Diagnostic and prognostic biomarkers of malignant pleural mesothelioma

Tsim, Selina (2018) Diagnostic and prognostic biomarkers of malignant pleural mesothelioma. PhD thesis, University of Glasgow.

Full text available as:
[thumbnail of 2018TsimPhD.pdf] PDF
Download (9MB)
Printed Thesis Information: https://eleanor.lib.gla.ac.uk/record=b3320720


Malignant Pleural Mesothelioma (MPM) is an aggressive intrathoracic malignancy with an overall poor prognosis. MPM is associated with asbestos exposure but has a long latency period between exposure and disease development. Incidence of MPM in the UK is therefore still rising, predicted to reach a peak in 2020. The majority of patients with MPM present with breathlessness, frequently due to a pleural effusion and/or chest pain. Diagnosis of MPM can be difficult. Radiological detection of early stage MPM in particular can be challenging, as pleural tumour, nodularity or significant pleural thickening may not be evident. Diagnosis is further complicated by the low yield of pleural fluid cytology examination in MPM and pleural biopsy is therefore usually required to allow definitive diagnosis. This can be achieved under image guidance, at surgical thoracoscopy or at local anaesthetic thoracoscopy (LAT). A significant number of patients are either elderly or have co-morbidity precluding general anaesthesia and surgical thoracoscopy. Image-guided pleural biopsy is not always feasible, particularly in the absence of significant pleural thickening. LAT remains a limited resource in the UK.

A non-invasive biomarker of MPM, which could be performed early in the patient’s presentation, and that could be available to most hospitals, would therefore be a major clinical advance, allowing clinicians to direct appropriate patients to specialist centres with access to LAT and specialist MDT input where MPM appears likely. There have been several potential blood biomarkers identified in the mesothelioma literature, including the most widely studied, Mesothelin, and more recently Fibulin-3 and SOMAscan™. Unfortunately study results have been variably limited by retrospective study design, inconsistent sampling time points, inconsistent results and lack of external validation, therefore despite initial promising results, none of these biomarkers have entered routine clinical practice for diagnosis. Similarly, utility of imaging biomarkers such as perfusion Computed Tomography (CT), Positron Emission Tomography (PET) and Dynamic Contrast Enhanced Magnetic Resonance Imaging (DCE-MRI) has been limited by high radiation dose, limited availability, and requirement for bulky (and therefore late stage) disease for assessment respectively.

In chapter 2, study design, recruitment and preliminary results of the DIAPHRAGM (Diagnostic and Prognostic Biomarkers in the Rational Assessment of Mesothelioma) study are reported. A prospective, multi-centre study was designed, recruiting patients with suspected pleural malignancy (SPM) at initial presentation to secondary care services, from a mixture of academic and more clinical units in the UK and Ireland, in addition to asbestos-exposed control subjects. In one of the largest biomarker studies in mesothelioma to date, 639 patients with SPM and 113 asbestos-exposed control subjects were recruited over three years. Data cleaning is being finalised by the Cancer Research UK Clinical Trials Unit Glasgow at the time of writing. Preliminary results reveal that 26% (n=154) patients recruited to the SPM cohort were diagnosed with MPM, 33% (n=209) had secondary pleural malignancy and 34% (n=218) were diagnosed with benign pleural disease. A final diagnosis is awaited in 7% (n=47) at the time of writing. SOMAscan™ and Fibulin-3 biomarker analyses are ongoing and DIAPHRAGM will definitively answer the question of diagnostic utility of these blood biomarkers in routine clinical practice, in a ‘real-life’ MPM population, relative to that of Mesothelin.

In chapter 3, contrast-enhanced MRI was performed in patients with suspected MPM and a novel MRI biomarker of pleural malignancy defined (Early Contrast Enhancement – ECE). ECE was defined as a peak in pleural signal intensity at or before 4.5 minutes after intravenous Gadobutrol administration. ECE assessment was successfully performed in all patients who underwent contrast-enhanced MRI. This included patients with pleural thickening <10mm (49/58 (84%)), the mean pleural thickness of all patients was 5mm. ECE demonstrated good overall diagnostic performance for the detection of pleural malignancy (sensitivity 83% (95% CI 61 – 94), specificity 83% (95% CI 68 - 91%), positive predictive value 68% (95% CI 47 – 84%), negative predictive value 92% (95% CI 78 – 97%)), comparable to morphology assessment at CT morphology and MRI morphology by experienced thoracic radiologists. In addition, ECE demonstrated good reproducibility (inter-observer κ = 0.864), superior to subjective morphology assessment at CT and MRI. Mean signal intensity gradient (MSIG), a marker of patient’s contrast enhancement pattern, correlated with tumour Microvessel Density (MVD) using Factor VII immunostain (Spearman’s rho = 0.43, p=0.02). Additionally, a high MSIG (>0.533AU/s), indicative of high tumour vascularity, was associated with poor median overall survival (12 months vs. 20 months, p=0.047).

Staging of MPM represents an additional challenge to clinicians. This is due to the complex morphology and often rind-like growth pattern of MPM. In addition, delineation of pleural disease from adjacent structures such as intercostal muscle and diaphragm can be difficult to assess, particularly at CT, which is the most commonly used imaging modality for diagnostic and staging assessment in MPM. Current clinical staging frequently underestimates extent of disease, with a significant proportion of patients being upstaged at time of surgery, and is limited by high inter-observer variability. Recent studies have reported the prognostic significance of CT-derived tumour volume; however, many of these studies have been limited by the laborious or complex nature of tumour segmentation, significant inter-observer variability or challenges encountered in separating pleural tumour from adjacent structures, which are often of similar density. MRI is superior to CT in the detection of invasion of the chest wall and diaphragm in MPM.

In Chapter 4, MRI was used to quantitatively assess pleural tumour volume in 31 patients with MPM using novel semi-automated segmentation methodology. Four different segmentation methodologies, using Myrian® segmentation software were developed and examined. Optimum methodology was defined, based on the accuracy of volume estimates of an MRI phantom, visual-based analysis, intra-observer agreement and analysis time. Using the optimum methodology, there was acceptable error around the MRI phantom volume (3.6%), a reasonable analysis time (approximately 14 minutes), good intra-observer agreement (intra-class correlation coefficient (ICC) 0.875) and excellent inter-observer agreement (ICC 0.962). Patients with a high MRI-estimated tumour volume (≥300cm3) had a significantly poorer median overall survival (8.5 months vs. 20 months) and was a statistically significant prognostic variable on univariate (HR 2.273 (95% CI 1.162 – 4.446), p=0.016) and multi-variate Cox proportional hazards model (HR 2.114 (95% CI 1.046 – 4.270), p=0.037).

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Keywords: Malignant pleural mesothelioma, magnetic resonance imaging, biomarkers, pleural effusion, pleural malignancy.
Subjects: Q Science > QR Microbiology
R Medicine > RC Internal medicine > RC0254 Neoplasms. Tumors. Oncology (including Cancer)
Colleges/Schools: College of Medical Veterinary and Life Sciences > School of Cancer Sciences
Funder's Name: NHS Greater Glasgow and Clyde Endowment Funds (NHSGGCEN)
Supervisor's Name: Chalmers, Professor Anthony and Blyth, Dr. Kevin
Date of Award: 2018
Depositing User: Dr Selina Tsim
Unique ID: glathesis:2018-30687
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 20 Jul 2018 15:09
Last Modified: 11 Sep 2018 07:06
URI: https://theses.gla.ac.uk/id/eprint/30687
Related URLs:

Actions (login required)

View Item View Item


Downloads per month over past year