[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Download PDF
Figure 1.
Overall survival (A) and overall peritoneal metastases-free survival (B) rates for all patients with positive (n = 36) and negative (n = 111) cytologic findings (P<.001 for both).

Overall survival (A) and overall peritoneal metastases-free survival (B) rates for all patients with positive (n = 36) and negative (n = 111) cytologic findings (P<.001 for both).

Figure 2.
Survival (A) and peritoneal metastases-free survival (B) rates for patients in the resectable group with positive (n = 5) and negative (n = 56) cytologic findings (P<.001 for both).

Survival (A) and peritoneal metastases-free survival (B) rates for patients in the resectable group with positive (n = 5) and negative (n = 56) cytologic findings (P<.001 for both).

Figure 3.
Survival (A) and peritoneal metastases-free survival (B) rates for patients in the locally advanced group with positive (n = 8) and negative (n = 45) cytologic findings (P = .64 and .30, respectively).

Survival (A) and peritoneal metastases-free survival (B) rates for patients in the locally advanced group with positive (n = 8) and negative (n = 45) cytologic findings (P = .64 and .30, respectively).

Figure 4.
Survival (A) and peritoneal metastases-free survival (B) rates for patients in the visible metastases group with positive (n = 23) and negative (n = 10) cytologic findings (P = .17 and .18, respectively).

Survival (A) and peritoneal metastases-free survival (B) rates for patients in the visible metastases group with positive (n = 23) and negative (n = 10) cytologic findings (P = .17 and .18, respectively).

Figure 5.
Survival (A) and peritoneal metastases-free survival (B) rates for patients with positive cytologic findings in each group. Survival times in the locally advanced group were significantly better than those in the other groups.

Survival (A) and peritoneal metastases-free survival (B) rates for patients with positive cytologic findings in each group. Survival times in the locally advanced group were significantly better than those in the other groups.

Table 1. 
Patient Distribution and Positive Cytologic Findings by Intraoperative Findings of Cancer Spread
Patient Distribution and Positive Cytologic Findings by Intraoperative Findings of Cancer Spread
Table 2. 
Pathologic Stage and Curability in the Resectable Group*
Pathologic Stage and Curability in the Resectable Group*
Table 3. 
Type of Therapy in the Locally Advanced and Visible Metastases Groups*
Type of Therapy in the Locally Advanced and Visible Metastases Groups*
1.
Griffin  JFSmalley  SRJewel  W Patterns of failure after curative resection of pancreatic carcinoma. Cancer. 1990;6656- 61Article
2.
Vogel  PRuschoff  JKummel  S  et al.  Prognostic value of microscopic peritoneal dissemination: comparison between colon and gastric cancer. Dis Colon Rectum. 2000;4392- 100Article
3.
Wu  C-CChen  J-TChang  M-C  et al.  Optimal surgical strategy for potentially curable serosa-involved gastric carcinoma with intraperitoneal free cancer cells. J Am Coll Surg. 1997;184611- 617
4.
Creaseman  WTRutledge  F The prognostic value of peritoneal cytology in gynecologic malignant disease. Am J Obstet Gynecol. 1971;110773- 781
5.
Turner  DAGershenson  DMAtkinson  NSneige  NWharton  AT The prognostic significance of peritoneal cytology for stage I endometrial cancer. Obstet Gynecol. 1989;74775- 780
6.
Sobun  LHedWittekind  Ced UICC TNM Classification of Malignant Tumors. 5th ed. New York, NY Wiley-Liss Inc1997;87- 90
7.
Warshaw  AL Implications of peritoneal cytology for staging of early pancreatic cancer. Am J Surg. 1991;16126- 29Article
8.
Merchant  NBConlon  KCSaigo  PDougherty  EBrennan  MF Positive peritoneal cytology predicts unresectability of pancreatic adenocarcinoma. J Am Coll Surg. 1999;188421- 426Article
9.
Jimenez  REWarshaw  ALCastillo  CF Laparoscopy and peritoneal cytology in the staging of pancreatic cancer. J Hepatobiliary Pancreat Surg. 2000;715- 20Article
10.
Martin  JKGoellner  JR Abdominal fluid cytology in patients with gastrointestinal malignant lesions. Mayo Clin Proc. 1986;61467- 471Article
11.
Lei  SKini  JKim  KHoward  JM Pancreatic cancer: cytologic study of peritoneal washings. Arch Surg. 1994;129639- 642Article
12.
Leach  SDRose  JALowy  AM  et al.  Significance of peritoneal cytology in patients with potentially resectable adenocarcinoma of the pancreatic head. Surgery. 1995;118472- 478Article
13.
Makary  MAWarshaw  ALCenteno  BAWillett  CGRattner  DWCastillo  CF Implications of peritoneal cytology for pancreatic cancer management. Arch Surg. 1998;133361- 365Article
14.
Nakatsuka  AYamaguchi  KShimizu  S  et al.  Positive washing cytology in patients with pancreatic cancer indicates a contraindication of pancreatectomy. Int J Surg Invest. 1999;1311- 317
15.
Juhl  HStritzel  MWroblewski  A  et al.  Immunocytology detection of micrometastatic cells: comparative evaluation of findings in the peritoneal cavity and the bone marrow of gastric, colorectal and pancreatic cancer patients. Int J Cancer. 1994;57330- 335Article
16.
Nomoto  SNakao  AKasai  Y  et al.  Peritoneal washing cytology combined with immunocytochemical staining and detecting mutant K-ras in pancreatic cancer: comparison of the sensitivity and availability of various methods. Pancreas. 1997;14126- 132Article
17.
Castillo  CFWarshaw  AL Laparoscopic staging and peritoneal cytology. Surg Oncol Clin N Am. 1998;7135- 142
Original Article
April 2002

Prognostic Value of Cytologic Examination of Peritoneal Washings in Pancreatic Cancer

Author Affiliations

From the Department of Surgery, National Cancer Center Hospital East, Kashiwa, Japan.

Arch Surg. 2002;137(4):475-480. doi:10.1001/archsurg.137.4.475
Abstract

Hypothesis  Findings from cytologic examination of peritoneal washings affect the staging, prognosis, and management of pancreatic cancer.

Design  Retrospective review.

Setting  Tertiary care center.

Patients  A total of 151 patients with invasive ductal adenocarcinoma of the pancreas and without distant metastases or ascites, as determined by preoperative radiologic examinations, underwent peritoneal washings. Based on intraoperative findings, patients were subdivided into 3 groups according to the extent of cancer spread: 65 were assigned to a resectable group, 53 to a locally advanced group, and 33 to a visible metastases group.

Main Outcome Measures  Survival and peritoneal metastases-free survival.

Results  The incidence of positive cytologic findings was 23.8% (36/151). Positive rates increased significantly with disease progression (P<.001). In the resectable group, survival and peritoneal metastases-free survival were significantly shorter for patients with positive vs negative cytologic findings. In the remaining 2 groups, no differences were observed between patients with positive vs negative cytologic findings in survival or peritoneal metastases-free survival. In patients with positive cytologic findings without visible metastases, survival and peritoneal metastases-free survival were significantly better in the locally advanced group undergoing chemoradiotherapy than in the resectable group.

Conclusions  Positive peritoneal cytologic findings are not independent factors that determine survival and peritoneal metastases-free survival; rather, they are associated with advanced disease. In patients with visible metastases, cytologic factors are not correlated with survival or peritoneal recurrence. In patients without visible metastases, chemoradiotherapy may be beneficial for those with positive cytologic findings.

DESPITE IMPROVED preoperative staging that excludes patients with distant macroscopic metastases, more than 80% of patients with pancreatic cancer undergoing curative resection die within 5 years of surgery. Because the peritoneum is one of the most frequent sites of metastasis in pancreatic cancer,1 it seems likely that most of these patients have microscopic metastases at the time of surgery. However, it is impossible to detect microscopic lesions in the peritoneum, and even small, visible lesions can go undetected. This results in the presence of free cancer cells in the peritoneal cavity that later cause tumors to spread throughout the peritoneum. Exfoliated cancer cells are detectable by cytologic examination of peritoneal washings, and the procedure has been widely used in the staging of ovarian, endometrial, and gastric malignant tumors.25 The role of exfoliated cancer cells in pancreatic cancer is poorly understood. Therefore, we attempted to assess the suitability and efficacy of cytologic examination of peritoneal washings with respect to the staging, prognosis, and management of pancreatic cancer.

PATIENTS AND METHODS
PATIENTS

Of 313 patients with invasive ductal adenocarcinoma of the pancreas, 165 underwent laparotomy at National Cancer Center Hospital East between July 1, 1992, and December 31, 2000. These 165 patients were without distant metastases or ascites, as determined by preoperative radiologic examinations (ultrasonography and dynamic, contrast-enhanced computed tomography). Peritoneal washings were performed in 151 patients (85 men and 66 women; mean age, 61 years; age range, 31-83 years) to investigate the prevalence of malignant cells within the peritoneal cavity. Fourteen patients were excluded from the study because of adhesions in the lower peritoneal cavity. Ninety-six patients had pancreatic tumors in the head and 55 had pancreatic tumors in the body and tail. All patients were confirmed to have adenocarcinoma of the pancreas by histologic testing of a resected specimen or needle biopsy at laparotomy. None of the patients underwent antecedent fine-needle aspiration.

Based on the intraoperative findings, the 151 patients were subdivided into 3 groups according to the extent of cancer spread (Table 1). The resectable group comprised 65 patients who had no detectable distant metastases or major vessel invasion; 48 underwent pylorus-preserving pancreatoduodenectomy, 16 underwent distal pancreatectomy, and 1 underwent total pancreatectomy. The locally advanced group comprised 53 patients who had major vessel invasion; all underwent intraoperative (2500 rad [25 Gy]) and external beam (4000 rad [40 Gy]) irradiation with or without 5-fluorouracil. The visible metastases group comprised 33 patients confirmed to have hepatic metastases, visible peritoneal metastases, or both; 14 had hepatic metastases, 14 had peritoneal metastases, and 5 had both hepatic and peritoneal metastases. All 33 patients underwent intraoperative irradiation (2500 rad [25 Gy]) with or without external beam irradiation (4000 rad [40 Gy]) and 5-fluorouracil. The cytologic interpretation of peritoneal washings did not affect the treatment strategy in any of the 151 patients.

Patient follow-up included physical examination, abdominal computed tomography, and ultrasonography every 3 months or until a diagnosis of disease recurrence was made. The presence of peritoneal metastases was confirmed by the development of large amounts of ascites, the presence of peritoneal nodules on radiologic examination, or histocytologic findings.

CYTOLOGIC EXAMINATION

Specimens for cytologic examination were obtained at the beginning of laparotomy. On opening the abdomen, 100 mL of isotonic sodium chloride solution was instilled into the pouch of Douglas. After manual agitation, the washings were retrieved by aspiration. After centrifugation for 3 minutes, direct smears were prepared and fixed in 95% ethanol. Two slides were prepared for each patient and were stained using the Papanicolaou method. All slides were reviewed by 3 cytopathologists.

DATA ANALYSIS

To confirm the effectiveness of cytologic examination of peritoneal washings for detecting invisible microperitoneal dissemination in patients with pancreatic cancer, survival and peritoneal metastases-free survival were analyzed in 147 patients; 4 patients whose deaths were related to resection were excluded. Because survival is affected by tumor spread and type of therapy, we compared these rates for each group according to intraoperative findings of cancer spread. Moreover, we compared patients with positive and negative cytologic findings in the resectable group with respect to tumor stage and curability, which were classified using the Union Internationale Contre le Cancer TNM classification criteria.6 In the locally advanced and visible metastases groups, type of therapy was compared between patients with positive and negative cytologic findings. The distribution of the patients was measured using the χ2 test. Survival curves were calculated using the Kaplan-Meier method. In peritoneal metastases-free survival, patients without peritoneal metastases at death were dealt with as censored cases. The differences between curves were measured using the log-rank test. P<.05 was defined as statistically significant.

RESULTS

The incidence of positive cytologic findings was 23.8% (36/151). The positive cytologic rate for each group divided by intraoperative findings of cancer spread was 8% (5/65) of the patients in the resectable group, 15% (8/53) in the locally advanced group, and 70% (23/33) in the visible metastases group. Positive cytologic findings occurred in 11.0% (13/118) of the patients without visible metastases. Positive rates increased significantly with disease progression (P<.001) (Table 1).

Overall survival and overall peritoneal metastases-free survival were significantly worse for patients with positive vs negative cytologic findings (P<.001 for both) (Figure 1). Because positive cytologic findings were closely correlated with disease progression, survival and peritoneal metastases-free survival were compared in each group. In the resectable group, no significant differences were observed between patients with positive vs negative cytologic findings in either tumor stage or curability (Table 2). In the same group, survival and peritoneal metastases-free survival for patients with positive cytologic findings were significantly worse than for those with negative findings (Figure 2). However, in the remaining 2 groups, no differences in either survival or peritoneal metastases-free survival were observed between patients with positive vs negative cytologic findings (Figure 3 and Figure 4). In addition, the type of therapy in these 2 groups did not differ significantly between those with positive vs negative cytologic findings (Table 3).

In patients with positive cytologic findings without visible metastases, survival and peritoneal metastases-free survival for those in the locally advanced group were significantly better than for those in the resectable group (Figure 5). Survival for patients with positive cytologic findings in the resectable group was similar to that for patients with positive findings in the visible metastases group (median survival, 5.2 vs 5.1 months). In contrast, survival in those with positive vs negative cytologic findings in the locally advanced group was similar (median survival, 11.1 vs 9.9 months).

COMMENT

Given that microscopic occult peritoneal metastases are believed to precede the appearance of malignant cells in the abdominal cavity, cytologic examination of peritoneal washings has been widely used in the staging of gastric and ovarian malignancies.24 To reduce the number of open laparotomies in patients with unresectable disease, the combination of laparoscopy and cytologic examination of peritoneal washings is frequently performed for the staging of pancreatic cancer.79 However, the role of peritoneal cytologic examination in pancreatic cancer is controversial.

Previous studies812 have reported that malignant cells are present in 7% to 21% of peritoneal washings obtained from patients with potentially resectable pancreatic cancer. We detected malignant cells in a slightly higher percentage (23.8%) of cytologic findings from patients without distant metastases during preoperative examinations, which is likely because our data included patients who were preoperatively diagnosed as having locally advanced cancer. In previous studies, positive cytologic findings were an indicator of unresectable, aggressive disease characterized by early metastasis and short survival. These studies also determined that patients with negative cytologic findings had a better prognosis. Jimenez et al9 proposed that patients with positive peritoneal cytologic findings be classified as M1 in the TNM system. However, positive cytologic findings are confirmed to be closely correlated with disease progression, with positive findings occurring in 45% of patients with visible metastases but in only 14% of those without.9 Our findings are similar in that positive cytologic findings occurred in 70% of our patients with visible metastases but in only 11.0% of cytologic findings from those without. Moreover, positive cytologic findings are rare (8%) in patients with potentially resectable pancreatic cancer. Because most patients with positive cytologic findings had distant metastases in our study, it follows that in an analysis of the entire group of patients we would indeed reveal worse survival for patients with positive vs negative cytologic findings. We did not detect any difference in survival between patients with positive and negative cytologic findings in either the visible metastases group or the locally advanced group, indicating that positive cytologic findings do not correlate with prognosis in patients with unresectable or advanced-stage pancreatic cancer.

Merchant et al8 found that peritoneal cytologic findings were not independent factors in determining survival for patients with pancreatic cancer, as it seems to depend on tumor stage and resectability. The correlation of survival with positive findings of free cancer cells in peritoneal washings seems to vary with different intra-abdominal malignancies. In contrast to such a correlation in gastric and ovarian cancer, Vogel et al2 found no significant correlation in colon cancer. One possible reason for the somewhat different results with respect to the association between survival and the presence of peritoneal free cancer cells could be that survival times differ among tumors. Survival for patients with pancreatic cancer tends to be shorter than that for patients with gastric or ovarian cancer. The median survival time for patients with gastric cancer and positive cytologic findings is reported to be longer than 1 year,2,3 which is similar to that for patients with pancreatic cancer and negative cytologic findings. We also found that positive cytologic findings had no correlation with peritoneal metastases-free survival in patients with unresectable or advanced-stage pancreatic cancer. Most of these patients showed major vessel invasion or hepatic metastases. Therefore, it seems that other significant prognostic variables are correlated more closely with prognosis than are positive cytologic findings in these patients.

Several studies have evaluated the significance of positive peritoneal cytologic findings in the absence of visible metastases. Makary et al13 reported that among patients with positive cytologic findings, there was no difference in survival between those with vs without visible metastases (median survival, 7.8 and 8.6 months). This survival time was notably shorter than that of treatment-matched patients with negative cytologic findings (median survival, 13.5 months). In the present study, although no differences were observed between positive and negative cytologic findings with respect to either tumor stage or curability in the resectable group, survival and peritoneal metastases-free survival were significantly worse for patients with positive vs negative cytologic findings. Survival for the resectable group with positive cytologic findings was similar to that for the visible metastases group with positive findings (median survival, 5.2 vs 5.1 months). However, among patients with positive cytologic findings, survival and peritoneal metastases-free survival for the locally advanced group were significantly better than those for the resectable group. It is conceivable that operative stress induced by pancreatectomy may promote the development of overt peritoneal metastases in patients with positive cytologic findings. Among the entire series of published studies, only 8 of 73 patients with positive cytologic findings underwent resection. Of these 8 patients, only 1 survived longer term (37 months) without evidence of disease; the remaining 7 patients had a short survival time.8,14

Because positive cytologic findings are rare, it is difficult to determine the most appropriate therapy for patients with potentially resectable tumors and positive cytologic findings. The results of the present study suggest that resection should proceed only with the knowledge that if cytologic findings are positive, the patient is unlikely to benefit significantly from surgical resection. Among patients with positive cytologic findings in the absence of visible metastasis, survival was significantly better in the locally advanced group than in the resectable group. Because survival was similar among patients in the locally advanced group with positive or negative cytologic findings (median survival, 11.1 vs 9.9 months), chemoradiotherapy (the treatment given to the locally advanced group in the present study) may be beneficial for patients with positive cytologic findings without visible metastases.

In recent years, immunohistochemical stains have been used to detect intraperitoneal free cancer cells. This method can be a useful diagnostic tool to help distinguish benign and malignant cells and to increase the likelihood of identifying malignant cells in peritoneal cytologic examination. Some researchers1517 have reported that immunohistochemical stains, such as carcinoembryonic antigen or carbohydrate antigen 19-9, can increase tumor cell detection in pancreatic cancer. However, the increase in the sensitivity of peritoneal cytologic examination did not improve prognostic significance, as positive cytologic rates seem to depend on tumor stage. In addition, Merchant et al8 found that immunohistochemical stains did not improve the yield over conventional stains. Further investigations are necessary to determine the usefulness of the immunohistochemical staining method.

In summary, positive peritoneal cytologic findings are not independent factors that determine survival and peritoneal metastases-free survival; rather, positive cytologic findings are associated with advanced disease. In patients with visible metastases, cytologic factors are not correlated with survival or peritoneal recurrence. In patients without visible metastases, chemoradiotherapy may be beneficial for those with positive cytologic findings.

Back to top
Article Information

Corresponding author and reprints: Masaru Konishi, MD, Department of Surgery, National Cancer Center Hospital East, 6-5-1 Kashiwanoha, Kashiwa 277-8577, Japan (e-mail: mkonishi@east.ncc.go.jp).

References
1.
Griffin  JFSmalley  SRJewel  W Patterns of failure after curative resection of pancreatic carcinoma. Cancer. 1990;6656- 61Article
2.
Vogel  PRuschoff  JKummel  S  et al.  Prognostic value of microscopic peritoneal dissemination: comparison between colon and gastric cancer. Dis Colon Rectum. 2000;4392- 100Article
3.
Wu  C-CChen  J-TChang  M-C  et al.  Optimal surgical strategy for potentially curable serosa-involved gastric carcinoma with intraperitoneal free cancer cells. J Am Coll Surg. 1997;184611- 617
4.
Creaseman  WTRutledge  F The prognostic value of peritoneal cytology in gynecologic malignant disease. Am J Obstet Gynecol. 1971;110773- 781
5.
Turner  DAGershenson  DMAtkinson  NSneige  NWharton  AT The prognostic significance of peritoneal cytology for stage I endometrial cancer. Obstet Gynecol. 1989;74775- 780
6.
Sobun  LHedWittekind  Ced UICC TNM Classification of Malignant Tumors. 5th ed. New York, NY Wiley-Liss Inc1997;87- 90
7.
Warshaw  AL Implications of peritoneal cytology for staging of early pancreatic cancer. Am J Surg. 1991;16126- 29Article
8.
Merchant  NBConlon  KCSaigo  PDougherty  EBrennan  MF Positive peritoneal cytology predicts unresectability of pancreatic adenocarcinoma. J Am Coll Surg. 1999;188421- 426Article
9.
Jimenez  REWarshaw  ALCastillo  CF Laparoscopy and peritoneal cytology in the staging of pancreatic cancer. J Hepatobiliary Pancreat Surg. 2000;715- 20Article
10.
Martin  JKGoellner  JR Abdominal fluid cytology in patients with gastrointestinal malignant lesions. Mayo Clin Proc. 1986;61467- 471Article
11.
Lei  SKini  JKim  KHoward  JM Pancreatic cancer: cytologic study of peritoneal washings. Arch Surg. 1994;129639- 642Article
12.
Leach  SDRose  JALowy  AM  et al.  Significance of peritoneal cytology in patients with potentially resectable adenocarcinoma of the pancreatic head. Surgery. 1995;118472- 478Article
13.
Makary  MAWarshaw  ALCenteno  BAWillett  CGRattner  DWCastillo  CF Implications of peritoneal cytology for pancreatic cancer management. Arch Surg. 1998;133361- 365Article
14.
Nakatsuka  AYamaguchi  KShimizu  S  et al.  Positive washing cytology in patients with pancreatic cancer indicates a contraindication of pancreatectomy. Int J Surg Invest. 1999;1311- 317
15.
Juhl  HStritzel  MWroblewski  A  et al.  Immunocytology detection of micrometastatic cells: comparative evaluation of findings in the peritoneal cavity and the bone marrow of gastric, colorectal and pancreatic cancer patients. Int J Cancer. 1994;57330- 335Article
16.
Nomoto  SNakao  AKasai  Y  et al.  Peritoneal washing cytology combined with immunocytochemical staining and detecting mutant K-ras in pancreatic cancer: comparison of the sensitivity and availability of various methods. Pancreas. 1997;14126- 132Article
17.
Castillo  CFWarshaw  AL Laparoscopic staging and peritoneal cytology. Surg Oncol Clin N Am. 1998;7135- 142
×