CRYOSURGERY AS A TREATMENT FOR ADVANCED STAGE HEPATOCELLULAR CARCINOMA
Results, Complications, and Alcohol Ablation
Wilson S. Wong, M.D.; Shirish C. Patel, M.D.; Federico S. Cruz, M.D.; Kirit V. Gala, M.D., A. Franklin Turner, M.D.
The objective of this study was to investigate the use of cryosurgery and to determine whether there is a role for combined therapy with alcohol ablation in the treatment of patients with hepatocellular carcinoma.
Twelve patients with biopsy proven hepatocellular carcinoma underwent ultrasound-guided cryosurgical ablation of their liver tumor. Postoperative alcohol ablation was performed on those patients who were found to have residual tumor or recurrence after the cryosurgical procedure.
Of the 12 patients (9 males, 3 females) the size of the primary tumor ranged from 3-13 cm with average size of 7 cm in greatest dimension. Most patients had advanced disease according to the TNM staging system: 9 patients had Stage IVa disease, 2 stage III, and 1 stage II. Three patients had residual tumors after the cryosurgical procedure. The residual tumor was treated with alcohol ablation. The 1-Year survival rate for the entire group was 50% (5 of 10) and the 2-year survival rate was 30% (3 of 10). At last follow-up, 1 patient with an 8-cm tumor was disease free for 3 years and another patient with a 13-cm tumor was disease free for 2.5 years. Both of these patients had stage IVa disease.
The authors found cryosurgery to be promising in the treatment of this extremely aggressive form of cancer, with the ability to prolong patient survival. Follow-up treatment with alcohol ablation is an important adjunct in treating residual tumor and controlling recurrence. (Cancer 1998; 82:1268-78)
Keywords: hepatocellular carcinoma, hepatoma, cryosurgery, alcohol ablation, cryoablation, liver cancer
Hepatocellular carcinoma is an extremely virulent form of cancer. Its onset is often insidious. Most patients are asymptomatic until the tumor has advanced to an extremely large size and has spread to other parts of the liver or other organ systems. Because of the late presentation, the prognosis of patients with this tumor is grim. The median survival rate from onset of symptoms is only 3-6 months.
Cryosurgery employs extremely low temperature to destroy cancerous tissue. It has been shown to be as effective as surgical resection for treatment of metastatic colon cancer to the liver. Because it is a focal treatment, it has the advantage over surgical resection of being able to destroy only the necessary amount of liver tissue while sparing more noninvolved liver. This is of particular importance to patients with hepatocellular carcinoma, because the majority of these patients have cirrhosis and compromised liver functions. By sparing more "normal" liver, the patients will have greater liver reserve. Cryosurgery is able to treat disease involving multiple, noncontiguous segments of the liver. Because of the warming effect of flowing blood, large blood vessels, such as the inferior vena cava and portal vein, are somewhat impervious to the effect of freezing. Hence, tumors close to these venous systems can be treated with cryosurgery, whereas resection of tumors close to major vascular structures is not possible, like liver resection, cryosurgery is unable to treat disseminated disease that has already spread throughout the liver or to other parts of the body.
Whou et al. have reported the successful use of cryosurgery in treating 107 patients with hepatocellular carcinoma. This is the only large series reported to date. Although alcohol ablation has been reported to be effective against small hepatocellular carcinomas (<3), there has not been any report concerning the combined use of alcohol and cryosurgical ablation in treating this extremely aggressive tumor. With these factors in mind, we have undertaken a study in the treatment of hepatocellular carcinoma by using cryosurgery. The following is a report of our experience using cryosurgery in ablating hepatocellular carcinoma in 12 patients. In addition, we are reporting the results of follow-up with alcohol ablation in three patients.
Between July 1993 and February 1997, 12 patients underwent liver cryosurgery at Alhambra Hospital. Informed consents were obtained from all patients undergoing liver cryosurgery. All patients had hepatocellular carcinoma diagnosed by computed tomography (CT)-guided liver biopsies. All patients were evaluated initially by a team consisting of a surgeon, an oncologist, and a radiologist to determine whether they were surgical candidates and whether their tumors were amenable to cryosurgery.
All patients were rigorously worked up to rule out any extrahepatic involvement. These studies included CT scan of chest, abdomen, and pelvis and bone scan. Ct portography was performed on all patients to define the size, location, and number of their liver tumors.
This procedure was performed by placing an angiographic catheter in the proximal portion of the superior mesenteric artery. Contrast material was continuously infused through the catheter during CT scanning. A delayed scan was performed 4 hours after the initial scan.
Preoperatively, patients with prolonged prothrombine time (PT) were transfused with fresh-frozen plasma (FFP) to correct the PT to an International Normalized Ratio (INR) value of 1.2 or less. If the patients'PT could not be corrected to the above level with FFP transfusion, then the patient was excluded from the cryotherapy procedure because of the significant increased risk of bleeding. Platelets were transfused for those patients with counts below 60,000. Alfa-Fetoprotein (AFT) levels were obtained in all patients. Every patient received complete bowel preparation and oral neomycin and erythromycin 1 day prior to the surgery.
A bilateral subcostal incision was made in all patients. Cholecystectomy was performed to facilitate subsequent insertion of cryoprobes. The liver was mobilized by division of the ligamentum teres, falciform, and triangular ligaments. The right lobe was mobilized to expose the "bare area" and vena cava. Bimanual palpation of the liver was performed. Intraoperative ultrasound using the Aloka 650 ultrasound scanner (Aloka, Wallingford, CT) was then used to closely examine the entire liver. A 7-MHz, T-shaped, linear array probe and a 7-MHz, linear laporoscopic probe were initially used to look for any lesions that were not seen on the CT portogram.
The cryosurgical procedures were performed by using techniques described by Onik at al. and Lee et al. The lesions were localized by using a biplane transrectal probe (5-MHz transaxial and 7-MHz longitudinal; Aloka). The general approache was to avoid direct puncture of the tumor. We tried to have some noncancerous liver tissue interposed between entry point and the tumor tissue. An 18-gouge needle was placed in the lesion under ultrasound guidance. The needle position was confirmed in both imaging planes on ultrasound. A J-shaped guide wire was positioned in the lesion through the needle. A sheathed dilator (3 mm or 8 mm) was introduced into the lesion over the guide wire. The dilator was removed, and the sheath was left in place. Cryogenic probes (3 mm or 8 mm) were then placed into the lesion through the sheath. The sheath was then pulled back. The cryosurgical procedures of the first ten patients were performed with the AccuProbe System (Cryomedical Sciences, Inc., Rockville, MD) by using liquid nitrogen and the last two procedures were performed with the Cryocare System (Endocare, Irvine, CA) by using Argon gas. The tumors were frozen at maximum flow rate for 15 minutes. The lesions were thawed for 5 minutes and then refrozen for another 10 minutes. If the initial ice balls were not large enough to encompass the entire lenght of the tumors, then the probes were pulled back for 2-5 cm depending on the length of the tumors. The cryogenic probes were turned on at maximum flow rate for an additional 15 minutes, thawed for 5 minutes, and then refrozen for 10 minutes. The entire process may be repeated for very large tumors.
For tumors < 1.5 in diameter, a 3-mm probe was placed at the center of the tumor. For tumors <2.5 cm in diameter, an 8-mm probe was used. For tumors larger than 3 cm, multiple combinations of 3-mm and 8-mm cryogenic probes were placed around the periphery and in the center of tumor to insure freezing of the entire tumor. The goal was to destroy the tumor plus a 1-cm margin around the tumor.
A "Bear-hugger" (Kentec, Irvine, CA) warming blanket continuously warmed the patients throughout the entire cryosurgery procedure. Body temperature was monitored constantly with a esophageal thermometer. Mannitol was infused intravenously to insure adequate urine output. Sodium bicarbonate was given intravenously to keep the urine pH neutral or slightly alkaline.
Postoperatively, all patients were placed on broadspectrum antibiotic coverage. If necessary, the patients were transfused with fresh-frozen plasma to keep the PT to an INR value of 1.2 or less. Platelets were transfused for patients with platelet count below 60,000. The fluid balance was closely monitored and patients were kept in slightly negative balance to prevent pulmonary congestion or adult respiratory distress syndrome (ARDS)
AFP levels were obtained at 3-months after the procedure in the majority of the patients. Three patients were found to have residual tumor on the follow-up CT scan. These patients underwent alcohol ablation of the residual tumor tissue.
Alcohol ablation was performed under CT guidance by using a 20-gouge Bernadino needle. This needle has two side holes at its tip, which allows better dispersion of the alcohol into the tumor tissue. The needle tip was placed close to the peripheral margin of the tumor. Absolute alcohol (100%) was slowly injected under real-time ultrasound control. If the alcohol ran off quickly into the blood stream rather than "staining" the tumor tissue, as observed by real-time ultrasound, then this indicated that the needle tip was not in the substance of the tumor. The needle was then repositioned. The goal of this procedure was to achieve a "black-stain" in the tumor tissue, as evidenced on CT scan. A maximim of 10 ml of 100% alcohol was injected per site, with a maximum of 20 ml per session. At the end of the procedure, 1% Lidocaine was injected along the needle track as the needle was slowly being withdrawn. The injection appeared to lessen the amount of discomfort the patient experienced following the session. The session was terminated prematurely, that is by using less than 10 ml of alcohol, if the patient experienced significant discomfort or if the tumor tissue appeared to be completely ablated. The number of sessions per patient was dependent upon the amount of residual tumor identified on CT.
Of the 12 patients who underwent cryosurgery of their primary hepatocellular carcinoma, nine patients were males, and three were females. There were 10 Asians and two Caucasians. Their ages ranged from 40 years to 72 years, with a mean age of 59 years. Nine patients had a history of hepatitis B infection, and three had hepatitis C.
By using Pugh's modification of the Child score in assessing the severity of cirrhosis, seven patients were classified as Class A, and five were classified as Class B. There was no patient in the Class C classification.
All patients had a dominant mass ranging in size from 3 cm to 13 cm in diameter. The average size of the dominant mass was 7 cm in diameter. Nine patients had one to five satellite lesions scattered throughout the remainder of the liver. The average number of satellite lesions for the nine patients was three. The sizes of the satellite tumors ranged from 1 cm to 5 cm, with an average size of 2 cm. Three patients had only one mass in the liver. One of these three patients had undergone previous surgical resection. He had a recurrent tumor located at the surgical margin and immediately adjacent to the inferior vena cava (IVC). The second patient had an infiltrating type of tumor with a poorly defined border and that had invaded the right right portal venous system. The third patient had a large tumor that was located in between the right and middle hepatic veins and close to the IVC. In all 12 patients, tumors were considered to be surgically unresectable.
Because of the large size of the primary tumor and multiplicity of tumors involving more than one lobe, most of the tumors were classified as T4 according to the TNM system of classification. Nine out of 12 patient were classified as having Stage IVa (T4 N0 M0) disease. Two patients had Stage III (T3 N0 M0). The only patient that was classified as Stage II (T2 N0 M0) had a T2 tumor that was discovered incidentally during a routine ultrasound scan for abdominal discomfort.
Eight of 12 patients required transfusion of 6-8 units of FFP, and six patients required transfusion of platelets prior to the surgery. The average operative time was 3 hours, with a range of 2.0-5.5 hours. The median number of days of hospital stay was 7 days, with a range of 4-30 days. Initially, patients were observed in the intensive care units for the first 2-3 postoperative days.
All except one patient had more advanced cirrhosis of the liver than anticipated from preoperative evaluation. All patients' livers were very firm, and many of the patients had livers with a nodular surface. The cirrhotic changes of liver attenuated the sound waves. The combination of these factors made ultrasound imaging very difficult. In general, most tumors were echogenic compared with the surrounding liver.
The median preoperative AFP level was 32 ng/ml, with a range of 2-15,774 ng/ml. The median 3-6 months postoperative AFP level was 13 ng/ml, with a range of 4-8,494 ng/ml. The difference was statistically significant (P<0.001).
Postoperatively, the tumors shrank in size. For the eight patients who had a 6-month follow-up CT scan available for review, the average size of the dominant tumor changed from a preoperative size of 7.5 cm (3-13 cm) to a 3-month postoperative size of 3.2 cm (1-6 cm). This change in size was not statistically significant (P=1.000). However, residual tumor size does not reflect tumor activity. Five patients had biopsies of thier residual tumor mass under CT guidance. All biopsies showed only dead or scan tissues. The mass continues to shrink in size with later follow up CT studies.
There were three patients that required postoperative alcohol injection to ablate residual tumor tissues. The inadequate destruction of tumor tissues was either be cause of inexperience with cryosurgery or because the tumors were too large or too numerous for adequate freezing. Our cryosurgery technique changed as we gained experience with this procedure. We were hesitant to be aggressive with our freezing technique with our first patient, because we did not know the full potential damage of cryosurgery to a patient with limited liver reserve. In this patient, we limited freezing to only 60% of the tumor area and to a single freeze. This patient recovered from the cryosurgery quite rapidly. Because of the limited freezing, we believed that he would have residual tumor. We tried to ablate the remaining tumor tissue with absolute alcohol. The 3-month, CT-guided needle biopsy was negative. However, this patient died 1 year after his cryosurgery due to liver failure. He was found to have new tumors in a different lobe of the liver. We became more agressive with freezing in our third patient. His dominant tumor was 9 cm, and he had five smaller (1-cm) satellite tumors. To freeze his dominant mass, we had to place four of the 8-mm cryoprobes around the periphery of the tumor. Because the length of ice ball was only 4.5 cm, we needed to pull back the probes by 2.5 cm to ensure freezing of the entire tumor. Portions of the tumor were frozen two to three times because of the overlapping of the ice balls between the initial freeze and the pull back. Because of the extended period of freezing of his primary tumor, we did not attempt to freeze the five small satellite lesions scattered throughout the liver. Postoperatively, we noted that the patient was much sicker than the first two patients. He developed ARDS involving both lungs and required prolonged hospitalization to treat his pulmonary complication.
After he recovered from his pulmonary complication, we ablated his remaining five satellite lesions with absolute alcohol injection. Six months after his cryosurgery, multiple CT-guided needle biopsies of his tumors, including the primary tumor, were negative. He was then followed with routine CT of the abdomen at a 4-6 month intervals. He was well until he presented 2 years after the cryosurgical procedure with bleeding from his esophageal varicies. Repeat biopsy showed recurrence of tumor infiltrating the liver. He died of hepatic failure.
Of the 12 patients, the one with the most remarkable result was Patient 9. He was a 40-years-old Chinese gentleman who presented with a 13-cm tumor replacing the entire right lobe of his liver. This tumor was only 6 cm on a CT scan that was performed in Taiwan 4 months prior to his presentation. CT portogram demonstrated an additional 2-cm tumor in the lateral segment of the left lobe of the liver. Because of the large tumor size and the satellite lesion, he was not a candidate for surgical excision. Because of his young age and good physical condition, we decided to attempt freezing of his tumors.
The small lesion in the left lobe was frozen with relative ease. However, freezing of the 13-cm tumor was performed with much difficulty. We had to use six of the 8-mm cryoprobes to freeze this tumor. The probes had to be pulled back twice in order to freeze the entire mass. Because the large mass was located posteriorly and superiorly, abutting against the diaphragm, ultrasound imaging was poor because of the limited access. One of the cryoprobes inadvertently penetrated through the diaphragm and froze a small portion of the lung. Postoperative chest radiograph demonstrated a right pneumothorax and an infiltrate in the right lung base. This complication was treated conservatively with a chest tube. The remaining postoperative course was unremarkable.
The patient's preoperative AFP level was 7,090 ng/ml. Two months after the cryosurgery procedure, the AFP level dropped to 450 ng/ml. However, the AFP level rose to 620 ng/ml 3 months after the surgery. A repeat CT of the abdomen demonstrated a reduction of the tumor size from 13 cm to 8 cm. Most of the tumor had a very low "fluid-like" density, indicating necrosis of tumor tissue. However, there was tissue located in the inferior, anterior, and medial aspects of the tumor that had an intermediate density, suggesting live tumor tissue. CT-guided biopsy of these areas confirmed the presence of live tumor.
Alcohol ablation of the residual tumor was then attempted by using the techniques described above. Because of the relatively large amount of residual tumor tissue, the patient underwent seven sessions of alcohol injection within a 2-month period. The total volume of alcohol injected was approximately 100 cc. The injection was stopped after CT showed that there was no residual tumor tissue. The AFP level dropped to 40 ng/ml at the end of alcohol ablation procedures.
The patient was doing well untill 13 months after his cryosurgical procedure, when his AFP level suddenly rose to 90 ng/ml. He underwent repeat CT-guided biopsy of his original tumor site. Pathological examination of the specimen revealed only fibrotic tissue without any live tumor. Because of the rise in the AFP level, despite his negative biopsy he underwent a second laporatomy procedure with the thought of undergoing repeat cryosurgery for any recurrent tumor. Intraoperative ultrasound demonstrated numerous small, echogenc nodules scattered throughout the entire liver. These nodules were believed to represent disseminated metastatic nodules. Initial frozen section examination of the tissue obtained by needle biopsy was negative for live tumor tissue. A 4-cm wedge biopsy of the left lobe of the liver was performed. Ultrasound confirmed the presence of echogenic nodules within the specimen. Because we felt that the patient had disseminated disease, we terminated the operation without performing any additional cryosurgery.
To our amazement, the pathological examination of the tissue showed that the echogenic nodules were regenerating nodules. The patient is currently living, 2.5 years after his initial operation, and is tumor-free at the time of this writing.
Patient 8 was a 48-years-old Chinese gentleman who had a tumor that was very difficult to visualize on CT or ultrasound. His AFP level at the time of presentation was 15,774 ng/ml. Because of this high level, he was felt to have diffuse tumor throughout the liver. However, there was no definite tumor identified on CT. Two CT-guided biopsies of the liver were negative for tumor. Ultrasound revealed thrombus in the right portal vein and a 5-cm, oval-shaped echogenic lesion around the portal vein. Ultrasound-guided biopsies of the thrombus and the surrounding tissues around the portal vein were positive for moderately differentiated hepatocellular carcinoma.
It was decided to freeze the entire echogenic lesion, including the portion of the partially thrombosed right portal vein. The patient was found to have very advanced cirrhosis at the time of surgery. The surgical procedure and the immediate postoperative period were relatively uneventful. However, his AFP level dropped only to 8,494-ng/ml 1 month after the surgery. It started to climb to 30,000 ng/ml 3 months after the procedure. Unlike most other patients, he did not recover well from his surgery. He continued to feel tired and to have abdominal discomfort. He died from liver failure 5 months after the procedure. The tumor involvement of his portal vein may have contributed to the failure of the cryosurgical procedure. The freezing of a portion of his portal vein may have further injured his already compromised liver and may have accelerated his liver failure.
Two of the patients, Patients 2 and 4, had previous surgical resection of their primary liver tumor. Both patients had recurrent tumor at the surgical margin. Patient 4 had a recurrent tumor in the posterior aspect of the right lobe of the liver. Preoperative workup did not show any evidence of disease elsewhere. One month after the cryosurgery, he developed disseminated disease throughout the entire liver, as shown on CT scan. It was likely that he had disseminated disease at the time of surgery, but our diagnostic examinations were not sensitive enough to discover it.
Three additional patients had tumor recurrence after the cryosurgery procedure. One recurred locally 2 months after the procedure in the patient with portal venous involvement. Two patients had tumors that developed in other portions of the liver, away from the sites of the primary tumors.
Bleeding was the major complication of this procedure. The average amount of blood loss was 700 ml. With a range of 100-2500 ml. The amount of blood loss was found to be related directly to the severity of the underlying liver disease and to the presence or absence of portal hypertension. It did not seem to be related to the size or number of cryoprobes used. Only four patients required blood transfusion during and after the surgery. Two patients received 2 units and one patient received 5 units of packed red blood cells. Two of these patients received 10 units of platelets. The fourth patient, who had very poor liver function with an albumen level of 2.5 mg/ml, died of uncontrollable bleeding 4 days after the cryosurgical procedure despite transfusions of large amounts of blood products. He appeared to have suffered from disseminated intravascular coagulopathy (DIC).
Because of the close proximity of the liver to the right lung, injury to the right lung and right hemidiaphragm were the most common complications of the procedure. Nearly every patient developed some form of atelectesis in the right lung base in the immediate postoperative period. Patient 3 developed a mild form of ARDS several days after the surgery.
Because of the technical difficulties, as described above, a small portion of the right lower lung of patient 9 was inadvertently frozen. To prevent recurrence of this complication, we currently place a rubber mat around the dome of the liver prior to any freezing of tumor that extends close to the dome. This protects any inadvertent penetration of the diaphragm by a needle or a cryogenic probe.
Patient 2 developed a fistula between the liver and colon after the cryosurgical procedure. This patient had recurrence of tumor in the left lobe 6 months after previous liver resection. This complication was most likely due to the adherence of the transverse colon to the liver from scar tissue after the primary resection. Postoperatively, he had recurrent abscess formation and died 9 months later from hepatic failure and sepsis. It was not known whether he had recurrent tumor.
Another patient developed a large biloma, which required repeated drainage procedures. This patient also had previous surgical resection of a tumor that was located in the posterior aspect of the right lobe of the liver. The patient underwent cryosurgery of his recurrent tumor, which was located at the surgical margin. This complication was most likely related to the fact that the cryosurgical lesion was located at the surface of the liver. The tumor underwent necrosis and may have sloughed after the procedure. The "weeping", exposed surface may have produced the biloma.
Unlike the portal or hepatic venous systems that were relatively immune to the freezing process because of the warm, flowing blood, the bile duct system was very sensitive to thermal injury. Patient 3 developed strictures of his bile ducts after the procedure. This problem was partly corrected with placement of several stents into his bile ducts.
The 1-year survival rate was 50% (5/10), and the 2-year survival rate was 30% (3/10). Most patients died of hepatic failure due to their underlying liver disease. Four patients are currently disease free, according to CT scanning and biopsy criteria. One of these four patients is disease free 3 years after her surgery. She had a CT-guided biopsy 2 years after the surgery and was found to be negative for tumor. Her current AFP level was normal. The second patient was the 40-years-old Chinese gentleman with the 13-cm tumor described above. His surgery was 2.5 years ago. Both of these patients, who had the longest survival, had Stage IVa disease.
Hepatocellular carcinoma is one of the most malignant forms of cancer. It is three to eight times more common in men than in women. Although it is relatively rare in the United States (US) and other western countries, it is very common in Southeast Asia. The reported incidence in th US and Northern Europe is 2 per 100,000 per year. In comparison, the reported incidence is 30 per 100,000 per year in South-east Asia. The higher incidence of this tumor in South-east Asia is related to the higher incidence of hepatitis B virus (HBV) infection in these countries. HBV infection could be a primary carcinogen in as many as 80% of patients with hepatocellular carcinoma worldwide. In the US, the most common type of hepatitis transmitted by transfusion is hepatitis C (HCV), which can also cause hepatocellular carcinoma.
Patients with hepatocellular carcinoma often present very late with symptoms of abdominal pain, weight loss, weakness, and abdominal swelling. Patients with HBV or HCV infections are considered to be high risk and should be screened for this disease. Serum AFP level is one of the most sensitive markers for hepatocellular carcinoma. It is elevated in 80-90% of individuals from the Orient bearing hepatocellular carcinoma. However, it is elevated in only 60-70% of hepatocellular carcinoma patients from the US and Europe. The best method for diagnosing this tumor is contrast CT scan of the abdomen. Ultrasound may not be as sensitive as CT in detecting the tumor because of poor sound penetration of cirrhotic livers.
Because of its late presentation, the prognosis for patients with hepatocellular carcinoma is extremely poor. The treatment options for patients with hepatocellular carcinoma are quite limited. These tumors are generally resistant to systemic chemotherapy and local radiation theraphy. If the tumors are relatively small in size (<3cm) and few in number, then they can be ablated by direct injection with 100% absolute alcohol. Liver transplantation is not a viable option for most patients because of the generally large size of most tumors and the limited availability of donor livers. Because of the aggressiveness of the tumors, most patients would have died of their disease prior to finding a donor liver.
If the tumor is of sufficient size but is limited to one or two segments of the liver, then surgical resection remains the best treatment option. More oftern, the tumors are not surgically resectable, because they frequently are extremely large and involve more than two contiguous segments. They may be multicentric, with several tumors located in more than one or two noncontiguous segments. In addition, surgical resection tends, to follow segmental anatomy of the liver. This often leads to removal of large amounts of noncancerous liver tissue.
The majority of cryosurgeries of the liver performed in the US are for treatment of metastatic tumors from colon carcinoma. This is due to the fact that colon cancer is a common disease in the US, unlike hepatocellular carcinoma. This is in contrast to the situation in Southeast Asia.
There are several major differences between the approaches to the cryosurgeries of hepatocellular carcinoma and metastatic disease to the liver. 1) Almost all patients with hepatocellular carcinoma have cirrhosis secondary to some underlying liver disease, such as HBV or HCV infection, alcoholism, or hemochromatosis. The liver reserves of these patients are generally poor. They have greater risk of developing liver failure when major portions of liver are destroyed or removed. These patients must be evaluated clinically to determine whether they can survive the surgery. 2) Many of these patients have bleeding diasthesis related to their underlying liver disease. In combination with portal hyper tension, these patients are at much greater risk of significant bleeding and DIC. The bleeding parameters must be monitored closely and corrected if necessary with blood or platelet transfusion before and after the surgery. Extra supplies of blood, platelets, and FFP must be available prior to the procedure. This is the reason why we include a hepatologist/oncologist as part of our team. 3) Because almost all of these patients have cirrhosis, intraoperative ultrasound imaging is often difficult to perform. The nodular surface of the liver prevents good contact with the transducer. Cirrhotic liver significantly attenuates sound waves and decreases ultrasound penetration. These factors lead to some difficulty in demonstrating the tumors intraoperatively. 4) Needles, guide wires, dilators, and cryogenic probes are more difficult to place in these patients because of the severe firmness of the cirrhotic liver substances. We have frequently bent guide wires while trying to push the dilators into position. The 8-mm dilators were clearly more difficult to place than the 3-mm dilators, because more tissue has to be spread apart. However, 8-mm cryogenic probes are often needed because of the large size of the primary tumor. 5) Unlike most metastatic colon cancers of the liver, hepatocellular carcinoma frequently has a very large, dominant mass, as seen in our series. These tumors required the placement of multiple cryogenic probes in a circular fashion in order to produce an adequate size and shape of the ice balls. Anyone who is contemplating freezing hepatocellular carcinoma should have a cryogenic system that is capable of operating at least three cryogenic probes simultaneously. 6) Because most patients are infected with HBV or HCV, the surgical staff should take extra precautions throughout the procedure. Careful handling of needles and tissues are mandatory. The entire surgical team should be vaccinated against HBV. Unfortunately, there is no vaccine currently available for HCV.
Zhou et al. have produced the only reports available so far on the effectiveness of cryosurgery in the treatment of hepatocellular carcinoma. They reported that the 5-year survival rate for subclinical disease patients was 37.5%. Their results showed a 5-years survival of 11.4% for all patients in their series, including those with more advanced disease. These are truly remarkable results in view of the aggressive nature of this tumor.
Our initial inexperience with this procedure may help to explain the difference in results. In the beginning, we were less aggressive with our freezing technique. We were afraid that multiple or prolonged freezing cycles might cause hepatic failure. With more experience, we found that more aggressive freezing is associated with better results, although it is also associated with slightly higher morbidity.
We found that postcryosurgery alcohol ablation was an effective means of controlling the disease. This was shown in the three patients in whom we used this combination therapy. Often, the tumor may be so large that cryosurgery alone may not be able to destroy it completely. This was the case for patient 9, who had a 13-cm tumor. The combination of the cryotherapy and alcohol ablation clearly helped in controlling his tumor.
There was no significant complication associated with alcohol ablation. This low complication rate is most likely related to our limitation of the injection per site to 10 ml of 100% alcohol, with a maximum of 20 ml per session. Redvanly et al. founded that the complication rate increased proportionately with increasing the amount of alcohol injected. In their series, complication rates were reported as 28% with less than 10 ml, 35% with 11-20 ml, 73% with 21-30 ml and 71% with more than 30 ml.
Tumor recurrence is a major problem with hepatocellular carcinoma. Frequently, the tumor recurred at the margin of the original tumor. This was seen in the two patients who had previous surgical resections. It is important to note that, in the four patients who had recurrent tumor, only one had recurrence at the original site. This patient had portal venous involvement, and the tumor recurred in the portal vein. The remaining three patients had recurrence in other parts of the liver. It is debatable in such circumstances whether this was a "recurrence" versus a "new tumor". If the entire liver is cirrhotic, then any part of the liver can develop new tumor. Perhaps liver cryosurgery is more effective than surgical resection in treating local disease. More study is needed in this regard. It is imperative that all patients should be closely monitored, perhaps every 3 months, with CT scanning and serum AFP level. Any new tumor can be treated with alcohol ablation.
In summary, we find that cryosurgery is a promising therapy for this extremely aggressive form of cancer. No treatment short of liver transplantation, as discussed above, can truly cure this disease. This is because the underlying liver disease may give rise to new tumors. By using cryosurgery in combination with alcohol ablation, we have found an effective treatment for controlling this cancer and perhaps prolonging the patient's life. It is important to note that, by using this procedure, we were able to salvage more than 10% of patients with advanced Stage Iva disease. However, the performance of cryosurgery of the liver is not without risk. Despite our best efforts, we had one surgically related death and had major complications, as noted above. All patients should be screened carefully to determine whether they are truly good candidates for this procedure. We believe that a team of physicians, which should include two surgeons who are versatile with liver resection methods, a radiologist who is skilled in interventional and intraoperative ultrasound techniques, and a hematologist/oncologist who is familiar with treating bleeding disorders, is necessary for the success of this procedure.