An integrated approach to pancreaticobiliary cancer diagnosis

This thesis is subdivided into three parts; it explores the current best methods for establishing a pancreatobiliary cancer diagnosis and examines which steps can be taken to improve this process.
Part I inventories the daily practice of histopathological assessment of pancreatic and hepatobiliary resection specimens. We describe which uncertainties still exist, and how the introduction of neoadjuvant treatment may affect the assessment. In a randomized trial, we demonstrate that two commonly used dissection methods for pancreatic head resections (i.e. axial slicing and bivalving) perform equally with regard to primary outcome measures.
Part 2 explores the potential benefit of minimally invasive DNA based techniques in the establishment of a pancreatobiliary cancer diagnosis, as it may be difficult to procure tumour tissue in some patients with suspected malignancy. We show that circulating tumour DNA could theoretically aid in diagnosing metastatic pancreatic cancer. In a systematic review, we asses which mutations commonly occur in biliary tract cancers, in order to explore if DNA sequencing of brush cytology could increase diagnostic sensitivity. Unfortunately, we did not find any benefit.
Part 3 describes how transcriptomic based subtyping of pancreatic tumours can potentially identify prognostically divergent patient subgroups. In a single centre patient cohort, we identify a poor prognosis subgroup. In a subsequent multicentre patient cohort, we validate these findings. Finally, we provide a critical note regarding the reliability of published molecular studies, as errors in sample selection appear to be quite common.