Considerations for the use of plasma cytokeratin 18 as a biomarker in pancreatic cancer

Abstract

Background: Enzyme-linked immunoassays of full-length (M65) and/or caspase-cleaved (M30) cytokeratin 18 (CK18) released from epithelial cells undergoing necrosis and/or apoptosis, respectively, may have prognostic or predictive biomarker utility in a range of solid tumour types. Characterisation of baseline levels of circulating full length and cleaved CK18 specifically in patients with pancreatic cancer.

Methods: Plasma samples from 103 patients with pancreatic cancer stored at -80 degrees C were assayed for M65 and M30 levels. The median (inter-quartile range (IQR)) duration of plasma storage was 34 (23-57) months. Patients with metastatic disease (n=19) were found to have greater median (IQR) M65 levels (1145 (739-1698) U l(-1)) compared with the locally advanced (n=20; 748 (406-1150) U l(-1)) and resected (n=64; 612 (331-987) U l(-1)) patients (P=0.002). Elevated M65 levels were associated with poorer overall survival on univariate (P<0.001) but not multivariate (P=0.202) analysis. M65 concentrations also exhibited significant associations with concurrent serum-bilirubin levels (P<0.001) and the duration of plasma storage (P<0.001).

Conclusions: Baseline plasma CK18 levels in pancreatic cancer are affected by the presence of obstructive jaundice and prolonged plasma storage. Clinical biomarker studies utilising serial CK18 levels are warranted in pancreatic cancer, provided consideration is given to these potentially confounding factors.

Figure 1

(A) Example of shed cells within a cancer gland exhibiting positive staining for activated caspase-3 on immunohistochemistry (arrowed) in a core of pancreatic ductal adenocarcinoma. (B) Example of sporadic nature of activated caspase-3 staining within the malignant ductal epithelium. (C) Haematoxylin and eosin (H&E) stained slide of pancreatic ductal adenocarcinoma exhibiting widespread tumour necrosis.

Figure 2

Box plot illustrating the effect of tumour resectability and presence of metastasis on plasma M65 levels. The dotted lines represent the inter quartile range (IQR) reference values of 191 and 339 U l⁻¹ for healthy controls (Greystoke et al, 2008).

Figure 3

Kaplan–Meier curves to illustrate survival trends in overall patient group. (A) The expected pattern of worsening survival associated with more advanced disease was observed. The survival curves according to plasma M65 levels (B) were comparable with those seen for serum carbohydrate antigen (CA)19-9 levels (C) in the overall patient group.