Detection of early-stage pancreatic adenocarcinoma

Abstract

Background: Pancreatic adenocarcinoma is an almost universally lethal disease, in large part, due to our inability to detect early-stage disease. Monoclonal antibody PAM4 is reactive with a unique biomarker expressed by >85% of pancreatic adenocarcinomas. In this report, we examined the ability of a PAM4-based immunoassay to detect early-stage disease.

Materials and methods: The PAM4-based immunoassay was used to quantitate antigen in the serum of healthy volunteers (n = 19), patients with known pancreatic adenocarcinoma (n = 68), and patients with a primary diagnosis of chronic pancreatitis (n = 29).

Results: Sensitivity for detection of pancreatic adenocarcinoma was 82%, with a false-positive rate of 5% for healthy controls. Patients with advanced disease had significantly higher antigen levels than those with early-stage disease (P < 0.01), with a diagnostic sensitivity of 91%, 86%, and 62% for stage 3/stage 4 advanced disease, stage 2, and stage 1, respectively. We also evaluated chronic pancreatitis sera, finding 38% positive for antigen; however, this was discordant with immunohistochemical findings that suggest the PAM4 antigen is not produced by inflamed pancreatic tissue. Furthermore, several of the serum-positive pancreatitis patients, for whom tissue specimens were available for pathologic interpretation, had evidence of neoplastic precursor lesions.

Conclusions: These results suggest the use of the PAM4 serum assay to detect early-stage pancreatic adenocarcinoma and that positive levels of PAM4 antigen are not derived from inflamed pancreatic tissues but rather may provide evidence of subclinical pancreatic neoplasia. EFFECT: The ability to detect pancreatic adenocarcinoma at an early stage could provide for early therapeutic intervention with potentially improved patient outcomes.

©2010 AACR.

Figures

Figure 1

Accuracy of the PAM4-immunoassay was determined to be within 10% of the nominal concentrations examined at or above the cutoff value of 2.40 units/mL. A linear trend was calculated with an equation of y = 0.965x + 0.174, and goodness of fit r2 = 0.999.

Figure 2

Frequency distribution of PAM4-reactive antigen in patient sera by stage of disease. Cutoff value = 2.4 units/mL (red line). The median values (units/mL) are shown for each study group.

Figure 3

Receiver Operator Characteristics (ROC) curve for the performance of the PAM4-based immunoassay; pancreatic adenocarcinoma vs healthy adults. Values for the area under the curves (AUC) and 95% confidence limits are provided.

Figure 4

Immunohistology of a PAM4-reactive PanIN-2-3 lesion identified within a biopsy section from a patient with primary diagnosis of chronic pancreatitis. (100x)

Figure 5

Immunohistology of tissue specimens derived from patients with primary diagnoses of chronic pancreatitis and pancreatic adenocarcinoma. The upper panels (A & B) are from a single patient specimen and are representative of 90% of the pancreatitis specimens (18 of 20 evaluable). The MA5 anti-MUC1 (peptide core) antibody, employed as a positive control, was reactive with acinar, ductal and ADM cells (A), whereas the PAM4 antibody was negative for all cell types (B). The middle panels (C & D) are from a single patient specimen where the MA5 antibody gave an intense and diffuse labeling of the acinar, ductal and ADM cells (C), whereas the PAM4 antibody gave only a focal, weak reactivity with ADM (D). The bottom panels (E & F) are two individual specimens of pancreatic adenocarcinoma, each labeled with PAM4 antibody. Panel E, representative of the majority of pancreatic adenocarcinomas, shows an intense, diffuse labeling of adenocarcinoma cells and secreted mucin, whereas panel F shows a weak, focal labeling of the adenocarcinoma cells. The arrows in panels D and F point to tissues that are weakly labeled with PAM4. (200x for all magnifications)