honjo and Kozaka described what appears to be the first planned set of sequential liver resections. These operations were performed on a patient with gastric cancer and extensive liver metastases. The initial operation consisted of a gastric resection, partial left hepatectomy, and right portal vein ligation.54 The right lobe was resected 3 weeks later. Although this initial experience was met with failure because the patient died from operative complications, it demonstrated the feasibility and potential utility of 2-stage resections. Bengmark et al subsequently performed a 2-staged hepatectomy for metastatic colon carcinoma.55 Adam et al used this surgical strategy for patients with colorectal carcinoma in whom removal of all liver metastases in a single operation did not leave behind sufficient residual liver function to avoid complications or death.56 Subsequent standardization of techniques for the prediction of liver remnant volumes have helped with the assessment of whether a patient can undergo a single complex resection versus staged, sequential hepatectomies.15
Preoperative induction of liver hypertrophy
On the basis of experimental studies, Tillmans reported in the 19th century the propensity for the liver to regenerate after major resection. In the 1920s, Rous and Larimore demonstrated in rabbits that ligation of a portal vein branch produces compensatory hypertrophy.58 Over the ensuing years, compensatory liver hypertrophy was considered a beneficial consequence of liver resection, tumor replacement of a liver lobe, or tumor invasion into portal vasculature. Makuuchi et al were the first to report a strategy of deliberate portal occlusion to induce atrophy and contralateral hypertrophy to allow for subsequent major liver resection.59, 60 They demonstrated that these techniques reduce the likelihood of postoperative liver insufficiency, provide for greater safety of extensive resections, and allow for tumor extirpations in patients who otherwise are considered unresectable because of inadequate remnant liver volume. Both portal vein ligation and portal vein embolization have been described as strategies for induction of hypertrophy of the future remnant liver.
Laparoscopic liver resections
The rapid advances of laparoscopic surgery in the 1980s were certain to influence the field of liver surgery as much as they influenced other surgical disciplines. The initial therapeutic laparoscopic liver operations were simple procedures involving fenestration of symptomatic liver cysts. The first laparoscopic resection is attributed to Gagne et al and was performed in 1992 for a focal nodular hyperplasia.61 In 1995, Ferzli et al laparoscopically resected a 9-cm adenoma.62 The first laparoscopic anatomic liver resection, a left lateral sectionectomy, was reported by Azagra et al in 1996.63 Similar to what is observed with other laparoscopic operations, the primary benefit of laparoscopic hepatectomy is the reductions in hospital stay and postoperative analgesic requirements. Although, today, the field of laparoscopic liver resection is in its infancy, it is likely only a matter of years before a majority of liver resections are performed laparoscopically.
Tumor involvement of the retrohepatic cava was long considered a contraindication for liver resection for fear of either uncontrollable hemorrhage or air embolism. It is not surprising that the first attempt at liver resection combined with caval replacement was performed by an experienced liver transplantation surgery team.64 Veno-veno bypass served as an important strategy to reduce both hemodynamic instability and splanchnic congestion associated with prolonged portal clamping.38, 39 And the additional measure of inducing liver hypothermia increases the liver's tolerance for prolonged periods of hepatic ischemia. Various materials have been used for caval substitution and repair, including synthetic materials (eg, polytetrafluoroethylene), allograft, autograft (eg, saphenous vein), and xenograft. And, with greater experience in techniques for liver transplantation and vascular surgery, hepatic venous reconstructions also became more common. These operations use techniques similar to those used for combined liver and caval resections and allow for resections when tumor involves all of the hepatic veins. Resection of the portal vein also was reported, followed by repair using a patch graft harvested from a hepatic vein in the resected liver, portal vein from the resected liver, saphenous vein, prosthetic material, or umbilical vein. These are particularly useful techniques in patients with hilar cholangiocarcinomas.65, 66
With the reductions in morbidity and mortality observed for liver resection in recent decades,67, 68 it is not surprising that indications for liver resection continue to evolve. In the 1970s and 1980s, patients were selected for resection primarily on the basis of whether their tumor technically was resectable, with little regard for tumor biology. During that era, Hughes and colleagues created a registry of patients with secondary liver tumors who had undergone resection, and they performed a retrospective analysis of factors to determine which were correlated with prognosis. They reported on a set of relative contraindications to liver resection, including noncolorectal histology, >3 metastases, extrahepatic metastases, and distribution into both liver lobes.69 However, analyses performed on more contemporary datasets suggested the absence of an absolute number or distribution of metastases, histology, or size that preclude the potential for long-term survival after successful resection,70, 71 and Adam et al convincingly demonstrated that patients with unresectable metastases could be converted to resectable status by chemotherapy.72 Those authors also demonstrated that an objective response to neoadjuvant chemotherapy provided prognostic information regarding the likelihood of benefit after resection.73 Surgeons pushed the frontiers further by more often combining resection of pulmonary and hepatic colorectal carcinoma metastases.74 Perhaps the greater use of PET emboldened such treatment recommendations. Reported benefits of PET for the detection of radiographically occult extrahepatic metastases aided in the selection of patients appropriate for liver resection.18 In addition, the aging of the overall population has not had a significant influence on the selection of patients for liver surgery, because it is recognized liver resections can be performed in older patients with acceptable safety.75
Similar advances in understanding which patients may experience long-term survival after resection of hepatocellular carcinoma have been noteworthy. Resection of hepatocellular carcinomas that are large,76 associated with hepatitis B or C virus,77 ruptured,78 or involve portal vein thrombus79 have been associated with benefit in selected patient subsets. With the evolution in indications for liver resections in both primary and secondary liver tumors, it is not surprising that, today, liver resection is performed more frequently than in prior decades.68
Radiofrequency ablation (RFA) was applied first to patients for the obliteration of pathologic myocardial conduction pathways. Subsequently, Rosenthal and colleagues used RFA to treat painful osteomas.80 Following on the heels of its first use for liver tumors in Italy, the current article's author, together with radiologist Nahum Goldberg, performed the first RFA of a liver tumor in the United States in 1996.81 This form of tumor ablation essentially has replaced alcohol injection into hepatocellular carcinoma and currently is used extensively for liver metastases.82 The evolution of RFA in primary and secondary liver tumors followed different paths. In patients with liver metastases from colorectal carcinoma, studies indicated that RFA was inferior to liver resection.83 Its use in colorectal carcinoma liver metastases has remained limited primarily to patients with unresectable liver metastases.84 However, prospective randomized trial data suggest equivalence of RFA to surgery for the management of hepatocellular carcinoma in patients with cirrhosis,85 and ablation remains a primary treatment option in these patients. Similarly, RFA has found useful application in patients awaiting liver transplantation, serving as a ‘bridge to transplantation.’
At the time the journal Cancer was launched in 1948, liver transplantation was a concept under active investigation that had yet to be applied to humans. Alexis Carrel is considered by many the founding father of experimental organ transplantation because of his pioneering work with vascular techniques. A half century later, in 1955, Welch published the first description of a technique for liver transplantation in animals,86 and the first experimental animal liver transplantation subsequently was performed by Moore and colleagues in 1959.87 That work set the stage for the first human liver transplantation in 1963.88 Unfortunately, the child aged 3 years with biliary atresia who received this first liver transplantation survived for only hours after receiving the graft. However, that transplantation team subsequently led the world in pioneering advances in liver transplantation. And the introduction of cyclosporine led to rapid and dramatic advances in liver transplantation in the ensuing decades.89 Americans published on numerous advances in liver transplantation the 1970s and 1980s. But it was in France that the first reduced-size liver transplantation was performed by Bismuth and Houssin,90 and the first split-liver transplantation was performed in Germany by Pichlmayr et al.91
Early experiences with liver transplantation for hepatocellular carcinoma were met with poor long-term survival results.92 These results reflected the treatment strategy at that time, when liver transplantation was offered for unresectable cancer irregardless of tumor stage. The Organ Procurement and Transplantation Network and the United Network for Organ Sharing in the United States subsequently established strict criteria for the allocation of cadaveric liver grafts to patients with liver cancer to reflect a better balance of the urgency in need for liver cancer patients and the desire to have long-term outcomes more closely approximate those observed after transplantation in patients with end-stage liver diseases without cancer. The basis for current transplantation criteria was published by Mazzaferra et al.93
Raia et al are credited with the first publication on an attempt at living-donor transplantation94; however, it was Strong and colleagues who reported the first successful liver donor transplantation—from mother to son.95 These important advances broadened the potential donor pool. But the use of living donor grafts in particular raised numerous safety and ethical issues.96 Belzer and Southard developed a preservation solution for improved cadaveric graft preservation that also enhanced the feasibility of transporting grafts from the proximity of their harvest to recipients located far away.97 However, with a growing population of patients with end-stage liver disease seeking livers, liver graft shortages are expected to continue for decades.
For this reason, liver xenotransplantation remains an area of active investigation. Several investigators have focused on miniature swine as donors for liver xenografts, because they have many advantages over nonhuman primates for this purpose.98 The risk of cross-species disease transmission is less with swine than with nonhuman primates because of their increased phylogenetic distance from humans. In addition, swine organs are comparable in size anatomically to humans, and these animals are readily available. Transgenic swine and gene therapy have been used in attempts to avoid major histocompatibility antigenic reactivity.99 Whether successes in this field lead to a ‘bridge to transplantation’ strategy or a real long-term solution remains to be seen.
The anniversary of the first publication of Cancer provides an opportunity to review the progress in liver cancer surgery in the past 60 years. Indeed, the past half century has been witness to remarkable advances in liver surgery that have provided benefit to many patients afflicted with primary or secondary liver tumors. However, no matter how astounding these past achievements may be, it is clear that the future holds even greater promise.