The aims of this review were to determine whether positive peritoneal lavage cytology (CY+) precludes radical resection in pancreatic cancer and to propose prospections for future studies.

MEDLINE, Embase, and Cochrane Central were searched for related articles. Dichotomous variables and survival outcomes were analyzed with the estimation of odds ratio and hazards ratio (HR), respectively.

A total of 4905 patients were included, of which 7.8% were CY+. Positive peritoneal lavage cytology was correlated with poor overall survival (univariate survival analysis [HR, 2.35; P < 0.00001]; multivariate analysis [HR, 1.62; P < 0.00001]), poor recurrence-free survival (univariate survival analysis [HR, 2.50; P < 0.00001]; multivariate analysis [HR, 1.84; P < 0.00001]), and higher initial peritoneal recurrence rate (odds ratio, 5.49; P < 0.00001).

Although CY+ predicts poor prognosis and a higher risk of peritoneal metastasis after curative resection, it is not sufficient to preclude curative resection based on the current evidence, and high-quality trials should be conducted to assess the prognostic impact of operation among resectable CY+ patients. In addition, more sensitive and accurate methods to detect peritoneal exfoliated tumor cells and more effective comprehensive treatment for resectable CY+ pancreatic cancer patients are clearly warranted.

The aims of this study were to clarify the type of intrapancreatic spread of cancer of the pancreatic body and tail and to assess whether a 2-cm transection margin is adequate to ensure negative margins.

We selected 66 patients who underwent distal pancreatectomy for cancer of the pancreatic body and tail. We investigated intrapancreatic cancer spread in these patients histopathologically and analyzed the relationship between 2-cm-margin positivity and other clinicopathological characteristics.

Two-centimeter-margin positivity was observed in 17 cases. In these, tumors had a tendency to spread toward the pancreatic head along the main pancreatic duct. As a result of statistical analysis, we considered venous invasion (odds ratio [OR], 15.48; 95% confidence interval [CI], 1.61-148.94; P = 0.0177), 2-cm-margin fibrosis (OR, 173.88; 95% CI, 8.96-3375.03; P = 0.0007), and 2-cm-margin hardness (OR, 5.97; 95% CI, 1.07-33.46; P = 0.0420) as being independently related to 2-cm-margin positivity.

The results suggest that 2 cm is not a safe length to ensure a negative margin. In the future, preoperative and intraoperative evaluation of the degree of fibrosis of pancreatic parenchyma could lead to cancer-free pancreatic cut-end margins.

Despite the development of more sophisticated diagnostic techniques, it remains difficult to detect pancreatic carcinoma in the early stage. However, the resection rate has been increasing due to recent advances in surgical techniques and the application of extensive surgery. In 1995, statistics of the Pancreatic Cancer Register Committee of Japan Pancreas Society showed a 44.5% rate of resection. Further, according to Japan Pancreas Society staging, the 5-year survival rate of Stage I, II, III, and IV after surgical resection was 46.3, 27.5, 20.4, and 8.3%, respectively. This paper will review the Japanese experience of pancreatic carcinoma and several problems in pancreatic cancer surgery.

Preoperative and postoperative serum levels of carbohydrate antigen-(CA)19-9 in 148 patients with carcinoma of the pancreas were studied. All 18 patients with carcinoma of the pancreas of Stage I, II, and III were resectable, and their pre-operative serum levels of CA-19-9 were under 1,344 U/ml. Pre-operative serum levels of CA-19-9 in patients with Stage IV ranged widely between 5 and 32,240 U/ml. The postoperative survival rate was significantly superior in patients (n = 15) whose CA-19-9 pre-operative serum levels were less than 2,000 U/ml compare to those (n = 64) whose levels were over that level. Fifteen resectable patients who showed serum levels of CA-19-9 over 2,000 U/ml pre-operatively died within 2 years postoperatively due to recurrence, especially by liver metastasis in spite of aggressive surgery. Intraoperative quick immunostaining of CA-19-9 and carcinoembryonic antigen (CEA) was useful to diagnose intrapancreatic carcinoma development on frozen sections of cut margin of the pancreas and also useful in abdominal washing cytology combined with conventional staining.

The rapid immunocytochemical staining of gastrointestinal cancer-associated antigens, carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9), for intraoperative peritoneal washing cytology of pancreatic or biliary carcinoma can be completed in only 25 minutes by using microwave irradiation to accelerate incubation of the primary antibody. Only 3-second irradiation at 500 W for fresh smears produced specific antigen staining of malignant cells similar to that obtained with longer incubation, and the sensitivity of intraoperative cytology was greater than by the conventional staining method. The rapid immunocytochemical staining method of peritoneal washings at laparotomy can provide sensitive information for determining more aggressive therapy for intraperitoneal spread of malignant cells before closing the abdominal wall.

The purpose of the present study was to clarify the indication of aggressive surgery for pancreatic head carcinoma.

Laparatomy was performed in 153 patients with carcinoma, 101 of whom underwent resection of the carcinoma. With histologic examination the degree of carcinoma invasion into the portal vein was classified into grades 0, I, or II according to the depth of invasion by the carcinoma. Macroscopic carcinoma invasion into portal vein was classified into types A, B, C, or D according to preoperative findings on the portal phase of superior mesenteric angiography or intraoperative portography.

Macroscopic findings correlated with the histologic invasion grades. The 1-year survival rate was 39.6% in grade 0, 11.3% in grade I, and 5.5% in grade II cases. The survival rates of patients with type A (p < 0.01), B (p < 0.05), and C invasion (p < 0.01) were higher than those of patients who did not undergo resection; however, no significant difference in the survival rates between patients who did not undergo resection and patients with type D invasion was observed.

For locally advanced carcinoma of the pancreatic head or entire pancreas, patients with type D invasion have no indication of aggressive surgery.

Seven murine monoclonal antibodies (MoAb) detecting blood group antigens of the Lewis system and their sialylated derivatives were used to define their immunohistochemical distribution in normal and malignant pancreatic tissues. The specific blood group antigens studied included Lewis(a) (Le(a)), Lewisb (Leb), Lewisx (Le(x)), Lewis(y) (Le(y)), sialyl-Lewis(a) (s-Le(a)), sialyl-lacto-N-tetraose (s-LNT), and sialyl-Lewisx (s-Le(x)).

The expression of these antigens was analyzed by immunoperoxidase technique in pancreatic tissue of patients with (n = 27) and without (n = 19) pancreatic cancer. The genetic background of their secretor status and their Lewis phenotypes were determined by the enzyme-linked immunosorbent assay using purified salivary glycoprotein and MoAb against eight different blood group-related antigens (Le(a), Leb, Le(x), Le(y), H1, H2, A, B2), and the putative genotypes of the patients were classified as follows: SeLe, Sele, seLe, and sele.

The following results were obtained: (1) In normal pancreas, Le(a), s-Le(a), s-LNT, and s-Le(x) were expressed in ducts, but their expression was restricted to the luminal side of the cytoplasm (cytoplasmic type with positive polarity). Leb and Le(y) showed similar patterns of expression in ducts and some parts of the acinus. In contrast, Le(x) was absent in most cases. (2) Expression of Leb was observed in normal pancreatic ducts in all 46 patients, regardless of their genotypes. Le(a) and s-Le(a) were not expressed in specimens from patients with the seLe and sele saliva phenotypes. (3) In pancreatic cancer, the following antigens were expressed cytoplasmically in the proportion of cases indicated: Le(a) (78%), Leb (85%), Le(x) (33%), Le(y) (44%), s-Le(a) (78%), s-LNT (89%), and s-Le(x) (85%). They also were detected in the surrounding stroma. This pattern of expression is distinct from that found in normal pancreatic tissue. Patients with Sele and sele genotypes did not express Le(a) or s-Le(a), except in one case. (4) Serum levels of carbohydrate antigen 19-9 (CA 19-9) were not elevated in patients with pancreatic cancer with the Sele and sele genotypes but were elevated in most patients with SeLe and seLe genotypes.

The overall findings indicate that Lewis-related antigens act as pancreatic tissue-related antigens, depending in part on salivary phenotypes of the patient. With proper antigen selection and the determination of secretor status, these anti-blood group MoAb may be of clinical utility in the diagnosis of pancreatic cancer.

Paraffin wax embedded, formalin fixed sections of 22 adenocarcinomas of the exocrine pancreas were stained with four mouse monoclonal antibodies: DD9-E7, an antibody raised against a human pancreatic tumour xenograft; carcino-embryonic antigen (CEA); epithelial membrane antigen (EMA); and cytokeratin (CAM 5.2). An indirect immunoperoxidase technique without enzyme pre-digestion and an affinity-purified sheep anti-mouse peroxidase conjugate were used. All of the tumours were positive for DD9-E7, EMA, and CAM 5.2. Twenty out of 22 were focally positive for CEA and the staining was often weak. As all of these adenocarcinomas were DD9-E7 positive, absence of staining for DD9-E7 in a tumour makes the diagnosis of adenocarcinoma of the exocrine pancreas very unlikely, and this is of value in distinction from endocrine carcinomas with a marked acinar pattern. The weak CEA staining distinguished pancreatic carcinomas from colorectal tumours. Because the distribution of staining for EMA and CAM 5.2 was no different from that previously seen in adenocarcinomas from other sites, these markers are likely to be of limited value in the differential diagnosis of abdominal adenocarcinomas of uncertain origin.

The rapid immunostaining of gastrointestinal cancer-associated antigens, CA19-9, CEA, DUPAN2, and CA50, for intraoperative pathological diagnosis of pancreatic cancer can be completed in only 13 min by applying microwave irradiation, which was used to accelerate the incubation of the primary antibody. Only 3 s irradiation at 500 W for fresh-frozen sections produced specific antigen staining of greater intensity than that obtained with longer incubation, and tissue preservation was satisfactory with minimal non-specific background staining. This rapid immunoperoxidase staining method on the cryostat section of pancreatic biopsy specimens can provide important information for determining an appropriate operative method for pancreatic cancer patients.