Research

The Johns Hopkins Hospital has been a leader in advancing the current knowledge of the optimal treatments for bile duct cancers. Meticulous studies at Hopkins have allowed the development of specific criteria to determine whether a complete resection should be attempted. Currently, tumors that involve both lobes of the liver, that involve either of the main blood vessels in this area (the portal vein or the common hepatic artery), or that have spread (metastasized) are considered unresectable-such patients receive palliative surgical therapy that has proven to be safe and effective. This includes a nerve (celiac axis) block to relieve pain and bypass procedures (gastrojejunostomy, hepaticojejunostomy) to prevent obstruction of the gastrointestinal tract.

The resectablity rate is highest for tumors of the distal common bile duct (91%) versus those of intrahepatic and perihilar locations (approximately 50% each). The distal common bile duct tumors are treated by the Whipple resection. Such tumors are often difficult to distinguish preoperatively from other tumors that are centered in this area, such as pancreatic adenocarcinomas, duodenal adenocarcinomas that have extended into this area, and true carcinomas of the ampulla of Vater. However, all of these tumors are optimally treated by Whipple resection, after which pathologic analysis of the resection specimen can determine the site of origin and the extent of the tumor. The 10 year survival rates are highest for those arising in the duodenum (59%), followed by those of the ampulla (25%), distal common bile duct (21%), and pancreas (5%). Factors that predict outcome in these patients are the microscopic status of the resection margins (that is, did the cancer extend to the edge of the tissue where the surgeon cut) and lymph node metastases, as well as the degree of differentiation of the tumor. Tumors that microscopically extend to the edges of the resection (positive margin) or those that have spread to regional lymph nodes fare worse. Tumors that closely resemble microscopically the normal glandular epithelium (well-differentiated) fare better than those that do not (poorly differentiated).

Not long ago, the Whipple resection was associated with a high (>10%) operative death rate, and there was considerable debate if the procedure's risks justified the small potential benefits that resection might confer upon patients with very aggressive tumors. Currently the operative mortality of this procedure is a remarkably low 1.4% at Johns Hopkins, which is significantly lower than the rate at other institutions at which this operation is less frequently performed. The operation has been shown to be safe and effective in patients ranging in age up to their 80's, in whom the morbidity and mortality rates approach those of younger patients. Current studies include a prospective randomized trial assessing the effects of additional extended retroperitoneal lymphadenectomy (radical Whipple resection) on the morbidity/mortality of the operation, and the survival of these patients.

Other studies have examined the clinical utility of biochemical tumor markers in following patients at risk for bile duct cancer. A study of patients with primary sclerosing cholangitis found that elevated serum CA19-9 levels (CA19-9 is a blood marker) accurately predicted the presence of cholangiocarcinoma. All patients who proved to have cholangiocarcinoma had elevated serum CA19-9, while no patient without cholangiocarcinoma had an elevated serum CA19-9. Another study found that the levels of carcinoembryonic antigen (CEA) excreted in the bile was elevated in patients with cholangiocarcinoma and in those with intrahepatic gallstones, compared to those with benign bile duct strictures. Those patients with primary sclerosing cholangitis or choledochal cysts had intermediate levels of biliary CEA.

Bile duct cancers have been traditionally considered among the most resistant of all cancers to chemotherapy. Oncologists throughout the world are searching for new agents which will prove more effective. One recent study found that therapy with 5-Fluorouracil combined with subcutaneous injection of interferon 2-alpha produced some effect, though the results are preliminary. Other methods of delivering as much therapy as possible to the tumor while minimizing systemic toxicity, such as intra-arterial pumps, are being explored.

It has become apparent over the past decade that cancer is, in essence, a genetic disease. Cancer results from an accumulation of changes in the DNA of a specific cell, giving that cell the ability to grow uncontrollably and spread. It is hoped that the study of the genetic alterations associated with biliary tract cancers will serve many purposes. First, this knowledge may help us choose better conventional chemotherapeutic agents that will be more effective against these tumors. If we learn what makes these cells grow, we should be better able to choose drugs that block this process. Second, knowledge of these genetic changes might eventually allow for these tumors to be treated by reinsertion of normal DNA into the cells ("gene therapy"). For example, if a gene which normally blocks cell proliferation is mutated and nonfunctional in these tumors, restoration of a normal copy of this gene might induce the cell to stop growing and eliminate its malignant potential. Third, it may be possible to use our knowledge of these genetic alterations to design diagnostic tests that are more sensitive and specific than the conventional means of diagnosis outlined above. This would allow potentially lethal tumors to be caught earlier when they can be cured by surgery.

Significant progress has been made at Johns Hopkins in these areas. A study published in Clinical Cancer Research examined the utility of detecting mutations in the K-ras gene in biliary brush cytology specimens for the diagnosis of bile duct carcinomas. The K-ras gene codes for a growth factor that has been shown to be mutated in the majority of bile duct cancers. The mutations are thought to confer a growth advantage to the cells affected. The study showed that the mutations in this gene could be detected in the same specimens collected for conventional cytology, using highly sensitive polymerase chain reaction (PCR) technology. [PCR is a laboratory technique that allows researchers to take a very small sample of DNA and amplify (copy) the sample exponentially. The result of the PCR is a sample large enough to study by genetic testing.] The sensitivity of the genetic test proved to be in fact greater than that of conventional cytology in this series (42% versus 36%), and used together the sensitivity of the combined tests rose to 62%. These results support the idea that genetic tests will soon play a major role in the diagnosis of biliary tract cancer.

Another study, published in Genes, Chromosomes, and Cancer, analyzed the overall genetic composition of distal bile duct cancers using two complimentary techniques. The first, conventional cytogenetics, is performed on fresh tumors, and requires intact dividing cells ("metaphases"). The second technique, comparative genomic hybridization, can be performed on frozen, non-dividing tumor cells and compares the chomosome content of the tumor with that of a normal cell. Both techniques allow gross gains and losses of genetic material to be detected. The analysis of bile duct tumors revealed frequent loss of genetic material on the long arms of chromosomes 18, 6, and 12 (18q, 6q, and 12q), and the short arms of chromosomes 10, 8, and 17 (10p, 8p, and 17p). Of great interest is the fact that many of these changes are similar to those previously identified in pancreatic cancers, suggesting that these two tumors share a number of genetic changes. Some of the sites identified suggest specific genes that might be involved in bile duct cancers. For example, the p53 gene is located on chromosome 17p and is known to be mutated in the majority of pancreatic cancers. Several small studies have strongly suggested that p53 is involved in bile duct cancers, and the frequent deletion of 17p in this study supports this notion. We are currently working to confirm this hypothesis, and identify the other genes involved.