Hepatectomy for Hepatocellular Carcinoma: Title and subTitle BreakThe Surgeon's Role in Long-term Survival FREE

HEPATECTOMY remains the treatment of choice for hepatocellular carcinoma (HCC), especially in situations where liver transplantation is not readily available. The 5-year survival rate of patients undergoing hepatectomy for HCC has improved from 12% to 20% in earlier years^{1 - 2} to 35% to 50%^{3 - 7} in recent years. Numerous studies have been performed to identify factors that affect survival. Almost all of these studies analyzed the prognostic values of clinical,⁴ pathological,^{3 ,8} and biological factors,^{9 - 10} either singly or in combination,¹¹ but the role of the surgeon in improving the survival rate after hepatectomy for HCC is unknown. This study was performed to identify the role of the surgeon in the long-term survival of patients undergoing hepatectomy for HCC.

From April 1989 to December 1995, 211 consecutive patients with HCC confirmed by histological examination underwent hepatectomies in the Department of Surgery at Queen Mary Hospital, Hong Kong, China. All data were collected prospectively by a single research assistant.

Preoperative evaluation of the tumor was done by ultrasonography, computed tomography, and hepatic arteriography. Hepatic function assessment, including the modification by Pugh et al¹² of the Child hepatic function classification system and the indocyanine green (ICG) clearance test,¹³ were performed after the HCC was considered resectable as determined by imaging studies. The ICG clearance test results are expressed as the percentage of ICG retained in the circulation 15 minutes after intravenous injection of 0.5 g/kg of body weight of ICG.

Surgery was performed through a bilateral subcostal incision with an upward midline extension. Thoracic extension was required in 11 patients, because the HCC in the right lobe was large or the tumor was in the right posterior segment.

Intraoperative ultrasonography was performed routinely to detect tumor nodules in the contralateral lobe and/or invasion of the tumor into the major blood vessels and to mark the line of parenchymal transection on the liver surface to obtain the optimal tumor-free resection margin and to avoid a major hepatic vein at the transection plane.

Ipsilateral hilar dissection was performed to control or divide the inflow vasculature. The liver lobe bearing the tumor was then mobilized to control the outflow vasculature. Rotation of the right liver lobe containing the tumor was intermittent to avoid prolonged twisting of the inflow and outflow vasculature of the left lobe.

Liver transection was performed by the finger-fracture technique from 1989 to 1992 and by the ultrasonic dissector after 1992. The Pringle maneuver was employed in 20-minute cycles during parenchymal transection in 111 patients. In the other patients, the Pringle maneuver was not applied. Hemostasis was performed promptly by diathermy coagulation, argon beam coagulation, or fine suturing during parenchymal transection to decrease the volume of blood lost.

After parenchymal transection, bile leakage was identified by injecting methylene blue into the common bile duct via the cystic duct or by pressing the transection surface gently with a piece of clean gauze. A leak, if present, was closed with 5-0 sutures. The falciform ligament was reconstructed for right hepatectomy or lobectomy to prevent the left lobe from rotating into the right subphrenic space, which would produce torsion of the left hepatic vein.¹⁴ The transected liver surface was covered, as far as possible, with the greater omentum. A closed vacuum suction drain was inserted into the subphrenic space.

During the operation, blood transfusion was initiated by the anesthetist when the hemoglobin concentration fell below 80 g/L, or in elderly patients when the electrocardiogram showed features of myocardial ischemia.

Postoperatively, patients, particularly those with cirrhosis, were monitored in the intensive care unit with attention to oxygenation, tissue perfusion, and fluid balance. Parenteral nutrition consisting of branched-chain amino acid–enriched solution, low-dose dextrose, medium-chain triglycerides, and phosphates was given to the cirrhotic patients immediately after the hepatectomy.¹⁵ To avoid fluid overload, elevation of central venous pressure, and congestion of the liver remnant, no other intravenous fluid was given once parenteral nutrition was started. All patients received intravenous albumin (50 g/L per day) for 3 to 5 days. Oral feeding was encouraged once bowel activity returned.

All patients underwent monthly follow-up examinations during the first year. Serum α-fetoprotein was measured, and ultrasonography of the liver was performed regularly to detect recurrence. Hepatic arteriography with or without lipiodol computed tomography was performed to confirm a diagnosis. A diagnosis of intrahepatic recurrence was based on the combination of typical radiographic appearance of HCC on 2 imaging studies with or without elevated serum α-fetoprotein levels. Unless liver function was not satisfactory, patients with intrahepatic recurrence were treated aggressively by re-resection, transarterial oily chemoembolization, or intralesional alcohol injection, depending on the location of the tumor. The patients with pulmonary metastasis received pulmonary resection if the metastasis was solitary or epirubicin if multiple.

Hospital mortality was defined as death during the period of hospitalization for the hepatectomy. Major hepatectomy was defined as resection of 3 or more liver segments according to the Couinaud nomenclature.¹⁶ Minor hepatectomy was defined as resection of 2 or fewer liver segments. In terms of anatomical region, the classification of hepatectomy was also made according to the Couinaud nomenclature.

All continuous variables were expressed as median (range) and compared between groups by the Mann-Whitney test. A χ² test was used to compare discrete variables. The overall cumulative survival rate, including hospital mortality and disease-free survival rate, were estimated by the Kaplan-Meier survival method. Disease-free survival time was calculated from the date of hepatectomy to the date when recurrence was diagnosed. Statistical comparison of survival distributions was analyzed by log-rank tests with respect to clinical, operative, and pathological factors. The clinical factors included age, sex, hepatitis B status, the Child hepatic function classification, preoperative ICG retention value at 15 minutes, liver and renal biochemistry, platelet count, prothrombin time, and use of perioperative nutritional support. Operative factors included type of hepatectomy, operation time, blood loss volume, blood and fresh frozen plasma replacement, findings of preexisting rupture of the HCC and/or iatrogenic rupture of HCC on mobilization, and the application of the Pringle maneuver or ultrasonic dissector. Preexisting rupture of the HCC was diagnosed by the presence of blood in the peritoneal cavity or blood clots covering the tumor. The pathological parameters included tumor size, number of tumor nodules, tumor capsule, microscopic venous permeation, macroscopic portal vein involvement, shortest resection margin, status of the uninvolved liver, histological involvement of the resection margin, and TNM stage.¹⁷ Multivariate analysis by the Cox proportional hazards regression model was used to identify independent prognostic factors in predicting disease-free and overall cumulative survival. Significance was defined as P<.05. All statistical calculations were made using commercial computer software (SPSS/PC+; SPSS, Chicago, Ill).

There were 179 male and 32 female patients with a median age of 53 years. The size of their HCCs ranged from 0.5 to 25 cm (median, 8 cm). Ninety-eight patients (46%) had underlying cirrhosis, and 71 (34%) had underlying chronic hepatitis.

The overall hospital mortality rate was 9.5% (n=20). However, the hospital mortality rate decreased from 32% in 1989 to 5% in 1995. Recurrence of HCC was detected in 122 patients with a median time of 10.6 months (range, 1-98 months). The 1-, 3-, and 5-year disease-free survival rates were 60%, 38%, and 27% respectively (Figure 1). The median follow-up period was 21 months (range, 1-101 months). One hundred eight patients, exclusive of hospital mortality, were confirmed dead when this article was completed. The causes of death included malignant cachexia (n=93), gastrointestinal bleeding (n=4), liver failure (n=3), renal failure (n=1), chest infection (n=1), sepsis (n=1), and unknown (n=4). The overall cumulative survival rates (inclusive of hospital mortality) at 1, 3, and 5 years were 66.6%, 50%, and 37%, respectively (Figure 1). The median overall survival period was 35 months.

By univariate analysis, prolonged prothrombin time (Table 1), prolonged operation time, blood transfusion, fresh frozen plasma infusion, presence of preexisting rupture (Table 2), multiple tumor nodules, involvement of resection margin by HCC, presence of venous permeation, macroscopic portal vein invasion, adjacent organ invasion, and TNM stage (Table 3) were factors that adversely affected the disease-free survival period. By Cox regression analysis, only blood transfusion (relative risk, 1.21; 95% confidence interval, 1.053-1.396) and TNM stage (relative risk, 1.90; 95% confidence interval, 1.467-2.467) were independent prognostic factors (Table 4).

For the overall cumulative survival, the significant adverse factors included preoperative ICG retention value at 15 minutes, serum alkaline phosphatase level, prothrombin time (Table 5), blood loss volume, blood transfusion, fresh frozen plasma infusion, operation time, finding of preexisting rupture of HCC, iatrogenic rupture of the tumor, application of ultrasonic dissector (Table 6), size of the tumor, number of tumor nodules, involvement of the resection margin by the tumor, the presence of venous permeation, portal vein involvement, invasion of adjacent organs, and TNM stage (Table 7). By Cox regression analysis, ICG retention value at 15 minutes, blood transfusion, tumor rupture (including iatrogenic and spontaneous rupture), and TNM stage were independent significant factors that influenced the overall cumulative survival (Table 4).

This study demonstrated that long-term survival of patients with HCC after hepatectomy depends on the tumor staging, preoperative hepatic functional reserve, history of blood transfusion, and rupture of HCC. Because preoperative liver function and tumor staging cannot be altered, the surgeon can help improve the prognosis if blood transfusion and iatrogenic tumor rupture can be avoided and if the function of the liver remnant can be preserved.

Tumor-node-metastasis staging¹⁷ incorporates all the pathologic information that adversely affects survival, namely, tumor size (>2 cm), multiplicity, presence of venous permeation, involvement of major branches of the portal and hepatic veins, and lymph node and distant metastases. A significant difference in survival time has been demonstrated in between different TNM stages.^{18 - 19} By itself, TNM stage should provide the simplest, yet most accurate, information about long-term survival.

Because HCC often develops on top of cirrhosis, hepatic function reserve is frequently suboptimal. Preoperative evaluation of liver function is, therefore, an important aspect of selecting patients with adequate function for hepatectomy. Of the types of hepatic function tests, the ICG clearance test alone^{20 - 21} or in combination with other parameters^{22 - 23} is accepted as a means to select patients for hepatectomy. An ICG retention level of 10% or 14% at 15 minutes was used as a cutoff value for a major hepatectomy,^{20 - 24} but a higher threshold was adopted at an experienced center.²⁵ However, a higher ICG cutoff should only be adopted when the surgeon is experienced and the volume of nontumorous liver to be resected is minimal because, even if the patient has an ICG retention value of less than 14% at 15 minutes, liver failure, ascites, infection, and bleeding esophageal varices would still occur if the liver remnant sustains hypoxic injury during and after the hepatectomy. Such hypoxic injury may be a result of massive bleeding, prolonged rotation of the liver for exposure and division of the hepatic veins, prolonged and continuous Pringle maneuver, prolonged inadvertent compression of the portal vein by a Deaver retractor, and rotation of the left lateral segment into the right subphrenic cavity after trisegmentectomy.¹⁴ With poor postoperative liver function, effective treatment for intrahepatic recurrence, such as re-resection, transarterial oily chemoembolization, or intralesional alcohol injection, cannot be given. Thus, patients with poor postoperative liver function succumb earlier to both tumor dissemination and liver failure than patients with satisfactory liver function.

Poor liver function has been shown in previous studies to affect long-term survival.^{3 ,26} Postoperative liver function is dependent on the volume of uninvolved liver resected, the degree of insult to the liver remnant during the operation, and the postoperative management. Liver regeneration after hepatectomy, though possible in patients with mild cirrhosis,²⁷ is unlikely in patients with moderate cirrhosis or may be insufficient if major damage to the liver remnant has occurred during the hepatectomy. During the follow-up periods, further deterioration of liver function may occur as the degree of cirrhosis increases. It seems that there are no effective measures to prevent progression of liver cirrhosis apart from the suggestion that oral supplementation with branched-chain amino acid–enriched solution can preserve liver function.²⁸ Another possible approach is to control the hepatitis B virus with nucleoside analogs²⁹ after hepatectomy, but a study of this approach has not been done. Interferon may be effective in controlling hepatitis C virus,³⁰ but its use in cirrhotic patients can be dangerous.

Bleeding remains a major problem in hepatectomy. Bleeding may occur during mobilization of the tumor from tearing of the collateral that supplies the tumor, rough handling of the tumor leading to iatrogenic rupture, or laceration of the hepatic vein or during parenchymal transection from intrahepatic branches of the hepatic vein. Previously bleeding of 4 L or more was shown to be detrimental.^{31 - 32} Although some patients can survive even if the blood loss is as much as 10 L,²⁰ this should not be regarded as an advance in hepatic surgery, because blood transfusion of any amount, as shown by this and other studies,^{33 - 34} is related to a shortening of the disease-free and overall cumulative survival periods. Thus, the primary aim of the surgeon during hepatectomy should be the avoidance of bleeding and blood transfusion.

To reduce bleeding during mobilization of the tumor, adequate and wide exposure by a generous incision and adequate illumination of the operative field (eg, by wearing a headlamp) are essential. Adequate exposure is likely to be difficult for a large right-lobe HCC that occupies the entire right subphrenic cavity and is soft owing to extensive necrosis. Apart from the risk of cancer cells disseminating into the systemic circulation,³⁵ forceful mobilization of a large right-lobe HCC may lead to iatrogenic rupture of the tumor or tearing of the right hepatic vein. In such a circumstance, a thoracotomy incision will allow adequate space for mobilization and easy access to the space between the inferior vena cava and posterior surface of the liver, and this will reduce the chance of bleeding. In addition, for segmentectomy of segment 6, 7, or 8, thoracotomy allows the liver transection plane to be perpendicular to the wound, making liver transection more expeditious.

During parenchymal transection, bleeding may occur from branches of the inflow or outflow vasculature. Intermittent use of the Pringle maneuver³⁶ and the ultrasonic dissector³⁷ are effective in reducing blood loss, but cannot prevent bleeding entirely because the ultrasonic dissector may damage the hepatic vein and the Pringle maneuver cannot control bleeding from the hepatic vein. To provide a completely avascular field, total vascular occlusion has been suggested, but total vascular occlusion was shown to be harmful to the liver when compared with the Pringle maneuver in a prospective randomized trial.³⁸ The difference between these 2 vascular control methods is that the liver cells continue to be nourished by venous backflow during the Pringle maneuver.³⁹ Thus, function of liver cells is not seriously impaired. Therefore, slow and cautious application of the ultrasonic dissector with intermittent use of the Pringle maneuver is recommended.

To reduce venous bleeding, reduction of central venous pressure to less than 6 mm Hg would be beneficial.^{40 - 41} However, our patients did not benefit from such a maneuver, because the anesthetists worried that an air embolism was more difficult to manage than the bleeding itself. To reduce the volume of bleeding from the transection surface, prompt hemostasis by suturing is essential. Suturing of individual bleeding points from branches of the hepatic vein is more effective than argon beam coagulation, but suturing, if not performed carefully, can lead to tearing of the hepatic vein and liver parenchyma, and therefore more bleeding. Accurate placement of sutures and cautious knotting are required to prevent iatrogenic induction of further bleeding. Forceful insertion of gauze to pack the space between the transection planes to stop venous bleeding is another iatrogenic cause of bleeding from tearing of branches of the hepatic vein and should be avoided.

To our knowledge, this study shows for the first time that rupture of HCC adversely affects the outcome of hepatectomy. Rupture may be spontaneous, occurring before surgery or during tumor mobilization. The cause of spontaneous rupture is not known, but iatrogenic rupture is likely due to rough handling or forceful mobilization of the tumor in the presence of inadequate wound exposure. Rupture of the tumor can lead to massive bleeding, hypoperfusion of the liver, and seeding of cancer cells into the peritoneal cavity. In an effort to remove a ruptured tumor quickly, a hepatectomy becomes rushed, which leads to even more bleeding and an inadequate tumor-free resection margin. Iatrogenic rupture is more likely if a large right lobe tumor infiltrates into the diaphragm. To prevent such a situation, an added thoracotomy incision will improve the exposure and excision of the diaphragm together with the main bulk of the tumor.

Another important aspect that can influence the long-term survival period is the width of the resection margin. The optimal width of the resection margin in hepatectomy for HCC is controversial and was shown to have an adverse effect⁴² or none at all.⁷ A curative hepatectomy is defined as resection with a tumor-free margin of more than 1 cm.⁴³ However, recurrence of HCC tends to be multicentric in the liver remnant far from the resection margin and is related to the carcinogenic potential of all hepatocytes affected by the hepatitis B or C virus. Attempts to increase the tumor-free resection margin will necessitate resection of more nontumorous liver. A fine balance and sound judgement by the surgeon is necessary. During the course of liver transection, repeated intraoperative ultrasound examination of the tumor in relation to the transection plane is useful in determining the appropriate margin.

In conclusion, surgical technique is as important as preoperative liver function and tumor staging in the long-term survival of patients with HCC after hepatectomy. The surgeon's ability to avoid blood transfusion and iatrogenic rupture of the tumor and to preserve liver function will improve the outcome of hepatectomy for HCC. With increased experience and improved surgical technique, many surgeons can perform hepatectomies for HCC without blood transfusions and with a very low hospital mortality rate.^{7 ,33 ,38 ,44 - 45} Apart from careful selection of patients through liver function screening, refined surgical technique is most likely responsible for the improved outcome of hepatectomy for HCC in recent years.

This study was supported by the Croucher Foundation, Hong Kong, China, Sheung-Tat Fan, MS, MD, FRCS(Glasg and Edin).

Corresponding author: Sheung-Tat Fan, MS, MD, FRCS(Glasg and Edin), Department of Surgery, The University of Hong Kong, Queen Mary Hospital, 102 Pokfulam Rd, Hong Kong, China (e-mail: hrmsfst@hkucc.hku.hk).