Investigation and Implementation of Clinically Relevant Prognostic Biomarkers in Human Malignancies

The discovery and validation of prognostic biomarkers is a central focus of current medical research that underpins the practice of personalised medicine. In the field of surgical pathology, there is ample opportunity to identify histomorphological variables that can help guide therapeutic decision making. This thesis explores several novel pathological biomarkers in various malignancies, with the aim of translating knowledge gained at the basic science level into techniques applicable to routine diagnostic surgical pathology practice. This thesis is broadly divided into four parts, each of which explores a novel prognostic tool through one or more studies published in peer-reviewed international pathology journals. First, the role of tumour infiltrating lymphocytes as a prognostic biomarker is examined in the contexts of colorectal carcinoma, triple negative breast carcinoma, and mesothelioma. The second part of this thesis explores various prognostic factors in medullary thyroid carcinoma and proposes a novel histopathological grading scheme for predicting overall survival that can be easily implemented in routine surgical pathology practice. The third part of the thesis explores existing grading systems for diffuse pleural mesothelioma and proposes a novel grading scheme that overcomes many of the limitations associated with other proposed systems. The final chapter of this thesis examines a comprehensive range of prognostic markers in colorectal carcinoma to develop a novel prognostic tool for predicting five-year overall survival after primary tumour resection. This prognostic nomogram was constructed and validated using a large cohort of patients with detailed clinicopathological data, is user-friendly, and can be easily accessed online. Each of these topics represents a practical application of our improved understanding of the mechanisms of tumour progression, with the aim of providing more detailed prognostic and predictive information to treating clinicians