Dr. Chakravarti and Dr. Morris are in the Department of Radiation Oncology at Massachusetts General Hospital, Harvard Medical School in Boston, MA. Dr. Madre-Bell is with the United States Army Reserves. Dr. Constable and Dr. Kelly are with the Department of Therapeutic Radiology and Oncology at the University of Virginia Health Sciences Center in Charlottesville, VA. Dr. Tegtmeyer was with the Department of Radiology at the University of Virginia Health Sciences Center.

E xtrahepatic bile duct cancer is uncommon, with an annual incidence of approximately 2500 cases in the United States. Ninety-eight percent of these tumors are adenocarcinomas, with squamous cell carcinomas representing the remainder. ¹ The prognosis in patients with this type of cancer is often grim: 5-year survival ranges from 0% to 31%. Distant metastases are less frequently reported with this tumor. Instead, this tumor behaves very aggressively in a local manner, often involving such organs as the liver, pancreas, gallbladder, and duodenum. Unfortunately, the ultimate goal of treatment is often no more than palliation of pain and obstructive jaundice symptoms.

It has been widely believed that surgery is the only treatment that may offer a cure. However, only 20 to 40% of these tumors are resectable for cure, and locoregional recurrences occur in 60 to 80% of cases. ² Distal duct lesions traditionally have been managed with a pancreaticoduodenal resection (Whipple procedure). Proximal lesions involving the region extending from the right and left hepatic ducts to the common hepatic duct inferiorly to the level of the cystic duct are referred to as Klatskin tumors. Fifty to 70% of Klatskin tumors are unresectable; the remainder is managed by surgical resection and Roux-en-Y hepaticojejunostomy bypass. ³

Radiation therapy traditionally has been considered to be a palliative option in treating bile duct cancers in cases of unresectable disease. It also has been used as an adjuvant for microscopic residual disease after surgery. Radiation therapy options currently available include external beam radiotherapy, intraoperative radiotherapy using orthovoltage equipment or electrons, and intracavitary irradiation using Ir-192 applied through a carefully placed "decompression" tube. Recent studies have shown a survival benefit in patients with bile duct cancers treated with brachytherapy and stenting compared to patients treated with stents alone. ⁴ Studies also have shown improved survival in patients treated with radiotherapy following either complete or incomplete surgical resection of bile duct tumors. ⁵

Our purpose in this study is to examine retrospectively the results of treating primary bile duct neoplasms with Ir-192 brachytherapy versus radical surgery, based on our experience between 1981 and 1994 at the University of Virginia Medical Center.

Materials and methods

Forty-six patients were diagnosed with primary extrahepatic bile duct carcinoma at the University of Virginia Hospital from 1981 to 1994. Of these, 21 received radiotherapy only, 14 were treated surgically with curative intent, 3 were treated palliatively with stent placement, and 8 patients either expired prior to treatment or refused treatment. Preoperative staging evaluations included physical examination, laboratory tests, and imaging studies including CT,
percutaneous cholangiography (PTC)
and, in some cases, angiography. The histopathologic diagnosis of bile duct adenocarcinoma was made by cytological brushings, needle biopsy, or by tissue obtained at surgery.

Three of the 21 radiotherapy patients were treated by external beam irradiation (EBRT) alone and were excluded from further analysis. Eighteen of the 21 radio-
therapy patients were treated by either brachytherapy alone or in combination with EBRT and constituted Group I in this study.

Table 1 presents demographic data, modality of treatment, and follow-up information on all Group I radiotherapy patients. There were 9 females (50%) and 9 males (50%) in this group. The mean age of Group I patients at the time of diagnosis was 66.2 years (SD±12.0 years). The mean total radiation dose delivered to this group was 3997 cGy (SD±1185 cGy). The mean brachytherapy dose rate was 45 cGy/hour (SD±6.2 cGy/hour) at 0.5 cm from the linear source. All patients in this group had percutaneously placed stents. Two patients received external beam irradiation in addition to Ir-192. Patient 1 received 30 Gy of EBRT and 30 Gy
Ir-192. Patient 18 received 20 Gy EBRT plus 20 Gy Ir-192.

Group II in our study was comprised of 14 patients treated surgically with curative intent. Of these patients, nine underwent the Whipple procedure (pancreaticoduodenectomy) and five underwent surgical excision of tumor with
an accompanying bypass procedure (hepaticojejunostomy). All patients in this group initially had percutaneous stents placed to relieve obstruction. Table 2 presents demographic and treatment information on all Group II patients. There were eight males (57%) and six females (43%) in this group. The mean age of Group II patients at the time of diagnosis was 64.6 years (SD±12.9 years).

The criteria for unresectability included findings of encasement or occlusion of the common hepatic artery or main portal vein on CT or angiography. Extensive invasion of adjacent structures such as the liver, stomach, colon, or gallbladder, or findings of peritoneal metastases on laparoscopy also deemed patients unresectable. All patients in Group I were judged to have clinically staged, unresectable, locally advanced disease. All Group II patients, based on preoperative evaluation, were deemed to have resectable disease. Operative and pathologic findings, however, revealed that five patients (#2, 5, 7, 10, and 14) had tumors with invasion into adjacent structures. Hence, 64% of patients in Group II had localized disease, and 36% had locally advanced disease based on operative and pathologic findings.

Results

Tables 1 and 2 present follow-up information for patients in Groups I and II, respectively. Of the 18 Group I patients treated with brachytherapy alone or in combination with external beam radiotherapy, three were alive at their latest follow-ups, ranging from 5 to 24 months after treatment. One patient in Group I (patient #17 in table 1) died of myocardial infarction but was free of disease at latest follow-up. Fourteen patients in Group I ultimately died of disease. All of these patients had progressive local disease with invasion into critical adjacent structures and died
of resulting complications, including infection/sepsis.

Life table survival probabilities are shown in figure 1 for both groups. Actuarial analysis of survival reveals a median survival of 13 months with a standard error (SE) of ±6.09 months. The Brookmeyer-Crowley 95th percent confidence interval for median survival ranged from 9 to 15 months. The 1- and 2-year life table survivals were 53% and 20%, respectively, for Group I patients (table 3). There appeared to be no dose-response effect. Of the nine patients receiving 30 Gy, three were alive at their latest follow-up. None of the nine patients receiving doses of greater than 30 Gy were alive at their latest follow-up.

Of the 14 patients treated with radical surgical resection (Group II), four were alive at their latest available follow-ups, ranging from 7 to 86 months. Three of these four patients had no evidence of disease (NED); one of these patients had a local recurrence and is presently being treated with salvage chemotherapy. Nine of the remaining 10 patients died of complications arising from recurrent local disease or hepatic metastases. One patient died of postoperative sepsis. Actuarial analysis of survival reveals a median survival of 14 months (SE±2.48 months). The Brookmeyer-Crowley 95th percent confidence interval for median survival ranged from 12 to 18 months. The 1- and 2-year life table survivals for Group II patients were 70% and 31%, respectively (table 3).

Any differences in median survival between the two groups were relatively minor and were not statistically significant (p=0.32). The mean age of the Group I patients was slightly higher than that of Group II patients; however, this was not found to be statistically significant (p>0.05). Mean age, therefore, does not appear to have been a confounding factor in this study.

With the exception of minor infections around the immediate catheter site, there were no reports of serious complications resulting from treatment in Group I patients. However, there were three reported cases of serious postoperative infection/sepsis in Group II patients. As previously mentioned, one of these patients died of post-surgical complications after a procedure involving surgical resection and bypass. The other two patients recovered fully from their infections with intensive IV antibiotic therapy. Although they were not analyzed, out of the three patients receiving only external beam therapy, one died of complications from a liver abscess and sepsis one year following treatment.

All patients in Groups I and II received palliation of their obstructive symptoms (jaundice and pain) after treatment. Table 4 summarizes complication rates and palliative benefits by treatment group.

Discussion

Primary bile duct adenocarcinomas are aggressive tumors with very poor long-term prognoses. It is widely believed that the only means of cure is radical surgery. Resectability in these cases should meet three major criteria for optimal success: 1) there should be an absence of extensive vascular invasion, in which tumor invades the main portal vein and/or both the right and left hepatic veins and arteries; 2) there should be a complete absence of metastatic disease, growth into surrounding structures, and peritoneal metastases; and 3) there should be a complete absence of tumor within the second-order biliary radicles. The patient also must be medically stable to tolerate radical surgery. ⁶ Only 30 to 40% of all extrahepatic biliary tumors meet all three aforementioned criteria and are considered resectable. Recurrences after resection have been found in up to 81% of cases, ⁷ and the survival prognosis is still grim after treatment, with some reports suggesting 5-year survival averages close to 5%. ³

Radical surgery for bile duct tumors commonly entails one of two procedures: 1) resection of tumor and Roux-en-Y hepaticojejunostomy as a bypass, which is more often used for proximal bile duct lesions (e.g. Klatskin tumors); 2) pancreaticoduodenectomy (Whipple procedure), which is preferred for malignant lesions involving the middle and lower thirds of the common bile duct. ³ Most patients with hilar bile duct carcinoma are technically unresectable, but a significant number of these patients undergo radical resection regardless and ultimately do very poorly.

Both surgical procedures can produce substantial morbidity. The average operative mortality rate is reported to range from 5% to as high as 20%. ³ The most serious complication resulting from the Whipple procedure is disruption of the pancreaticojejunostomy, which occurs in 10% of all cases. ³ This may result in an upper abdominal abscess and/or external pancreatic fistula. In the worst case, necrotizing retroperitoneal infection may result in the erosion of the major vessels of the upper abdomen, including the exposed portal vein or the remnant of the gastroduodenal artery, ultimately leading to death. Additionally, following surgical resection, recurrences are common and often involve peritoneal implants, malignant ascites, and malignant pleural effusions due to hematogenous spread. In the many series in the literature, 2-year survival following surgical resection approaches 25%, and 5-year survival is reported to be close to 5%. ⁸ As the mean age of onset is 60 years of age, surgical complications in many cases contribute to the high mortality rates in this group.

Nonsurgical palliative stenting and brachytherapy traditionally has been reserved for those patients with unresectable disease. Stents can be placed through the obstructing tumor either during endoscopic retrograde cholangiopancreatography (ERCP), or they can be implanted by the percutaneous transhepatic approach. ⁸ The latter approach was used initially to relieve obstructive symptoms in all Group I and II patients in our study.

The types of ensuing complications from stent placement are classified according to when they occur. Early complications generally occur within a week or two after placement. ⁹ Sepsis and bleeding represent the majority of early complications and ultimately cause death in 1.5% of all external stent placements; that rate is somewhat lower with endoscopic catheters. Late complications generally are much less severe and primarily involve local infection around the catheter site. These can be treated with simple catheter exchanges in the majority of cases.

Recently it has been reported that combined external beam radiotherapy and intramural Ir-192 brachytherapy significantly increased 2-year survival in patients who received this combined modality of treatment, compared to those patients who did not receive radiation in the course of their treatment. Alden et al hypothesize that the effectiveness of brachytherapy resides in its ability to maintain ductal stent patency, delaying the onset of severe complications from biliary obstruction. ⁴ Equivalent survival rates have been reported
in a study comparing brachytherapy
with more radical surgical procedures. ⁴ Veeze-Kuijpers et al have reported that the most accurate prognostic factors in bile duct cancers are the status of lymph nodes and of hepatic involvement. ¹⁰ These researchers demonstrated higher survival rates in those patients treated with a combination of surgical resection and postoperative radiotherapy compared to those patients who received radiotherapy alone. Fletcher et al, in the early 1980s, found equivalent survival rates in those patients treated with Ir-192 brachytherapy compared to other treatment modalities.

Our data suggests that a significant difference in median survival does not exist between those patients treated
surgically and those treated with brachytherapy (p=0.32). With either treatment modality the prognosis appears grim, with few long-term survivors. In our series, the 3-year survival rate for the brachytherapy patients was 12%, compared to 22% for patients treated with surgical resection. There was no significant survival difference between the two groups despite the presence of more advanced tumors in the radiotherapy group.

It is unclear whether surgery is the only modality that offers curative potential. The patient in our study who has lived out to 86 months in the surgical arm and is presently alive and free of disease may represent a curative success. Brachytherapy appears to prolong survival; however, it remains unclear whether it offers hope of definitive cure (one patient in Group I of our study did live out to 81 months, though the patient ultimately died of disease).

An important consideration is effectiveness of the treatment modality in palliating pain and obstructive jaundice symptoms while minimizing the rate of serious complications. All patients treated with brachytherapy and surgery received palliation of their obstructive symptoms. However, the rates of serious complications differed significantly. With the exception of minor infections around the immediate catheter site, there were no reports of serious complications in the group of patients in our study treated with brachytherapy. In the surgical group, there were three reported cases of serious postoperative sepsis, and in one of these cases a patient died of post-surgical complications. Radical surgical procedures such as the Whipple procedure are known to carry significant risks, with mortality rates between 5 and 20%. ³ The procedure itself is long and technically complicated for the surgeon, and it is often not well tolerated in the age group commonly presenting with primary extrahepatic bile duct tumors.

Conclusion

Surgery may offer curative potential for a very limited number of cases of bile duct tumors, as evidenced by one patient in our series who has lived out
to 86 months disease free. Although brachytherapy with stent placement appears to result in comparable median survival, it remains to be seen if lasting cures can result from this treatment approach. It must be emphasized that the probability of long-term survival with either surgery or brachytherapy appears to be minimal, with 2-year survivals of 31% and 20%, respectively, in our series. We would therefore recommend surgery with curative intent for those younger patients with resectable tumors who are medically fit to tolerate surgery.

With high palliation rates and what appears to be a far lower rate of serious complications than radical surgery, brachytherapy may be better suited for older patients, especially in view of the general medical condition of this patient population. Future prospective randomized studies comparing brachytherapy with radical surgical resection in the management of primary extrahepatic bile duct tumors may be needed to better resolve this issue. AR