Biomarkers for pancreatic cancer: identification, validation and clinical application

Ali, Asif (2015) Biomarkers for pancreatic cancer: identification, validation and clinical application. PhD thesis, University of Glasgow.

Abstract

Pancreatic cancer is common and aggressive: the main type is pancreatic ductal adenocarcinoma (PDAC). It is the fifth most common cause of cancer death in the UK with an overall five year survival of only 2-3%. Establishing the diagnosis of PDAC is important for optimal patient management but can be difficult and relies on imaging and cytology/pathology. Although imaging may be highly suggestive of PDAC, a pathological diagnosis is preferred prior to definitive treatment; therefore tissue samples are required. Cytology samples are obtained at endoscopy. Cytological analysis requires the identification of different cell types and in particular the distinction of malignant pancreatic epithelial cells from reactive pancreatic cells and other gastrointestinal contaminants. This requires experience and expertise and can be difficult. A tissue diagnosis is not achieved in a significant proportion of PDAC cases. Hence, an unmet clinical need exists for the diagnosis of PDAC from cytological samples. One potential way of improving the cytological diagnosis is to use immunohistochemistry (IHC) biomarkers as an adjunct to cytology in difficult to diagnose cases. Diagnostic IHC biomarkers have been investigated, but to date none has entered into routine clinical practice.
The aim of this study was to improve the diagnosis of PDAC from cytology samples. It is hoped that the identification and validation of IHC biomarkers in PDAC will help their clinical translation. For biomarker identification a meta-analysis of potential IHC diagnostic biomarkers investigated in PDAC was performed. Sixteen biomarkers were quantified in meta-analysis and the highest ranked biomarkers were KOC, maspin, S100P, mesothelin and MUC1. These biomarkers have not entered into clinical practice partly because they were investigated in separate studies with relatively small sample sizes and without a uniform and clinically appropriate cut-off.
Biomarkers identified in the meta-analysis were validated in a resection cohort from patients with pancreatico-biliary adenocarcinoma. The aim was to identify better biomarkers and cut-offs that could potentially be investigated in cytology samples. KOC, S100P, mesothelin and MUC1 were investigated in one set of tissue microarrays, while maspin was investigated in another set. Five cut-offs were carefully chosen for sensitivity/specificity analysis using receiver operating characteristics curve analysis. Using 20% positive cells as a cut-off achieved higher sensitivity/specificity values: KOC 84%/100%; S100P 83%/100%; mesothelin 88%/92%; MUC1 89%/63%; and Maspin 96%/99%. Analysis of a panel of KOC, S100P and mesothelin achieved almost 100% sensitivity and specificity if at least two biomarkers were positive for both 10% and 20% cut-offs.
Clinical translation of biomarker requires a reliable and reproducible cut-off for interpretation of IHC. We identified three cut-offs for investigation to establish a consensus based cut-off(s) that could potentially be used by pathologists for PDAC and other cancers. A series of IHC images of microarray cores were used to investigate observer agreements for 10%, 20% and +2/+3 cut-offs. Seven pathologists participated in this study. The inter- and intra-observer agreements using the three cut-offs were reasonably good. For all three cut-offs a positive correlation was observed with perceived ease of interpretations (p<0.01 for all cut-offs). Finally, cytoplasmic only staining achieved higher agreement than cytoplasmic/nuclear and cytoplasmic/membranous staining. All three cut-offs investigated achieve reasonable strength of agreement modestly decreasing inter and intra-observer variability in IHC interpretation but 10% is slightly better than 20% and +2/+3 cut-offs.
Finally, KOC, maspin, mesothelin and S100P were investigated in archival cytology samples with the aim to generate a diagnostic panel which could potentially help as an adjunct to cytology. Using 10% cut-off achieved higher sensitivity/specificity values: KOC 92%/100%; maspin 54%/100% and mesothelin 72%/100%. But no staining was observed for S100P. In addition, analysis of a panel of KOC, maspin and mesothelin achieved 82% sensitivity and 100% specificity for 10% cut-off if at least two biomarkers in the panel were positive. The inter-observer agreement for 10% positive cells as IHC cut-off in cytology samples was very good for all three biomarkers.
In conclusion, a panel of KOC, maspin and mesothelin is a suitable diagnostic panel and 10% cut-off is a reasonable cut-off achieving high observer agreement. Their diagnostic accuracies approach those of optimal conventional cytology. These markers may be appropriate for further clinical validation and potentially routine use in difficult cases.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Keywords: Pancreatic cancer, diagnosis, biomarkers, cytology.
• Subjects: R Medicine > RB Pathology
• Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Cancer Sciences
• Supervisor's Name: Oien, Dr. Karin
• Date of Award: 2015
• Depositing User: Dr Asif Ali
• Unique ID: glathesis:2015-6272
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 19 Jun 2015 09:12
• Last Modified: 29 Jun 2017 11:03
• URI: https://theses.gla.ac.uk/id/eprint/6272