Methods

Top Methods Results Discussion Author & Article Info References

Study Design

For this prospective, double-blind study, pancreatic juice and brush washings were collected and sampled in the Department of Endoscopy at Centre Hospitalier de Rangueil, Toulouse, France. K-ras mutation identification was done in the Unite 151, Institut National de la Sante et de la Recherche Medicale, Toulouse, France. To avoid bias, the clinical diagnoses and molecular results were compared after analysis had been completed and diagnosis had been clearly assessed. All patients had at least biannual clinical follow-up.

Patients

After informed consent was obtained, endoscopic ductal aspiration of cells or brush cytology was done on all patients (n = 141) referred to our study group who had endoscopic retrograde cholangiopancreatography (complete procedure with both cholangiography and pancreatography) for diagnostic or therapeutic purposes between January 1991 and June 1994. The main symptoms were obstructive jaundice (n = 47), cholangitis (n = 29), abdominal pain (n = 27), acute pancreatitis (n = 16), loss of weight (n = 8), residual stone of the common bile duct (n = 8), and miscellaneous (n = 6). During the study, we excluded 66 patients. Forty-nine were excluded at the time endoscopy was done (13 because of failure of the endoscopic procedure, 6 because of tumor-induced obstruction of the main pancreatic duct, 8 because of unsuccessful K-ras analysis, and 22 because they had biliary, ampullary, or neuroendocrine tumors), and 17 were excluded during the follow-up period (6 because it was impossible to determine the pancreatic origin of the tumor, 8 because follow-up was too short to exclude pancreatic cancer, and 3 because they were lost to follow-up). The 75 remaining patients were classified into 3 groups according to standard clinical and pathologic criteria (Table 1).

Group 1 (n = 24) was composed of patients with no pancreatic disease. Endoscopic retrograde cholangiopancreatography was done for biliary therapeutic indications (20 patients had cholangitis; 4 had residual stone of the common bile duct). Each patient had a normal pancreatogram and no pancreatic mass on at least three imaging techniques. The clinical course (mean, 22 months; range, 6 to 38 months) was free of any pancreatic clinical events.

Group 2 (n = 29) was composed of patients with nontumoral pancreatic disease. Ten patients had biliary acute pancreatitis, 13 had chronic pancreatitis shown by abdominal pain, 5 had chronic pancreatitis shown by loss of weight, and 1 had recurrent drug-induced pancreatitis. The subsequent clinical course (mean, 19 months; range, 6 to 40 months) was in agreement with the nonmalignant nature of the diseases.

Group 3 (n = 22) was composed of patients who had or were developing a pancreatic tumor. Eighteen patients had obstructive jaundice, 3 patients had abdominal pain and loss of weight, and 1 patient had acute pancreatitis. All but 2 patients had a pancreatic mass on at least one imaging technique when K-ras was first studied (see Results). Diagnosis of pancreatic cancer was based on the clinical data shown in Table 1.

Study Methods

Intraductal aspiration of pure pancreatic juice or cytologic brushing was done with fluoroscopic guidance. Samples were collected in sterile tubes and stored at –20°C.

Pancreatic juice was centrifuged, and the DNA in the pellet was extracted according to the method of Saiki [9]. Polymerase chain reaction amplifications were done as described by Jiang and colleagues [10]. Aliquots of polymerase chain reaction samples were digested using BstN1 (New England Biolabs, Beverly, Massachusetts) (Figure 1). Direct sequencing of amplified polymerase chain reaction product was done using a double-stranded DNA cycle sequencing system (BRL, Gaithersburg, Maryland).

View larger version (32K): [in this window] [in a new window]   Figure 1. Detection of K-ras 12th codon mutation in pancreatic juice by polymerase chain reaction-mediated restriction length fragment polymorphism analysis. Polymerase chain reaction amplification produced a 165-bp fragment that contained a BstN1 cleavage site. This site was absent in mutant amplified K-ras fragment. Arrows indicate the positions of the nondigested (165 bp) and the digested wild-type (136 bp) fragment. (1, 1',2, 2') = normal K-ras gene fragment; (3, 3') = DNA sample showing both the mutant (165 bp) and the wild-type (136 bp) fragments after BstN1 digestion.

Results

Top Methods Results Discussion Author & Article Info References

All patients from groups 1 and 2 (n = 53) had normal sequences for the K-ras 12th codon. The polymerase chain reaction amplification product was found to be partly resistant for 2 patients who had no evidence of pancreatic cancer (follow-up, 15 and 32 months, respectively). Seventeen patients with pancreatic carcinoma (17 of the 22 patients in group 3) had an activating point mutation of the K-ras 12th codon.

Two patients had no evidence of pancreatic cancer at first examination, but were found to have a mutation of the K-ras 12th codon. Both developed pancreatic tumors several months later.

One of these two patients (male, 65 years of age) presented with chronic abdominal pain and narrowing of the main pancreatic duct on pancreatography, which was associated with a pseudocyst of the pancreatic tail (diameter, 4 cm). Preoperative explorations and laparotomy, including surgical biopsy of the pancreas, showed no evidence of malignancy. A cystojejunostomy was done, and the disease was ascribed to nonalcoholic chronic pancreatitis; however, K-ras was subsequently found to be mutated (GGTGCT). After a pain-free period of 1 year, pain recurred. After another 6 months, the patient had a second surgical examination because of acute hemorrhage related to the pseudocyst; this examination was followed by resection of the pancreatic tail. Macroscopic examination of the pancreas still favored nonmalignant disease, whereas histopathologic analysis showed a 1.5-cm anaplastic carcinoma. The patient died of recurrent hemorrhage on the sixth day after surgery.

The other patient (male, 71 years of age) presented with acute pancreatitis unrelated to alcohol abuse or any other cause (idiopathic pancreatitis). Endoscopic retrograde cholangiopancreatography showed slight narrowing of the main pancreatic duct, whereas other imaging techniques, including computed tomography and endosonography, showed normal results. Cytologic examination of pure pancreatic juice showed it to be normal. Again, the K-ras 12th codon was found to be mutated (GGTGTT). Clinical follow-up had been free of any pancreatic events for 40 months, at which time the patient presented with obstructive jaundice. Endoscopic retrograde cholangiopancreatography showed an extensive stricture of the main pancreatic duct, with distal dilatation and short stricture of the common bile duct. Cytologic examination of pure pancreatic juice confirmed the diagnosis of pancreatic cancer. The patient had had cardiac bypass surgery in 1992; thus, his general status did not allow pancreatic resection. He is still alive 10 months later.

Discussion

Top Methods Results Discussion Author & Article Info References

Our findings confirm that samples collected during endoscopic retrograde cholangiopancreatography contain enough DNA for polymerase chain reaction amplification and point mutation characterization. The overall sensitivity of the test is within the range of previous studies [2-7].

Our study provides the first direct evidence that K-ras mutations occur early in the development of pancreatic cancer: Two patients were found to have mutations 18 and 40 months, respectively, before the clinical diagnosis of pancreatic tumor was made. Recent studies suggest that K-ras mutation is an early event in pancreatic carcinoma [11, 12] and that it may occur years before clinical presentation in both lung and colon carcinoma [13, 14].

Further studies need to be done in larger groups of patients to determine the diagnostic value of this test. The size of our cohort does not reflect the proportions seen in gastroenterology practices in the general population, because we are a specialized biliopancreatic endoscopy unit. Two groups of patients are overrepresented in our cohort: 1) old patients with obstructive jaundice who were unfit for palliative surgical treatment and were sent for biliary stenting and 2) patients sent for differential diagnosis of chronic pancreatitis and pancreatic carcinoma. We would recommend this invasive test not for screening, but as a complement to radiologic and pathologic techniques in selected subgroups of patients. Several groups of patients could potentially benefit from this test: Epidemiologic studies indicate that chronic pancreatitis [15], non-insulin-dependent diabetes of recent onset [16], and idiopathic acute pancreatitis (one case in our study) may be associated with pancreatic cancer. Differentiating pancreatic cancer from chronic pancreatitis is still difficult [1] and might be made easier by the use of K-ras mutation identification. Patients with the mutation would be identified and would have close radiologic follow-up adapted to their ability to withstand surgical resection. Because the prognosis of pancreatic cancer is clearly related to tumor size [17], such patients could benefit from curative pancreatic resection.

In two patients in our study, K-ras mutations were found to be present early in the development of pancreatic cancer. No pancreatic mass was identified in either of these patients, even though surgical exploration was done in one of them. K-ras mutation is an early event in pancreatic carcinogenesis and has been identified in intraductal and in situ carcinoma [11, 12]. However K-ras mutations have also been reported in pancreatic ductal mucinous hyperplasia, which is considered to be a preneoplastic lesion. Ductal mucinous hyperplasia is present in chronic pancreatitis and is known to have a low rate of transformation into cancer (1.8% [CI, 1% to 2.6%]) 10 years after the diagnosis of chronic pancreatitis [15].

Whether the pancreas should be resected when this mutation is discovered has not been settled. Further observations similar to those reported here are needed before we can recommend surgical resection of a mass-free pancreas presenting an oncogenic activation process.