Peritoneal washing cytology combined with immunocytochemical staining and detecting mutant K-ras in pancreatic cancer: comparison of the sensitivity and availability of various methods.

Peritoneal metastases are the second most common site of involvement, following the liver, in pancreatic cancer. Thus, we performed peritoneal washing cytology at laparotomy to diagnose accurately the intraperitoneal spread of carcinoma cells to determine the appropriate therapy. Peritoneal washings were collected at laparotomy from 20 Japanese pancreatic carcinoma patients at Nagoya University Hospital between April 1993 and December 1994. From centrifuged deposits, we examined the cytology by three methods as follows. The first method was conventional cytology, including May-Grünwald and Giemsa, Papanicolaou, periodic acid-Schiff, and Alcian blue. The second method was immunocytochemical staining, using antibodies to carbohydrate antigen (CA19-9) and carcinoembryonic antigen. After extracting DNA from the remaining pellet, we studied the last method, detecting K-ras point mutation, by two-step polymerase chain reaction and restriction fragment length polymorphism analysis. In two cases, peritoneal metastases were macroscopically recognized, and the results of all three methods were positive. In the two other cases, where peritoneal dissemination was not macroscopically recognized, the judgments of conventional cytological study and detecting K-ras point mutation were negative. However, a few malignant cells were found by the immunocytochemical staining method. Judging from their clinical course, the positively stained cells were suggestive of malignancy. At present, the immunocytochemical staining method is the most sensitive of these three methods in peritoneal washing cytology. However, preserving DNA is suitable for repeated examination, and a modified method can be applied. If the sensitivity increases, the method of detecting K-ras has the potential to become the standard for peritoneal washing cytology in pancreatic cancer.