Recently, a highly sensitive fluorescent imaging technique was developed for the real-time identification of hepatic tumors. The authors applied this procedure for the intraoperative detection of radiographically occult hepatic micrometastases from pancreatic cancer.

Forty-nine consecutive patients with pancreatic cancer who underwent surgical intervention were examined. Preoperative clinical images had not revealed any hepatic metastases. On the day before surgery, indocyanine green was injected intravenously. During the operation, the liver was observed with a near-infrared camera system, and abnormal fluorescent foci were examined by frozen-section histology. The patients with hepatic micrometastases were judged to have unresectable disease and underwent only palliative surgery followed by systemic chemotherapy using gemcitabine.

Abnormal hepatic fluorescence at least 1.5 mm in greatest dimension without any apparent tumor was observed in 13 patients. Among them, histologic examination confirmed micrometastases in 8 of 49 patients (16%). All patients with hepatic micrometastases had clinical T3 or T4 disease and high serum CA19-9 levels (P = .042). On follow-up computed tomography images that were obtained within 6 months after surgery, the patients with hepatic micrometastases manifested hepatic overt metastases (7 of 8 patients; 88%) more frequently than the patients without hepatic micrometastases (4 of 41 patients; 10%; P < .001). Regardless of histologic confirmation, the positive predictive value of abnormal fluorescence for the manifestation of hepatic relapse within 6 months was 77% (10 of 13 patients), and the negative predictive value was 97% (35 of 36 patients).

Indocyanine green-fluorescent imaging can detect hepatic micrometastases of pancreatic cancer during surgery. The hepatic micrometastases seem to have an adverse clinical impact identical to that of evident distant metastases.

Cancers of the pancreas are identified in 11 800 to 13 500 patients each year in Germany. Epidemiological studies prove smoking and chronic alcohol consumption as causes of about 30% of pancreatic cancers.

Selective literature review.

Only patients within TNM stage I and II have after oncologic tumor extirpation a chance for long term survival. Controlled prospective clinical trials demonstrated adjuvant chemotherapy yielding an additional significant survival benefit. The 3- and 5-year-survival after R0-resection and adjuvant chemotherapy are about 30% and below 15% respectively. Using the criteria of observed 5-year-survival less than 2% of all pancreatic cancer patients are alive. After R0-resection the median survival time is between 17 and 28 months, after R1/2-resection between 8 and 22 months.

Pancreatic cancer is even today for more than 95% of the patients incurable. Strategies to prevent pancreatic cancer are intended to stop smoking and chronic alcohol consumption and early surgical extirpation of cystic neoplastic lesions. For patients with established pancreatic cancer risk a follow-up protocol is discussed.

Isolated tumor cells (ITCs) in cancer patients are retrieved mostly using immunohistochemistry with antibodies directed against antiepithelial antigens (for example Ber-EP4), which are supposed not to be present in metastatic-free tissue. To date, there has been ongoing controversy whether those cells have biologic significance and are linked with tumor progression and impaired patient's prognosis. Therefore, the aim of this study was to further characterize Ber-EP4-positive cells in various tissues, with special emphasis on their tumorigenic origin.

The frequency and prognostic impact of ITCs in lymph nodes displayed by means of monoclonal antibody Ber-EP4 were evaluated in retrospective (n = 292) and prospective (n = 100) collectives of various gastrointestinal carcinomas free of metastatic disease in conventional histopathology (pN0). Furthermore, the frequency of ITCs in the peritoneal cavity and bone marrow was analyzed in case of absence of overt distant metastasis (pM0) in the prospective collective. Ber-EP4-immunoreactive cells were further characterized for tumorigenic origin using morphological criteria and immunohistochemical double staining for Ber-EP4 and p53.

Ber-EP4-positive cells could be revealed in lymph nodes in 44.3% of pN0-gastrointestinal carcinomas, in the peritoneal cavity in 19%, and in the bone marrow in 10%. In lymph nodes, BerEP4-immunoreactive cells exhibited a metastatic-atypical morphology in 59%; however, it was always typical for true tumor cells in the peritoneal cavity or bone marrow. The cumulative 5-year survival rate was adversely affected by Ber-EP4-immunoreactive cells in uni- and multivariate analysis, irrespective of the underlying cell morphology (68% for Ber-EP4 negative, 41% for Ber-EP4 positive with atypical and typical morphology each). In the case of a p53-positive primary tumor, 70% of the corresponding ITCs also overexpressed p53, while the remainder was deemed p53 negative (p = 0.002).

ITCs detected by the antiepithelial antibody Ber-EP4 are present in a substantial proportion of apparently tumor-free lymph nodes. These cells impair patients' prognoses, irrespective of the underlying cell morphology. As approximately one third of Ber-EP4-positive cells in p53-positive primary tumors do not overexpress p53; their true tumorigenic origin needs to be further investigated.

New understanding of the dynamic of acute pancreatitis, the clinical impact of local pathology in chronic pancreatitis and cystic neoplastic lesions bearing high potential for malignant transformation has changed the management of pancreatic diseases.

In acute pancreatitis, risk factors independently determining outcome in severe acute pancreatitis are early and persistent multiorgan failure, infected necrosis and extended sterile necrosis. The management of severe acute pancreatitis is based on early intensive-care treatment and late surgical debridement. In chronic pancreatitis, recent data from randomized controlled clinical trials have demonstrated duodenum-preserving pancreatic head resection with an inflammatory mass of the head as superior to pylorus-preserving Whipple resection. Cystic neoplasms are local lesions of the pancreas with high malignant potential. Local organ-preserving resection techniques have been applied with low morbidity and mortality, replacing a Whipple-type resection. Resection of pancreatic cancer is ineffective to cure patients. After an R0-resection, a significant survival benefit has been achieved when adjuvant chemotherapy has additionally been applied.

New knowledge about the nature of inflammatory diseases, cystic neoplastic lesions and malignant pancreatic tumours has changed the indication for surgical treatment and the application of organ-preserving surgical techniques.

To determine whether early primary pancreatic tumor resection can prevent liver metastases of intrapancreatic transplantation in a hamster model.

Cells from the PGHAM-1 cell line were transplanted into the pancreases of 30 Syrian golden hamsters. A suspension of 5 x 10(6) cells was injected into the splenic lobe of each pancreas. The primary pancreatic tumor was resected in 15 of the hamsters 10 days after transplantation (resection group). Fifteen other animals with transplantation but without resection served as controls (control group). All hamsters were killed 21 days after transplantation. The primary pancreatic tumors were measured for size and volume and examined histologically and immunohistologically for angiogenesis and tumor proliferation.

In the resection group, small pancreatic tumors 4.7 +/- 0.94 mm in diameter were found and resected 10 days after transplantation. Neither pancreatic tumors nor liver metastases were found in the resection group at the end of the experiment. All animals in the control group had pancreatic tumors 12.3 +/- 3.29 mm in size, and 11 of 15 (73.3%) had liver metastases. The primary pancreatic tumors in the group with liver metastasis were significantly larger in diameter and volume than those in this group without liver metastasis (p<0.01). In the control group, proliferation of the primary pancreatic tumor, evaluated according to argyrophilic nucleolar organizer region, showed no differences within the pancreatic tumor group. On the other hand, the microvessel density of pancreatic tumors with liver metastases was significantly higher than that of tumors without liver metastases.

Our results suggest that 10 days after transplantation, the pancreatic tumors were small in size and volume and ready to proliferate but not yet ready to begin metastasizing through angiogenesis. This is one reason why early resection of the primary tumor prevents liver metastasis.

The liver is a common site of metastasis from a variety of solid malignancies. This is due to disseminated tumour cells (DTC) that have spread prior to or during surgery from the primary carcinoma. This article gives a short overview of the data published on the detection of DTC in the liver and describes the commonly used detection methods and respective markers.

A literature survey was performed in public medical databases comprising the last 15 years with focus on DTC detection in liver tissue of cancer patients.

Although the liver is a preferred site of metastasis, only a few studies have analysed the DTC incidence in inconspicuous liver tissue. The available reports include only patients with pancreatic and colorectal carcinomas. In patients with pancreatic cancer the DTC incidence varied from 5 to 76%. No follow-up data has been reported so far. In patients with colorectal carcinoma hepatic DTC were found in 5-69% of cases. A negative prognostic influence of hepatic DTC was reported in all but one studies with follow-up information.

The detection of DTC in the liver can contribute to identify patients with increased risk who could benefit from an intensified follow-up or new treatment strategies.

Despite recent advances in early diagnosis and surgical treatment, the clinical outcome of patients with pancreatic cancer has not been improved markedly. One of the reasons for the dismal outcome is early dissemination of tumor cells. Sensitive immunohistocytochemical and nucleic acid-based assays have detected disseminated tumor cells in the lymph nodes, bone marrow, peritoneal cavity or peripheral blood. Formation of the metastatic disease depends on the nature of the disseminated tumor cells. Standardization of protocols is mandatory to detect occult tumor cells in clinical practice. We present an overview of recent studies on the incidence, prognostic values and some characteristics of occult tumor cells disseminated in the secondary sites of patients with pancreatic cancer.

To clarify the relationship between intratumoral dihydropyrimidine dehydrogenase (DPD) expression and response to 5-fluorouracil (5-FU) liver perfusion chemotherapy (LPC) in pancreatic cancer patients, we evaluated DPD expression immunohistochemically in resected pancreatic cancer tissues.

Pancreatic cancer is considered a disease with a poor prognosis even if aggressive resection is performed. One of the main causes of death is hepatic metastasis soon after surgery. As a treatment, we have assessed adjuvant LPC via the portal vein using 5-FU just after pancreatectomy for advanced pancreatic cancer since 1994. However, the results remain unsatisfying.

Sixty-eight resected specimens were obtained from patients with pancreatic cancer from 1988 to 2000. Formalin-fixed paraffin-embedded tissues were immunostained with polyclonal anti-DPD antibody. The relation between intratumoral DPD expression and the prognoses of pancreatic cancer patients was investigated statistically.

Of the 68 tumors studied, 27 carcinomas (39.7%) were DPD(+), and 41 (60.3%) were DPD(-). In the DPD(+) group, there was no significant difference between the LPC(+) and LPC(-) subgroups, whereas in the DPD(-) group the LPC(+) subgroup showed a significantly higher survival rate than the LPC(-) subgroup. Moreover, in the LPC(+) group, overall survival in the DPD(-) subgroup was significantly better than in the DPD(+) subgroup.

An immunohistochemical evaluation of intratumoral DPD expression might be useful in predicting responsiveness to 5-FU-based chemotherapy in pancreatic cancer patients. In the DPD(-) group, liver perfusion chemotherapy using 5-FU via the portal vein is effective adjuvant therapy for pancreatic cancer once pancreatectomy has been performed.

To determine the frequency and prognostic impact of isolated tumor cells (ITC) in regional lymph nodes judged to be tumor free in conventional histopathology among gastric cancer patients.

Among 161 patients who underwent gastrectomy and D2-lymphadenectomy, 56 were staged pN0(35%). Archival paraffin blocks of 1148 resected regional lymph nodes of those pN0 patients were reevaluated for ITC using monoclonal antibody Ber-EP4. Patients with and without ITC were compared with regard to the distribution of various clinicopathological factors. Prognostic impact of ITC was tested in uni- and multivariate analysis.

Of 56 pN0 patients, 33 (59%) exhibited single Ber-Ep4 immunoreactive cells or small cell clusters in at least one lymph node. The occurrence of ITC was not dependent on other clinicopathological factors. ITC impaired patients' prognoses significantly in uni- as well as multivariate analyses [estimated 5-year survival rate: 82% for pN0((i-))vs 58% for pN0((i+))(p = 0.059) and 15% for pN1/2 (p = 0.0005 and p < 0.0001, respectively)].

ITC are a frequent event in apparently tumor-free lymph nodes of gastric cancer patients and are overlooked by conventional histopathology. They are encountered even in limited stages of disease and impair patients' prognoses. This should be borne in mind when advocating local resection for early gastric cancer.

Adenocarcinoma of the pancreas is associated with the worst survival of any form of gastrointestinal malignancy. In spite of the progress in surgical treatment, resulting in increasing resection rates and a decrease in treatment-related morbidity and mortality, the true figures of cure are even today below 3%. The dissemination of pancreatic cancer behind the local tissue compartments restricts the short-term (< 3 years) and long-term outcome for patients who have undergone resection. By histological evaluation, less than 15% of the patients undergoing R(0) resection have a pN(0) status, more than 60% suffer from lymph angiosis carcinomatosa, and more than 50% suffer extrapancreatic nerve plexus infiltration. Hematoxylin and eosin-negative lymph nodes were found to be cancer positive when reverse transcriptase polymerase chain reaction (RT- PCR) or immunostaining was applied to the HE-negative lymph nodes. Cancer of the uncinate process has a very poor prognosis because there are no early symptoms; vessel wall involvement occurs early and frequently; a high association of liver metastasis exists as well. Surgery offers a low success rate, but it provides the only chance of cure. Ductal pancreatic cancer is diagnosed in more than 95% of the cases in an advanced stage; potentially curative resection can be performed only in about 10%-15% of these patients. Major contributions of surgery to improved treatment results are the reduction of surgical morbidity--e.g., early postoperative local and systemic complications--and a decrease of hospital mortality below 3%-5%. In most recently published prospective trials, R(0) resection has been reported to result in an increase in short-term survival beyond that recorded for patients with residual tumor. However, R(0) resection fails to improve long-term survival. In many published R(0) series, standard tissue resection of pancreatic head cancer with the Kausch-Whipple procedure failed to include remote cancer cell-positive tissues in the operative specimen; e.g., N(2)-lymph nodes, nerve plexus, and perivascular extrapancreatic and retropancreatic tissues were not excised. Cancer recurrence after so-called R(0) resection with curative intent is frequently the consequence of cancer left behind. Thus, long-term survival (> 5 years) is observed in a very small group of patients, contradicting the published 5-year actuarial survival rates of 20%-45% for resected patients. The assessment of clinical benefit from surgical or medical cancer treatment should therefore be based on several end points, not only on actuarial survival. Publication of actuarial survival figures must include the number of observed (actual) survivals, the definition of the subset of patients followed after resection, and the total number of patients in the study group; anything less is misleading. In reporting pancreatic cancer treatment trial results after oncological resections, more convincing primary end points to evaluate treatment efficacy are median survival (in months), actual survival at 1-5 years, and progression-free survival (in months). In series with multimodality treatment, clinical benefit response as well as quality of life measurements using the EORTC Quality of Life index C30 (QLQ-C30) are of importance in evaluating survival data. Adjuvant treatment improves survival after oncological resection; however, the short-term and long-term benefit after adjuvant chemotherapy in R(0) as well as in R(1)-(2) resected patients has not yet been underscored by data from controlled clinical trials. The survival benefit (median survival time) of adjuvant chemotherapy or radiochemotherapy has been demonstrated to be 6-10 months. Therefore, after oncological resection of pancreatic cancer each patient should be offered adjuvant treatment. A neoadjuvant treatment protocol for pancreatic cancer, however, has not been established.

Metastatic disease determines the prognosis of patients with pancreatic cancer. Current routine staging methods often underestimate the tumor stage because they do not include the search for disseminated tumor cells that spread early in different compartments of the body. Immunohistochemical and molecular methods developed recently are able to detect these cells in multiple compartments of the body.

The current status of the detection and the prognostic impact of disseminated tumor cells detected in lymph nodes, bone marrow, blood and peritoneal lavage of patients with pancreatic carcinoma are reviewed.

Disseminated tumor cells can be detected in different compartments of the body even in early tumor stages and when a resection of the primary tumor in curative intention was performed. Furthermore, the detection of these cells has importance for the prognosis and therefore will have therapeutic implications. Standardization of the methods is a prerequisite for further studies.

The detection of disseminated tumor cells should be included into studies to reveal that this increased staging has an prognostic impact and can be useful for therapeutic decisions in patients with pancreatic carcinoma.

Tumor progression after curative resection of gastrointestinal carcinomas is probably caused by pre- or intraoperative tumor cell dissemination. Disseminated tumor cells are generally detected by immunohistochemistry- or PCR-based molecular-biology methods. A consensus on which is the most adequate detection method has not yet been found, which makes the comparison of data difficult. The prognostic relevance of disseminated cells has been shown, at least in part, for esophageal, gastric, pancreatic, and colonic cancer. The data regarding hepatocellular cancer is conflicting. This article gives a critical review of tumor cell detection in gastrointestinal cancer.

Pancreatic cancer remains a highly malignant disease. Curative treatment is only possible for patients diagnosed at a very early stage. Therefore, the vast majority of pancreatic cancer patients receive palliative treatment. Surgical palliation is offered to patients who are found not to have a resectable tumor. The treatment of obstructive jaundice is managed by stenting of the common bile duct or by a surgical bypass. The best possible surgical procedure should be based on the factors that influence hospital mortality, length of survival, and quality of life. In patients with a life expectancy of longer than 3 months, surgical bypass is recommended, with hepaticojejunostomy the treatment of choice. In the same surgical procedure, the relief of duodenal obstruction with a gastroenteric bypass should be achieved. Chemotherapy, radiotherapy, or a combination of both is employed as a neoadjuvant measure, as an adjuvant treatment, or, in most patients, as palliation. As palliative chemotherapy alone, 5-fluorouracil (5-FU) plus folinic acid is still the treatment of choice; however, newer drugs, such as gemcitabine, seem to have similar or marginally better results. Palliative radiochemotherapy with external-beam radiation plus 5-FU and folinic acid seems to lead to better local control of tumor progression but not to better survival, for which distant metastases are the limiting factor.

Frequent P53 mutations and Ki-ras codon 12 point mutations have been reported in pancreatic cancer. Pancreatic cancer often recurs in the liver and/or lymph nodes shortly after a surgical resection. The purpose of this study is to elucidate the occurrence of microcirculating cancer cells and micrometastasis in pancreatic cancer.

P53 mutations and Ki-ras codon 12 point mutations were examined in the main tumor, liver, portal vein, and peripheral arterial blood, and para-aortic lymph nodes of patients with pancreatic cancer using molecular examinations.

P53 mutations in the main tumor were present in nine (29%) of 31 patients with pancreatic cancer, whereas a Ki-ras codon 12 point mutation was evident in 18 (62%) of 29 examined patients. The peripheral arterial and portal vein blood and liver were positive for gene abnormalities in one (5%) of 21, in none (0%) of 19, and in one (1%) of 20, respectively. A P53 mutation in the main tumor was evident in none (0%) of seven stage I or II carcinomas and in nine (38%) of 24 stage III or IV cases, whereas a Ki-ras codon 12 point mutation was present in four (67%) of six stage I or II cases and in 14 (61%) of 23 stage III or IV cases. In addition, 15 (71%) of 21 patients with gene abnormalities (Ki-ras codon 12 point and/or p53 mutation) in the main tumor showed lymph node metastasis at surgery, whereas five (42%) of 12 without gene abnormalities did not demonstrate lymph node metastasis. Two (29%) of six patients with gene abnormalities in the main tumor and without metastatic disease at surgery developed liver metastasis within 6 months after surgery, whereas all five (100%) without the gene abnormalities and metastatic disease at surgery did not develop the metastasis, with the sensitivity being 100%, specificity 44%, the predictive value of the positive test 36%, and the predictive value of the negative test 100%. Two patients who had gene abnormalities in the para-aortic lymph node were free from histopathological metastasis and these two patients developed para-aortic lymph node metastasis within 6 months after surgery.

A molecular examination of Ki-ras codon 12 and p53 mutations therefore enables us to predict, to some degree, the occurrence of liver and lymph node metastasis in pancreatic carcinoma.

The liver is the most common site of metastasis in pancreatic cancer, and there are no promising strategies to treat it. Angiostatin, a kringle-containing fragment of plasminogen, is a potent inhibitor of angiogenesis. The effect of angiostatin on liver metastasis in pancreatic cancer was investigated by using our established hamster model of liver metastasis. Pancreatic cancer cells (PGHAM-1, 1 x 10(6)) derived from N-nitrosobis(2-oxopropyl)amine (BOP)-induced pancreatic tumor in Syrian golden hamsters were transplanted into the spleen of female hamsters, and the animals were subcutaneously injected with angiostatin and saline. Subsequently, the macroscopic appearance of liver surface metastases was evaluated. In addition, histological sections of the liver metastases were analyzed for neovascularization, proliferation, and apoptosis on the basis of von Willebrand factor, argyrophilic nucleolar organizer region (Ag-NOR), and TdT-mediated dUTP-biotin nick end labeling (TUNEL) staining, respectively. The results showed significant tumor growth retardation and inhibition of angiogenesis in metastatic liver tumors in response to treatment with angiostatin. Moreover, the metastases remained in a nearly dormant state due to a balance between apoptosis and proliferation of the tumor, with no detectable side effects. This is the first experimental trial of angiostatin on pancreatic cancer and liver metastasis. The results suggest that angiostatin therapy could be effective against liver metastases of pancreatic cancer.

The prognosis of pancreatic adenocarcinoma after radical pancreatectomy is poor, especially in advanced-stage disease.

To determine the survival rates and evaluate the effectiveness of multimodality treatment for advanced pancreatic cancer.

From November 1983 to January 1993, 30 patients with pancreatic adenocarcinoma including 9 with carcinoma of the body and tail were treated by a multimodal approach consisting of extended pancreatectomy, intraoperative radiotherapy (IORT), and hepatic artery or portal vein infusion of mitomycin C (MMC) followed by systemic bolus injection. All surviving patients were followed for more than 8 yr and survival rates were calculated by the Kaplan-Meier method.

There were no operative or hospital deaths. Eight patients survived for more than 5 yr, 3 of whom survived more than 10 yr. The 5-yr survival rate for 27 patients excluding 3 with metastasis to the liver, peritoneum, or lung was 31%, with a median survival of 31.1 mo. Among them, the 1-, 3-, and 5-yr survival rates for 19 patients with regional nodal metastasis were 95, 50, and 28%, respectively, with a median survival of 36.0 mo.

The multimodality treatment combined with IORT and MMC chemotherapy appeared to have a benefit for prognosis of advanced pancreatic adenocarcinoma.

In pancreatic cancer, the mutation of c-K-ras is a critical event of tumor growth and metastasis. We have previously demonstrated a dominant negative effect of N116Y on the growth of pancreatic cancer cells. To evaluate the potential of N116Y for suppressing the metastatic growth of pancreatic tumor cells, we made a replication-deficient recombinant N116Y adenovirus driven by the carcinoembryonic antigen (CEA) promoter (Ad CEA-N116Y). We demonstrated that the expression of N116Y, growth inhibition, and apoptotic death induction were all specific to pancreatic cancer cell lines (PCI-35 and PCI-43) that were promoter positive, whereas no growth retardation was observed in human embryonic pancreas-derived cell line 1C3D3 after Ad CEA-N116Y infection. We examined the effect of Ad CEA-N116Y on the metastatic growth of PCI-43 colonies in liver, which was generated by tumor injection into the spleen of nude mice. The results showed that Ad CEA-N116Y effectively reduced the number of metastatic colonies without any complication by injecting intrasplenically 5 days after tumor cell inoculation. Thus, N116Y can selectively suppress the metastatic growth of pancreatic tumor cell by using the CEA promoter-driven adenovirus vector indicating that N116Y gene therapy may be potentially useful for the treatment of pancreatic cancer patients with liver micrometastasis.

The accurate detection of low-level disease in patients with cancer is essential to improve the staging of disease and consequently to define appropriate treatment strategies. Most methods currently used for staging are based on imaging studies and histological and immunocytochemical analysis of tissues such as bone marrow aspirates, or antibody assays for marker proteins secreted into the circulation. These methods have limited sensitivity. However, assays for nucleic acid-based markers may be valuable tools for the sensitive detection, assessment, and monitoring of disease status in asymptomatic cancer patients. Application of these methods may allow the early detection of cancer, when the tumour burden is smaller and the disease potentially more curable. The last decade has seen the application of polymerase chain reaction (PCR)-based methods to the detection of tumour in a wide variety of compartments, including peripheral blood, bone marrow, lymph nodes, urine, sputum, faeces, pancreatic juice, and more recently plasma. Molecular detection of disease by PCR has targeted DNA and RNA markers, including mutations, microsatellites, and tissue-specific gene expression. It is likely that these molecular methods will provide important clinical information, though their current clinical utility remains unclear. The current status of nucleic acid-based assays for the detection and assessment of disease status in the management of patients with solid tumours is reviewed.

Metastatic relapse in patients with solid tumors is caused by systemic preoperative or perioperative dissemination of tumor cells. The presence of individual tumor cells in bone marrow and in peripheral blood can be detected by immunologic or molecular methods and is being regarded increasingly as a clinically relevant prognostic factor. Because the goal of adjuvant therapy is the eradication of occult micrometastatic tumor cells before metastatic disease becomes clinically evident, the early detection of micrometastases could identify the patients who are most (and least) likely to benefit from adjuvant therapy. In addition, more sensitive methods for detecting such cells should increase knowledge about the biologic mechanisms of metastasis and improve the diagnosis and treatment of micrometastatic disease. In contrast to solid metastatic tumors, micrometastatic tumor cells are appropriate targets for intravenously applied agents because macromolecules and immunocompetent effector cells should have access to the tumor cells. Because the majority of micrometastatic tumor cells may be nonproliferative (G0 phase), standard cytotoxic chemotherapies aimed at proliferating cells may be less effective, which might explain, in part, the failure of chemotherapy. Thus, adjuvant therapies that are aimed at dividing and quiescent cells, such as antibody-based therapies, are of considerable interest. From a literature search that used the databases MEDLINE(R), CANCERLIT(R), Biosis(R), Embase(R), and SciSearch(R), we discuss the current state of research on minimal residual cancer in patients with epithelial tumors and the diagnostic and clinical implications of these findings.

Reports of improved survival rates for patients with resected adenocarcinoma of the pancreas coincide with the adoption of adjuvant chemoradiation protocols. The impact of nodal micrometastases demonstrated by molecular assays and adjuvant therapy on survival of patients with stage I pancreatic cancer has not been adequately assessed.

A retrospective analysis of postoperative chemoradiation on survival in 61 patients undergoing resection of pancreatic adenocarcinomas from 1984 to 1997 was performed. Archival tumors and regional nodes from 25 patients with stage I cancers were tested for a Kiras oncogene mutation using polymerase chain reaction and analysis for restriction fragment length polymorphisms (PCR/RFLP).

Adjuvant chemoradiation was associated with improved survival for stage I (P < .01), but not stage III, disease. Seventeen (68%) of 25 patients with stage I disease tested had evidence of mutant Kiras in one or more regional nodes. Survival did not differ for patients with molecular micrometastases. Six of 17 (35%) patients with micrometastases received adjuvant chemoradiation and had improved survival (P < .05).

The majority of patients with stage I pancreatic cancer have PCR/RFLP evidence of lymph node micrometastases. Adjuvant chemoradiation improves survival in these patients by treating micrometastases not detected by histology. Adjuvant chemoradiation should be used for patients with stage I pancreatic cancers.

Midkine (MK) is a growth factor identified as a product of a retinoic acid-responsive gene. A truncated form of MK mRNA, which lacks a sequence encoding the N-terminally located domain, was recently found in cancer cells. We investigated the expression of the truncated MK mRNA in specimens of 47 surgically removed human gastrointestinal organs using polymerase chain reaction. Truncated MK was not detected in all of the 46 corresponding non-cancerous regions. On the other hand, this short MK mRNA was expressed in the primary tumours in 12 of 16 gastric cancers, 8 of 13 colorectal carcinomas, five of nine hepatocellular carcinomas, two of two oesophageal carcinomas and one ampullary duodenal cancer. In addition, truncated MK was detectable in all of the 14 lymph node metastases but in none of three metastatic sites in the liver, suggesting that truncated MK mRNA could become a good marker of nodal metastases in gastrointestinal tract.

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.

It has been approximately 7 years since the introduction of gene therapy. Since conventional procedures such as surgery, chemotherapy, and radiotherapy have had limited therapeutic value for cancer patients, the evolution of gene therapy seems a promising alternative to many researchers and clinicians. Indeed, about half of all gene therapy treatments are administered to patients with cancer. However, there are relatively few studies of using gene therapy with pancreatic cancer. We will review the general concept of gene therapy for cancer and its accomplishments to date.

Pancreatic adenocarcinomas are known to have a high incidence of K-ras gene mutation. We summarize our efforts to detect micrometastases through a search for mutated K-ras oncogene in liver tissues, peritoneal washings, para-aortic lymph nodes, and perioperative peripheral blood. Two-stage polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) analysis were used to detect K-ras oncogene mutations at codon 12. Our results suggest that PCR/RFLP analysis is potentially highly sensitive for the detection of micrometastases in various tissues and might be of value in the diagnosis, treatment and follow-up of metastases in other organs with pancreatic adenocarcinoma.

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.

Most cancer detection tests currently performed are based on either antibody assays to a marker protein with altered expression in cancer patients or on imaging studies to identify characteristic lesions. Generally, for a positive result, these detection assays require that a tumor have a significant volume of cancer cells. Advances in diagnostic techniques and technology may allow for cancer detection at earlier stages, when the tumor burden is smaller and potentially more curable. The molecular techniques of polymerase chain reaction (PCR) and reverse transcriptase PCR (RT-PCR) are highly sensitive methods for detecting a small number of cancer cells. Over the past few years, numerous clinical studies have used PCR techniques to detect physical alterations of genes, such as mutations, deletions, translocations and amplification, the presence of oncogenic viruses, and the expression of genes specific to tissue, cancer, and metastasis. The current status of PCR as a method for detecting marker genes in the management of solid tumors is reviewed.

A review of the literature on the clinical utility of PCR and RT-PCR in the detection of solid tumor micrometastasis was conducted.

Amplification by PCR is a highly sensitive method to determine gene expression. A single cell expressing a tumor marker among 10-100 million lymphocytes can be detected by the PCR assay. This approach has been used to detect tumor cells in approximately 18 different solid tumor types, with melanoma and carcinoma of the breast and prostate the most widely investigated to date. PCR-based assays have been used to detect cancer cells in biopsies of solid tissue, lymph nodes, bone marrow, peripheral blood, and other body fluids. Several studies have reported a high specificity and sensitivity of tumor marker detection and a high correlation between PCR results and the presence of metastatic disease. However, in a few studies, PCR assays have not consistently demonstrated a higher sensitivity and specificity of detection than traditional modalities for many types of cancer. There has been a wide range in sensitivity and specificity among the studies, which may be partly attributed to the lack of uniformity among the PCR protocols used in different studies.

PCR can detect tumor marker-expressing cells that are otherwise undetectable by other means in patients with localized or metastatic cancer. Reports from various study groups have lacked uniformity in their protocols, and this has prevented adequate comparison. The clinical utility of this assay as a tool for the prognosis and management of cancer patients remains and area of active investigation. PCR is a powerful tool in the study of the biology of cancer metastasis and will likely serve as a useful adjunct to clinical decision-making in the future.

Most cancer detection tests currently performed are based on either antibody assays to a marker protein with altered expression in cancer patients or on imaging studies to identify characteristic lesions. Generally, for a positive result, these detection assays require that a tumor have a significant volume of cancer cells. Advances in diagnostic techniques and technology may allow for cancer detection at earlier stages, when the tumor burden is smaller and potentially more curable. The molecular techniques of polymerase chain reaction (PCR) and reverse transcriptase PCR (RT-PCR) are highly sensitive methods for detecting a small number of cancer cells. Over the past few years, numerous clinical studies have used PCR techniques to detect physical alterations of genes, such as mutations, deletions, translocations and amplification, the presence of oncogenic viruses, and the expression of genes specific to tissue, cancer, and metastasis. The current status of PCR as a method for detecting marker genes in the management of solid tumors is reviewed.

A review of the literature on the clinical utility of PCR and RT-PCR in the detection of solid tumor micrometastasis was conducted.

Amplification by PCR is a highly sensitive method to determine gene expression. A single cell expressing a tumor marker among 10-100 million lymphocytes can be detected by the PCR assay. This approach has been used to detect tumor cells in approximately 18 different solid tumor types, with melanoma and carcinoma of the breast and prostate the most widely investigated to date. PCR-based assays have been used to detect cancer cells in biopsies of solid tissue, lymph nodes, bone marrow, peripheral blood, and other body fluids. Several studies have reported a high specificity and sensitivity of tumor marker detection and a high correlation between PCR results and the presence of metastatic disease. However, in a few studies, PCR assays have not consistently demonstrated a higher sensitivity and specificity of detection than traditional modalities for many types of cancer. There has been a wide range in sensitivity and specificity among the studies, which may be partly attributed to the lack of uniformity among the PCR protocols used in different studies.

PCR can detect tumor marker-expressing cells that are otherwise undetectable by other means in patients with localized or metastatic cancer. Reports from various study groups have lacked uniformity in their protocols, and this has prevented adequate comparison. The clinical utility of this assay as a tool for the prognosis and management of cancer patients remains and area of active investigation. PCR is a powerful tool in the study of the biology of cancer metastasis and will likely serve as a useful adjunct to clinical decision-making in the future.

The recently introduced genetic diagnosis of cancer micrometastasis is quite attractive because of its high detection sensitivity. Not infrequently, however, there are marked discrepancies between genetic and conventional histologic diagnoses, especially concerning lymph nodes from colon carcinoma patients. Because the histologic approach has long been relied on in the clinic, the reasons for the differences in results need to be elucidated.

Serial sections of genetic diagnosis positive but histologic negative lymph nodes of colon carcinoma patients were prepared for histologic and immunohistochemical studies. To investigate the possible contaminating influence of DNA sequences derived from degraded carcinoma cells from the primary site through the lymphatics, the authors also injected purified DNA of human colon carcinoma cells (SW480) into the foot pads of rats and sequentially examined lymph nodes using genetic diagnosis methodology.

Careful histologic examination of genetic diagnosis positive, histologically negative lymph nodes of colon carcinoma patients confirmed the absence of cancer cells, whereas p53 protein was immunohistochemically demonstrated to be present in the cytoplasm of sinus histiocytes. In the rat experiment, positive reactions were obtained with the inguinal lymph nodes beginning 30 minutes after the injection, and lymph nodes at various sites were subsequently affected, even after a 72-hour period.

The current study thus suggests that positive results with genetic diagnosis may simply indicate the presence of tumor DNA and do not necessarily mean that viable cancer cells are present. Although the genetic approach may still hold promise for the detection of cancer micrometastases, its predictive value should be carefully assessed clinicopathologically, because its supersensitivity may be associated with a greatly increased false-positive rate.

PCR assays for the presence of mutant K-ras or p53 sequences are potentially useful as sensitive tests for tumor diagnosis. The technical challenge is to design assays sensitive enough to detect a few molecules of mutant DNA yet sufficiently specific that a false positive signal is not produced by a 10(5)- or 10(6)-fold excess of normal DNA. We determined the detection limit of allele-specific PCR (ASA) as a function of the particular mismatch involved using all 12 possible mismatches in two different DNA sequence contexts (K-ras codon 12 and p53 codon 273). Depending on the identity of the mismatch, mismatched template was amplified 10(2)-10(4)-fold less than perfectly matched template. In other words, a mutant allele could be detected by ASA if it represented > 1-0.01% of the total DNA from that locus. Peptide nucleic acid (PNA) clamping was used to improve the K-ras ASA assay. Selective amplification of mutant sequences was achieved using a PNA complementary to the normal sequence to inhibit the amplification of wild-type DNA. PNA clamping followed by ASA resulted in significant improvement in sensitivity and specificity, permitting the detection of tumor DNA diluted with a 300,000-fold excess of normal human DNA.

Surgeons wish to know of any correlation between an operation and the incidence of metastasis. In perioperative periods, pancreatic cancer cells were identified by detecting mutant K-ras gene by two-step PCR and RFLP analysis in blood samples taken from peripheral blood. In no case was K-ras point mutation detected in blood before operation, although the mutant hand was observed in all cases at the time the lesion was resected. Surprisingly, in five of ten cases, positive bands were identified just after laparotomy, before we had reached the primary lesion. In almost all cases, mutant K-ras was detected until the fourteenth postoperative day. These findings suggest that cancer cells exist in the circulation, and have a potential for hematogenous metastasis during the perioperative period. In conclusion, surgical stress causes hematogenous dissemination of pancreatic cancer cells, and surgeons should employ the appropriate anti-metastasis therapy in the perioperative period.